freebsd-skq/sys/dev/pccbb/pccbb.c

1605 lines
47 KiB
C
Raw Normal View History

/*-
* Copyright (c) 2002-2004 M. Warner Losh.
* Copyright (c) 2000-2001 Jonathan Chen.
* All rights reserved.
*
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
*
*/
2005-01-11 05:34:14 +00:00
/*-
* Copyright (c) 1998, 1999 and 2000
* HAYAKAWA Koichi. All rights reserved.
*
* 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 HAYAKAWA Koichi.
* 4. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
*/
/*
* Driver for PCI to CardBus Bridge chips
* and PCI to PCMCIA Bridge chips
* and ISA to PCMCIA host adapters
* and C Bus to PCMCIA host adapters
*
* References:
* TI Datasheets:
* http://www-s.ti.com/cgi-bin/sc/generic2.cgi?family=PCI+CARDBUS+CONTROLLERS
*
* Written by Jonathan Chen <jon@freebsd.org>
* The author would like to acknowledge:
* * HAYAKAWA Koichi: Author of the NetBSD code for the same thing
* * Warner Losh: Newbus/newcard guru and author of the pccard side of things
* * YAMAMOTO Shigeru: Author of another FreeBSD cardbus driver
* * David Cross: Author of the initial ugly hack for a specific cardbus card
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/bus.h>
#include <sys/condvar.h>
#include <sys/errno.h>
#include <sys/kernel.h>
#include <sys/module.h>
#include <sys/kthread.h>
#include <sys/interrupt.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/rman.h>
#include <sys/sysctl.h>
#include <sys/systm.h>
#include <machine/bus.h>
#include <machine/resource.h>
#include <dev/pci/pcireg.h>
#include <dev/pci/pcivar.h>
#include <dev/pci/pcib_private.h>
#include <dev/pccard/pccardreg.h>
#include <dev/pccard/pccardvar.h>
#include <dev/exca/excareg.h>
#include <dev/exca/excavar.h>
#include <dev/pccbb/pccbbreg.h>
#include <dev/pccbb/pccbbvar.h>
#include "power_if.h"
#include "card_if.h"
#include "pcib_if.h"
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
#define DPRINTF(x) do { if (cbb_debug) printf x; } while (0)
#define DEVPRINTF(x) do { if (cbb_debug) device_printf x; } while (0)
#define PCI_MASK_CONFIG(DEV,REG,MASK,SIZE) \
pci_write_config(DEV, REG, pci_read_config(DEV, REG, SIZE) MASK, SIZE)
#define PCI_MASK2_CONFIG(DEV,REG,MASK1,MASK2,SIZE) \
pci_write_config(DEV, REG, ( \
pci_read_config(DEV, REG, SIZE) MASK1) MASK2, SIZE)
#define CBB_CARD_PRESENT(s) ((s & CBB_STATE_CD) == 0)
2002-08-15 08:01:24 +00:00
#define CBB_START_MEM 0x88000000
#define CBB_START_32_IO 0x1000
#define CBB_START_16_IO 0x100
devclass_t cbb_devclass;
/* sysctl vars */
static SYSCTL_NODE(_hw, OID_AUTO, cbb, CTLFLAG_RD, 0, "CBB parameters");
/* There's no way to say TUNEABLE_LONG to get the right types */
2002-08-15 08:01:24 +00:00
u_long cbb_start_mem = CBB_START_MEM;
SYSCTL_ULONG(_hw_cbb, OID_AUTO, start_memory, CTLFLAG_RWTUN,
2002-08-15 08:01:24 +00:00
&cbb_start_mem, CBB_START_MEM,
"Starting address for memory allocations");
2002-08-15 08:01:24 +00:00
u_long cbb_start_16_io = CBB_START_16_IO;
SYSCTL_ULONG(_hw_cbb, OID_AUTO, start_16_io, CTLFLAG_RWTUN,
2002-08-15 08:01:24 +00:00
&cbb_start_16_io, CBB_START_16_IO,
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
"Starting ioport for 16-bit cards");
2002-08-15 08:01:24 +00:00
u_long cbb_start_32_io = CBB_START_32_IO;
SYSCTL_ULONG(_hw_cbb, OID_AUTO, start_32_io, CTLFLAG_RWTUN,
2002-08-15 08:01:24 +00:00
&cbb_start_32_io, CBB_START_32_IO,
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
"Starting ioport for 32-bit cards");
int cbb_debug = 0;
SYSCTL_INT(_hw_cbb, OID_AUTO, debug, CTLFLAG_RWTUN, &cbb_debug, 0,
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
"Verbose cardbus bridge debugging");
2002-08-15 08:01:24 +00:00
static void cbb_insert(struct cbb_softc *sc);
static void cbb_removal(struct cbb_softc *sc);
static uint32_t cbb_detect_voltage(device_t brdev);
static int cbb_cardbus_reset_power(device_t brdev, device_t child, int on);
2002-08-15 08:01:24 +00:00
static int cbb_cardbus_io_open(device_t brdev, int win, uint32_t start,
uint32_t end);
2002-08-15 08:01:24 +00:00
static int cbb_cardbus_mem_open(device_t brdev, int win,
uint32_t start, uint32_t end);
2002-08-15 08:01:24 +00:00
static void cbb_cardbus_auto_open(struct cbb_softc *sc, int type);
static int cbb_cardbus_activate_resource(device_t brdev, device_t child,
int type, int rid, struct resource *res);
2002-08-15 08:01:24 +00:00
static int cbb_cardbus_deactivate_resource(device_t brdev,
device_t child, int type, int rid, struct resource *res);
2002-08-15 08:01:24 +00:00
static struct resource *cbb_cardbus_alloc_resource(device_t brdev,
device_t child, int type, int *rid, u_long start,
u_long end, u_long count, u_int flags);
2002-08-15 08:01:24 +00:00
static int cbb_cardbus_release_resource(device_t brdev, device_t child,
int type, int rid, struct resource *res);
static int cbb_cardbus_power_enable_socket(device_t brdev,
device_t child);
static int cbb_cardbus_power_disable_socket(device_t brdev,
device_t child);
static int cbb_func_filt(void *arg);
static void cbb_func_intr(void *arg);
static void
2002-08-15 08:01:24 +00:00
cbb_remove_res(struct cbb_softc *sc, struct resource *res)
{
2002-08-15 08:01:24 +00:00
struct cbb_reslist *rle;
SLIST_FOREACH(rle, &sc->rl, link) {
if (rle->res == res) {
2002-08-15 08:01:24 +00:00
SLIST_REMOVE(&sc->rl, rle, cbb_reslist, link);
free(rle, M_DEVBUF);
return;
}
}
}
static struct resource *
2002-08-15 08:01:24 +00:00
cbb_find_res(struct cbb_softc *sc, int type, int rid)
{
2002-08-15 08:01:24 +00:00
struct cbb_reslist *rle;
SLIST_FOREACH(rle, &sc->rl, link)
if (SYS_RES_MEMORY == rle->type && rid == rle->rid)
return (rle->res);
return (NULL);
}
static void
2002-08-15 08:01:24 +00:00
cbb_insert_res(struct cbb_softc *sc, struct resource *res, int type,
int rid)
{
2002-08-15 08:01:24 +00:00
struct cbb_reslist *rle;
/*
* Need to record allocated resource so we can iterate through
* it later.
