freebsd-skq/sys/dev/usb/ohci.c

3632 lines
92 KiB
C
Raw Normal View History

/* $NetBSD: ohci.c,v 1.138 2003/02/08 03:32:50 ichiro Exp $ */
/* Also, already ported:
* $NetBSD: ohci.c,v 1.140 2003/05/13 04:42:00 gson Exp $
* $NetBSD: ohci.c,v 1.141 2003/09/10 20:08:29 mycroft Exp $
* $NetBSD: ohci.c,v 1.142 2003/10/11 03:04:26 toshii Exp $
* $NetBSD: ohci.c,v 1.143 2003/10/18 04:50:35 simonb Exp $
* $NetBSD: ohci.c,v 1.144 2003/11/23 19:18:06 augustss Exp $
* $NetBSD: ohci.c,v 1.145 2003/11/23 19:20:25 augustss Exp $
* $NetBSD: ohci.c,v 1.146 2003/12/29 08:17:10 toshii Exp $
* $NetBSD: ohci.c,v 1.147 2004/06/22 07:20:35 mycroft Exp $
* $NetBSD: ohci.c,v 1.148 2004/06/22 18:27:46 mycroft Exp $
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*-
* Copyright (c) 1998 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
2000-05-14 16:43:10 +00:00
* by Lennart Augustsson (lennart@augustsson.net) at
* Carlstedt Research & Technology.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the NetBSD
* Foundation, Inc. and its contributors.
* 4. Neither the name of The NetBSD Foundation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
/*
* USB Open Host Controller driver.
*
2002-04-01 21:13:22 +00:00
* OHCI spec: http://www.compaq.com/productinfo/development/openhci.html
* USB spec: http://www.usb.org/developers/docs/usbspec.zip
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/malloc.h>
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#include <sys/kernel.h>
#if defined(__NetBSD__) || defined(__OpenBSD__)
#include <sys/device.h>
#include <sys/select.h>
#elif defined(__FreeBSD__)
#include <sys/endian.h>
#include <sys/module.h>
#include <sys/bus.h>
#if defined(DIAGNOSTIC) && defined(__i386__) && defined(__FreeBSD__)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#include <machine/cpu.h>
#endif
#endif
#include <sys/proc.h>
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#include <sys/queue.h>
#include <sys/sysctl.h>
#include <machine/bus.h>
#include <machine/endian.h>
#include <dev/usb/usb.h>
#include <dev/usb/usbdi.h>
#include <dev/usb/usbdivar.h>
#include <dev/usb/usb_mem.h>
#include <dev/usb/usb_quirks.h>
#include <dev/usb/ohcireg.h>
#include <dev/usb/ohcivar.h>
#if defined(__FreeBSD__)
#include <machine/clock.h>
#define delay(d) DELAY(d)
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#if defined(__OpenBSD__)
struct cfdriver ohci_cd = {
NULL, "ohci", DV_DULL
};
#endif
#ifdef USB_DEBUG
#define DPRINTF(x) if (ohcidebug) logprintf x
#define DPRINTFN(n,x) if (ohcidebug>(n)) logprintf x
int ohcidebug = 0;
SYSCTL_NODE(_hw_usb, OID_AUTO, ohci, CTLFLAG_RW, 0, "USB ohci");
SYSCTL_INT(_hw_usb_ohci, OID_AUTO, debug, CTLFLAG_RW,
&ohcidebug, 0, "ohci debug level");
#ifndef __NetBSD__
#define bitmask_snprintf(q,f,b,l) snprintf((b), (l), "%b", (q), (f))
#endif
#else
#define DPRINTF(x)
#define DPRINTFN(n,x)
#endif
/*
* The OHCI controller is little endian, so on big endian machines
* the data strored in memory needs to be swapped.
*/
#if defined(__OpenBSD__)
#if BYTE_ORDER == BIG_ENDIAN
#define htole32(x) (bswap32(x))
#define le32toh(x) (bswap32(x))
#else
#define htole32(x) (x)
#define le32toh(x) (x)
#endif
#endif
struct ohci_pipe;
Static ohci_soft_ed_t *ohci_alloc_sed(ohci_softc_t *);
Static void ohci_free_sed(ohci_softc_t *, ohci_soft_ed_t *);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
Static ohci_soft_td_t *ohci_alloc_std(ohci_softc_t *);
Static void ohci_free_std(ohci_softc_t *, ohci_soft_td_t *);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
Static ohci_soft_itd_t *ohci_alloc_sitd(ohci_softc_t *);
Static void ohci_free_sitd(ohci_softc_t *,ohci_soft_itd_t *);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#if 0
Static void ohci_free_std_chain(ohci_softc_t *, ohci_soft_td_t *,
ohci_soft_td_t *);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#endif
Static usbd_status ohci_alloc_std_chain(struct ohci_pipe *,
ohci_softc_t *, int, int, usbd_xfer_handle,
ohci_soft_td_t *, ohci_soft_td_t **);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
#if defined(__NetBSD__) || defined(__OpenBSD__)
Static void ohci_shutdown(void *v);
Static void ohci_power(int, void *);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
#endif
Static usbd_status ohci_open(usbd_pipe_handle);
Static void ohci_poll(struct usbd_bus *);
Static void ohci_softintr(void *);
Static void ohci_waitintr(ohci_softc_t *, usbd_xfer_handle);
Static void ohci_add_done(ohci_softc_t *, ohci_physaddr_t);
Static void ohci_rhsc(ohci_softc_t *, usbd_xfer_handle);
Static usbd_status ohci_device_request(usbd_xfer_handle xfer);
Static void ohci_add_ed(ohci_soft_ed_t *, ohci_soft_ed_t *);
Static void ohci_rem_ed(ohci_soft_ed_t *, ohci_soft_ed_t *);
Static void ohci_hash_add_td(ohci_softc_t *, ohci_soft_td_t *);
Static void ohci_hash_rem_td(ohci_softc_t *, ohci_soft_td_t *);
Static ohci_soft_td_t *ohci_hash_find_td(ohci_softc_t *, ohci_physaddr_t);
Static void ohci_hash_add_itd(ohci_softc_t *, ohci_soft_itd_t *);
Static void ohci_hash_rem_itd(ohci_softc_t *, ohci_soft_itd_t *);
Static ohci_soft_itd_t *ohci_hash_find_itd(ohci_softc_t *, ohci_physaddr_t);
Static usbd_status ohci_setup_isoc(usbd_pipe_handle pipe);
Static void ohci_device_isoc_enter(usbd_xfer_handle);
Static usbd_status ohci_allocm(struct usbd_bus *, usb_dma_t *, u_int32_t);
Static void ohci_freem(struct usbd_bus *, usb_dma_t *);
Static usbd_xfer_handle ohci_allocx(struct usbd_bus *);
Static void ohci_freex(struct usbd_bus *, usbd_xfer_handle);
Static usbd_status ohci_root_ctrl_transfer(usbd_xfer_handle);
Static usbd_status ohci_root_ctrl_start(usbd_xfer_handle);
Static void ohci_root_ctrl_abort(usbd_xfer_handle);
Static void ohci_root_ctrl_close(usbd_pipe_handle);
Static void ohci_root_ctrl_done(usbd_xfer_handle);
Static usbd_status ohci_root_intr_transfer(usbd_xfer_handle);
Static usbd_status ohci_root_intr_start(usbd_xfer_handle);
Static void ohci_root_intr_abort(usbd_xfer_handle);
Static void ohci_root_intr_close(usbd_pipe_handle);
Static void ohci_root_intr_done(usbd_xfer_handle);
Static usbd_status ohci_device_ctrl_transfer(usbd_xfer_handle);
Static usbd_status ohci_device_ctrl_start(usbd_xfer_handle);
Static void ohci_device_ctrl_abort(usbd_xfer_handle);
Static void ohci_device_ctrl_close(usbd_pipe_handle);
Static void ohci_device_ctrl_done(usbd_xfer_handle);
Static usbd_status ohci_device_bulk_transfer(usbd_xfer_handle);
Static usbd_status ohci_device_bulk_start(usbd_xfer_handle);
Static void ohci_device_bulk_abort(usbd_xfer_handle);
Static void ohci_device_bulk_close(usbd_pipe_handle);
Static void ohci_device_bulk_done(usbd_xfer_handle);
Static usbd_status ohci_device_intr_transfer(usbd_xfer_handle);
Static usbd_status ohci_device_intr_start(usbd_xfer_handle);
Static void ohci_device_intr_abort(usbd_xfer_handle);
Static void ohci_device_intr_close(usbd_pipe_handle);
Static void ohci_device_intr_done(usbd_xfer_handle);
Static usbd_status ohci_device_isoc_transfer(usbd_xfer_handle);
Static usbd_status ohci_device_isoc_start(usbd_xfer_handle);
Static void ohci_device_isoc_abort(usbd_xfer_handle);
Static void ohci_device_isoc_close(usbd_pipe_handle);
Static void ohci_device_isoc_done(usbd_xfer_handle);
Static usbd_status ohci_device_setintr(ohci_softc_t *sc,
struct ohci_pipe *pipe, int ival);
2001-07-05 10:09:22 +00:00
Static int ohci_str(usb_string_descriptor_t *, int, const char *);
Static void ohci_timeout(void *);
Static void ohci_timeout_task(void *);
Static void ohci_rhsc_able(ohci_softc_t *, int);
Static void ohci_rhsc_enable(void *);
Static void ohci_close_pipe(usbd_pipe_handle, ohci_soft_ed_t *);
Static void ohci_abort_xfer(usbd_xfer_handle, usbd_status);
Static void ohci_device_clear_toggle(usbd_pipe_handle pipe);
Static void ohci_noop(usbd_pipe_handle pipe);
Static usbd_status ohci_controller_init(ohci_softc_t *sc);
#ifdef USB_DEBUG
Static void ohci_dumpregs(ohci_softc_t *);
Static void ohci_dump_tds(ohci_soft_td_t *);
Static void ohci_dump_td(ohci_soft_td_t *);
Static void ohci_dump_ed(ohci_soft_ed_t *);
Static void ohci_dump_itd(ohci_soft_itd_t *);
Static void ohci_dump_itds(ohci_soft_itd_t *);
#endif
#define OBARR(sc) bus_space_barrier((sc)->iot, (sc)->ioh, 0, (sc)->sc_size, \
BUS_SPACE_BARRIER_READ|BUS_SPACE_BARRIER_WRITE)
#define OWRITE1(sc, r, x) \
do { OBARR(sc); bus_space_write_1((sc)->iot, (sc)->ioh, (r), (x)); } while (0)
#define OWRITE2(sc, r, x) \
do { OBARR(sc); bus_space_write_2((sc)->iot, (sc)->ioh, (r), (x)); } while (0)
#define OWRITE4(sc, r, x) \
do { OBARR(sc); bus_space_write_4((sc)->iot, (sc)->ioh, (r), (x)); } while (0)
#define OREAD1(sc, r) (OBARR(sc), bus_space_read_1((sc)->iot, (sc)->ioh, (r)))
#define OREAD2(sc, r) (OBARR(sc), bus_space_read_2((sc)->iot, (sc)->ioh, (r)))
#define OREAD4(sc, r) (OBARR(sc), bus_space_read_4((sc)->iot, (sc)->ioh, (r)))
/* Reverse the bits in a value 0 .. 31 */
Static u_int8_t revbits[OHCI_NO_INTRS] =
{ 0x00, 0x10, 0x08, 0x18, 0x04, 0x14, 0x0c, 0x1c,
0x02, 0x12, 0x0a, 0x1a, 0x06, 0x16, 0x0e, 0x1e,
0x01, 0x11, 0x09, 0x19, 0x05, 0x15, 0x0d, 0x1d,
0x03, 0x13, 0x0b, 0x1b, 0x07, 0x17, 0x0f, 0x1f };
struct ohci_pipe {
struct usbd_pipe pipe;
ohci_soft_ed_t *sed;
u_int32_t aborting;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
union {
ohci_soft_td_t *td;
ohci_soft_itd_t *itd;
} tail;
/* Info needed for different pipe kinds. */
union {
/* Control pipe */
struct {
usb_dma_t reqdma;
u_int length;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_soft_td_t *setup, *data, *stat;
} ctl;
/* Interrupt pipe */
struct {
int nslots;
int pos;
} intr;
/* Bulk pipe */
struct {
u_int length;
int isread;
} bulk;
/* Iso pipe */
struct iso {
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
int next, inuse;
} iso;
} u;
};
#define OHCI_INTR_ENDPT 1
Static struct usbd_bus_methods ohci_bus_methods = {
ohci_open,
ohci_softintr,
ohci_poll,
ohci_allocm,
ohci_freem,
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ohci_allocx,
ohci_freex,
};
Static struct usbd_pipe_methods ohci_root_ctrl_methods = {
ohci_root_ctrl_transfer,
ohci_root_ctrl_start,
ohci_root_ctrl_abort,
ohci_root_ctrl_close,
ohci_noop,
ohci_root_ctrl_done,
};
Static struct usbd_pipe_methods ohci_root_intr_methods = {
ohci_root_intr_transfer,
ohci_root_intr_start,
ohci_root_intr_abort,
ohci_root_intr_close,
ohci_noop,
ohci_root_intr_done,
};
Static struct usbd_pipe_methods ohci_device_ctrl_methods = {
ohci_device_ctrl_transfer,
ohci_device_ctrl_start,
ohci_device_ctrl_abort,
ohci_device_ctrl_close,
ohci_noop,
ohci_device_ctrl_done,
};
Static struct usbd_pipe_methods ohci_device_intr_methods = {
ohci_device_intr_transfer,
ohci_device_intr_start,
ohci_device_intr_abort,
ohci_device_intr_close,
ohci_device_clear_toggle,
ohci_device_intr_done,
};
Static struct usbd_pipe_methods ohci_device_bulk_methods = {
ohci_device_bulk_transfer,
ohci_device_bulk_start,
ohci_device_bulk_abort,
ohci_device_bulk_close,
ohci_device_clear_toggle,
ohci_device_bulk_done,
};
Static struct usbd_pipe_methods ohci_device_isoc_methods = {
ohci_device_isoc_transfer,
ohci_device_isoc_start,
ohci_device_isoc_abort,
ohci_device_isoc_close,
ohci_noop,
ohci_device_isoc_done,
};
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#if defined(__NetBSD__) || defined(__OpenBSD__)
int
ohci_activate(device_ptr_t self, enum devact act)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
{
struct ohci_softc *sc = (struct ohci_softc *)self;
int rv = 0;
switch (act) {
case DVACT_ACTIVATE:
return (EOPNOTSUPP);
case DVACT_DEACTIVATE:
if (sc->sc_child != NULL)
rv = config_deactivate(sc->sc_child);
sc->sc_dying = 1;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
break;
}
return (rv);
}
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
int
ohci_detach(struct ohci_softc *sc, int flags)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
{
int i, rv = 0;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#if defined(__NetBSD__) || defined(__OpenBSD__)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sc->sc_child != NULL)
rv = config_detach(sc->sc_child, flags);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (rv != 0)
return (rv);
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_uncallout(sc->sc_tmo_rhsc, ohci_rhsc_enable, sc);
#if defined(__NetBSD__) || defined(__OpenBSD__)
powerhook_disestablish(sc->sc_powerhook);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
shutdownhook_disestablish(sc->sc_shutdownhook);
#endif
OWRITE4(sc, OHCI_INTERRUPT_DISABLE, OHCI_ALL_INTRS);
OWRITE4(sc, OHCI_CONTROL, OHCI_HCFS_RESET);
usb_delay_ms(&sc->sc_bus, 300); /* XXX let stray task complete */
for (i = 0; i < OHCI_NO_EDS; i++)
ohci_free_sed(sc, sc->sc_eds[i]);
ohci_free_sed(sc, sc->sc_isoc_head);
ohci_free_sed(sc, sc->sc_bulk_head);
ohci_free_sed(sc, sc->sc_ctrl_head);
usb_freemem(&sc->sc_bus, &sc->sc_hccadma);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
return (rv);
}
ohci_soft_ed_t *
ohci_alloc_sed(ohci_softc_t *sc)
{
ohci_soft_ed_t *sed;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
int i, offs;
usb_dma_t dma;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sc->sc_freeeds == NULL) {
DPRINTFN(2, ("ohci_alloc_sed: allocating chunk\n"));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = usb_allocmem(&sc->sc_bus, OHCI_SED_SIZE * OHCI_SED_CHUNK,
OHCI_ED_ALIGN, &dma);
if (err)
return (NULL);
for(i = 0; i < OHCI_SED_CHUNK; i++) {
offs = i * OHCI_SED_SIZE;
sed = KERNADDR(&dma, offs);
sed->physaddr = DMAADDR(&dma, offs);
sed->next = sc->sc_freeeds;
sc->sc_freeeds = sed;
}
}
sed = sc->sc_freeeds;
sc->sc_freeeds = sed->next;
memset(&sed->ed, 0, sizeof(ohci_ed_t));
sed->next = 0;
return (sed);
}
void
ohci_free_sed(ohci_softc_t *sc, ohci_soft_ed_t *sed)
{
sed->next = sc->sc_freeeds;
sc->sc_freeeds = sed;
}
ohci_soft_td_t *
ohci_alloc_std(ohci_softc_t *sc)
{
ohci_soft_td_t *std;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
int i, offs;
usb_dma_t dma;
int s;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sc->sc_freetds == NULL) {
DPRINTFN(2, ("ohci_alloc_std: allocating chunk\n"));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = usb_allocmem(&sc->sc_bus, OHCI_STD_SIZE * OHCI_STD_CHUNK,
OHCI_TD_ALIGN, &dma);
if (err)
return (NULL);
s = splusb();
for(i = 0; i < OHCI_STD_CHUNK; i++) {
offs = i * OHCI_STD_SIZE;
std = KERNADDR(&dma, offs);
std->physaddr = DMAADDR(&dma, offs);
std->nexttd = sc->sc_freetds;
sc->sc_freetds = std;
}
splx(s);
}
s = splusb();
std = sc->sc_freetds;
sc->sc_freetds = std->nexttd;
memset(&std->td, 0, sizeof(ohci_td_t));
std->nexttd = NULL;
std->xfer = NULL;
ohci_hash_add_td(sc, std);
splx(s);
return (std);
}
void
ohci_free_std(ohci_softc_t *sc, ohci_soft_td_t *std)
{
int s;
s = splusb();
ohci_hash_rem_td(sc, std);
std->nexttd = sc->sc_freetds;
sc->sc_freetds = std;
splx(s);
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status
ohci_alloc_std_chain(struct ohci_pipe *opipe, ohci_softc_t *sc,
int alen, int rd, usbd_xfer_handle xfer,
ohci_soft_td_t *sp, ohci_soft_td_t **ep)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
{
ohci_soft_td_t *next, *cur;
ohci_physaddr_t dataphys;
u_int32_t tdflags;
int offset = 0;
int len, curlen;
usb_dma_t *dma = &xfer->dmabuf;
u_int16_t flags = xfer->flags;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTFN(alen < 4096,("ohci_alloc_std_chain: start len=%d\n", alen));
len = alen;
cur = sp;
tdflags = htole32(
(rd ? OHCI_TD_IN : OHCI_TD_OUT) |
(flags & USBD_SHORT_XFER_OK ? OHCI_TD_R : 0) |
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
OHCI_TD_NOCC | OHCI_TD_TOGGLE_CARRY | OHCI_TD_SET_DI(6));
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
for (;;) {
next = ohci_alloc_std(sc);
if (next == NULL)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
goto nomem;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
dataphys = DMAADDR(dma, offset);
/*
* The OHCI hardware can handle at most one 4k crossing.