*/
2002-08-15 08:01:24 +00:00
rle = malloc(sizeof(struct cbb_reslist), M_DEVBUF, M_NOWAIT);
if (rle == NULL)
2002-08-15 08:01:24 +00:00
panic("cbb_cardbus_alloc_resource: can't record entry!");
rle->res = res;
rle->type = type;
rle->rid = rid;
SLIST_INSERT_HEAD(&sc->rl, rle, link);
}
static void
2002-08-15 08:01:24 +00:00
cbb_destroy_res(struct cbb_softc *sc)
{
2002-08-15 08:01:24 +00:00
struct cbb_reslist *rle;
while ((rle = SLIST_FIRST(&sc->rl)) != NULL) {
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
device_printf(sc->dev, "Danger Will Robinson: Resource "
"left allocated! This is a bug... "
"(rid=%x, type=%d, addr=%lx)\n", rle->rid, rle->type,
rman_get_start(rle->res));
SLIST_REMOVE_HEAD(&sc->rl, link);
free(rle, M_DEVBUF);
}
}
/*
* Disable function interrupts by telling the bridge to generate IRQ1
* interrupts. These interrupts aren't really generated by the chip, since
* IRQ1 is reserved. Some chipsets assert INTA# inappropriately during
* initialization, so this helps to work around the problem.
*
* XXX We can't do this workaround for all chipsets, because this
* XXX causes interference with the keyboard because somechipsets will
* XXX actually signal IRQ1 over their serial interrupt connections to
* XXX the south bridge. Disable it it for now.
*/
void
cbb_disable_func_intr(struct cbb_softc *sc)
{
2003-11-08 03:28:43 +00:00
#if 0
uint8_t reg;
reg = (exca_getb(&sc->exca[0], EXCA_INTR) & ~EXCA_INTR_IRQ_MASK) |
EXCA_INTR_IRQ_RESERVED1;
exca_putb(&sc->exca[0], EXCA_INTR, reg);
#endif
}
/*
* Enable function interrupts. We turn on function interrupts when the card
* requests an interrupt. The PCMCIA standard says that we should set
* the lower 4 bits to 0 to route via PCI. Note: we call this for both
* CardBus and R2 (PC Card) cases, but it should have no effect on CardBus
* cards.
*/
static void
cbb_enable_func_intr(struct cbb_softc *sc)
{
uint8_t reg;
reg = (exca_getb(&sc->exca[0], EXCA_INTR) & ~EXCA_INTR_IRQ_MASK) |
EXCA_INTR_IRQ_NONE;
exca_putb(&sc->exca[0], EXCA_INTR, reg);
}
int
2002-08-15 08:01:24 +00:00
cbb_detach(device_t brdev)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
device_t *devlist;
int tmp, tries, error, numdevs;
/*
* Before we delete the children (which we have to do because
* attach doesn't check for children busses correctly), we have
* to detach the children. Even if we didn't need to delete the
* children, we have to detach them.
*/
error = bus_generic_detach(brdev);
if (error != 0)
return (error);
/*
* Since the attach routine doesn't search for children before it
* attaches them to this device, we must delete them here in order
* for the kldload/unload case to work. If we failed to do that, then
* we'd get duplicate devices when cbb.ko was reloaded.
*/
tries = 10;
do {
error = device_get_children(brdev, &devlist, &numdevs);
if (error == 0)
break;
/*
* Try hard to cope with low memory.
*/
if (error == ENOMEM) {
pause("cbbnomem", 1);
continue;
}
} while (tries-- > 0);
for (tmp = 0; tmp < numdevs; tmp++)
device_delete_child(brdev, devlist[tmp]);
free(devlist, M_TEMP);
/* Turn off the interrupts */
cbb_set(sc, CBB_SOCKET_MASK, 0);
/* reset 16-bit pcmcia bus */
exca_clrb(&sc->exca[0], EXCA_INTR, EXCA_INTR_RESET);
/* turn off power */
cbb_power(brdev, CARD_OFF);
/* Ack the interrupt */
cbb_set(sc, CBB_SOCKET_EVENT, 0xffffffff);
/*
* Wait for the thread to die. kproc_exit will do a wakeup
* on the event thread's struct proc * so that we know it is
* safe to proceed. IF the thread is running, set the please
* die flag and wait for it to comply. Since the wakeup on
* the event thread happens only in kproc_exit, we don't
* need to loop here.
*/
bus_teardown_intr(brdev, sc->irq_res, sc->intrhand);
mtx_lock(&sc->mtx);
2002-08-15 08:01:24 +00:00
sc->flags |= CBB_KTHREAD_DONE;
while (sc->flags & CBB_KTHREAD_RUNNING) {
DEVPRINTF((sc->dev, "Waiting for thread to die\n"));
wakeup(&sc->intrhand);
msleep(sc->event_thread, &sc->mtx, PWAIT, "cbbun", 0);
}
mtx_unlock(&sc->mtx);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
bus_release_resource(brdev, SYS_RES_IRQ, 0, sc->irq_res);
bus_release_resource(brdev, SYS_RES_MEMORY, CBBR_SOCKBASE,
sc->base_res);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
mtx_destroy(&sc->mtx);
return (0);
}
int
2002-08-15 08:01:24 +00:00
cbb_setup_intr(device_t dev, device_t child, struct resource *irq,
int flags, driver_filter_t *filt, driver_intr_t *intr, void *arg,
void **cookiep)
{
struct cbb_intrhand *ih;
struct cbb_softc *sc = device_get_softc(dev);
int err;
if (filt == NULL && intr == NULL)
return (EINVAL);
ih = malloc(sizeof(struct cbb_intrhand), M_DEVBUF, M_NOWAIT);
if (ih == NULL)
return (ENOMEM);
*cookiep = ih;
ih->filt = filt;
ih->intr = intr;
ih->arg = arg;
ih->sc = sc;
/*
* XXX need to turn on ISA interrupts, if we ever support them, but
* XXX for now that's all we need to do.
*/
err = BUS_SETUP_INTR(device_get_parent(dev), child, irq, flags,
filt ? cbb_func_filt : NULL, intr ? cbb_func_intr : NULL, ih,
&ih->cookie);
if (err != 0) {
free(ih, M_DEVBUF);
return (err);
}
cbb_enable_func_intr(sc);
sc->cardok = 1;
return 0;
}
int
2002-08-15 08:01:24 +00:00
cbb_teardown_intr(device_t dev, device_t child, struct resource *irq,
void *cookie)
{
struct cbb_intrhand *ih;
int err;
/* XXX Need to do different things for ISA interrupts. */
ih = (struct cbb_intrhand *) cookie;
err = BUS_TEARDOWN_INTR(device_get_parent(dev), child, irq,
ih->cookie);
if (err != 0)
return (err);
free(ih, M_DEVBUF);
return (0);
}
void
2002-08-15 08:01:24 +00:00
cbb_driver_added(device_t brdev, driver_t *driver)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
device_t *devlist;
device_t dev;
int tmp;
int numdevs;
int wake = 0;
DEVICE_IDENTIFY(driver, brdev);
tmp = device_get_children(brdev, &devlist, &numdevs);
if (tmp != 0) {
device_printf(brdev, "Cannot get children list, no reprobe\n");
return;
}
for (tmp = 0; tmp < numdevs; tmp++) {
dev = devlist[tmp];
if (device_get_state(dev) == DS_NOTPRESENT &&
device_probe_and_attach(dev) == 0)
wake++;
}
free(devlist, M_TEMP);
if (wake > 0)
wakeup(&sc->intrhand);
}
void
2002-08-15 08:01:24 +00:00
cbb_child_detached(device_t brdev, device_t child)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
/* I'm not sure we even need this */
if (child != sc->cbdev && child != sc->exca[0].pccarddev)
device_printf(brdev, "Unknown child detached: %s\n",
device_get_nameunit(child));
}
/************************************************************************/
/* Kthreads */
/************************************************************************/
void
2002-08-15 08:01:24 +00:00
cbb_event_thread(void *arg)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = arg;
uint32_t status;
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
int err;
int not_a_card = 0;
/*
* We need to act as a power sequencer on startup. Delay 2s/channel
* to ensure the other channels have had a chance to come up. We likely
* should add a lock that's shared on a per-slot basis so that only
* one power event can happen per slot at a time.
*/
pause("cbbstart", hz * device_get_unit(sc->dev) * 2);
mtx_lock(&sc->mtx);
2002-08-15 08:01:24 +00:00
sc->flags |= CBB_KTHREAD_RUNNING;
while ((sc->flags & CBB_KTHREAD_DONE) == 0) {
mtx_unlock(&sc->mtx);
/*
* We take out Giant here because we need it deep,
* down in the bowels of the vm system for mapping the
* memory we need to read the CIS. In addition, since
* we are adding/deleting devices from the dev tree,
* and that code isn't MP safe, we have to hold Giant.
*/
mtx_lock(&Giant);
status = cbb_get(sc, CBB_SOCKET_STATE);
DPRINTF(("Status is 0x%x\n", status));
if (!CBB_CARD_PRESENT(status)) {
not_a_card = 0; /* We know card type */
cbb_removal(sc);
} else if (status & CBB_STATE_NOT_A_CARD) {
/*
* Up to 10 times, try to rescan the card when we see
* NOT_A_CARD. 10 is somehwat arbitrary. When this
* pathology hits, there's a ~40% chance each try will
* fail. 10 tries takes about 5s and results in a
* 99.99% certainty of the results.
*/
if (not_a_card++ < 10) {
DEVPRINTF((sc->dev,
"Not a card bit set, rescanning\n"));
cbb_setb(sc, CBB_SOCKET_FORCE, CBB_FORCE_CV_TEST);
} else {
device_printf(sc->dev,
"Can't determine card type\n");
}
} else {
not_a_card = 0; /* We know card type */
cbb_insert(sc);
}
mtx_unlock(&Giant);
/*
* First time through we need to tell mountroot that we're
* done.
*/
if (sc->sc_root_token) {
root_mount_rel(sc->sc_root_token);
sc->sc_root_token = NULL;
}
/*
* Wait until it has been 250ms since the last time we
* get an interrupt. We handle the rest of the interrupt
* at the top of the loop. Although we clear the bit in the
* ISR, we signal sc->cv from the detach path after we've
* set the CBB_KTHREAD_DONE bit, so we can't do a simple
* 250ms sleep here.
*
* In our ISR, we turn off the card changed interrupt. Turn
* them back on here before we wait for them to happen. We
* turn them on/off so that we can tolerate a large latency
* between the time we signal cbb_event_thread and it gets
* a chance to run.
*/
mtx_lock(&sc->mtx);
cbb_setb(sc, CBB_SOCKET_MASK, CBB_SOCKET_MASK_CD | CBB_SOCKET_MASK_CSTS);
msleep(&sc->intrhand, &sc->mtx, 0, "-", 0);
err = 0;
while (err != EWOULDBLOCK &&
(sc->flags & CBB_KTHREAD_DONE) == 0)
err = msleep(&sc->intrhand, &sc->mtx, 0, "-", hz / 5);
}
DEVPRINTF((sc->dev, "Thread terminating\n"));
2002-08-15 08:01:24 +00:00
sc->flags &= ~CBB_KTHREAD_RUNNING;
mtx_unlock(&sc->mtx);
kproc_exit(0);
}
/************************************************************************/
/* Insert/removal */
/************************************************************************/
static void
2002-08-15 08:01:24 +00:00
cbb_insert(struct cbb_softc *sc)
{
uint32_t sockevent, sockstate;
sockevent = cbb_get(sc, CBB_SOCKET_EVENT);
sockstate = cbb_get(sc, CBB_SOCKET_STATE);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
DEVPRINTF((sc->dev, "card inserted: event=0x%08x, state=%08x\n",
sockevent, sockstate));
if (sockstate & CBB_STATE_R2_CARD) {
if (device_is_attached(sc->exca[0].pccarddev)) {
sc->flags |= CBB_16BIT_CARD;
exca_insert(&sc->exca[0]);
} else {
device_printf(sc->dev,
"16-bit card inserted, but no pccard bus.\n");
}
} else if (sockstate & CBB_STATE_CB_CARD) {
if (device_is_attached(sc->cbdev)) {
2002-08-15 08:01:24 +00:00
sc->flags &= ~CBB_16BIT_CARD;
CARD_ATTACH_CARD(sc->cbdev);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
} else {
device_printf(sc->dev,
"CardBus card inserted, but no cardbus bus.\n");
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
}
} else {
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
/*
* We should power the card down, and try again a couple of
* times if this happens. XXX
*/
device_printf(sc->dev, "Unsupported card type detected\n");
}
}
static void
2002-08-15 08:01:24 +00:00
cbb_removal(struct cbb_softc *sc)
{
sc->cardok = 0;
if (sc->flags & CBB_16BIT_CARD) {
exca_removal(&sc->exca[0]);
} else {
if (device_is_attached(sc->cbdev))
CARD_DETACH_CARD(sc->cbdev);
}
2002-08-15 08:01:24 +00:00
cbb_destroy_res(sc);
}
/************************************************************************/
/* Interrupt Handler */
/************************************************************************/
static int
cbb_func_filt(void *arg)
{
struct cbb_intrhand *ih = (struct cbb_intrhand *)arg;
struct cbb_softc *sc = ih->sc;
/*
* Make sure that the card is really there.
*/
if (!sc->cardok)
return (FILTER_STRAY);
if (!CBB_CARD_PRESENT(cbb_get(sc, CBB_SOCKET_STATE))) {
sc->cardok = 0;
return (FILTER_HANDLED);
}
/*
* nb: don't have to check for giant or not, since that's done in the
* ISR dispatch and one can't hold Giant in a filter anyway...
*/
return ((*ih->filt)(ih->arg));
}
static void
cbb_func_intr(void *arg)
{
struct cbb_intrhand *ih = (struct cbb_intrhand *)arg;
struct cbb_softc *sc = ih->sc;
/*
* While this check may seem redundant, it helps close a race
* condition. If the card is ejected after the filter runs, but
* before this ISR can be scheduled, then we need to do the same
* filtering to prevent the card's ISR from being called. One could
* argue that the card's ISR should be able to cope, but experience
* has shown they can't always. This mitigates the problem by making
* the race quite a bit smaller. Properly written client ISRs should
* cope with the card going away in the middle of the ISR. We assume
* that drivers that are sophisticated enough to use filters don't
* need our protection. This also allows us to ensure they *ARE*
* called if their filter said they needed to be called.