* XXX - currently we only allocate contigous buffers, but
* the OHCI spec says: If during the data transfer the buffer
* address contained in the HC's working copy of
* CurrentBufferPointer crosses a 4K boundary, the upper 20
* bits of Buffer End are copied to the working value of
* CurrentBufferPointer causing the next buffer address to
* be the 0th byte in the same 4K page that contains the
* last byte of the buffer (the 4K boundary crossing may
* occur within a data packet transfer.)
*
* If/when dma has multiple segments, this will need to
* properly handle fragmenting TD's.
*
* Note that if we are gathering data from multiple SMALL
* segments, e.g. mbufs, we need to do special gymnastics,
* e.g. bounce buffering or data aggregation,
* BEFORE WE GET HERE because a bulk USB transfer must
* consist of maximally sized packets right up to the end.
* A shorter than maximal packet means that it is the end
* of the transfer. If the data transfer length is a
* multiple of the packet size, then a 0 byte
* packet will be the signal of the end of transfer.
* Since packets can't cross TDs this means that
* each TD except the last one must cover an exact multiple
* of the maximal packet length.
*/
if (OHCI_PAGE_OFFSET(dataphys) + len <= (2 * OHCI_PAGE_SIZE)) {
/* We can handle all that remains in this TD */
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
curlen = len;
} else {
/* must use multiple TDs, fill as much as possible. */
curlen = 2 * OHCI_PAGE_SIZE -
OHCI_PAGE_OFFSET(dataphys);
/* the length must be a multiple of the max size */
curlen -= curlen %
UGETW(opipe->pipe.endpoint->edesc->wMaxPacketSize);
KASSERT((curlen != 0), ("ohci_alloc_std: curlen == 0"));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
}
DPRINTFN(4,("ohci_alloc_std_chain: dataphys=0x%08x "
"len=%d curlen=%d\n",
dataphys, len, curlen));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
len -= curlen;
cur->td.td_flags = tdflags;
cur->td.td_cbp = htole32(dataphys);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
cur->nexttd = next;
cur->td.td_nexttd = htole32(next->physaddr);
cur->td.td_be = htole32(DMAADDR(dma, offset + curlen - 1));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
cur->len = curlen;
cur->flags = OHCI_ADD_LEN;
cur->xfer = xfer;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTFN(10,("ohci_alloc_std_chain: cbp=0x%08x be=0x%08x\n",
dataphys, dataphys + curlen - 1));
if (len == 0)
break;
if (len < 0)
panic("Length went negative: %d curlen %d dma %p offset %08x", len, curlen, dma, (int)0);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTFN(10,("ohci_alloc_std_chain: extend chain\n"));
offset += curlen;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
cur = next;
}
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
if ((flags & USBD_FORCE_SHORT_XFER) &&
alen % UGETW(opipe->pipe.endpoint->edesc->wMaxPacketSize) == 0) {
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/* Force a 0 length transfer at the end. */
cur = next;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
next = ohci_alloc_std(sc);
if (next == NULL)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
goto nomem;
cur->td.td_flags = tdflags;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
cur->td.td_cbp = 0; /* indicate 0 length packet */
cur->nexttd = next;
cur->td.td_nexttd = htole32(next->physaddr);
cur->td.td_be = ~0;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
cur->len = 0;
cur->flags = 0;
cur->xfer = xfer;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
DPRINTFN(2,("ohci_alloc_std_chain: add 0 xfer\n"));
}
*ep = cur;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
return (USBD_NORMAL_COMPLETION);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
nomem:
/* XXX free chain */
return (USBD_NOMEM);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
}
#if 0
Static void
ohci_free_std_chain(ohci_softc_t *sc, ohci_soft_td_t *std,
ohci_soft_td_t *stdend)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
{
ohci_soft_td_t *p;
for (; std != stdend; std = p) {
p = std->nexttd;
ohci_free_std(sc, std);
}
}
#endif
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ohci_soft_itd_t *
ohci_alloc_sitd(ohci_softc_t *sc)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
ohci_soft_itd_t *sitd;
usbd_status err;
int i, s, offs;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
usb_dma_t dma;
if (sc->sc_freeitds == NULL) {
DPRINTFN(2, ("ohci_alloc_sitd: allocating chunk\n"));
err = usb_allocmem(&sc->sc_bus, OHCI_SITD_SIZE * OHCI_SITD_CHUNK,
OHCI_ITD_ALIGN, &dma);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
if (err)
return (NULL);
s = splusb();
for(i = 0; i < OHCI_SITD_CHUNK; i++) {
offs = i * OHCI_SITD_SIZE;
sitd = KERNADDR(&dma, offs);
sitd->physaddr = DMAADDR(&dma, offs);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
sitd->nextitd = sc->sc_freeitds;
sc->sc_freeitds = sitd;
}
splx(s);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
s = splusb();
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
sitd = sc->sc_freeitds;
sc->sc_freeitds = sitd->nextitd;
memset(&sitd->itd, 0, sizeof(ohci_itd_t));
sitd->nextitd = NULL;
sitd->xfer = NULL;
ohci_hash_add_itd(sc, sitd);
splx(s);
#ifdef DIAGNOSTIC
sitd->isdone = 0;
#endif
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
return (sitd);
}
void
ohci_free_sitd(ohci_softc_t *sc, ohci_soft_itd_t *sitd)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
int s;
DPRINTFN(10,("ohci_free_sitd: sitd=%p\n", sitd));
#ifdef DIAGNOSTIC
if (!sitd->isdone) {
panic("ohci_free_sitd: sitd=%p not done", sitd);
return;
}
/* Warn double free */
sitd->isdone = 0;
#endif
s = splusb();
ohci_hash_rem_itd(sc, sitd);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
sitd->nextitd = sc->sc_freeitds;
sc->sc_freeitds = sitd;
splx(s);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
usbd_status
ohci_init(ohci_softc_t *sc)
{
ohci_soft_ed_t *sed, *psed;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
int i;
u_int32_t rev;
DPRINTF(("ohci_init: start\n"));
#if defined(__OpenBSD__)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
printf(",");
#else
printf("%s:", USBDEVNAME(sc->sc_bus.bdev));
#endif
rev = OREAD4(sc, OHCI_REVISION);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
printf(" OHCI version %d.%d%s\n", OHCI_REV_HI(rev), OHCI_REV_LO(rev),
OHCI_REV_LEGACY(rev) ? ", legacy support" : "");
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (OHCI_REV_HI(rev) != 1 || OHCI_REV_LO(rev) != 0) {
printf("%s: unsupported OHCI revision\n",
USBDEVNAME(sc->sc_bus.bdev));
sc->sc_bus.usbrev = USBREV_UNKNOWN;
return (USBD_INVAL);
}
sc->sc_bus.usbrev = USBREV_1_0;
for (i = 0; i < OHCI_HASH_SIZE; i++)
LIST_INIT(&sc->sc_hash_tds[i]);
for (i = 0; i < OHCI_HASH_SIZE; i++)
LIST_INIT(&sc->sc_hash_itds[i]);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
SIMPLEQ_INIT(&sc->sc_free_xfers);
/* XXX determine alignment by R/W */
/* Allocate the HCCA area. */
err = usb_allocmem(&sc->sc_bus, OHCI_HCCA_SIZE,
OHCI_HCCA_ALIGN, &sc->sc_hccadma);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (err)
return (err);
sc->sc_hcca = KERNADDR(&sc->sc_hccadma, 0);
memset(sc->sc_hcca, 0, OHCI_HCCA_SIZE);
sc->sc_eintrs = OHCI_NORMAL_INTRS;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/* Allocate dummy ED that starts the control list. */
sc->sc_ctrl_head = ohci_alloc_sed(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sc->sc_ctrl_head == NULL) {
err = USBD_NOMEM;
goto bad1;
}
sc->sc_ctrl_head->ed.ed_flags |= htole32(OHCI_ED_SKIP);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/* Allocate dummy ED that starts the bulk list. */
sc->sc_bulk_head = ohci_alloc_sed(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sc->sc_bulk_head == NULL) {
err = USBD_NOMEM;
goto bad2;
}
sc->sc_bulk_head->ed.ed_flags |= htole32(OHCI_ED_SKIP);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/* Allocate dummy ED that starts the isochronous list. */
sc->sc_isoc_head = ohci_alloc_sed(sc);
if (sc->sc_isoc_head == NULL) {
err = USBD_NOMEM;
goto bad3;
}
sc->sc_isoc_head->ed.ed_flags |= htole32(OHCI_ED_SKIP);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/* Allocate all the dummy EDs that make up the interrupt tree. */
for (i = 0; i < OHCI_NO_EDS; i++) {
sed = ohci_alloc_sed(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sed == NULL) {
while (--i >= 0)
ohci_free_sed(sc, sc->sc_eds[i]);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_NOMEM;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
goto bad4;
}
/* All ED fields are set to 0. */
sc->sc_eds[i] = sed;
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
if (i != 0)
psed = sc->sc_eds[(i-1) / 2];
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
else
psed= sc->sc_isoc_head;
sed->next = psed;
sed->ed.ed_nexted = htole32(psed->physaddr);
}
/*
* Fill HCCA interrupt table. The bit reversal is to get
* the tree set up properly to spread the interrupts.
*/
for (i = 0; i < OHCI_NO_INTRS; i++)
sc->sc_hcca->hcca_interrupt_table[revbits[i]] =
htole32(sc->sc_eds[OHCI_NO_EDS-OHCI_NO_INTRS+i]->physaddr);
#ifdef USB_DEBUG
if (ohcidebug > 15) {
for (i = 0; i < OHCI_NO_EDS; i++) {
printf("ed#%d ", i);
ohci_dump_ed(sc->sc_eds[i]);
}
printf("iso ");
ohci_dump_ed(sc->sc_isoc_head);
}
#endif
err = ohci_controller_init(sc);
if (err != USBD_NORMAL_COMPLETION)
goto bad5;
/* Set up the bus struct. */
sc->sc_bus.methods = &ohci_bus_methods;
sc->sc_bus.pipe_size = sizeof(struct ohci_pipe);
#if defined(__NetBSD__) || defined(__OpenBSD__)
sc->sc_control = sc->sc_intre = 0;
sc->sc_powerhook = powerhook_establish(ohci_power, sc);
sc->sc_shutdownhook = shutdownhook_establish(ohci_shutdown, sc);
#endif
usb_callout_init(sc->sc_tmo_rhsc);
return (USBD_NORMAL_COMPLETION);
bad5:
for (i = 0; i < OHCI_NO_EDS; i++)
ohci_free_sed(sc, sc->sc_eds[i]);
bad4:
ohci_free_sed(sc, sc->sc_isoc_head);
bad3:
ohci_free_sed(sc, sc->sc_bulk_head);
bad2:
ohci_free_sed(sc, sc->sc_ctrl_head);
bad1:
usb_freemem(&sc->sc_bus, &sc->sc_hccadma);
return (err);
}
Static usbd_status
ohci_controller_init(ohci_softc_t *sc)
{
int i;
u_int32_t s, ctl, ival, hcr, fm, per, desca;
/* Determine in what context we are running. */
ctl = OREAD4(sc, OHCI_CONTROL);
if (ctl & OHCI_IR) {
/* SMM active, request change */
DPRINTF(("ohci_init: SMM active, request owner change\n"));
s = OREAD4(sc, OHCI_COMMAND_STATUS);
OWRITE4(sc, OHCI_COMMAND_STATUS, s | OHCI_OCR);
for (i = 0; i < 100 && (ctl & OHCI_IR); i++) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_delay_ms(&sc->sc_bus, 1);
ctl = OREAD4(sc, OHCI_CONTROL);
}
if ((ctl & OHCI_IR) == 0) {
printf("%s: SMM does not respond, resetting\n",
USBDEVNAME(sc->sc_bus.bdev));
OWRITE4(sc, OHCI_CONTROL, OHCI_HCFS_RESET);
goto reset;
}
#if 0
/* Don't bother trying to reuse the BIOS init, we'll reset it anyway. */
} else if ((ctl & OHCI_HCFS_MASK) != OHCI_HCFS_RESET) {
/* BIOS started controller. */
DPRINTF(("ohci_init: BIOS active\n"));
if ((ctl & OHCI_HCFS_MASK) != OHCI_HCFS_OPERATIONAL) {
OWRITE4(sc, OHCI_CONTROL, OHCI_HCFS_OPERATIONAL);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_delay_ms(&sc->sc_bus, USB_RESUME_DELAY);
}
#endif
} else {
DPRINTF(("ohci_init: cold started\n"));
reset:
/* Controller was cold started. */
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_delay_ms(&sc->sc_bus, USB_BUS_RESET_DELAY);
}
/*
* This reset should not be necessary according to the OHCI spec, but
* without it some controllers do not start.
*/
DPRINTF(("%s: resetting\n", USBDEVNAME(sc->sc_bus.bdev)));
OWRITE4(sc, OHCI_CONTROL, OHCI_HCFS_RESET);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_delay_ms(&sc->sc_bus, USB_BUS_RESET_DELAY);
/* We now own the host controller and the bus has been reset. */
ival = OHCI_GET_IVAL(OREAD4(sc, OHCI_FM_INTERVAL));
OWRITE4(sc, OHCI_COMMAND_STATUS, OHCI_HCR); /* Reset HC */
/* Nominal time for a reset is 10 us. */
for (i = 0; i < 10; i++) {
delay(10);
hcr = OREAD4(sc, OHCI_COMMAND_STATUS) & OHCI_HCR;
if (!hcr)
break;
}
if (hcr) {
printf("%s: reset timeout\n", USBDEVNAME(sc->sc_bus.bdev));
return (USBD_IOERROR);
}
#ifdef USB_DEBUG
if (ohcidebug > 15)
ohci_dumpregs(sc);
#endif
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/* The controller is now in SUSPEND state, we have 2ms to finish. */
/* Set up HC registers. */
OWRITE4(sc, OHCI_HCCA, DMAADDR(&sc->sc_hccadma, 0));
OWRITE4(sc, OHCI_CONTROL_HEAD_ED, sc->sc_ctrl_head->physaddr);
OWRITE4(sc, OHCI_BULK_HEAD_ED, sc->sc_bulk_head->physaddr);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* disable all interrupts and then switch on all desired interrupts */
OWRITE4(sc, OHCI_INTERRUPT_DISABLE, OHCI_ALL_INTRS);
OWRITE4(sc, OHCI_INTERRUPT_ENABLE, sc->sc_eintrs | OHCI_MIE);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* switch on desired functional features */
ctl = OREAD4(sc, OHCI_CONTROL);
ctl &= ~(OHCI_CBSR_MASK | OHCI_LES | OHCI_HCFS_MASK | OHCI_IR);
ctl |= OHCI_PLE | OHCI_IE | OHCI_CLE | OHCI_BLE |
OHCI_RATIO_1_4 | OHCI_HCFS_OPERATIONAL;
/* And finally start it! */
OWRITE4(sc, OHCI_CONTROL, ctl);
/*
* The controller is now OPERATIONAL. Set a some final
* registers that should be set earlier, but that the
* controller ignores when in the SUSPEND state.
*/
fm = (OREAD4(sc, OHCI_FM_INTERVAL) & OHCI_FIT) ^ OHCI_FIT;
fm |= OHCI_FSMPS(ival) | ival;
OWRITE4(sc, OHCI_FM_INTERVAL, fm);
per = OHCI_PERIODIC(ival); /* 90% periodic */
OWRITE4(sc, OHCI_PERIODIC_START, per);
/* Fiddle the No OverCurrent Protection bit to avoid chip bug. */
desca = OREAD4(sc, OHCI_RH_DESCRIPTOR_A);
OWRITE4(sc, OHCI_RH_DESCRIPTOR_A, desca | OHCI_NOCP);
OWRITE4(sc, OHCI_RH_STATUS, OHCI_LPSC); /* Enable port power */
usb_delay_ms(&sc->sc_bus, OHCI_ENABLE_POWER_DELAY);
OWRITE4(sc, OHCI_RH_DESCRIPTOR_A, desca);
/*
* The AMD756 requires a delay before re-reading the register,
* otherwise it will occasionally report 0 ports.