*/
if (ih->filt == NULL) {
if (!sc->cardok)
return;
if (!CBB_CARD_PRESENT(cbb_get(sc, CBB_SOCKET_STATE))) {
sc->cardok = 0;
return;
}
}
/*
* Call the registered ithread interrupt handler. This entire routine
* will be called with Giant if this isn't an MP safe driver, or not
* if it is. Either way, we don't have to worry.
*/
ih->intr(ih->arg);
}
/************************************************************************/
/* Generic Power functions */
/************************************************************************/
static uint32_t
2002-08-15 08:01:24 +00:00
cbb_detect_voltage(device_t brdev)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
uint32_t psr;
uint32_t vol = CARD_UKN_CARD;
2002-08-15 08:01:24 +00:00
psr = cbb_get(sc, CBB_SOCKET_STATE);
if (psr & CBB_STATE_5VCARD && psr & CBB_STATE_5VSOCK)
vol |= CARD_5V_CARD;
if (psr & CBB_STATE_3VCARD && psr & CBB_STATE_3VSOCK)
vol |= CARD_3V_CARD;
if (psr & CBB_STATE_XVCARD && psr & CBB_STATE_XVSOCK)
vol |= CARD_XV_CARD;
if (psr & CBB_STATE_YVCARD && psr & CBB_STATE_YVSOCK)
vol |= CARD_YV_CARD;
return (vol);
}
static uint8_t
cbb_o2micro_power_hack(struct cbb_softc *sc)
{
uint8_t reg;
/*
* Issue #2: INT# not qualified with IRQ Routing Bit. An
2005-09-22 06:01:44 +00:00
* unexpected PCI INT# may be generated during PC Card
* initialization even with the IRQ Routing Bit Set with some
2005-09-22 06:01:44 +00:00
* PC Cards.
*
* This is a two part issue. The first part is that some of
* our older controllers have an issue in which the slot's PCI
* INT# is NOT qualified by the IRQ routing bit (PCI reg. 3Eh
* bit 7). Regardless of the IRQ routing bit, if NO ISA IRQ
* is selected (ExCA register 03h bits 3:0, of the slot, are
* cleared) we will generate INT# if IREQ# is asserted. The
2005-09-22 06:01:44 +00:00
* second part is because some PC Cards prematurally assert
* IREQ# before the ExCA registers are fully programmed. This
* in turn asserts INT# because ExCA register 03h bits 3:0
* (ISA IRQ Select) are not yet programmed.
*
* The fix for this issue, which will work for any controller
* (old or new), is to set ExCA register 03h bits 3:0 = 0001b
* (select IRQ1), of the slot, before turning on slot power.
* Selecting IRQ1 will result in INT# NOT being asserted
* (because IRQ1 is selected), and IRQ1 won't be asserted
* because our controllers don't generate IRQ1.
*
* Other, non O2Micro controllers will generate irq 1 in some
* situations, so we can't do this hack for everybody. Reports of
* keyboard controller's interrupts being suppressed occurred when
* we did this.
*/
reg = exca_getb(&sc->exca[0], EXCA_INTR);
exca_putb(&sc->exca[0], EXCA_INTR, (reg & 0xf0) | 1);
return (reg);
}
/*
* Restore the damage that cbb_o2micro_power_hack does to EXCA_INTR so
* we don't have an interrupt storm on power on. This has the efect of
* disabling card status change interrupts for the duration of poweron.
*/
static void
cbb_o2micro_power_hack2(struct cbb_softc *sc, uint8_t reg)
{
exca_putb(&sc->exca[0], EXCA_INTR, reg);
}
int
2002-08-15 08:01:24 +00:00
cbb_power(device_t brdev, int volts)
{
uint32_t status, sock_ctrl, reg_ctrl, mask;
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int cnt, sane;
int retval = 0;
int on = 0;
uint8_t reg = 0;
2002-08-15 08:01:24 +00:00
sock_ctrl = cbb_get(sc, CBB_SOCKET_CONTROL);
sock_ctrl &= ~CBB_SOCKET_CTRL_VCCMASK;
switch (volts & CARD_VCCMASK) {
case 5:
sock_ctrl |= CBB_SOCKET_CTRL_VCC_5V;
on++;
break;
case 3:
sock_ctrl |= CBB_SOCKET_CTRL_VCC_3V;
on++;
break;
case XV:
sock_ctrl |= CBB_SOCKET_CTRL_VCC_XV;
on++;
break;
case YV:
sock_ctrl |= CBB_SOCKET_CTRL_VCC_YV;
on++;
break;
case 0:
break;
default:
return (0); /* power NEVER changed */
}
/* VPP == VCC */
sock_ctrl &= ~CBB_SOCKET_CTRL_VPPMASK;
sock_ctrl |= ((sock_ctrl >> 4) & 0x07);
2002-08-15 08:01:24 +00:00
if (cbb_get(sc, CBB_SOCKET_CONTROL) == sock_ctrl)
return (1); /* no change necessary */
DEVPRINTF((sc->dev, "cbb_power: %dV\n", volts));
if (volts != 0 && sc->chipset == CB_O2MICRO)
reg = cbb_o2micro_power_hack(sc);
/*
* We have to mask the card change detect interrupt while we're
* messing with the power. It is allowed to bounce while we're
* messing with power as things settle down. In addition, we mask off
* the card's function interrupt by routing it via the ISA bus. This
* bit generally only affects 16-bit cards. Some bridges allow one to
* set another bit to have it also affect 32-bit cards. Since 32-bit
* cards are required to be better behaved, we don't bother to get
* into those bridge specific features.
*
* XXX I wonder if we need to enable the READY bit interrupt in the
* EXCA CSC register for 16-bit cards, and disable the CD bit?
*/
mask = cbb_get(sc, CBB_SOCKET_MASK);
mask |= CBB_SOCKET_MASK_POWER;
mask &= ~CBB_SOCKET_MASK_CD;
cbb_set(sc, CBB_SOCKET_MASK, mask);
PCI_MASK_CONFIG(brdev, CBBR_BRIDGECTRL,
|CBBM_BRIDGECTRL_INTR_IREQ_ISA_EN, 2);
2002-08-15 08:01:24 +00:00
cbb_set(sc, CBB_SOCKET_CONTROL, sock_ctrl);
if (on) {
mtx_lock(&sc->mtx);
cnt = sc->powerintr;
/*
* We have a shortish timeout of 500ms here. Some bridges do
* not generate a POWER_CYCLE event for 16-bit cards. In
* those cases, we have to cope the best we can, and having
* only a short delay is better than the alternatives. Others
* raise the power cycle a smidge before it is really ready.
* We deal with those below.
*/
sane = 10;
while (!(cbb_get(sc, CBB_SOCKET_STATE) & CBB_STATE_POWER_CYCLE) &&
cnt == sc->powerintr && sane-- > 0)
msleep(&sc->powerintr, &sc->mtx, 0, "-", hz / 20);
mtx_unlock(&sc->mtx);
/*
* Relax for 100ms. Some bridges appear to assert this signal
* right away, but before the card has stabilized. Other
* cards need need more time to cope up reliabily.
* Experiments with troublesome setups show this to be a
* "cheap" way to enhance reliabilty. We need not do this for
* "off" since we don't touch the card after we turn it off.
*/
pause("cbbPwr", min(hz / 10, 1));
/*
* The TOPIC95B requires a little bit extra time to get its
* act together, so delay for an additional 100ms. Also as
* documented below, it doesn't seem to set the POWER_CYCLE
* bit, so don't whine if it never came on.
*/
if (sc->chipset == CB_TOPIC95)
pause("cbb95B", hz / 10);
else if (sane <= 0)
device_printf(sc->dev, "power timeout, doom?\n");
}
/*
* After the power is good, we can turn off the power interrupt.
* However, the PC Card standard says that we must delay turning the
* CD bit back on for a bit to allow for bouncyness on power down
* (recall that we don't wait above for a power down, since we don't
* get an interrupt for that). We're called either from the suspend
* code in which case we don't want to turn card change on again, or
* we're called from the card insertion code, in which case the cbb
* thread will turn it on for us before it waits to be woken by a
* change event.
*
* NB: Topic95B doesn't set the power cycle bit. we assume that
* both it and the TOPIC95 behave the same.
*/
cbb_clrb(sc, CBB_SOCKET_MASK, CBB_SOCKET_MASK_POWER);
2002-08-15 08:01:24 +00:00
status = cbb_get(sc, CBB_SOCKET_STATE);
if (on && sc->chipset != CB_TOPIC95) {
if ((status & CBB_STATE_POWER_CYCLE) == 0)
device_printf(sc->dev, "Power not on?\n");
}
if (status & CBB_STATE_BAD_VCC_REQ) {
device_printf(sc->dev, "Bad Vcc requested\n");
/*
* Turn off the power, and try again. Retrigger other
* active interrupts via force register. From NetBSD
* PR 36652, coded by me to description there.
*/
sock_ctrl &= ~CBB_SOCKET_CTRL_VCCMASK;
sock_ctrl &= ~CBB_SOCKET_CTRL_VPPMASK;
cbb_set(sc, CBB_SOCKET_CONTROL, sock_ctrl);
status &= ~CBB_STATE_BAD_VCC_REQ;
status &= ~CBB_STATE_DATA_LOST;
status |= CBB_FORCE_CV_TEST;
cbb_set(sc, CBB_SOCKET_FORCE, status);
goto done;
}
if (sc->chipset == CB_TOPIC97) {
reg_ctrl = pci_read_config(sc->dev, TOPIC_REG_CTRL, 4);
reg_ctrl &= ~TOPIC97_REG_CTRL_TESTMODE;
if (on)
reg_ctrl |= TOPIC97_REG_CTRL_CLKRUN_ENA;
else
reg_ctrl &= ~TOPIC97_REG_CTRL_CLKRUN_ENA;
pci_write_config(sc->dev, TOPIC_REG_CTRL, reg_ctrl, 4);
}
PCI_MASK_CONFIG(brdev, CBBR_BRIDGECTRL,
& ~CBBM_BRIDGECTRL_INTR_IREQ_ISA_EN, 2);
retval = 1;
done:;
if (volts != 0 && sc->chipset == CB_O2MICRO)
cbb_o2micro_power_hack2(sc, reg);
return (retval);
}
static int
cbb_current_voltage(device_t brdev)
{
struct cbb_softc *sc = device_get_softc(brdev);
uint32_t ctrl;
ctrl = cbb_get(sc, CBB_SOCKET_CONTROL);
switch (ctrl & CBB_SOCKET_CTRL_VCCMASK) {
case CBB_SOCKET_CTRL_VCC_5V:
return CARD_5V_CARD;
case CBB_SOCKET_CTRL_VCC_3V:
return CARD_3V_CARD;
case CBB_SOCKET_CTRL_VCC_XV:
return CARD_XV_CARD;
case CBB_SOCKET_CTRL_VCC_YV:
return CARD_YV_CARD;
}
return 0;
}
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
/*
* detect the voltage for the card, and set it. Since the power
* used is the square of the voltage, lower voltages is a big win
* and what Windows does (and what Microsoft prefers). The MS paper
* also talks about preferring the CIS entry as well, but that has
* to be done elsewhere. We also optimize power sequencing here
* and don't change things if we're already powered up at a supported
* voltage.
*
* In addition, we power up with OE disabled. We'll set it later
* in the power up sequence.
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
*/
static int
2002-08-15 08:01:24 +00:00
cbb_do_power(device_t brdev)
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
{
struct cbb_softc *sc = device_get_softc(brdev);
uint32_t voltage, curpwr;
uint32_t status;
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
/* Don't enable OE (output enable) until power stable */
exca_clrb(&sc->exca[0], EXCA_PWRCTL, EXCA_PWRCTL_OE);
2002-08-15 08:01:24 +00:00
voltage = cbb_detect_voltage(brdev);
curpwr = cbb_current_voltage(brdev);
status = cbb_get(sc, CBB_SOCKET_STATE);
if ((status & CBB_STATE_POWER_CYCLE) && (voltage & curpwr))
return 0;
/* Prefer lowest voltage supported */
cbb_power(brdev, CARD_OFF);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
if (voltage & CARD_YV_CARD)
cbb_power(brdev, CARD_VCC(YV));
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
else if (voltage & CARD_XV_CARD)
cbb_power(brdev, CARD_VCC(XV));
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
else if (voltage & CARD_3V_CARD)
cbb_power(brdev, CARD_VCC(3));
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
else if (voltage & CARD_5V_CARD)
cbb_power(brdev, CARD_VCC(5));
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
else {
device_printf(brdev, "Unknown card voltage\n");
return (ENXIO);
}
return (0);
}
/************************************************************************/
/* CardBus power functions */
/************************************************************************/
static int
cbb_cardbus_reset_power(device_t brdev, device_t child, int on)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
uint32_t b, h;
int delay, count, zero_seen, func;
/*
* Asserting reset for 20ms is necessary for most bridges. For some
* reason, the Ricoh RF5C47x bridges need it asserted for 400ms. The
* root cause of this is unknown, and NetBSD does the same thing.
*/
delay = sc->chipset == CB_RF5C47X ? 400 : 20;
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
PCI_MASK_CONFIG(brdev, CBBR_BRIDGECTRL, |CBBM_BRIDGECTRL_RESET, 2);
pause("cbbP3", hz * delay / 1000);
/*
* If a card exists and we're turning it on, take it out of reset.
* After clearing reset, wait up to 1.1s for the first configuration
* register (vendor/product) configuration register of device 0.0 to
* become != 0xffffffff. The PCMCIA PC Card Host System Specification
* says that when powering up the card, the PCI Spec v2.1 must be
* followed. In PCI spec v2.2 Table 4-6, Trhfa (Reset High to first
* Config Access) is at most 2^25 clocks, or just over 1s. Section
* 2.2.1 states any card not ready to participate in bus transactions
* must tristate its outputs. Therefore, any access to its
* configuration registers must be ignored. In that state, the config
* reg will read 0xffffffff. Section 6.2.1 states a vendor id of
* 0xffff is invalid, so this can never match a real card. Print a
* warning if it never returns a real id. The PCMCIA PC Card
* Electrical Spec Section 5.2.7.1 implies only device 0 is present on
* a cardbus bus, so that's the only register we check here.