*/
sc->sc_noport = 0;
for (i = 0; i < 10 && sc->sc_noport == 0; i++) {
usb_delay_ms(&sc->sc_bus, OHCI_READ_DESC_DELAY);
sc->sc_noport = OHCI_GET_NDP(OREAD4(sc, OHCI_RH_DESCRIPTOR_A));
}
#ifdef USB_DEBUG
if (ohcidebug > 5)
ohci_dumpregs(sc);
#endif
return (USBD_NORMAL_COMPLETION);
}
usbd_status
ohci_allocm(struct usbd_bus *bus, usb_dma_t *dma, u_int32_t size)
{
return (usb_allocmem(bus, size, 0, dma));
}
void
ohci_freem(struct usbd_bus *bus, usb_dma_t *dma)
{
usb_freemem(bus, dma);
}
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
usbd_xfer_handle
ohci_allocx(struct usbd_bus *bus)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_softc *sc = (struct ohci_softc *)bus;
usbd_xfer_handle xfer;
xfer = SIMPLEQ_FIRST(&sc->sc_free_xfers);
if (xfer != NULL) {
SIMPLEQ_REMOVE_HEAD(&sc->sc_free_xfers, next);
#ifdef DIAGNOSTIC
if (xfer->busy_free != XFER_FREE) {
printf("ohci_allocx: xfer=%p not free, 0x%08x\n", xfer,
xfer->busy_free);
}
#endif
} else {
xfer = malloc(sizeof(struct ohci_xfer), M_USB, M_NOWAIT);
}
if (xfer != NULL) {
memset(xfer, 0, sizeof (struct ohci_xfer));
usb_init_task(&OXFER(xfer)->abort_task, ohci_timeout_task,
xfer);
OXFER(xfer)->ohci_xfer_flags = 0;
#ifdef DIAGNOSTIC
xfer->busy_free = XFER_BUSY;
#endif
}
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
return (xfer);
}
void
ohci_freex(struct usbd_bus *bus, usbd_xfer_handle xfer)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_softc *sc = (struct ohci_softc *)bus;
struct ohci_xfer *oxfer = (struct ohci_xfer *)xfer;
ohci_soft_itd_t *sitd;
if (oxfer->ohci_xfer_flags & OHCI_ISOC_DIRTY) {
for (sitd = xfer->hcpriv; sitd != NULL && sitd->xfer == xfer;
sitd = sitd->nextitd)
ohci_free_sitd(sc, sitd);
}
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
#ifdef DIAGNOSTIC
if (xfer->busy_free != XFER_BUSY) {
printf("ohci_freex: xfer=%p not busy, 0x%08x\n", xfer,
xfer->busy_free);
return;
}
xfer->busy_free = XFER_FREE;
#endif
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
SIMPLEQ_INSERT_HEAD(&sc->sc_free_xfers, xfer, next);
}
/*
* Shut down the controller when the system is going down.
*/
void
ohci_shutdown(void *v)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
ohci_softc_t *sc = v;
DPRINTF(("ohci_shutdown: stopping the HC\n"));
OWRITE4(sc, OHCI_CONTROL, OHCI_HCFS_RESET);
}
/*
* Handle suspend/resume.
*
* We need to switch to polling mode here, because this routine is
* called from an intterupt context. This is all right since we
* are almost suspended anyway.
*/
void
ohci_power(int why, void *v)
{
ohci_softc_t *sc = v;
u_int32_t ctl;
int s;
#ifdef USB_DEBUG
DPRINTF(("ohci_power: sc=%p, why=%d\n", sc, why));
ohci_dumpregs(sc);
#endif
s = splhardusb();
if (why != PWR_RESUME) {
sc->sc_bus.use_polling++;
ctl = OREAD4(sc, OHCI_CONTROL) & ~OHCI_HCFS_MASK;
if (sc->sc_control == 0) {
/*
* Preserve register values, in case that APM BIOS
* does not recover them.
*/
sc->sc_control = ctl;
sc->sc_intre = OREAD4(sc, OHCI_INTERRUPT_ENABLE);
}
ctl |= OHCI_HCFS_SUSPEND;
OWRITE4(sc, OHCI_CONTROL, ctl);
usb_delay_ms(&sc->sc_bus, USB_RESUME_WAIT);
sc->sc_bus.use_polling--;
} else {
sc->sc_bus.use_polling++;
/* Some broken BIOSes never initialize Controller chip */
ohci_controller_init(sc);
if (sc->sc_intre)
OWRITE4(sc, OHCI_INTERRUPT_ENABLE,
sc->sc_intre & (OHCI_ALL_INTRS | OHCI_MIE));
if (sc->sc_control)
ctl = sc->sc_control;
else
ctl = OREAD4(sc, OHCI_CONTROL);
ctl |= OHCI_HCFS_RESUME;
OWRITE4(sc, OHCI_CONTROL, ctl);
usb_delay_ms(&sc->sc_bus, USB_RESUME_DELAY);
ctl = (ctl & ~OHCI_HCFS_MASK) | OHCI_HCFS_OPERATIONAL;
OWRITE4(sc, OHCI_CONTROL, ctl);
usb_delay_ms(&sc->sc_bus, USB_RESUME_RECOVERY);
sc->sc_control = sc->sc_intre = 0;
sc->sc_bus.use_polling--;
}
splx(s);
}
#ifdef USB_DEBUG
void
ohci_dumpregs(ohci_softc_t *sc)
{
DPRINTF(("ohci_dumpregs: rev=0x%08x control=0x%08x command=0x%08x\n",
OREAD4(sc, OHCI_REVISION),
OREAD4(sc, OHCI_CONTROL),
OREAD4(sc, OHCI_COMMAND_STATUS)));
DPRINTF((" intrstat=0x%08x intre=0x%08x intrd=0x%08x\n",
OREAD4(sc, OHCI_INTERRUPT_STATUS),
OREAD4(sc, OHCI_INTERRUPT_ENABLE),
OREAD4(sc, OHCI_INTERRUPT_DISABLE)));
DPRINTF((" hcca=0x%08x percur=0x%08x ctrlhd=0x%08x\n",
OREAD4(sc, OHCI_HCCA),
OREAD4(sc, OHCI_PERIOD_CURRENT_ED),
OREAD4(sc, OHCI_CONTROL_HEAD_ED)));
DPRINTF((" ctrlcur=0x%08x bulkhd=0x%08x bulkcur=0x%08x\n",
OREAD4(sc, OHCI_CONTROL_CURRENT_ED),
OREAD4(sc, OHCI_BULK_HEAD_ED),
OREAD4(sc, OHCI_BULK_CURRENT_ED)));
DPRINTF((" done=0x%08x fmival=0x%08x fmrem=0x%08x\n",
OREAD4(sc, OHCI_DONE_HEAD),
OREAD4(sc, OHCI_FM_INTERVAL),
OREAD4(sc, OHCI_FM_REMAINING)));
DPRINTF((" fmnum=0x%08x perst=0x%08x lsthrs=0x%08x\n",
OREAD4(sc, OHCI_FM_NUMBER),
OREAD4(sc, OHCI_PERIODIC_START),
OREAD4(sc, OHCI_LS_THRESHOLD)));
DPRINTF((" desca=0x%08x descb=0x%08x stat=0x%08x\n",
OREAD4(sc, OHCI_RH_DESCRIPTOR_A),
OREAD4(sc, OHCI_RH_DESCRIPTOR_B),
OREAD4(sc, OHCI_RH_STATUS)));
DPRINTF((" port1=0x%08x port2=0x%08x\n",
OREAD4(sc, OHCI_RH_PORT_STATUS(1)),
OREAD4(sc, OHCI_RH_PORT_STATUS(2))));
DPRINTF((" HCCA: frame_number=0x%04x done_head=0x%08x\n",
le32toh(sc->sc_hcca->hcca_frame_number),
le32toh(sc->sc_hcca->hcca_done_head)));
}
#endif
Static int ohci_intr1(ohci_softc_t *);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
int
ohci_intr(void *p)
{
ohci_softc_t *sc = p;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sc == NULL || sc->sc_dying)
return (0);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* If we get an interrupt while polling, then just ignore it. */
2000-01-10 22:33:43 +00:00
if (sc->sc_bus.use_polling) {
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
#ifdef DIAGNOSTIC
printf("ohci_intr: ignored interrupt while polling\n");
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
return (0);
2000-01-10 22:33:43 +00:00
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
return (ohci_intr1(sc));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
}
Static int
ohci_intr1(ohci_softc_t *sc)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
{
1999-01-22 00:51:12 +00:00
u_int32_t intrs, eintrs;
ohci_physaddr_t done;
DPRINTFN(14,("ohci_intr1: enter\n"));
/* In case the interrupt occurs before initialization has completed. */
if (sc == NULL || sc->sc_hcca == NULL) {
#ifdef DIAGNOSTIC
printf("ohci_intr: sc->sc_hcca == NULL\n");
#endif
return (0);
}
intrs = 0;
done = le32toh(sc->sc_hcca->hcca_done_head);
2000-01-26 08:46:41 +00:00
/* The LSb of done is used to inform the HC Driver that an interrupt
* condition exists for both the Done list and for another event
* recorded in HcInterruptStatus. On an interrupt from the HC, the HC
* Driver checks the HccaDoneHead Value. If this value is 0, then the
* interrupt was caused by other than the HccaDoneHead update and the
* HcInterruptStatus register needs to be accessed to determine that
* exact interrupt cause. If HccaDoneHead is nonzero, then a Done list
* update interrupt is indicated and if the LSb of done is nonzero,
* then an additional interrupt event is indicated and
* HcInterruptStatus should be checked to determine its cause.
*/
if (done != 0) {
if (done & ~OHCI_DONE_INTRS)
intrs = OHCI_WDH;
2000-01-26 08:46:41 +00:00
if (done & OHCI_DONE_INTRS) {
intrs |= OREAD4(sc, OHCI_INTERRUPT_STATUS);
2000-01-26 08:46:41 +00:00
done &= ~OHCI_DONE_INTRS;
}
sc->sc_hcca->hcca_done_head = 0;
} else
intrs = OREAD4(sc, OHCI_INTERRUPT_STATUS) & ~OHCI_WDH;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (intrs == 0) /* nothing to be done (PCI shared interrupt) */
return (0);
2000-01-26 08:46:41 +00:00
intrs &= ~OHCI_MIE;
OWRITE4(sc, OHCI_INTERRUPT_STATUS, intrs); /* Acknowledge */
eintrs = intrs & sc->sc_eintrs;
if (!eintrs)
return (0);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
sc->sc_bus.intr_context++;
sc->sc_bus.no_intrs++;
DPRINTFN(7, ("ohci_intr: sc=%p intrs=0x%x(0x%x) eintrs=0x%x\n",
sc, (u_int)intrs, OREAD4(sc, OHCI_INTERRUPT_STATUS),
(u_int)eintrs));
if (eintrs & OHCI_SO) {
sc->sc_overrun_cnt++;
if (usbd_ratecheck(&sc->sc_overrun_ntc)) {
printf("%s: %u scheduling overruns\n",
USBDEVNAME(sc->sc_bus.bdev), sc->sc_overrun_cnt);
sc->sc_overrun_cnt = 0;
}
/* XXX do what */
eintrs &= ~OHCI_SO;
}
if (eintrs & OHCI_WDH) {
ohci_add_done(sc, done &~ OHCI_DONE_INTRS);
usb_schedsoftintr(&sc->sc_bus);
eintrs &= ~OHCI_WDH;
}
if (eintrs & OHCI_RD) {
printf("%s: resume detect\n", USBDEVNAME(sc->sc_bus.bdev));
/* XXX process resume detect */
}
if (eintrs & OHCI_UE) {
printf("%s: unrecoverable error, controller halted\n",
USBDEVNAME(sc->sc_bus.bdev));
OWRITE4(sc, OHCI_CONTROL, OHCI_HCFS_RESET);
/* XXX what else */
}
if (eintrs & OHCI_RHSC) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_rhsc(sc, sc->sc_intrxfer);
/*
* Disable RHSC interrupt for now, because it will be
* on until the port has been reset.
*/
ohci_rhsc_able(sc, 0);
/* Do not allow RHSC interrupts > 1 per second */
usb_callout(sc->sc_tmo_rhsc, hz, ohci_rhsc_enable, sc);
eintrs &= ~OHCI_RHSC;
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
sc->sc_bus.intr_context--;
if (eintrs != 0) {
/* Block unprocessed interrupts. XXX */
OWRITE4(sc, OHCI_INTERRUPT_DISABLE, eintrs);
sc->sc_eintrs &= ~eintrs;
printf("%s: blocking intrs 0x%x\n",
USBDEVNAME(sc->sc_bus.bdev), eintrs);
}
return (1);
}
void
ohci_rhsc_able(ohci_softc_t *sc, int on)
{
DPRINTFN(4, ("ohci_rhsc_able: on=%d\n", on));
if (on) {
sc->sc_eintrs |= OHCI_RHSC;
OWRITE4(sc, OHCI_INTERRUPT_ENABLE, OHCI_RHSC);
} else {
sc->sc_eintrs &= ~OHCI_RHSC;
OWRITE4(sc, OHCI_INTERRUPT_DISABLE, OHCI_RHSC);
}
}
void
ohci_rhsc_enable(void *v_sc)
{
ohci_softc_t *sc = v_sc;
int s;
s = splhardusb();
ohci_rhsc_able(sc, 1);
splx(s);
}
#ifdef USB_DEBUG
char *ohci_cc_strs[] = {
"NO_ERROR",
"CRC",
"BIT_STUFFING",
"DATA_TOGGLE_MISMATCH",
"STALL",
"DEVICE_NOT_RESPONDING",
"PID_CHECK_FAILURE",
"UNEXPECTED_PID",
"DATA_OVERRUN",
"DATA_UNDERRUN",
"BUFFER_OVERRUN",
"BUFFER_UNDERRUN",
"reserved",
"reserved",
"NOT_ACCESSED",
"NOT_ACCESSED"
};
#endif
void
ohci_add_done(ohci_softc_t *sc, ohci_physaddr_t done)
{
ohci_soft_itd_t *sitd, *sidone, **ip;
ohci_soft_td_t *std, *sdone, **p;
/* Reverse the done list. */
for (sdone = NULL, sidone = NULL; done != 0; ) {
std = ohci_hash_find_td(sc, done);
if (std != NULL) {
std->dnext = sdone;
done = le32toh(std->td.td_nexttd);
sdone = std;
DPRINTFN(10,("add TD %p\n", std));
continue;
}
sitd = ohci_hash_find_itd(sc, done);
if (sitd != NULL) {
sitd->dnext = sidone;
done = le32toh(sitd->itd.itd_nextitd);
sidone = sitd;
DPRINTFN(5,("add ITD %p\n", sitd));
continue;
}
panic("ohci_add_done: addr 0x%08lx not found", (u_long)done);
}
/* sdone & sidone now hold the done lists. */
/* Put them on the already processed lists. */
for (p = &sc->sc_sdone; *p != NULL; p = &(*p)->dnext)
;
*p = sdone;
for (ip = &sc->sc_sidone; *ip != NULL; ip = &(*ip)->dnext)
;
*ip = sidone;
}
void
ohci_softintr(void *v)
{
ohci_softc_t *sc = v;
ohci_soft_itd_t *sitd, *sidone, *sitdnext;
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
ohci_soft_td_t *std, *sdone, *stdnext, *p, *n;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_xfer_handle xfer;
struct ohci_pipe *opipe;
int len, cc, s;
int i, j, iframes;
DPRINTFN(10,("ohci_softintr: enter\n"));
sc->sc_bus.intr_context++;
s = splhardusb();
sdone = sc->sc_sdone;
sc->sc_sdone = NULL;
sidone = sc->sc_sidone;
sc->sc_sidone = NULL;
splx(s);
DPRINTFN(10,("ohci_softintr: sdone=%p sidone=%p\n", sdone, sidone));
#ifdef USB_DEBUG
if (ohcidebug > 10) {
DPRINTF(("ohci_process_done: TD done:\n"));
ohci_dump_tds(sdone);
}
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
for (std = sdone; std; std = stdnext) {
xfer = std->xfer;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
stdnext = std->dnext;
DPRINTFN(10, ("ohci_process_done: std=%p xfer=%p hcpriv=%p\n",
std, xfer, (xfer ? xfer->hcpriv : NULL)));
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
if (xfer == NULL) {
/*
* xfer == NULL: There seems to be no xfer associated
* with this TD. It is tailp that happened to end up on
* the done queue.
*/
continue;
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->status == USBD_CANCELLED ||
xfer->status == USBD_TIMEOUT) {
DPRINTF(("ohci_process_done: cancel/timeout %p\n",
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer));
/* Handled by abort routine. */
continue;
}
len = std->len;
if (std->td.td_cbp != 0)
len -= le32toh(std->td.td_be) -
le32toh(std->td.td_cbp) + 1;
DPRINTFN(10, ("ohci_process_done: len=%d, flags=0x%x\n", len,
std->flags));
if (std->flags & OHCI_ADD_LEN)
xfer->actlen += len;
cc = OHCI_TD_GET_CC(le32toh(std->td.td_flags));
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
if (cc != OHCI_CC_NO_ERROR) {
/*
* Endpoint is halted. First unlink all the TDs
* belonging to the failed transfer, and then restart
* the endpoint.
*/
opipe = (struct ohci_pipe *)xfer->pipe;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTFN(15,("ohci_process_done: error cc=%d (%s)\n",
OHCI_TD_GET_CC(le32toh(std->td.td_flags)),
ohci_cc_strs[OHCI_TD_GET_CC(le32toh(std->td.td_flags))]));
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
usb_uncallout(xfer->timeout_handle, ohci_timeout, xfer);
usb_rem_task(OXFER(xfer)->xfer.pipe->device,
&OXFER(xfer)->abort_task);
/* Remove all this xfer's TDs from the done queue. */
for (p = std; p->dnext != NULL; p = p->dnext) {
if (p->dnext->xfer != xfer)
continue;
p->dnext = p->dnext->dnext;
}
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
/* The next TD may have been removed. */
stdnext = std->dnext;
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
/* Remove all TDs belonging to this xfer. */
for (p = xfer->hcpriv; p->xfer == xfer; p = n) {
n = p->nexttd;
ohci_free_std(sc, p);
}
/* clear halt */
opipe->sed->ed.ed_headp = htole32(p->physaddr);
OWRITE4(sc, OHCI_COMMAND_STATUS, OHCI_CLF);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (cc == OHCI_CC_STALL)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->status = USBD_STALLED;
else
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->status = USBD_IOERROR;
s = splusb();
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_transfer_complete(xfer);
splx(s);
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
continue;
}
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
/*
* Skip intermediate TDs. They remain linked from
* xfer->hcpriv and we free them when the transfer completes.