*/
if (on && CBB_CARD_PRESENT(cbb_get(sc, CBB_SOCKET_STATE))) {
PCI_MASK_CONFIG(brdev, CBBR_BRIDGECTRL,
&~CBBM_BRIDGECTRL_RESET, 2);
b = pcib_get_bus(child);
count = 1100 / 20;
do {
pause("cbbP4", hz * 2 / 100);
} while (PCIB_READ_CONFIG(brdev, b, 0, 0, PCIR_DEVVENDOR, 4) ==
0xfffffffful && --count >= 0);
if (count < 0)
device_printf(brdev, "Warning: Bus reset timeout\n");
/*
* Some cards (so far just an atheros card I have) seem to
* come out of reset in a funky state. They report they are
* multi-function cards, but have nonsense for some of the
* higher functions. So if the card claims to be MFDEV, and
* any of the higher functions' ID is 0, then we've hit the
* bug and we'll try again.
*/
h = PCIB_READ_CONFIG(brdev, b, 0, 0, PCIR_HDRTYPE, 1);
if ((h & PCIM_MFDEV) == 0)
return 0;
zero_seen = 0;
for (func = 1; func < 8; func++) {
h = PCIB_READ_CONFIG(brdev, b, 0, func,
PCIR_DEVVENDOR, 4);
if (h == 0)
zero_seen++;
}
if (!zero_seen)
return 0;
return (EINVAL);
}
return 0;
}
static int
cbb_cardbus_power_disable_socket(device_t brdev, device_t child)
{
cbb_power(brdev, CARD_OFF);
cbb_cardbus_reset_power(brdev, child, 0);
return (0);
}
static int
2002-08-15 08:01:24 +00:00
cbb_cardbus_power_enable_socket(device_t brdev, device_t child)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int err, count;
if (!CBB_CARD_PRESENT(cbb_get(sc, CBB_SOCKET_STATE)))
return (ENODEV);
count = 10;
do {
err = cbb_do_power(brdev);
if (err)
return (err);
err = cbb_cardbus_reset_power(brdev, child, 1);
if (err) {
device_printf(brdev, "Reset failed, trying again.\n");
cbb_cardbus_power_disable_socket(brdev, child);
pause("cbbErr1", hz / 10); /* wait 100ms */
}
} while (err != 0 && count-- > 0);
return (0);
}
/************************************************************************/
/* CardBus Resource */
/************************************************************************/
static void
cbb_activate_window(device_t brdev, int type)
{
PCI_ENABLE_IO(device_get_parent(brdev), brdev, type);
}
static int
2002-08-15 08:01:24 +00:00
cbb_cardbus_io_open(device_t brdev, int win, uint32_t start, uint32_t end)
{
int basereg;
int limitreg;
if ((win < 0) || (win > 1)) {
DEVPRINTF((brdev,
2002-08-15 08:01:24 +00:00
"cbb_cardbus_io_open: window out of range %d\n", win));
return (EINVAL);
}
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
basereg = win * 8 + CBBR_IOBASE0;
limitreg = win * 8 + CBBR_IOLIMIT0;
pci_write_config(brdev, basereg, start, 4);
pci_write_config(brdev, limitreg, end, 4);
cbb_activate_window(brdev, SYS_RES_IOPORT);
return (0);
}
static int
2002-08-15 08:01:24 +00:00
cbb_cardbus_mem_open(device_t brdev, int win, uint32_t start, uint32_t end)
{
int basereg;
int limitreg;
if ((win < 0) || (win > 1)) {
DEVPRINTF((brdev,
2002-08-15 08:01:24 +00:00
"cbb_cardbus_mem_open: window out of range %d\n", win));
return (EINVAL);
}
2008-12-05 04:48:04 +00:00
basereg = win * 8 + CBBR_MEMBASE0;
limitreg = win * 8 + CBBR_MEMLIMIT0;
pci_write_config(brdev, basereg, start, 4);
pci_write_config(brdev, limitreg, end, 4);
cbb_activate_window(brdev, SYS_RES_MEMORY);
return (0);
}
#define START_NONE 0xffffffff
#define END_NONE 0
static void
2002-08-15 08:01:24 +00:00
cbb_cardbus_auto_open(struct cbb_softc *sc, int type)
{
uint32_t starts[2];
uint32_t ends[2];
2002-08-15 08:01:24 +00:00
struct cbb_reslist *rle;
int align, i;
uint32_t reg;
starts[0] = starts[1] = START_NONE;
ends[0] = ends[1] = END_NONE;
if (type == SYS_RES_MEMORY)
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
align = CBB_MEMALIGN;
else if (type == SYS_RES_IOPORT)
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
align = CBB_IOALIGN;
else
align = 1;
SLIST_FOREACH(rle, &sc->rl, link) {
if (rle->type != type)
continue;
if (rle->res == NULL)
continue;
if (!(rman_get_flags(rle->res) & RF_ACTIVE))
continue;
if (rman_get_flags(rle->res) & RF_PREFETCHABLE)
i = 1;
else
i = 0;
if (rman_get_start(rle->res) < starts[i])
starts[i] = rman_get_start(rle->res);
if (rman_get_end(rle->res) > ends[i])
ends[i] = rman_get_end(rle->res);
}
for (i = 0; i < 2; i++) {
if (starts[i] == START_NONE)
continue;
starts[i] &= ~(align - 1);
ends[i] = ((ends[i] + align - 1) & ~(align - 1)) - 1;
}
if (starts[0] != START_NONE && starts[1] != START_NONE) {
if (starts[0] < starts[1]) {
if (ends[0] > starts[1]) {
device_printf(sc->dev, "Overlapping ranges"
" for prefetch and non-prefetch memory\n");
return;
}
} else {
if (ends[1] > starts[0]) {
device_printf(sc->dev, "Overlapping ranges"
" for prefetch and non-prefetch memory\n");
return;
}
}
}
if (type == SYS_RES_MEMORY) {
2002-08-15 08:01:24 +00:00
cbb_cardbus_mem_open(sc->dev, 0, starts[0], ends[0]);
cbb_cardbus_mem_open(sc->dev, 1, starts[1], ends[1]);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
reg = pci_read_config(sc->dev, CBBR_BRIDGECTRL, 2);
reg &= ~(CBBM_BRIDGECTRL_PREFETCH_0 |
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
CBBM_BRIDGECTRL_PREFETCH_1);
if (starts[1] != START_NONE)
reg |= CBBM_BRIDGECTRL_PREFETCH_1;
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
pci_write_config(sc->dev, CBBR_BRIDGECTRL, reg, 2);
if (bootverbose) {
device_printf(sc->dev, "Opening memory:\n");
if (starts[0] != START_NONE)
device_printf(sc->dev, "Normal: %#x-%#x\n",
starts[0], ends[0]);
if (starts[1] != START_NONE)
device_printf(sc->dev, "Prefetch: %#x-%#x\n",
starts[1], ends[1]);
}
} else if (type == SYS_RES_IOPORT) {
2002-08-15 08:01:24 +00:00
cbb_cardbus_io_open(sc->dev, 0, starts[0], ends[0]);
cbb_cardbus_io_open(sc->dev, 1, starts[1], ends[1]);
if (bootverbose && starts[0] != START_NONE)
device_printf(sc->dev, "Opening I/O: %#x-%#x\n",
starts[0], ends[0]);
}
}
static int
2002-08-15 08:01:24 +00:00
cbb_cardbus_activate_resource(device_t brdev, device_t child, int type,
int rid, struct resource *res)
{
int ret;
ret = BUS_ACTIVATE_RESOURCE(device_get_parent(brdev), child,
type, rid, res);
if (ret != 0)
return (ret);
2002-08-15 08:01:24 +00:00
cbb_cardbus_auto_open(device_get_softc(brdev), type);
return (0);
}
static int
2002-08-15 08:01:24 +00:00
cbb_cardbus_deactivate_resource(device_t brdev, device_t child, int type,
int rid, struct resource *res)
{
int ret;
ret = BUS_DEACTIVATE_RESOURCE(device_get_parent(brdev), child,
type, rid, res);
if (ret != 0)
return (ret);
2002-08-15 08:01:24 +00:00
cbb_cardbus_auto_open(device_get_softc(brdev), type);
return (0);
}
static struct resource *
2002-08-15 08:01:24 +00:00
cbb_cardbus_alloc_resource(device_t brdev, device_t child, int type,
int *rid, u_long start, u_long end, u_long count, u_int flags)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int tmp;
struct resource *res;
u_long align;
switch (type) {
case SYS_RES_IRQ:
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
tmp = rman_get_start(sc->irq_res);
if (start > tmp || end < tmp || count != 1) {
device_printf(child, "requested interrupt %ld-%ld,"
2002-08-15 08:01:24 +00:00
"count = %ld not supported by cbb\n",
start, end, count);
return (NULL);
}
start = end = tmp;
flags |= RF_SHAREABLE;
break;
case SYS_RES_IOPORT:
2002-08-15 08:01:24 +00:00
if (start <= cbb_start_32_io)