*/
if ((std->flags & OHCI_CALL_DONE) == 0)
continue;
/* Normal transfer completion */
usb_uncallout(xfer->timeout_handle, ohci_timeout, xfer);
usb_rem_task(OXFER(xfer)->xfer.pipe->device,
&OXFER(xfer)->abort_task);
for (p = xfer->hcpriv; p->xfer == xfer; p = n) {
n = p->nexttd;
ohci_free_std(sc, p);
}
xfer->status = USBD_NORMAL_COMPLETION;
s = splusb();
usb_transfer_complete(xfer);
splx(s);
}
#ifdef USB_DEBUG
if (ohcidebug > 10) {
DPRINTF(("ohci_softintr: ITD done:\n"));
ohci_dump_itds(sidone);
}
#endif
for (sitd = sidone; sitd != NULL; sitd = sitdnext) {
xfer = sitd->xfer;
sitdnext = sitd->dnext;
sitd->flags |= OHCI_ITD_INTFIN;
DPRINTFN(1, ("ohci_process_done: sitd=%p xfer=%p hcpriv=%p\n",
sitd, xfer, xfer ? xfer->hcpriv : 0));
if (xfer == NULL)
continue;
if (xfer->status == USBD_CANCELLED ||
xfer->status == USBD_TIMEOUT) {
DPRINTF(("ohci_process_done: cancel/timeout %p\n",
xfer));
/* Handled by abort routine. */
continue;
}
if (xfer->pipe)
if (xfer->pipe->aborting)
continue; /*Ignore.*/
#ifdef DIAGNOSTIC
if (sitd->isdone)
printf("ohci_softintr: sitd=%p is done\n", sitd);
sitd->isdone = 1;
#endif
opipe = (struct ohci_pipe *)xfer->pipe;
if (opipe->aborting)
continue;
if (sitd->flags & OHCI_CALL_DONE) {
ohci_soft_itd_t *next;
opipe->u.iso.inuse -= xfer->nframes;
xfer->status = USBD_NORMAL_COMPLETION;
for (i = 0, sitd = xfer->hcpriv;;sitd = next) {
next = sitd->nextitd;
if (OHCI_ITD_GET_CC(sitd->itd.itd_flags) != OHCI_CC_NO_ERROR)
xfer->status = USBD_IOERROR;
if (xfer->status == USBD_NORMAL_COMPLETION) {
iframes = OHCI_ITD_GET_FC(sitd->itd.itd_flags);
for (j = 0; j < iframes; i++, j++) {
len = le16toh(sitd->itd.itd_offset[j]);
len =
(OHCI_ITD_PSW_GET_CC(len) ==
OHCI_CC_NOT_ACCESSED) ? 0 :
OHCI_ITD_PSW_LENGTH(len);
xfer->frlengths[i] = len;
}
}
if (sitd->flags & OHCI_CALL_DONE)
break;
}
s = splusb();
usb_transfer_complete(xfer);
splx(s);
}
}
#ifdef USB_USE_SOFTINTR
if (sc->sc_softwake) {
sc->sc_softwake = 0;
wakeup(&sc->sc_softwake);
}
#endif /* USB_USE_SOFTINTR */
sc->sc_bus.intr_context--;
DPRINTFN(10,("ohci_softintr: done:\n"));
}
void
ohci_device_ctrl_done(usbd_xfer_handle xfer)
{
DPRINTFN(10,("ohci_device_ctrl_done: xfer=%p\n", xfer));
#ifdef DIAGNOSTIC
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (!(xfer->rqflags & URQ_REQUEST)) {
panic("ohci_device_ctrl_done: not a request");
}
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->hcpriv = NULL;
}
void
ohci_device_intr_done(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
ohci_softc_t *sc = (ohci_softc_t *)opipe->pipe.device->bus;
ohci_soft_ed_t *sed = opipe->sed;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_soft_td_t *data, *tail;
DPRINTFN(10,("ohci_device_intr_done: xfer=%p, actlen=%d\n",
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer, xfer->actlen));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->hcpriv = NULL;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->pipe->repeat) {
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
data = opipe->tail.td;
tail = ohci_alloc_std(sc); /* XXX should reuse TD */
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (tail == NULL) {
xfer->status = USBD_NOMEM;
return;
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
tail->xfer = NULL;
data->td.td_flags = htole32(
OHCI_TD_IN | OHCI_TD_NOCC |
OHCI_TD_SET_DI(1) | OHCI_TD_TOGGLE_CARRY);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->flags & USBD_SHORT_XFER_OK)
data->td.td_flags |= htole32(OHCI_TD_R);
data->td.td_cbp = htole32(DMAADDR(&xfer->dmabuf, 0));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
data->nexttd = tail;
data->td.td_nexttd = htole32(tail->physaddr);
data->td.td_be = htole32(le32toh(data->td.td_cbp) +
xfer->length - 1);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
data->len = xfer->length;
data->xfer = xfer;
data->flags = OHCI_CALL_DONE | OHCI_ADD_LEN;
xfer->hcpriv = data;
xfer->actlen = 0;
sed->ed.ed_tailp = htole32(tail->physaddr);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
opipe->tail.td = tail;
}
}
void
ohci_device_bulk_done(usbd_xfer_handle xfer)
{
DPRINTFN(10,("ohci_device_bulk_done: xfer=%p, actlen=%d\n",
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer, xfer->actlen));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->hcpriv = NULL;
}
void
ohci_rhsc(ohci_softc_t *sc, usbd_xfer_handle xfer)
{
usbd_pipe_handle pipe;
u_char *p;
int i, m;
int hstatus;
hstatus = OREAD4(sc, OHCI_RH_STATUS);
DPRINTF(("ohci_rhsc: sc=%p xfer=%p hstatus=0x%08x\n",
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
sc, xfer, hstatus));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer == NULL) {
/* Just ignore the change. */
return;
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
pipe = xfer->pipe;
p = KERNADDR(&xfer->dmabuf, 0);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
m = min(sc->sc_noport, xfer->length * 8 - 1);
memset(p, 0, xfer->length);
for (i = 1; i <= m; i++) {
/* Pick out CHANGE bits from the status reg. */
if (OREAD4(sc, OHCI_RH_PORT_STATUS(i)) >> 16)
p[i/8] |= 1 << (i%8);
}
DPRINTF(("ohci_rhsc: change=0x%02x\n", *p));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->actlen = xfer->length;
xfer->status = USBD_NORMAL_COMPLETION;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_transfer_complete(xfer);
}
void
ohci_root_intr_done(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->hcpriv = NULL;
}
void
ohci_root_ctrl_done(usbd_xfer_handle xfer)
{
xfer->hcpriv = NULL;
}
/*
* Wait here until controller claims to have an interrupt.
* Then call ohci_intr and return. Use timeout to avoid waiting
* too long.
*/
void
ohci_waitintr(ohci_softc_t *sc, usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
int timo = xfer->timeout;
int usecs;
u_int32_t intrs;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->status = USBD_IN_PROGRESS;
for (usecs = timo * 1000000 / hz; usecs > 0; usecs -= 1000) {
usb_delay_ms(&sc->sc_bus, 1);
if (sc->sc_dying)
break;
intrs = OREAD4(sc, OHCI_INTERRUPT_STATUS) & sc->sc_eintrs;
DPRINTFN(15,("ohci_waitintr: 0x%04x\n", intrs));
#ifdef USB_DEBUG
if (ohcidebug > 15)
ohci_dumpregs(sc);
#endif
if (intrs) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_intr1(sc);
if (xfer->status != USBD_IN_PROGRESS)
return;
}
}
/* Timeout */
DPRINTF(("ohci_waitintr: timeout\n"));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->status = USBD_TIMEOUT;
usb_transfer_complete(xfer);
/* XXX should free TD */
}
void
ohci_poll(struct usbd_bus *bus)
{
ohci_softc_t *sc = (ohci_softc_t *)bus;
#ifdef USB_DEBUG
static int last;
int new;
new = OREAD4(sc, OHCI_INTERRUPT_STATUS);
if (new != last) {
DPRINTFN(10,("ohci_poll: intrs=0x%04x\n", new));
last = new;
}
#endif
if (OREAD4(sc, OHCI_INTERRUPT_STATUS) & sc->sc_eintrs)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_intr1(sc);
}
usbd_status
ohci_device_request(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
usb_device_request_t *req = &xfer->request;
usbd_device_handle dev = opipe->pipe.device;
ohci_softc_t *sc = (ohci_softc_t *)dev->bus;
int addr = dev->address;
ohci_soft_td_t *setup, *stat, *next, *tail;
ohci_soft_ed_t *sed;
int isread;
int len;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
int s;
isread = req->bmRequestType & UT_READ;
len = UGETW(req->wLength);
DPRINTFN(3,("ohci_device_control type=0x%02x, request=0x%02x, "
"wValue=0x%04x, wIndex=0x%04x len=%d, addr=%d, endpt=%d\n",
req->bmRequestType, req->bRequest, UGETW(req->wValue),
UGETW(req->wIndex), len, addr,
opipe->pipe.endpoint->edesc->bEndpointAddress));
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
setup = opipe->tail.td;
stat = ohci_alloc_std(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (stat == NULL) {
err = USBD_NOMEM;
goto bad1;
}
tail = ohci_alloc_std(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (tail == NULL) {
err = USBD_NOMEM;
goto bad2;
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
tail->xfer = NULL;
sed = opipe->sed;
opipe->u.ctl.length = len;
/* Update device address and length since they may have changed
during the setup of the control pipe in usbd_new_device(). */
/* XXX This only needs to be done once, but it's too early in open. */
/* XXXX Should not touch ED here! */
sed->ed.ed_flags = htole32(
(le32toh(sed->ed.ed_flags) & ~(OHCI_ED_ADDRMASK | OHCI_ED_MAXPMASK)) |
OHCI_ED_SET_FA(addr) |
OHCI_ED_SET_MAXP(UGETW(opipe->pipe.endpoint->edesc->wMaxPacketSize)));
next = stat;
/* Set up data transaction */
if (len != 0) {
ohci_soft_td_t *std = stat;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = ohci_alloc_std_chain(opipe, sc, len, isread, xfer,
std, &stat);
stat = stat->nexttd; /* point at free TD */
if (err)
goto bad3;
/* Start toggle at 1 and then use the carried toggle. */
std->td.td_flags &= htole32(~OHCI_TD_TOGGLE_MASK);
std->td.td_flags |= htole32(OHCI_TD_TOGGLE_1);
}
memcpy(KERNADDR(&opipe->u.ctl.reqdma, 0), req, sizeof *req);
setup->td.td_flags = htole32(OHCI_TD_SETUP | OHCI_TD_NOCC |
The ohci driver's processing of completed transfer descriptors (TDs) appeared to rely on all kinds of non-guaranteed behaviours: the transfer abort code assumed that TDs with no interrupt timeout configured would end up on the done queue within 20ms, the done queue processing assumed that all TDs from a transfer would appear at the same time, and there were access-after-free bugs triggered on failed transfers. Attempt to fix these problems by the following changes: - Use a maximum (6-frame) interrupt delay instead of no interrupt delay to ensure that the 20ms wait in ohci_abort_xfer() is enough for the TDs to have been taken off the hardware done queue. - Defer cancellation of timeouts and freeing of TDs until we either hit an error or reach the final TD. - Remove TDs from the done queue before freeing them so that it is safe to continue traversing the done queue. This appears to fix a hang that was reproducable with revision 1.67 or 1.68 of ulpt.c (earlier revisions had a different transfer pattern). With certain HP printers, the command "true > /dev/ulpt0" would cause ohci_add_done() to spin because the done queue had a loop. The list corruption was caused by a 3-TD transfer where the first TD completed but remained on the internal host controller done queue because it had no interrupt timeout. When the transfer timed out, the TD got freed and reused, so it caused a loop in the done queue when it was inserted a second time from a different transfer. Reported by: Alex Pivovarov MFC after: 1 week
2005-11-27 09:05:37 +00:00
OHCI_TD_TOGGLE_0 | OHCI_TD_SET_DI(6));
setup->td.td_cbp = htole32(DMAADDR(&opipe->u.ctl.reqdma, 0));
setup->nexttd = next;
setup->td.td_nexttd = htole32(next->physaddr);
setup->td.td_be = htole32(le32toh(setup->td.td_cbp) + sizeof *req - 1);
setup->len = 0;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
setup->xfer = xfer;
setup->flags = 0;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->hcpriv = setup;
stat->td.td_flags = htole32(
(isread ? OHCI_TD_OUT : OHCI_TD_IN) |
OHCI_TD_NOCC | OHCI_TD_TOGGLE_1 | OHCI_TD_SET_DI(1));
stat->td.td_cbp = 0;
stat->nexttd = tail;
stat->td.td_nexttd = htole32(tail->physaddr);
stat->td.td_be = 0;
stat->flags = OHCI_CALL_DONE;
stat->len = 0;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
stat->xfer = xfer;
#ifdef USB_DEBUG
if (ohcidebug > 5) {
DPRINTF(("ohci_device_request:\n"));
ohci_dump_ed(sed);
ohci_dump_tds(setup);
}
#endif
/* Insert ED in schedule */
s = splusb();
sed->ed.ed_tailp = htole32(tail->physaddr);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
opipe->tail.td = tail;
OWRITE4(sc, OHCI_COMMAND_STATUS, OHCI_CLF);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->timeout && !sc->sc_bus.use_polling) {
Huge merge from NetBSD: usbdi.c (1.61): =================================================================== revision 1.61 date: 2000/01/31 20:13:07; author: augustss; lines: +20 -4 Change the way the HC done method is invoked a little. =================================================================== usbdi.c (1.65): =================================================================== revision 1.65 date: 2000/03/08 15:34:10; author: augustss; lines: +4 -2 Get the status right when a polled transfer times out. =================================================================== ohci.c (1.79), uhci.c (1.89), uhcivar.h (1.24), usb_port.h (1.22), usbdivar.h (1.48): =================================================================== date: 2000/03/23 07:01:46; author: thorpej; New callout mechanism with two major improvements over the old timeout()/untimeout() API: - Clients supply callout handle storage, thus eliminating problems of resource allocation. - Insertion and removal of callouts is constant time, important as this facility is used quite a lot in the kernel. The old timeout()/untimeout() API has been removed from the kernel. =================================================================== uhci.c (1.80), usbdi.c (1.66): =================================================================== date: 2000/03/23 18:59:10; author: thorpej; Shake out some bugs from the callout changes. =================================================================== ohci.c (1.80), uhci.c (1.91), uhcivar.h (1.25), usb_port.h (1.23), usbdi.c (1.67), usbdivar.h (1.49): =================================================================== date: 2000/03/24 22:03:30; author: augustss; Some cleanup and renaming of the callouts used in USB drivers. =================================================================== uhci.c (1.92), uhcivar.h (1.26): =================================================================== date: 2000/03/24 22:57:58; author: augustss; Two major changes: Make each xfer have its own intr_info. This is necessary if we want to queue multiple xfers on an endpoint. This should get rid of the (mostly harmless) DIAGNOSTICs about intr_infos (not) being done. Change (again!) how xfers are aborted. Aborting a TD is a nightmare on the braindead UHCI controller. (Unless you stop the HC, thereby losing isoc traffic.) Hopefully I got it right this time. =================================================================== usbdivar.h (1.50): =================================================================== revision 1.50 date: 2000/03/25 00:10:19; author: augustss; lines: +4 -2 GC an unsued field and add some DIAGNOSTIC in xfer. =================================================================== ums.c: Use the callout functions instead of the timeout ones. uhci.c (1.93): =================================================================== revision 1.93 date: 2000/03/25 00:11:21; author: augustss; lines: +26 -1 Add more DIAGNOSTIC when aborting isoc. =================================================================== uhci.c (1.94), usbdivar.h (1.51): =================================================================== date: 2000/03/25 07:13:05; author: augustss; More DIAGNOSTIC. Initialize a callout handle I forgot. =================================================================== uhci.c (1.95): =================================================================== revision 1.95 date: 2000/03/25 07:23:12; author: augustss; Exp; lines: +24 -7 Improve uhci_dump_ii(). =================================================================== ohci.c (1.81), uhci.c (1.96), uhcivar.h (1.27), usb_subr.c (1.68), usbdi.c (1.68), usbdivar.h (1.52): =================================================================== date: 2000/03/25 18:02:33; author: augustss; Rename and move around callout handles to make it more sane. Add some DIAGNOSTIC. Fix buglet in isoc abort on UHCI. =================================================================== uhci.c (1.98): =================================================================== revision 1.98 date: 2000/03/27 07:39:48; author: augustss; lines: +12 -4 Make it compile without DIAGNOSTIC. =================================================================== uhci.c (1.99): =================================================================== revision 1.99 date: 2000/03/27 08:01:09; author: augustss; lines: +1 -5 Remove some debug nonsense. =================================================================== uhci.c (1.100): =================================================================== revision 1.100 date: 2000/03/27 09:41:36; author: augustss; lines: +13 -3 Don't mess with QH in bulk abort for the moment. =================================================================== uhci.c (1.102): =================================================================== revision 1.102 date: 2000/03/27 22:42:57; author: augustss; lines: +66 -26 Be a little more careful when aborting. Preallocate some TDs for large buffers. =================================================================== uhci.c (1.103): =================================================================== date: 2000/03/28 09:47:10; author: augustss; lines: +11 -1 Another patch for xfer abort... XXX The current xfer queueing and aborting semantics should really XXX be changed. It cannot be implemented in a sane way on UHCI. XXX One day when I have lots of time I'll redesign it... =================================================================== uhci.c (1.104): Correct a debug message. uhci.c (1.105): Be more defensive in a DIAGNOSTIC test. uhci.c (1.106): =================================================================== revision 1.106 date: 2000/03/29 01:49:13; author: augustss; lines: +14 -309 *SIGH* Revert back to the old method of aborting xfers. I had tested the new stuff for two months now, but as soon as I commited it the problems started to appear. Murphy, no doubt... =================================================================== usb_subr.c (1.70), usbdi.c (1.71), usbdivar.h (1.53): =================================================================== revision 1.70 date: 2000/03/29 01:45:20; author: augustss; lines: +2 -1 Do not accept new xfers for queuing while a pipe is aborting. ===================================================================
2002-03-16 12:06:01 +00:00
usb_callout(xfer->timeout_handle, MS_TO_TICKS(xfer->timeout),
ohci_timeout, xfer);
}
splx(s);
#ifdef USB_DEBUG
if (ohcidebug > 20) {
delay(10000);
DPRINTF(("ohci_device_request: status=%x\n",
OREAD4(sc, OHCI_COMMAND_STATUS)));
ohci_dumpregs(sc);
printf("ctrl head:\n");
ohci_dump_ed(sc->sc_ctrl_head);
printf("sed:\n");
ohci_dump_ed(sed);
ohci_dump_tds(setup);
}
#endif
return (USBD_NORMAL_COMPLETION);
bad3:
ohci_free_std(sc, tail);
bad2:
ohci_free_std(sc, stat);
bad1:
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
return (err);
}
/*
* Add an ED to the schedule. Called at splusb().