start = cbb_start_32_io;
if (end < start)
end = start;
if (count > (1 << RF_ALIGNMENT(flags)))
flags = (flags & ~RF_ALIGNMENT_MASK) |
rman_make_alignment_flags(count);
break;
case SYS_RES_MEMORY:
2002-08-15 08:01:24 +00:00
if (start <= cbb_start_mem)
start = cbb_start_mem;
if (end < start)
end = start;
if (count < CBB_MEMALIGN)
align = CBB_MEMALIGN;
else
align = count;
if (align > (1 << RF_ALIGNMENT(flags)))
flags = (flags & ~RF_ALIGNMENT_MASK) |
rman_make_alignment_flags(align);
break;
}
res = BUS_ALLOC_RESOURCE(device_get_parent(brdev), child, type, rid,
start, end, count, flags & ~RF_ACTIVE);
if (res == NULL) {
printf("cbb alloc res fail type %d rid %x\n", type, *rid);
return (NULL);
}
2002-08-15 08:01:24 +00:00
cbb_insert_res(sc, res, type, *rid);
if (flags & RF_ACTIVE)
if (bus_activate_resource(child, type, *rid, res) != 0) {
bus_release_resource(child, type, *rid, res);
return (NULL);
}
return (res);
}
static int
2002-08-15 08:01:24 +00:00
cbb_cardbus_release_resource(device_t brdev, device_t child, int type,
int rid, struct resource *res)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int error;
if (rman_get_flags(res) & RF_ACTIVE) {
error = bus_deactivate_resource(child, type, rid, res);
if (error != 0)
return (error);
}
2002-08-15 08:01:24 +00:00
cbb_remove_res(sc, res);
return (BUS_RELEASE_RESOURCE(device_get_parent(brdev), child,
type, rid, res));
}
/************************************************************************/
/* PC Card Power Functions */
/************************************************************************/
static int
2002-08-15 08:01:24 +00:00
cbb_pcic_power_enable_socket(device_t brdev, device_t child)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
int err;
2002-08-15 08:01:24 +00:00
DPRINTF(("cbb_pcic_socket_enable:\n"));
/* power down/up the socket to reset */
2002-08-15 08:01:24 +00:00
err = cbb_do_power(brdev);
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
if (err)
return (err);
exca_reset(&sc->exca[0], child);
return (0);
}
static int
2002-08-15 08:01:24 +00:00
cbb_pcic_power_disable_socket(device_t brdev, device_t child)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
2002-08-15 08:01:24 +00:00
DPRINTF(("cbb_pcic_socket_disable\n"));
/* Turn off the card's interrupt and leave it in reset, wait 10ms */
exca_putb(&sc->exca[0], EXCA_INTR, 0);
pause("cbbP1", hz / 100);
/* power down the socket */
cbb_power(brdev, CARD_OFF);
exca_putb(&sc->exca[0], EXCA_PWRCTL, 0);
/* wait 300ms until power fails (Tpf). */
pause("cbbP2", hz * 300 / 1000);
/* enable CSC interrupts */
exca_putb(&sc->exca[0], EXCA_INTR, EXCA_INTR_ENABLE);
return (0);
}
/************************************************************************/
/* POWER methods */
/************************************************************************/
int
2002-08-15 08:01:24 +00:00
cbb_power_enable_socket(device_t brdev, device_t child)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
2002-08-15 08:01:24 +00:00
if (sc->flags & CBB_16BIT_CARD)
return (cbb_pcic_power_enable_socket(brdev, child));
return (cbb_cardbus_power_enable_socket(brdev, child));
}
int
2002-08-15 08:01:24 +00:00
cbb_power_disable_socket(device_t brdev, device_t child)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
if (sc->flags & CBB_16BIT_CARD)
return (cbb_pcic_power_disable_socket(brdev, child));
return (cbb_cardbus_power_disable_socket(brdev, child));
}
static int
2002-08-15 08:01:24 +00:00
cbb_pcic_activate_resource(device_t brdev, device_t child, int type, int rid,
struct resource *res)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int error;
error = exca_activate_resource(&sc->exca[0], child, type, rid, res);
if (error == 0)
cbb_activate_window(brdev, type);
return (error);
}
static int
2002-08-15 08:01:24 +00:00
cbb_pcic_deactivate_resource(device_t brdev, device_t child, int type,
int rid, struct resource *res)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
return (exca_deactivate_resource(&sc->exca[0], child, type, rid, res));
}
static struct resource *
2002-08-15 08:01:24 +00:00
cbb_pcic_alloc_resource(device_t brdev, device_t child, int type, int *rid,
u_long start, u_long end, u_long count, u_int flags)
{
struct resource *res = NULL;
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int align;
int tmp;
switch (type) {
case SYS_RES_MEMORY:
2002-08-15 08:01:24 +00:00
if (start < cbb_start_mem)
start = cbb_start_mem;
if (end < start)
end = start;
if (count < CBB_MEMALIGN)
align = CBB_MEMALIGN;
else
align = count;
if (align > (1 << RF_ALIGNMENT(flags)))
flags = (flags & ~RF_ALIGNMENT_MASK) |
rman_make_alignment_flags(align);
break;
case SYS_RES_IOPORT:
2002-08-15 08:01:24 +00:00
if (start < cbb_start_16_io)
start = cbb_start_16_io;
if (end < start)
end = start;
break;
case SYS_RES_IRQ:
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
tmp = rman_get_start(sc->irq_res);
if (start > tmp || end < tmp || count != 1) {
device_printf(child, "requested interrupt %ld-%ld,"
2002-08-15 08:01:24 +00:00
"count = %ld not supported by cbb\n",
start, end, count);
return (NULL);
}
flags |= RF_SHAREABLE;
Bulk changes made during the BSDcon kernel summit and travel afterwards. Appologies for making this one bulk commit, but I have tested all these changes together and don't want to break anything by trying to disentangle it. o Make debugging a sysctl/tunable o Remove flags word from yenta chip info, it is unused o Make 16-bit card I/O range and 32-bit card I/O range tunables o Start the rename of pccbb to cbb to match NetBSD by misc renames. o Kill the now bogus list of softcs to create kthread. Instead, just create the kthread in the attach routine. o Remove sc_ from some structure names. It isn't needed. o Refine chipset lookup code. o Match generic PCI <-> CardBus bridges. We specifically don't generically match PCI PCMCIA bridges because they are not, with one exception, yenta devices. o Add some comments about the why we need to have a function table ala OLDCARD o The PCI interrupt routing by using the ExCA registers is needed for for all bridges, per the spec, not just TI ones. o Collapse TOPIC95 and TOPIC95B. o Using the ToPIC 97 and 100 datasheets, try to support these bridges better, but more work is needed. o Generally clarify some XXX comments and add them in a few places where things didn't look right to me. o Move interrupt generating register access until after we establish an ISR. o Add support for YV and XV cards. X and Y are numbers to be determined later (but maybe never). o factor powerup code for 16-bit and 32-bit cards. o When a card supports more than one voltage, prefer the lowest supported volage. Windows does this, and MS's design guides imply this is the right thing to do. o Document race between kthread_exit(0) and kldunload's unmapping of pages that John Baldwin and I discovered. o Debounce the CSC interrupt a little better. o When a 16-bit card is inserted when we don't have a pccard child, warn about it better. Ditto for 32-bit card. o Ack ALL the interrupt bits that we get, not just 0x1. o maybe a couple minor style nits corrected.