*/
void
ohci_add_ed(ohci_soft_ed_t *sed, ohci_soft_ed_t *head)
{
DPRINTFN(8,("ohci_add_ed: sed=%p head=%p\n", sed, head));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
SPLUSBCHECK;
sed->next = head->next;
sed->ed.ed_nexted = head->ed.ed_nexted;
head->next = sed;
head->ed.ed_nexted = htole32(sed->physaddr);
}
/*
* Remove an ED from the schedule. Called at splusb().
*/
void
ohci_rem_ed(ohci_soft_ed_t *sed, ohci_soft_ed_t *head)
{
ohci_soft_ed_t *p;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
SPLUSBCHECK;
/* XXX */
for (p = head; p != NULL && p->next != sed; p = p->next)
;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (p == NULL)
panic("ohci_rem_ed: ED not found");
p->next = sed->next;
p->ed.ed_nexted = sed->ed.ed_nexted;
}
/*
* When a transfer is completed the TD is added to the done queue by
* the host controller. This queue is the processed by software.
* Unfortunately the queue contains the physical address of the TD
* and we have no simple way to translate this back to a kernel address.
* To make the translation possible (and fast) we use a hash table of
* TDs currently in the schedule. The physical address is used as the
* hash value.
*/
#define HASH(a) (((a) >> 4) % OHCI_HASH_SIZE)
/* Called at splusb() */
void
ohci_hash_add_td(ohci_softc_t *sc, ohci_soft_td_t *std)
{
int h = HASH(std->physaddr);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
SPLUSBCHECK;
LIST_INSERT_HEAD(&sc->sc_hash_tds[h], std, hnext);
}
/* Called at splusb() */
void
ohci_hash_rem_td(ohci_softc_t *sc, ohci_soft_td_t *std)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
SPLUSBCHECK;
LIST_REMOVE(std, hnext);
}
ohci_soft_td_t *
ohci_hash_find_td(ohci_softc_t *sc, ohci_physaddr_t a)
{
int h = HASH(a);
ohci_soft_td_t *std;
2000-01-29 14:41:48 +00:00
/* if these are present they should be masked out at an earlier
* stage.
*/
KASSERT((a&~OHCI_HEADMASK) == 0, ("%s: 0x%b has lower bits set\n",
2000-01-29 11:50:44 +00:00
USBDEVNAME(sc->sc_bus.bdev),
(int) a, "\20\1HALT\2TOGGLE"));
for (std = LIST_FIRST(&sc->sc_hash_tds[h]);
std != NULL;
std = LIST_NEXT(std, hnext))
if (std->physaddr == a)
return (std);
DPRINTF(("%s: ohci_hash_find_td: addr 0x%08lx not found\n",
USBDEVNAME(sc->sc_bus.bdev), (u_long) a));
return (NULL);
}
/* Called at splusb() */
void
ohci_hash_add_itd(ohci_softc_t *sc, ohci_soft_itd_t *sitd)
{
int h = HASH(sitd->physaddr);
SPLUSBCHECK;
DPRINTFN(10,("ohci_hash_add_itd: sitd=%p physaddr=0x%08lx\n",
sitd, (u_long)sitd->physaddr));
LIST_INSERT_HEAD(&sc->sc_hash_itds[h], sitd, hnext);
}
/* Called at splusb() */
void
ohci_hash_rem_itd(ohci_softc_t *sc, ohci_soft_itd_t *sitd)
{
SPLUSBCHECK;
DPRINTFN(10,("ohci_hash_rem_itd: sitd=%p physaddr=0x%08lx\n",
sitd, (u_long)sitd->physaddr));
LIST_REMOVE(sitd, hnext);
}
ohci_soft_itd_t *
ohci_hash_find_itd(ohci_softc_t *sc, ohci_physaddr_t a)
{
int h = HASH(a);
ohci_soft_itd_t *sitd;
for (sitd = LIST_FIRST(&sc->sc_hash_itds[h]);
sitd != NULL;
sitd = LIST_NEXT(sitd, hnext))
if (sitd->physaddr == a)
return (sitd);
return (NULL);
}
void
ohci_timeout(void *addr)
{
struct ohci_xfer *oxfer = addr;
struct ohci_pipe *opipe = (struct ohci_pipe *)oxfer->xfer.pipe;
ohci_softc_t *sc = (ohci_softc_t *)opipe->pipe.device->bus;
DPRINTF(("ohci_timeout: oxfer=%p\n", oxfer));
if (sc->sc_dying) {
ohci_abort_xfer(&oxfer->xfer, USBD_TIMEOUT);
return;
}
/* Execute the abort in a process context. */
usb_add_task(oxfer->xfer.pipe->device, &oxfer->abort_task);
}
void
ohci_timeout_task(void *addr)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_xfer_handle xfer = addr;
int s;
DPRINTF(("ohci_timeout_task: xfer=%p\n", xfer));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
s = splusb();
ohci_abort_xfer(xfer, USBD_TIMEOUT);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
splx(s);
}
#ifdef USB_DEBUG
void
ohci_dump_tds(ohci_soft_td_t *std)
{
for (; std; std = std->nexttd)
ohci_dump_td(std);
}
void
ohci_dump_td(ohci_soft_td_t *std)
{
char sbuf[128];
bitmask_snprintf((u_int32_t)le32toh(std->td.td_flags),
"\20\23R\24OUT\25IN\31TOG1\32SETTOGGLE",
sbuf, sizeof(sbuf));
printf("TD(%p) at %08lx: %s delay=%d ec=%d cc=%d\ncbp=0x%08lx "
"nexttd=0x%08lx be=0x%08lx\n",
std, (u_long)std->physaddr, sbuf,
OHCI_TD_GET_DI(le32toh(std->td.td_flags)),
OHCI_TD_GET_EC(le32toh(std->td.td_flags)),
OHCI_TD_GET_CC(le32toh(std->td.td_flags)),
(u_long)le32toh(std->td.td_cbp),
(u_long)le32toh(std->td.td_nexttd),
(u_long)le32toh(std->td.td_be));
}
void
ohci_dump_itd(ohci_soft_itd_t *sitd)
{
int i;
printf("ITD(%p) at %08lx: sf=%d di=%d fc=%d cc=%d\n"
"bp0=0x%08lx next=0x%08lx be=0x%08lx\n",
sitd, (u_long)sitd->physaddr,
OHCI_ITD_GET_SF(le32toh(sitd->itd.itd_flags)),
OHCI_ITD_GET_DI(le32toh(sitd->itd.itd_flags)),
OHCI_ITD_GET_FC(le32toh(sitd->itd.itd_flags)),
OHCI_ITD_GET_CC(le32toh(sitd->itd.itd_flags)),
(u_long)le32toh(sitd->itd.itd_bp0),
(u_long)le32toh(sitd->itd.itd_nextitd),
(u_long)le32toh(sitd->itd.itd_be));
for (i = 0; i < OHCI_ITD_NOFFSET; i++)
printf("offs[%d]=0x%04x ", i,
(u_int)le16toh(sitd->itd.itd_offset[i]));
printf("\n");
}
void
ohci_dump_itds(ohci_soft_itd_t *sitd)
{
for (; sitd; sitd = sitd->nextitd)
ohci_dump_itd(sitd);
}
void
ohci_dump_ed(ohci_soft_ed_t *sed)
{
char sbuf[128], sbuf2[128];
bitmask_snprintf((u_int32_t)le32toh(sed->ed.ed_flags),
"\20\14OUT\15IN\16LOWSPEED\17SKIP\20ISO",
sbuf, sizeof(sbuf));
bitmask_snprintf((u_int32_t)le32toh(sed->ed.ed_headp),
"\20\1HALT\2CARRY", sbuf2, sizeof(sbuf2));
printf("ED(%p) at 0x%08lx: addr=%d endpt=%d maxp=%d flags=%s\ntailp=0x%08lx "
"headflags=%s headp=0x%08lx nexted=0x%08lx\n",
sed, (u_long)sed->physaddr,
OHCI_ED_GET_FA(le32toh(sed->ed.ed_flags)),
OHCI_ED_GET_EN(le32toh(sed->ed.ed_flags)),
OHCI_ED_GET_MAXP(le32toh(sed->ed.ed_flags)), sbuf,
(u_long)le32toh(sed->ed.ed_tailp), sbuf2,
(u_long)le32toh(sed->ed.ed_headp),
(u_long)le32toh(sed->ed.ed_nexted));
}
#endif
usbd_status
ohci_open(usbd_pipe_handle pipe)
{
usbd_device_handle dev = pipe->device;
ohci_softc_t *sc = (ohci_softc_t *)dev->bus;
usb_endpoint_descriptor_t *ed = pipe->endpoint->edesc;
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
u_int8_t addr = dev->address;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
u_int8_t xfertype = ed->bmAttributes & UE_XFERTYPE;
ohci_soft_ed_t *sed;
ohci_soft_td_t *std;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ohci_soft_itd_t *sitd;
ohci_physaddr_t tdphys;
u_int32_t fmt;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
int s;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
int ival;
DPRINTFN(1, ("ohci_open: pipe=%p, addr=%d, endpt=%d (%d)\n",
pipe, addr, ed->bEndpointAddress, sc->sc_addr));
Huge merge from NetBSD: usbdi.c (1.61): =================================================================== revision 1.61 date: 2000/01/31 20:13:07; author: augustss; lines: +20 -4 Change the way the HC done method is invoked a little. =================================================================== usbdi.c (1.65): =================================================================== revision 1.65 date: 2000/03/08 15:34:10; author: augustss; lines: +4 -2 Get the status right when a polled transfer times out. =================================================================== ohci.c (1.79), uhci.c (1.89), uhcivar.h (1.24), usb_port.h (1.22), usbdivar.h (1.48): =================================================================== date: 2000/03/23 07:01:46; author: thorpej; New callout mechanism with two major improvements over the old timeout()/untimeout() API: - Clients supply callout handle storage, thus eliminating problems of resource allocation. - Insertion and removal of callouts is constant time, important as this facility is used quite a lot in the kernel. The old timeout()/untimeout() API has been removed from the kernel. =================================================================== uhci.c (1.80), usbdi.c (1.66): =================================================================== date: 2000/03/23 18:59:10; author: thorpej; Shake out some bugs from the callout changes. =================================================================== ohci.c (1.80), uhci.c (1.91), uhcivar.h (1.25), usb_port.h (1.23), usbdi.c (1.67), usbdivar.h (1.49): =================================================================== date: 2000/03/24 22:03:30; author: augustss; Some cleanup and renaming of the callouts used in USB drivers. =================================================================== uhci.c (1.92), uhcivar.h (1.26): =================================================================== date: 2000/03/24 22:57:58; author: augustss; Two major changes: Make each xfer have its own intr_info. This is necessary if we want to queue multiple xfers on an endpoint. This should get rid of the (mostly harmless) DIAGNOSTICs about intr_infos (not) being done. Change (again!) how xfers are aborted. Aborting a TD is a nightmare on the braindead UHCI controller. (Unless you stop the HC, thereby losing isoc traffic.) Hopefully I got it right this time. =================================================================== usbdivar.h (1.50): =================================================================== revision 1.50 date: 2000/03/25 00:10:19; author: augustss; lines: +4 -2 GC an unsued field and add some DIAGNOSTIC in xfer. =================================================================== ums.c: Use the callout functions instead of the timeout ones. uhci.c (1.93): =================================================================== revision 1.93 date: 2000/03/25 00:11:21; author: augustss; lines: +26 -1 Add more DIAGNOSTIC when aborting isoc. =================================================================== uhci.c (1.94), usbdivar.h (1.51): =================================================================== date: 2000/03/25 07:13:05; author: augustss; More DIAGNOSTIC. Initialize a callout handle I forgot. =================================================================== uhci.c (1.95): =================================================================== revision 1.95 date: 2000/03/25 07:23:12; author: augustss; Exp; lines: +24 -7 Improve uhci_dump_ii(). =================================================================== ohci.c (1.81), uhci.c (1.96), uhcivar.h (1.27), usb_subr.c (1.68), usbdi.c (1.68), usbdivar.h (1.52): =================================================================== date: 2000/03/25 18:02:33; author: augustss; Rename and move around callout handles to make it more sane. Add some DIAGNOSTIC. Fix buglet in isoc abort on UHCI. =================================================================== uhci.c (1.98): =================================================================== revision 1.98 date: 2000/03/27 07:39:48; author: augustss; lines: +12 -4 Make it compile without DIAGNOSTIC. =================================================================== uhci.c (1.99): =================================================================== revision 1.99 date: 2000/03/27 08:01:09; author: augustss; lines: +1 -5 Remove some debug nonsense. =================================================================== uhci.c (1.100): =================================================================== revision 1.100 date: 2000/03/27 09:41:36; author: augustss; lines: +13 -3 Don't mess with QH in bulk abort for the moment. =================================================================== uhci.c (1.102): =================================================================== revision 1.102 date: 2000/03/27 22:42:57; author: augustss; lines: +66 -26 Be a little more careful when aborting. Preallocate some TDs for large buffers. =================================================================== uhci.c (1.103): =================================================================== date: 2000/03/28 09:47:10; author: augustss; lines: +11 -1 Another patch for xfer abort... XXX The current xfer queueing and aborting semantics should really XXX be changed. It cannot be implemented in a sane way on UHCI. XXX One day when I have lots of time I'll redesign it... =================================================================== uhci.c (1.104): Correct a debug message. uhci.c (1.105): Be more defensive in a DIAGNOSTIC test. uhci.c (1.106): =================================================================== revision 1.106 date: 2000/03/29 01:49:13; author: augustss; lines: +14 -309 *SIGH* Revert back to the old method of aborting xfers. I had tested the new stuff for two months now, but as soon as I commited it the problems started to appear. Murphy, no doubt... =================================================================== usb_subr.c (1.70), usbdi.c (1.71), usbdivar.h (1.53): =================================================================== revision 1.70 date: 2000/03/29 01:45:20; author: augustss; lines: +2 -1 Do not accept new xfers for queuing while a pipe is aborting. ===================================================================
2002-03-16 12:06:01 +00:00
if (sc->sc_dying)
return (USBD_IOERROR);
std = NULL;
sed = NULL;
if (addr == sc->sc_addr) {
switch (ed->bEndpointAddress) {
case USB_CONTROL_ENDPOINT:
pipe->methods = &ohci_root_ctrl_methods;
break;
case UE_DIR_IN | OHCI_INTR_ENDPT:
pipe->methods = &ohci_root_intr_methods;
break;
default:
return (USBD_INVAL);
}
} else {
sed = ohci_alloc_sed(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (sed == NULL)
goto bad0;
opipe->sed = sed;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
if (xfertype == UE_ISOCHRONOUS) {
sitd = ohci_alloc_sitd(sc);
if (sitd == NULL)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
goto bad1;
opipe->tail.itd = sitd;
opipe->aborting = 0;
tdphys = sitd->physaddr;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
fmt = OHCI_ED_FORMAT_ISO;
if (UE_GET_DIR(ed->bEndpointAddress) == UE_DIR_IN)
fmt |= OHCI_ED_DIR_IN;
else
fmt |= OHCI_ED_DIR_OUT;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
} else {
std = ohci_alloc_std(sc);
if (std == NULL)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
goto bad1;
opipe->tail.td = std;
tdphys = std->physaddr;
fmt = OHCI_ED_FORMAT_GEN | OHCI_ED_DIR_TD;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
sed->ed.ed_flags = htole32(
OHCI_ED_SET_FA(addr) |
OHCI_ED_SET_EN(UE_GET_ADDR(ed->bEndpointAddress)) |
(dev->speed == USB_SPEED_LOW ? OHCI_ED_SPEED : 0) |
fmt |
OHCI_ED_SET_MAXP(UGETW(ed->wMaxPacketSize)));
sed->ed.ed_headp = sed->ed.ed_tailp = htole32(tdphys);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
switch (xfertype) {
case UE_CONTROL:
pipe->methods = &ohci_device_ctrl_methods;
err = usb_allocmem(&sc->sc_bus,
sizeof(usb_device_request_t),
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
0, &opipe->u.ctl.reqdma);
if (err)
goto bad;
s = splusb();
ohci_add_ed(sed, sc->sc_ctrl_head);
splx(s);
break;
case UE_INTERRUPT:
pipe->methods = &ohci_device_intr_methods;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ival = pipe->interval;
if (ival == USBD_DEFAULT_INTERVAL)
ival = ed->bInterval;
return (ohci_device_setintr(sc, opipe, ival));
case UE_ISOCHRONOUS:
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
pipe->methods = &ohci_device_isoc_methods;
return (ohci_setup_isoc(pipe));
case UE_BULK:
pipe->methods = &ohci_device_bulk_methods;
s = splusb();
ohci_add_ed(sed, sc->sc_bulk_head);
splx(s);
break;
}
}
return (USBD_NORMAL_COMPLETION);
bad:
if (std != NULL)
ohci_free_std(sc, std);
bad1:
if (sed != NULL)
ohci_free_sed(sc, sed);
bad0:
return (USBD_NOMEM);
}
/*
* Close a reqular pipe.
* Assumes that there are no pending transactions.
*/
void
ohci_close_pipe(usbd_pipe_handle pipe, ohci_soft_ed_t *head)
{
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
ohci_soft_ed_t *sed = opipe->sed;
int s;
s = splusb();
#ifdef DIAGNOSTIC
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP);
if ((le32toh(sed->ed.ed_tailp) & OHCI_HEADMASK) !=
(le32toh(sed->ed.ed_headp) & OHCI_HEADMASK)) {
ohci_soft_td_t *std;
std = ohci_hash_find_td(sc, le32toh(sed->ed.ed_headp));
printf("ohci_close_pipe: pipe not empty sed=%p hd=0x%x "
"tl=0x%x pipe=%p, std=%p\n", sed,
(int)le32toh(sed->ed.ed_headp),
(int)le32toh(sed->ed.ed_tailp),
pipe, std);
#ifdef USB_DEBUG
usbd_dump_pipe(&opipe->pipe);
#endif
#ifdef USB_DEBUG
ohci_dump_ed(sed);
if (std)
ohci_dump_td(std);
#endif
usb_delay_ms(&sc->sc_bus, 2);
if ((le32toh(sed->ed.ed_tailp) & OHCI_HEADMASK) !=
(le32toh(sed->ed.ed_headp) & OHCI_HEADMASK))
printf("ohci_close_pipe: pipe still not empty\n");
}
#endif
ohci_rem_ed(sed, head);
/* Make sure the host controller is not touching this ED */
usb_delay_ms(&sc->sc_bus, 1);
splx(s);
ohci_free_sed(sc, opipe->sed);
}
/*
* Abort a device request.
* If this routine is called at splusb() it guarantees that the request
* will be removed from the hardware scheduling and that the callback
* for it will be called with USBD_CANCELLED status.
* It's impossible to guarantee that the requested transfer will not
* have happened since the hardware runs concurrently.
* If the transaction has already happened we rely on the ordinary
* interrupt processing to process it.
*/
void
ohci_abort_xfer(usbd_xfer_handle xfer, usbd_status status)
{
struct ohci_xfer *oxfer = OXFER(xfer);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
ohci_softc_t *sc = (ohci_softc_t *)opipe->pipe.device->bus;
ohci_soft_ed_t *sed = opipe->sed;
ohci_soft_td_t *p, *n;
ohci_physaddr_t headp;
int s, hit;
DPRINTF(("ohci_abort_xfer: xfer=%p pipe=%p sed=%p\n", xfer, opipe,sed));
if (sc->sc_dying) {
/* If we're dying, just do the software part. */
s = splusb();
xfer->status = status; /* make software ignore it */
usb_uncallout(xfer->timeout_handle, ohci_timeout, xfer);
usb_rem_task(xfer->pipe->device, &OXFER(xfer)->abort_task);
usb_transfer_complete(xfer);
splx(s);
}
if (xfer->device->bus->intr_context || !curproc)
panic("ohci_abort_xfer: not in process context");
/*
* If an abort is already in progress then just wait for it to
* complete and return.
*/
if (oxfer->ohci_xfer_flags & OHCI_XFER_ABORTING) {
DPRINTFN(2, ("ohci_abort_xfer: already aborting\n"));
/* No need to wait if we're aborting from a timeout. */
if (status == USBD_TIMEOUT)
return;
/* Override the status which might be USBD_TIMEOUT. */
xfer->status = status;
DPRINTFN(2, ("ohci_abort_xfer: waiting for abort to finish\n"));
oxfer->ohci_xfer_flags |= OHCI_XFER_ABORTWAIT;
while (oxfer->ohci_xfer_flags & OHCI_XFER_ABORTING)
tsleep(&oxfer->ohci_xfer_flags, PZERO, "ohciaw", 0);
return;
}
/*
* Step 1: Make interrupt routine and hardware ignore xfer.
*/
s = splusb();
oxfer->ohci_xfer_flags |= OHCI_XFER_ABORTING;
xfer->status = status; /* make software ignore it */
usb_uncallout(xfer->timeout_handle, ohci_timeout, xfer);
usb_rem_task(xfer->pipe->device, &OXFER(xfer)->abort_task);
splx(s);
DPRINTFN(1,("ohci_abort_xfer: stop ed=%p\n", sed));
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP); /* force hardware skip */
/*
* Step 2: Wait until we know hardware has finished any possible
* use of the xfer. Also make sure the soft interrupt routine
* has run.
*/
usb_delay_ms(opipe->pipe.device->bus, 20); /* Hardware finishes in 1ms */
s = splusb();
#ifdef USB_USE_SOFTINTR
sc->sc_softwake = 1;
#endif /* USB_USE_SOFTINTR */
usb_schedsoftintr(&sc->sc_bus);
#ifdef USB_USE_SOFTINTR
tsleep(&sc->sc_softwake, PZERO, "ohciab", 0);
#endif /* USB_USE_SOFTINTR */
splx(s);
/*
* Step 3: Remove any vestiges of the xfer from the hardware.
* The complication here is that the hardware may have executed
* beyond the xfer we're trying to abort. So as we're scanning
* the TDs of this xfer we check if the hardware points to
* any of them.
*/
s = splusb(); /* XXX why? */
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
p = xfer->hcpriv;
#ifdef DIAGNOSTIC
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (p == NULL) {
oxfer->ohci_xfer_flags &= ~OHCI_XFER_ABORTING; /* XXX */
splx(s);
printf("ohci_abort_xfer: hcpriv is NULL\n");
return;
}
#endif
#ifdef USB_DEBUG
if (ohcidebug > 1) {
DPRINTF(("ohci_abort_xfer: sed=\n"));
ohci_dump_ed(sed);
ohci_dump_tds(p);
}
#endif
headp = le32toh(sed->ed.ed_headp) & OHCI_HEADMASK;
hit = 0;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
for (; p->xfer == xfer; p = n) {
hit |= headp == p->physaddr;
n = p->nexttd;
ohci_free_std(sc, p);
}
/* Zap headp register if hardware pointed inside the xfer. */
if (hit) {
DPRINTFN(1,("ohci_abort_xfer: set hd=0x08%x, tl=0x%08x\n",
(int)p->physaddr, (int)le32toh(sed->ed.ed_tailp)));
sed->ed.ed_headp = htole32(p->physaddr); /* unlink TDs */
} else {
DPRINTFN(1,("ohci_abort_xfer: no hit\n"));
}
/*
* Step 4: Turn on hardware again.
*/
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP); /* remove hardware skip */
/*
* Step 5: Execute callback.
*/
/* Do the wakeup first to avoid touching the xfer after the callback. */
oxfer->ohci_xfer_flags &= ~OHCI_XFER_ABORTING;
if (oxfer->ohci_xfer_flags & OHCI_XFER_ABORTWAIT) {
oxfer->ohci_xfer_flags &= ~OHCI_XFER_ABORTWAIT;
wakeup(&oxfer->ohci_xfer_flags);
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_transfer_complete(xfer);
splx(s);
}
/*
* Data structures and routines to emulate the root hub.
*/
Static usb_device_descriptor_t ohci_devd = {
USB_DEVICE_DESCRIPTOR_SIZE,
UDESC_DEVICE, /* type */
{0x00, 0x01}, /* USB version */
UDCLASS_HUB, /* class */
UDSUBCLASS_HUB, /* subclass */
UDPROTO_FSHUB, /* protocol */
64, /* max packet */
{0},{0},{0x00,0x01}, /* device id */
1,2,0, /* string indicies */
1 /* # of configurations */
};
Static usb_config_descriptor_t ohci_confd = {
USB_CONFIG_DESCRIPTOR_SIZE,
UDESC_CONFIG,
{USB_CONFIG_DESCRIPTOR_SIZE +
USB_INTERFACE_DESCRIPTOR_SIZE +
USB_ENDPOINT_DESCRIPTOR_SIZE},
1,
1,
0,
UC_SELF_POWERED,
0 /* max power */
};
Static usb_interface_descriptor_t ohci_ifcd = {
USB_INTERFACE_DESCRIPTOR_SIZE,
UDESC_INTERFACE,
0,
0,
1,
UICLASS_HUB,
UISUBCLASS_HUB,
UIPROTO_FSHUB,
0
};
Static usb_endpoint_descriptor_t ohci_endpd = {
USB_ENDPOINT_DESCRIPTOR_SIZE,
UDESC_ENDPOINT,
UE_DIR_IN | OHCI_INTR_ENDPT,
UE_INTERRUPT,
{8, 0}, /* max packet */
255
};
Static usb_hub_descriptor_t ohci_hubd = {
USB_HUB_DESCRIPTOR_SIZE,
UDESC_HUB,
0,
{0,0},
0,
0,
{0},
};
Static int
ohci_str(usb_string_descriptor_t *p, int l, const char *s)
{
int i;
if (l == 0)
return (0);
p->bLength = 2 * strlen(s) + 2;
if (l == 1)
return (1);
p->bDescriptorType = UDESC_STRING;
l -= 2;
for (i = 0; s[i] && l > 1; i++, l -= 2)
USETW2(p->bString[i], 0, s[i]);
return (2*i+2);
}
/*
* Simulate a hardware hub by handling all the necessary requests.
*/
Static usbd_status
ohci_root_ctrl_transfer(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
/* Insert last in queue. */
err = usb_insert_transfer(xfer);
if (err)
return (err);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* Pipe isn't running, start first */
return (ohci_root_ctrl_start(SIMPLEQ_FIRST(&xfer->pipe->queue)));
}
Static usbd_status
ohci_root_ctrl_start(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_softc_t *sc = (ohci_softc_t *)xfer->pipe->device->bus;
usb_device_request_t *req;
void *buf = NULL;
int port, i;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
int s, len, value, index, l, totlen = 0;
usb_port_status_t ps;
usb_hub_descriptor_t hubd;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
u_int32_t v;
if (sc->sc_dying)
return (USBD_IOERROR);
#ifdef DIAGNOSTIC
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (!(xfer->rqflags & URQ_REQUEST))
/* XXX panic */
return (USBD_INVAL);
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
req = &xfer->request;
DPRINTFN(4,("ohci_root_ctrl_control type=0x%02x request=%02x\n",
req->bmRequestType, req->bRequest));
len = UGETW(req->wLength);
value = UGETW(req->wValue);
index = UGETW(req->wIndex);
if (len != 0)
buf = KERNADDR(&xfer->dmabuf, 0);
#define C(x,y) ((x) | ((y) << 8))
switch(C(req->bRequest, req->bmRequestType)) {
case C(UR_CLEAR_FEATURE, UT_WRITE_DEVICE):
case C(UR_CLEAR_FEATURE, UT_WRITE_INTERFACE):
case C(UR_CLEAR_FEATURE, UT_WRITE_ENDPOINT):
/*
* DEVICE_REMOTE_WAKEUP and ENDPOINT_HALT are no-ops
* for the integrated root hub.
*/
break;
case C(UR_GET_CONFIG, UT_READ_DEVICE):
if (len > 0) {
*(u_int8_t *)buf = sc->sc_conf;
totlen = 1;
}
break;
case C(UR_GET_DESCRIPTOR, UT_READ_DEVICE):
DPRINTFN(8,("ohci_root_ctrl_control wValue=0x%04x\n", value));
switch(value >> 8) {
case UDESC_DEVICE:
if ((value & 0xff) != 0) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
totlen = l = min(len, USB_DEVICE_DESCRIPTOR_SIZE);
USETW(ohci_devd.idVendor, sc->sc_id_vendor);
memcpy(buf, &ohci_devd, l);
break;
case UDESC_CONFIG:
if ((value & 0xff) != 0) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
totlen = l = min(len, USB_CONFIG_DESCRIPTOR_SIZE);
memcpy(buf, &ohci_confd, l);
buf = (char *)buf + l;
len -= l;
l = min(len, USB_INTERFACE_DESCRIPTOR_SIZE);
totlen += l;
memcpy(buf, &ohci_ifcd, l);
buf = (char *)buf + l;
len -= l;
l = min(len, USB_ENDPOINT_DESCRIPTOR_SIZE);
totlen += l;
memcpy(buf, &ohci_endpd, l);
break;
case UDESC_STRING:
if (len == 0)
break;
*(u_int8_t *)buf = 0;
totlen = 1;
switch (value & 0xff) {
case 1: /* Vendor */
totlen = ohci_str(buf, len, sc->sc_vendor);
break;
case 2: /* Product */
totlen = ohci_str(buf, len, "OHCI root hub");
break;
}
break;
default:
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
break;
case C(UR_GET_INTERFACE, UT_READ_INTERFACE):
if (len > 0) {
*(u_int8_t *)buf = 0;
totlen = 1;
}
break;
case C(UR_GET_STATUS, UT_READ_DEVICE):
if (len > 1) {
USETW(((usb_status_t *)buf)->wStatus,UDS_SELF_POWERED);
totlen = 2;
}
break;
case C(UR_GET_STATUS, UT_READ_INTERFACE):
case C(UR_GET_STATUS, UT_READ_ENDPOINT):
if (len > 1) {
USETW(((usb_status_t *)buf)->wStatus, 0);
totlen = 2;
}
break;
case C(UR_SET_ADDRESS, UT_WRITE_DEVICE):
if (value >= USB_MAX_DEVICES) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
sc->sc_addr = value;
break;
case C(UR_SET_CONFIG, UT_WRITE_DEVICE):
if (value != 0 && value != 1) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
sc->sc_conf = value;
break;
case C(UR_SET_DESCRIPTOR, UT_WRITE_DEVICE):
break;
case C(UR_SET_FEATURE, UT_WRITE_DEVICE):
case C(UR_SET_FEATURE, UT_WRITE_INTERFACE):
case C(UR_SET_FEATURE, UT_WRITE_ENDPOINT):
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
case C(UR_SET_INTERFACE, UT_WRITE_INTERFACE):
break;
case C(UR_SYNCH_FRAME, UT_WRITE_ENDPOINT):
break;
/* Hub requests */
case C(UR_CLEAR_FEATURE, UT_WRITE_CLASS_DEVICE):
break;
case C(UR_CLEAR_FEATURE, UT_WRITE_CLASS_OTHER):
DPRINTFN(8, ("ohci_root_ctrl_control: UR_CLEAR_PORT_FEATURE "
"port=%d feature=%d\n",
index, value));
if (index < 1 || index > sc->sc_noport) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
port = OHCI_RH_PORT_STATUS(index);
switch(value) {
case UHF_PORT_ENABLE:
OWRITE4(sc, port, UPS_CURRENT_CONNECT_STATUS);
break;
case UHF_PORT_SUSPEND:
OWRITE4(sc, port, UPS_OVERCURRENT_INDICATOR);
break;
case UHF_PORT_POWER:
/* Yes, writing to the LOW_SPEED bit clears power. */
OWRITE4(sc, port, UPS_LOW_SPEED);
break;
case UHF_C_PORT_CONNECTION:
OWRITE4(sc, port, UPS_C_CONNECT_STATUS << 16);
break;
case UHF_C_PORT_ENABLE:
OWRITE4(sc, port, UPS_C_PORT_ENABLED << 16);
break;
case UHF_C_PORT_SUSPEND:
OWRITE4(sc, port, UPS_C_SUSPEND << 16);
break;
case UHF_C_PORT_OVER_CURRENT:
OWRITE4(sc, port, UPS_C_OVERCURRENT_INDICATOR << 16);
break;
case UHF_C_PORT_RESET:
OWRITE4(sc, port, UPS_C_PORT_RESET << 16);
break;
default:
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
switch(value) {
case UHF_C_PORT_CONNECTION:
case UHF_C_PORT_ENABLE:
case UHF_C_PORT_SUSPEND:
case UHF_C_PORT_OVER_CURRENT:
case UHF_C_PORT_RESET:
/* Enable RHSC interrupt if condition is cleared. */
if ((OREAD4(sc, port) >> 16) == 0)
ohci_rhsc_able(sc, 1);
break;
default:
break;
}
break;
case C(UR_GET_DESCRIPTOR, UT_READ_CLASS_DEVICE):
if ((value & 0xff) != 0) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
v = OREAD4(sc, OHCI_RH_DESCRIPTOR_A);
hubd = ohci_hubd;
hubd.bNbrPorts = sc->sc_noport;
USETW(hubd.wHubCharacteristics,
(v & OHCI_NPS ? UHD_PWR_NO_SWITCH :
v & OHCI_PSM ? UHD_PWR_GANGED : UHD_PWR_INDIVIDUAL)
/* XXX overcurrent */
);
hubd.bPwrOn2PwrGood = OHCI_GET_POTPGT(v);
v = OREAD4(sc, OHCI_RH_DESCRIPTOR_B);
for (i = 0, l = sc->sc_noport; l > 0; i++, l -= 8, v >>= 8)
hubd.DeviceRemovable[i++] = (u_int8_t)v;
hubd.bDescLength = USB_HUB_DESCRIPTOR_SIZE + i;
l = min(len, hubd.bDescLength);
totlen = l;
memcpy(buf, &hubd, l);
break;
case C(UR_GET_STATUS, UT_READ_CLASS_DEVICE):
if (len != 4) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
memset(buf, 0, len); /* ? XXX */
totlen = len;
break;
case C(UR_GET_STATUS, UT_READ_CLASS_OTHER):
DPRINTFN(8,("ohci_root_ctrl_transfer: get port status i=%d\n",
index));
if (index < 1 || index > sc->sc_noport) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
if (len != 4) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
v = OREAD4(sc, OHCI_RH_PORT_STATUS(index));
DPRINTFN(8,("ohci_root_ctrl_transfer: port status=0x%04x\n",
v));
USETW(ps.wPortStatus, v);
USETW(ps.wPortChange, v >> 16);
l = min(len, sizeof ps);
memcpy(buf, &ps, l);
totlen = l;
break;
case C(UR_SET_DESCRIPTOR, UT_WRITE_CLASS_DEVICE):
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
case C(UR_SET_FEATURE, UT_WRITE_CLASS_DEVICE):
break;
case C(UR_SET_FEATURE, UT_WRITE_CLASS_OTHER):
if (index < 1 || index > sc->sc_noport) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
port = OHCI_RH_PORT_STATUS(index);
switch(value) {
case UHF_PORT_ENABLE:
OWRITE4(sc, port, UPS_PORT_ENABLED);
break;
case UHF_PORT_SUSPEND:
OWRITE4(sc, port, UPS_SUSPEND);
break;
case UHF_PORT_RESET:
DPRINTFN(5,("ohci_root_ctrl_transfer: reset port %d\n",
index));
OWRITE4(sc, port, UPS_RESET);
for (i = 0; i < 5; i++) {
usb_delay_ms(&sc->sc_bus,
USB_PORT_ROOT_RESET_DELAY);
if (sc->sc_dying) {
err = USBD_IOERROR;
goto ret;
}
if ((OREAD4(sc, port) & UPS_RESET) == 0)
break;
}
DPRINTFN(8,("ohci port %d reset, status = 0x%04x\n",
index, OREAD4(sc, port)));
break;
case UHF_PORT_POWER:
DPRINTFN(2,("ohci_root_ctrl_transfer: set port power "
"%d\n", index));
OWRITE4(sc, port, UPS_PORT_POWER);
break;
default:
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
break;
default:
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = USBD_IOERROR;
goto ret;
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->actlen = totlen;
err = USBD_NORMAL_COMPLETION;
ret:
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->status = err;
s = splusb();
usb_transfer_complete(xfer);
splx(s);
return (USBD_IN_PROGRESS);
}
/* Abort a root control request. */
Static void
ohci_root_ctrl_abort(usbd_xfer_handle xfer)
{
/* Nothing to do, all transfers are synchronous. */
}
/* Close the root pipe. */
Static void
ohci_root_ctrl_close(usbd_pipe_handle pipe)
{
DPRINTF(("ohci_root_ctrl_close\n"));
/* Nothing to do. */
}
Static usbd_status
ohci_root_intr_transfer(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
/* Insert last in queue. */
err = usb_insert_transfer(xfer);
if (err)
return (err);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* Pipe isn't running, start first */
return (ohci_root_intr_start(SIMPLEQ_FIRST(&xfer->pipe->queue)));
}
Static usbd_status
ohci_root_intr_start(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_pipe_handle pipe = xfer->pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
if (sc->sc_dying)
return (USBD_IOERROR);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
sc->sc_intrxfer = xfer;
return (USBD_IN_PROGRESS);
}
/* Abort a root interrupt request. */
Static void
ohci_root_intr_abort(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
int s;
if (xfer->pipe->intrxfer == xfer) {
DPRINTF(("ohci_root_intr_abort: remove\n"));
xfer->pipe->intrxfer = NULL;
}
xfer->status = USBD_CANCELLED;
s = splusb();
usb_transfer_complete(xfer);
splx(s);
}
/* Close the root pipe. */
Static void
ohci_root_intr_close(usbd_pipe_handle pipe)
{
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
DPRINTF(("ohci_root_intr_close\n"));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
sc->sc_intrxfer = NULL;
}
/************************/
Static usbd_status
ohci_device_ctrl_transfer(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* Insert last in queue. */
err = usb_insert_transfer(xfer);
if (err)
return (err);
/* Pipe isn't running, start first */
return (ohci_device_ctrl_start(SIMPLEQ_FIRST(&xfer->pipe->queue)));
}
Static usbd_status
ohci_device_ctrl_start(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_softc_t *sc = (ohci_softc_t *)xfer->pipe->device->bus;
usbd_status err;
if (sc->sc_dying)
return (USBD_IOERROR);
#ifdef DIAGNOSTIC
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (!(xfer->rqflags & URQ_REQUEST)) {
/* XXX panic */
printf("ohci_device_ctrl_transfer: not a request\n");
return (USBD_INVAL);
}
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
err = ohci_device_request(xfer);
if (err)
return (err);
if (sc->sc_bus.use_polling)