2002-02-17 03:11:11 +00:00
start = end = rman_get_start(sc->irq_res);
break;
}
res = BUS_ALLOC_RESOURCE(device_get_parent(brdev), child, type, rid,
start, end, count, flags & ~RF_ACTIVE);
if (res == NULL)
return (NULL);
2002-08-15 08:01:24 +00:00
cbb_insert_res(sc, res, type, *rid);
if (flags & RF_ACTIVE) {
if (bus_activate_resource(child, type, *rid, res) != 0) {
bus_release_resource(child, type, *rid, res);
return (NULL);
}
}
return (res);
}
static int
2002-08-15 08:01:24 +00:00
cbb_pcic_release_resource(device_t brdev, device_t child, int type,
int rid, struct resource *res)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
int error;
if (rman_get_flags(res) & RF_ACTIVE) {
error = bus_deactivate_resource(child, type, rid, res);
if (error != 0)
return (error);
}
2002-08-15 08:01:24 +00:00
cbb_remove_res(sc, res);
return (BUS_RELEASE_RESOURCE(device_get_parent(brdev), child,
type, rid, res));
}
/************************************************************************/
/* PC Card methods */
/************************************************************************/
int
2002-08-15 08:01:24 +00:00
cbb_pcic_set_res_flags(device_t brdev, device_t child, int type, int rid,
u_long flags)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
struct resource *res;
if (type != SYS_RES_MEMORY)
return (EINVAL);
2002-08-15 08:01:24 +00:00
res = cbb_find_res(sc, type, rid);
if (res == NULL) {
device_printf(brdev,
"set_res_flags: specified rid not found\n");
return (ENOENT);
}
return (exca_mem_set_flags(&sc->exca[0], res, flags));
}
int
2002-08-15 08:01:24 +00:00
cbb_pcic_set_memory_offset(device_t brdev, device_t child, int rid,
uint32_t cardaddr, uint32_t *deltap)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
struct resource *res;
2002-08-15 08:01:24 +00:00
res = cbb_find_res(sc, SYS_RES_MEMORY, rid);
if (res == NULL) {
device_printf(brdev,
"set_memory_offset: specified rid not found\n");
return (ENOENT);
}
return (exca_mem_set_offset(&sc->exca[0], res, cardaddr, deltap));
}
/************************************************************************/
/* BUS Methods */
/************************************************************************/
int
2002-08-15 08:01:24 +00:00
cbb_activate_resource(device_t brdev, device_t child, int type, int rid,
struct resource *r)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
2002-08-15 08:01:24 +00:00
if (sc->flags & CBB_16BIT_CARD)
return (cbb_pcic_activate_resource(brdev, child, type, rid, r));
else
2002-08-15 08:01:24 +00:00
return (cbb_cardbus_activate_resource(brdev, child, type, rid,
r));
}
int
2002-08-15 08:01:24 +00:00
cbb_deactivate_resource(device_t brdev, device_t child, int type,
int rid, struct resource *r)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
2002-08-15 08:01:24 +00:00
if (sc->flags & CBB_16BIT_CARD)
return (cbb_pcic_deactivate_resource(brdev, child, type,
rid, r));
else
2002-08-15 08:01:24 +00:00
return (cbb_cardbus_deactivate_resource(brdev, child, type,
rid, r));
}
struct resource *
2002-08-15 08:01:24 +00:00
cbb_alloc_resource(device_t brdev, device_t child, int type, int *rid,
u_long start, u_long end, u_long count, u_int flags)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
2002-08-15 08:01:24 +00:00
if (sc->flags & CBB_16BIT_CARD)
return (cbb_pcic_alloc_resource(brdev, child, type, rid,
start, end, count, flags));
else
2002-08-15 08:01:24 +00:00
return (cbb_cardbus_alloc_resource(brdev, child, type, rid,
start, end, count, flags));
}
int
2002-08-15 08:01:24 +00:00
cbb_release_resource(device_t brdev, device_t child, int type, int rid,
struct resource *r)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
2002-08-15 08:01:24 +00:00
if (sc->flags & CBB_16BIT_CARD)
return (cbb_pcic_release_resource(brdev, child, type,
rid, r));
else
2002-08-15 08:01:24 +00:00
return (cbb_cardbus_release_resource(brdev, child, type,
rid, r));
}
int
2002-08-15 08:01:24 +00:00
cbb_read_ivar(device_t brdev, device_t child, int which, uintptr_t *result)
{
2002-08-15 08:01:24 +00:00
struct cbb_softc *sc = device_get_softc(brdev);
switch (which) {
case PCIB_IVAR_DOMAIN:
*result = sc->domain;
return (0);
case PCIB_IVAR_BUS:
*result = sc->bus.sec;
return (0);
}
return (ENOENT);
}
int
2002-08-15 08:01:24 +00:00
cbb_write_ivar(device_t brdev, device_t child, int which, uintptr_t value)
{
switch (which) {
case PCIB_IVAR_DOMAIN:
return (EINVAL);
case PCIB_IVAR_BUS:
return (EINVAL);
}
return (ENOENT);
}
int
cbb_child_present(device_t parent, device_t child)
{
struct cbb_softc *sc = (struct cbb_softc *)device_get_softc(parent);
uint32_t sockstate;
sockstate = cbb_get(sc, CBB_SOCKET_STATE);
return (CBB_CARD_PRESENT(sockstate) && sc->cardok);
}