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_waitintr(sc, xfer);
return (USBD_IN_PROGRESS);
}
/* Abort a device control request. */
Static void
ohci_device_ctrl_abort(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTF(("ohci_device_ctrl_abort: xfer=%p\n", xfer));
ohci_abort_xfer(xfer, USBD_CANCELLED);
}
/* Close a device control pipe. */
Static void
ohci_device_ctrl_close(usbd_pipe_handle pipe)
{
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
DPRINTF(("ohci_device_ctrl_close: pipe=%p\n", pipe));
ohci_close_pipe(pipe, sc->sc_ctrl_head);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ohci_free_std(sc, opipe->tail.td);
}
/************************/
Static void
ohci_device_clear_toggle(usbd_pipe_handle pipe)
{
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
opipe->sed->ed.ed_headp &= htole32(~OHCI_TOGGLECARRY);
}
Static void
ohci_noop(usbd_pipe_handle pipe)
{
}
Static usbd_status
ohci_device_bulk_transfer(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* Insert last in queue. */
err = usb_insert_transfer(xfer);
if (err)
return (err);
/* Pipe isn't running, start first */
return (ohci_device_bulk_start(SIMPLEQ_FIRST(&xfer->pipe->queue)));
}
Static usbd_status
ohci_device_bulk_start(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
usbd_device_handle dev = opipe->pipe.device;
ohci_softc_t *sc = (ohci_softc_t *)dev->bus;
int addr = dev->address;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_soft_td_t *data, *tail, *tdp;
ohci_soft_ed_t *sed;
int s, len, isread, endpt;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
if (sc->sc_dying)
return (USBD_IOERROR);
#ifdef DIAGNOSTIC
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->rqflags & URQ_REQUEST) {
/* XXX panic */
printf("ohci_device_bulk_start: a request\n");
return (USBD_INVAL);
}
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
len = xfer->length;
endpt = xfer->pipe->endpoint->edesc->bEndpointAddress;
isread = UE_GET_DIR(endpt) == UE_DIR_IN;
sed = opipe->sed;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTFN(4,("ohci_device_bulk_start: xfer=%p len=%d isread=%d "
"flags=%d endpt=%d\n", xfer, len, isread, xfer->flags,
endpt));
opipe->u.bulk.isread = isread;
opipe->u.bulk.length = len;
/* Update device address */
sed->ed.ed_flags = htole32(
(le32toh(sed->ed.ed_flags) & ~OHCI_ED_ADDRMASK) |
OHCI_ED_SET_FA(addr));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* Allocate a chain of new TDs (including a new tail). */
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
data = opipe->tail.td;
err = ohci_alloc_std_chain(opipe, sc, len, isread, xfer,
data, &tail);
/* We want interrupt at the end of the transfer. */
tail->td.td_flags &= htole32(~OHCI_TD_INTR_MASK);
tail->td.td_flags |= htole32(OHCI_TD_SET_DI(1));
tail->flags |= OHCI_CALL_DONE;
tail = tail->nexttd; /* point at sentinel */
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (err)
return (err);
tail->xfer = NULL;
xfer->hcpriv = data;
DPRINTFN(4,("ohci_device_bulk_start: ed_flags=0x%08x td_flags=0x%08x "
"td_cbp=0x%08x td_be=0x%08x\n",
(int)le32toh(sed->ed.ed_flags),
(int)le32toh(data->td.td_flags),
(int)le32toh(data->td.td_cbp),
(int)le32toh(data->td.td_be)));
#ifdef USB_DEBUG
if (ohcidebug > 5) {
ohci_dump_ed(sed);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_dump_tds(data);
}
#endif
/* Insert ED in schedule */
s = splusb();
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
for (tdp = data; tdp != tail; tdp = tdp->nexttd) {
tdp->xfer = xfer;
}
sed->ed.ed_tailp = htole32(tail->physaddr);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
opipe->tail.td = tail;
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP);
OWRITE4(sc, OHCI_COMMAND_STATUS, OHCI_BLF);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->timeout && !sc->sc_bus.use_polling) {
Huge merge from NetBSD: usbdi.c (1.61): =================================================================== revision 1.61 date: 2000/01/31 20:13:07; author: augustss; lines: +20 -4 Change the way the HC done method is invoked a little. =================================================================== usbdi.c (1.65): =================================================================== revision 1.65 date: 2000/03/08 15:34:10; author: augustss; lines: +4 -2 Get the status right when a polled transfer times out. =================================================================== ohci.c (1.79), uhci.c (1.89), uhcivar.h (1.24), usb_port.h (1.22), usbdivar.h (1.48): =================================================================== date: 2000/03/23 07:01:46; author: thorpej; New callout mechanism with two major improvements over the old timeout()/untimeout() API: - Clients supply callout handle storage, thus eliminating problems of resource allocation. - Insertion and removal of callouts is constant time, important as this facility is used quite a lot in the kernel. The old timeout()/untimeout() API has been removed from the kernel. =================================================================== uhci.c (1.80), usbdi.c (1.66): =================================================================== date: 2000/03/23 18:59:10; author: thorpej; Shake out some bugs from the callout changes. =================================================================== ohci.c (1.80), uhci.c (1.91), uhcivar.h (1.25), usb_port.h (1.23), usbdi.c (1.67), usbdivar.h (1.49): =================================================================== date: 2000/03/24 22:03:30; author: augustss; Some cleanup and renaming of the callouts used in USB drivers. =================================================================== uhci.c (1.92), uhcivar.h (1.26): =================================================================== date: 2000/03/24 22:57:58; author: augustss; Two major changes: Make each xfer have its own intr_info. This is necessary if we want to queue multiple xfers on an endpoint. This should get rid of the (mostly harmless) DIAGNOSTICs about intr_infos (not) being done. Change (again!) how xfers are aborted. Aborting a TD is a nightmare on the braindead UHCI controller. (Unless you stop the HC, thereby losing isoc traffic.) Hopefully I got it right this time. =================================================================== usbdivar.h (1.50): =================================================================== revision 1.50 date: 2000/03/25 00:10:19; author: augustss; lines: +4 -2 GC an unsued field and add some DIAGNOSTIC in xfer. =================================================================== ums.c: Use the callout functions instead of the timeout ones. uhci.c (1.93): =================================================================== revision 1.93 date: 2000/03/25 00:11:21; author: augustss; lines: +26 -1 Add more DIAGNOSTIC when aborting isoc. =================================================================== uhci.c (1.94), usbdivar.h (1.51): =================================================================== date: 2000/03/25 07:13:05; author: augustss; More DIAGNOSTIC. Initialize a callout handle I forgot. =================================================================== uhci.c (1.95): =================================================================== revision 1.95 date: 2000/03/25 07:23:12; author: augustss; Exp; lines: +24 -7 Improve uhci_dump_ii(). =================================================================== ohci.c (1.81), uhci.c (1.96), uhcivar.h (1.27), usb_subr.c (1.68), usbdi.c (1.68), usbdivar.h (1.52): =================================================================== date: 2000/03/25 18:02:33; author: augustss; Rename and move around callout handles to make it more sane. Add some DIAGNOSTIC. Fix buglet in isoc abort on UHCI. =================================================================== uhci.c (1.98): =================================================================== revision 1.98 date: 2000/03/27 07:39:48; author: augustss; lines: +12 -4 Make it compile without DIAGNOSTIC. =================================================================== uhci.c (1.99): =================================================================== revision 1.99 date: 2000/03/27 08:01:09; author: augustss; lines: +1 -5 Remove some debug nonsense. =================================================================== uhci.c (1.100): =================================================================== revision 1.100 date: 2000/03/27 09:41:36; author: augustss; lines: +13 -3 Don't mess with QH in bulk abort for the moment. =================================================================== uhci.c (1.102): =================================================================== revision 1.102 date: 2000/03/27 22:42:57; author: augustss; lines: +66 -26 Be a little more careful when aborting. Preallocate some TDs for large buffers. =================================================================== uhci.c (1.103): =================================================================== date: 2000/03/28 09:47:10; author: augustss; lines: +11 -1 Another patch for xfer abort... XXX The current xfer queueing and aborting semantics should really XXX be changed. It cannot be implemented in a sane way on UHCI. XXX One day when I have lots of time I'll redesign it... =================================================================== uhci.c (1.104): Correct a debug message. uhci.c (1.105): Be more defensive in a DIAGNOSTIC test. uhci.c (1.106): =================================================================== revision 1.106 date: 2000/03/29 01:49:13; author: augustss; lines: +14 -309 *SIGH* Revert back to the old method of aborting xfers. I had tested the new stuff for two months now, but as soon as I commited it the problems started to appear. Murphy, no doubt... =================================================================== usb_subr.c (1.70), usbdi.c (1.71), usbdivar.h (1.53): =================================================================== revision 1.70 date: 2000/03/29 01:45:20; author: augustss; lines: +2 -1 Do not accept new xfers for queuing while a pipe is aborting. ===================================================================
2002-03-16 12:06:01 +00:00
usb_callout(xfer->timeout_handle, MS_TO_TICKS(xfer->timeout),
ohci_timeout, xfer);
}
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#if 0
/* This goes wrong if we are too slow. */
if (ohcidebug > 10) {
delay(10000);
DPRINTF(("ohci_device_intr_transfer: status=%x\n",
OREAD4(sc, OHCI_COMMAND_STATUS)));
ohci_dump_ed(sed);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_dump_tds(data);
}
#endif
splx(s);
return (USBD_IN_PROGRESS);
}
Static void
ohci_device_bulk_abort(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTF(("ohci_device_bulk_abort: xfer=%p\n", xfer));
ohci_abort_xfer(xfer, USBD_CANCELLED);
}
/*
* Close a device bulk pipe.
*/
Static void
ohci_device_bulk_close(usbd_pipe_handle pipe)
{
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
DPRINTF(("ohci_device_bulk_close: pipe=%p\n", pipe));
ohci_close_pipe(pipe, sc->sc_bulk_head);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ohci_free_std(sc, opipe->tail.td);
}
/************************/
Static usbd_status
ohci_device_intr_transfer(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usbd_status err;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/* Insert last in queue. */
err = usb_insert_transfer(xfer);
if (err)
return (err);
/* Pipe isn't running, start first */
return (ohci_device_intr_start(SIMPLEQ_FIRST(&xfer->pipe->queue)));
}
Static usbd_status
ohci_device_intr_start(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
usbd_device_handle dev = opipe->pipe.device;
ohci_softc_t *sc = (ohci_softc_t *)dev->bus;
ohci_soft_ed_t *sed = opipe->sed;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_soft_td_t *data, *tail;
int len;
int s;
if (sc->sc_dying)
return (USBD_IOERROR);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
DPRINTFN(3, ("ohci_device_intr_transfer: xfer=%p len=%d "
"flags=%d priv=%p\n",
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer, xfer->length, xfer->flags, xfer->priv));
#ifdef DIAGNOSTIC
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->rqflags & URQ_REQUEST)
panic("ohci_device_intr_transfer: a request");
#endif
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
len = xfer->length;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
data = opipe->tail.td;
tail = ohci_alloc_std(sc);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (tail == NULL)
return (USBD_NOMEM);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
tail->xfer = NULL;
data->td.td_flags = htole32(
OHCI_TD_IN | OHCI_TD_NOCC |
OHCI_TD_SET_DI(1) | OHCI_TD_TOGGLE_CARRY);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->flags & USBD_SHORT_XFER_OK)
data->td.td_flags |= htole32(OHCI_TD_R);
data->td.td_cbp = htole32(DMAADDR(&xfer->dmabuf, 0));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
data->nexttd = tail;
data->td.td_nexttd = htole32(tail->physaddr);
data->td.td_be = htole32(le32toh(data->td.td_cbp) + len - 1);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
data->len = len;
data->xfer = xfer;
data->flags = OHCI_CALL_DONE | OHCI_ADD_LEN;
xfer->hcpriv = data;
#ifdef USB_DEBUG
if (ohcidebug > 5) {
DPRINTF(("ohci_device_intr_transfer:\n"));
ohci_dump_ed(sed);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_dump_tds(data);
}
#endif
/* Insert ED in schedule */
s = splusb();
sed->ed.ed_tailp = htole32(tail->physaddr);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
opipe->tail.td = tail;
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP);
1999-05-03 23:11:41 +00:00
#if 0
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
/*
* This goes horribly wrong, printing thousands of descriptors,
* because false references are followed due to the fact that the
* TD is gone.
*/
if (ohcidebug > 5) {
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
usb_delay_ms(&sc->sc_bus, 5);
DPRINTF(("ohci_device_intr_transfer: status=%x\n",
OREAD4(sc, OHCI_COMMAND_STATUS)));
ohci_dump_ed(sed);
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
ohci_dump_tds(data);
}
#endif
splx(s);
return (USBD_IN_PROGRESS);
}
/* Abort a device control request. */
Static void
ohci_device_intr_abort(usbd_xfer_handle xfer)
{
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
if (xfer->pipe->intrxfer == xfer) {
DPRINTF(("ohci_device_intr_abort: remove\n"));
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
xfer->pipe->intrxfer = NULL;
}
ohci_abort_xfer(xfer, USBD_CANCELLED);
}
/* Close a device interrupt pipe. */
Static void
ohci_device_intr_close(usbd_pipe_handle pipe)
{
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
int nslots = opipe->u.intr.nslots;
int pos = opipe->u.intr.pos;
int j;
ohci_soft_ed_t *p, *sed = opipe->sed;
int s;
DPRINTFN(1,("ohci_device_intr_close: pipe=%p nslots=%d pos=%d\n",
pipe, nslots, pos));
s = splusb();
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP);
if ((le32toh(sed->ed.ed_tailp) & OHCI_HEADMASK) !=
(le32toh(sed->ed.ed_headp) & OHCI_HEADMASK))
usb_delay_ms(&sc->sc_bus, 2);
#ifdef DIAGNOSTIC
if ((le32toh(sed->ed.ed_tailp) & OHCI_HEADMASK) !=
(le32toh(sed->ed.ed_headp) & OHCI_HEADMASK))
panic("%s: Intr pipe %p still has TDs queued",
USBDEVNAME(sc->sc_bus.bdev), pipe);
#endif
for (p = sc->sc_eds[pos]; p && p->next != sed; p = p->next)
;
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#ifdef DIAGNOSTIC
if (p == NULL)
panic("ohci_device_intr_close: ED not found");
Synchronisation with NetBSD as of 1999/11/16: Cleaning up the code: - Declare many functions static - Change variable names to make them more self explanatory - Change usbd_request_handle -> usbd_xfer_handle - Syntactical changes - Remove some unused code - Other KNF changes Interrupt context handling - Change delay to usbd_delay_ms were possible (takes polling mode into account) - Change detection mechanism for interrupt context Add support for pre-allocation DMA-able memory by device driver Add preliminary support for isochronous to the UHCI driver (not for OHCI yet). usb.c, uhci.c, ohci.c - Initial attempt at detachable USB host controllers - Handle the use_polling flag with a lttle more care and only set it if we are cold booting. usb.c, uhci.c ohci.c, usbdi.c usbdi_util.c usb_subr.c - Make sure an aborted pipe is marked as not running. - Start queued request in the right order. - Insert some more DIAGNOSTIC sanity checks. - Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN. usb.c, usb_subr.c - Add an event mechanism so that a userland process can watch devices come and go. ohci.c - Handle the case when a USB transfer is so long that it crosses two page (4K) boundaries. OHCI cannot do that with a single TD so we make a chain. ulpt.c - Use a bigger buffer when transferring data. - Pre-allocate the DMA buffer. This makes the driver slightly more efficient. - Comment out the GET_DEVICE_ID code, because for some unknown reason it causes printing to fail sometimes. usb.h - Add a macro to extract the isoc type. - Add a macro to check whether the routine has been entered after splusb and if not, complain. usbdi.c - Fix a glitch in dequeueing and aborting requests on interrupt pipes. - Add a flag in the request to determine if the data copying is done by the driver or the usbdi layer.
1999-11-17 22:33:51 +00:00
#endif
p->next = sed->next;
p->ed.ed_nexted = sed->ed.ed_nexted;
splx(s);
for (j = 0; j < nslots; j++)
--sc->sc_bws[(pos * nslots + j) % OHCI_NO_INTRS];
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ohci_free_std(sc, opipe->tail.td);
ohci_free_sed(sc, opipe->sed);
}
Static usbd_status
ohci_device_setintr(ohci_softc_t *sc, struct ohci_pipe *opipe, int ival)
{
int i, j, s, best;
u_int npoll, slow, shigh, nslots;
u_int bestbw, bw;
ohci_soft_ed_t *hsed, *sed = opipe->sed;
DPRINTFN(2, ("ohci_setintr: pipe=%p\n", opipe));
if (ival == 0) {
printf("ohci_setintr: 0 interval\n");
return (USBD_INVAL);
}
npoll = OHCI_NO_INTRS;
while (npoll > ival)
npoll /= 2;
DPRINTFN(2, ("ohci_setintr: ival=%d npoll=%d\n", ival, npoll));
/*
* We now know which level in the tree the ED must go into.
* Figure out which slot has most bandwidth left over.
* Slots to examine:
* npoll
* 1 0
* 2 1 2
* 4 3 4 5 6
* 8 7 8 9 10 11 12 13 14
* N (N-1) .. (N-1+N-1)
*/
slow = npoll-1;
shigh = slow + npoll;
nslots = OHCI_NO_INTRS / npoll;
for (best = i = slow, bestbw = ~0; i < shigh; i++) {
bw = 0;
for (j = 0; j < nslots; j++)
bw += sc->sc_bws[(i * nslots + j) % OHCI_NO_INTRS];
if (bw < bestbw) {
best = i;
bestbw = bw;
}
}
DPRINTFN(2, ("ohci_setintr: best=%d(%d..%d) bestbw=%d\n",
best, slow, shigh, bestbw));
s = splusb();
hsed = sc->sc_eds[best];
sed->next = hsed->next;
sed->ed.ed_nexted = hsed->ed.ed_nexted;
hsed->next = sed;
hsed->ed.ed_nexted = htole32(sed->physaddr);
splx(s);
for (j = 0; j < nslots; j++)
++sc->sc_bws[(best * nslots + j) % OHCI_NO_INTRS];
opipe->u.intr.nslots = nslots;
opipe->u.intr.pos = best;
DPRINTFN(5, ("ohci_setintr: returns %p\n", opipe));
return (USBD_NORMAL_COMPLETION);
}
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
/***********************/
usbd_status
ohci_device_isoc_transfer(usbd_xfer_handle xfer)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
usbd_status err;
DPRINTFN(5,("ohci_device_isoc_transfer: xfer=%p\n", xfer));
/* Put it on our queue, */
err = usb_insert_transfer(xfer);
/* bail out on error, */
if (err && err != USBD_IN_PROGRESS)
return (err);
/* XXX should check inuse here */
/* insert into schedule, */
ohci_device_isoc_enter(xfer);
/* and start if the pipe wasn't running */
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
if (!err)
ohci_device_isoc_start(SIMPLEQ_FIRST(&xfer->pipe->queue));
return (err);
}
void
ohci_device_isoc_enter(usbd_xfer_handle xfer)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
usbd_device_handle dev = opipe->pipe.device;
ohci_softc_t *sc = (ohci_softc_t *)dev->bus;
ohci_soft_ed_t *sed = opipe->sed;
struct iso *iso = &opipe->u.iso;
struct ohci_xfer *oxfer = (struct ohci_xfer *)xfer;
ohci_soft_itd_t *sitd, *nsitd;
ohci_physaddr_t buf, offs, noffs, bp0, tdphys;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
int i, ncur, nframes;
int s;
DPRINTFN(1,("ohci_device_isoc_enter: used=%d next=%d xfer=%p "
"nframes=%d\n",
iso->inuse, iso->next, xfer, xfer->nframes));
if (sc->sc_dying)
return;
if (iso->next == -1) {
/* Not in use yet, schedule it a few frames ahead. */
iso->next = le32toh(sc->sc_hcca->hcca_frame_number) + 5;
DPRINTFN(2,("ohci_device_isoc_enter: start next=%d\n",
iso->next));
}
if (xfer->hcpriv) {
for (sitd = xfer->hcpriv; sitd != NULL && sitd->xfer == xfer;
sitd = sitd->nextitd)
ohci_free_sitd(sc, sitd); /* Free ITDs in prev xfer*/
if (sitd == NULL) {
sitd = ohci_alloc_sitd(sc);
if (sitd == NULL)
panic("cant alloc isoc");
opipe->tail.itd = sitd;
tdphys = sitd->physaddr;
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP); /* Stop*/
sed->ed.ed_headp =
sed->ed.ed_tailp = htole32(tdphys);
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP); /* Start.*/
}
}
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
sitd = opipe->tail.itd;
buf = DMAADDR(&xfer->dmabuf, 0);
bp0 = OHCI_PAGE(buf);
offs = OHCI_PAGE_OFFSET(buf);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
nframes = xfer->nframes;
xfer->hcpriv = sitd;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
for (i = ncur = 0; i < nframes; i++, ncur++) {
noffs = offs + xfer->frlengths[i];
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
if (ncur == OHCI_ITD_NOFFSET || /* all offsets used */
OHCI_PAGE(buf + noffs) > bp0 + OHCI_PAGE_SIZE) { /* too many page crossings */
/* Allocate next ITD */
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
nsitd = ohci_alloc_sitd(sc);
if (nsitd == NULL) {
/* XXX what now? */
printf("%s: isoc TD alloc failed\n",
USBDEVNAME(sc->sc_bus.bdev));
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
return;
}
/* Fill current ITD */
sitd->itd.itd_flags = htole32(
OHCI_ITD_NOCC |
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
OHCI_ITD_SET_SF(iso->next) |
OHCI_ITD_SET_DI(6) | /* delay intr a little */
OHCI_ITD_SET_FC(ncur));
sitd->itd.itd_bp0 = htole32(bp0);
sitd->nextitd = nsitd;
sitd->itd.itd_nextitd = htole32(nsitd->physaddr);
sitd->itd.itd_be = htole32(bp0 + offs - 1);
sitd->xfer = xfer;
sitd->flags = OHCI_ITD_ACTIVE;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
sitd = nsitd;
iso->next = iso->next + ncur;
bp0 = OHCI_PAGE(buf + offs);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
ncur = 0;
}
sitd->itd.itd_offset[ncur] = htole16(OHCI_ITD_MK_OFFS(offs));
offs = noffs;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
nsitd = ohci_alloc_sitd(sc);
if (nsitd == NULL) {
/* XXX what now? */
printf("%s: isoc TD alloc failed\n",
USBDEVNAME(sc->sc_bus.bdev));
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
return;
}
/* Fixup last used ITD */
sitd->itd.itd_flags = htole32(
OHCI_ITD_NOCC |
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
OHCI_ITD_SET_SF(iso->next) |
OHCI_ITD_SET_DI(0) |
OHCI_ITD_SET_FC(ncur));
sitd->itd.itd_bp0 = htole32(bp0);
sitd->nextitd = nsitd;
sitd->itd.itd_nextitd = htole32(nsitd->physaddr);
sitd->itd.itd_be = htole32(bp0 + offs - 1);
sitd->xfer = xfer;
sitd->flags = OHCI_CALL_DONE | OHCI_ITD_ACTIVE;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
iso->next = iso->next + ncur;
iso->inuse += nframes;
xfer->actlen = offs; /* XXX pretend we did it all */
xfer->status = USBD_IN_PROGRESS;
oxfer->ohci_xfer_flags |= OHCI_ISOC_DIRTY;
#ifdef USB_DEBUG
if (ohcidebug > 5) {
DPRINTF(("ohci_device_isoc_enter: frame=%d\n",
le32toh(sc->sc_hcca->hcca_frame_number)));
ohci_dump_itds(xfer->hcpriv);
ohci_dump_ed(sed);
}
#endif
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
s = splusb();
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
opipe->tail.itd = nsitd;
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP);
sed->ed.ed_tailp = htole32(nsitd->physaddr);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
splx(s);
#ifdef USB_DEBUG
if (ohcidebug > 5) {
delay(150000);
DPRINTF(("ohci_device_isoc_enter: after frame=%d\n",
le32toh(sc->sc_hcca->hcca_frame_number)));
ohci_dump_itds(xfer->hcpriv);
ohci_dump_ed(sed);
}
#endif
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
usbd_status
ohci_device_isoc_start(usbd_xfer_handle xfer)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
ohci_softc_t *sc = (ohci_softc_t *)opipe->pipe.device->bus;
ohci_soft_ed_t *sed;
int s;
DPRINTFN(5,("ohci_device_isoc_start: xfer=%p\n", xfer));
if (sc->sc_dying)
return (USBD_IOERROR);
#ifdef DIAGNOSTIC
if (xfer->status != USBD_IN_PROGRESS)
printf("ohci_device_isoc_start: not in progress %p\n", xfer);
#endif
/* XXX anything to do? */
s = splusb();
sed = opipe->sed; /* Turn off ED skip-bit to start processing */
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP); /* ED's ITD list.*/
splx(s);
return (USBD_IN_PROGRESS);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
void
ohci_device_isoc_abort(usbd_xfer_handle xfer)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_pipe *opipe = (struct ohci_pipe *)xfer->pipe;
ohci_softc_t *sc = (ohci_softc_t *)opipe->pipe.device->bus;
ohci_soft_ed_t *sed;
ohci_soft_itd_t *sitd, *tmp_sitd;
int s,undone,num_sitds;
s = splusb();
opipe->aborting = 1;
DPRINTFN(1,("ohci_device_isoc_abort: xfer=%p\n", xfer));
/* Transfer is already done. */
if (xfer->status != USBD_NOT_STARTED &&
xfer->status != USBD_IN_PROGRESS) {
splx(s);
printf("ohci_device_isoc_abort: early return\n");
return;
}
/* Give xfer the requested abort code. */
xfer->status = USBD_CANCELLED;
sed = opipe->sed;
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP); /* force hardware skip */
num_sitds = 0;
sitd = xfer->hcpriv;
#ifdef DIAGNOSTIC
if (sitd == NULL) {
splx(s);
printf("ohci_device_isoc_abort: hcpriv==0\n");
return;
}
#endif
for (; sitd != NULL && sitd->xfer == xfer; sitd = sitd->nextitd) {
num_sitds++;
#ifdef DIAGNOSTIC
DPRINTFN(1,("abort sets done sitd=%p\n", sitd));
sitd->isdone = 1;
#endif
}
splx(s);
/*
* Each sitd has up to OHCI_ITD_NOFFSET transfers, each can
* take a usb 1ms cycle. Conservatively wait for it to drain.
* Even with DMA done, it can take awhile for the "batch"
* delivery of completion interrupts to occur thru the controller.
*/
do {
usb_delay_ms(&sc->sc_bus, 2*(num_sitds*OHCI_ITD_NOFFSET));
undone = 0;
tmp_sitd = xfer->hcpriv;
for (; tmp_sitd != NULL && tmp_sitd->xfer == xfer;
tmp_sitd = tmp_sitd->nextitd) {
if (OHCI_CC_NO_ERROR ==
OHCI_ITD_GET_CC(le32toh(tmp_sitd->itd.itd_flags)) &&
tmp_sitd->flags & OHCI_ITD_ACTIVE &&
(tmp_sitd->flags & OHCI_ITD_INTFIN) == 0)
undone++;
}
} while( undone != 0 );
s = splusb();
/* Run callback. */
usb_transfer_complete(xfer);
if (sitd != NULL)
/*
* Only if there is a `next' sitd in next xfer...
* unlink this xfer's sitds.
*/
sed->ed.ed_headp = htole32(sitd->physaddr);
else
sed->ed.ed_headp = 0;
sed->ed.ed_flags &= htole32(~OHCI_ED_SKIP); /* remove hardware skip */
splx(s);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
void
ohci_device_isoc_done(usbd_xfer_handle xfer)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
/* This null routine corresponds to non-isoc "done()" routines
* that free the stds associated with an xfer after a completed
* xfer interrupt. However, in the case of isoc transfers, the
* sitds associated with the transfer have already been processed
* and reallocated for the next iteration by
* "ohci_device_isoc_transfer()".
*
* Routine "usb_transfer_complete()" is called at the end of every
* relevant usb interrupt. "usb_transfer_complete()" indirectly
* calls 1) "ohci_device_isoc_transfer()" (which keeps pumping the
* pipeline by setting up the next transfer iteration) and 2) then
* calls "ohci_device_isoc_done()". Isoc transfers have not been
* working for the ohci usb because this routine was trashing the
* xfer set up for the next iteration (thus, only the first
* UGEN_NISOREQS xfers outstanding on an open would work). Perhaps
* this could all be re-factored, but that's another pass...
*/
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}
usbd_status
ohci_setup_isoc(usbd_pipe_handle pipe)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
struct iso *iso = &opipe->u.iso;
int s;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
iso->next = -1;
iso->inuse = 0;
s = splusb();
ohci_add_ed(opipe->sed, sc->sc_isoc_head);
splx(s);
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
return (USBD_NORMAL_COMPLETION);
}
void
ohci_device_isoc_close(usbd_pipe_handle pipe)
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
{
struct ohci_pipe *opipe = (struct ohci_pipe *)pipe;
ohci_softc_t *sc = (ohci_softc_t *)pipe->device->bus;
ohci_soft_ed_t *sed;
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
DPRINTF(("ohci_device_isoc_close: pipe=%p\n", pipe));
sed = opipe->sed;
sed->ed.ed_flags |= htole32(OHCI_ED_SKIP); /* Stop device. */
ohci_close_pipe(pipe, sc->sc_isoc_head); /* Stop isoc list, free ED.*/
/* up to NISOREQs xfers still outstanding. */
#ifdef DIAGNOSTIC
opipe->tail.itd->isdone = 1;
#endif
ohci_free_sitd(sc, opipe->tail.itd); /* Next `avail free' sitd.*/
More USB ethernet tweaks: - Sync ohci, uhci and usbdi modules with NetBSD in order to obtain the following improvements: o New USBD_NO_TSLEEP flag can be used in place of UQ_NO_TSLEEP quirk. This allows drivers to specify busy waiting only for certain transfers (namely control transfers for reading/writing registers and stuff). o New USBD_FORCE_SHORT_XFER flag can be used to deal with devices like the ADMtek Pegasus that sense the end of bulk OUT transfers in a special way (if a transfer is exactly a multiple of 64 bytes in size, you need to send an extra empty packet to terminate the transfer). o usbd_open_pipe_intr() now accepts an interval argument which can be used to change the rate at which the interrupt callback routine is invoked. Specifying USBD_DEFAULT_INTERVAL uses the value specified in the device's config data, but drivers can override it if needed. - Change if_aue to use USBD_FORCE_SHORT_XFER for packet transmissions. - Change if_aue, if_kue and if_cue to use USBD_NO_TSLEEP for all control transfers. We no longer force the non-tsleep hack for bulk transfers since these are done asynchronously anyway. - Removed quirk entry fiddling from if_aue and if_kue since we don't need it anymore now that we have the USBD_NO_TSLEEP flag. - Tweak ulpt, uhid, ums and ukbd drivers to use the new arg to usbd_open_pipe_intr(). - Add a flag to the softc struct in the ethernet drivers to indicate when a device has been detached, and use this flag to perform tests to prevent the drivers from trying to do control transfers if this is the case. This is necessary because calling if_detach() with INET6 enabled will eventually result in a call to the driver's ioctl() routine to delete the multicast groups on the interface, which will result in attempts to perform control transfers. (It's possible this also happens even without INET6 support enabled.) This is pointless since we know that if the detach method has been called, the hardware has been unplugged. - Changed watchdog timeout routines to just call the driver init routines to initialize the device states without trying to close and re-open the pipes. This is partly because we don't want to frob things at interrupt context, but also because this doesn't seem to work right and I don't want to panic the system just because a USB device may have stopped responding. - Fix aue_rxeof() to be a little smarter about detecting when a double transfer is needed. Unfortunately, the design of the chip makes it hard to get this exactly right. Hopefully, this will go away once either Nick or Lennart finds the bug in the uhci driver that makes this ugly hack necessary. - Also sync usbdevs with NetBSD.
2000-01-20 07:38:33 +00:00
}