freebsd-dev/sys/dev/le/if_le_pci.c

509 lines
15 KiB
C
Raw Normal View History

/* $NetBSD: if_le_pci.c,v 1.43 2005/12/11 12:22:49 christos Exp $ */
/*-
* Copyright (c) 1997, 1998 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Charles M. Hannum and by Jason R. Thorpe of the Numerical Aerospace
* Simulation Facility, NASA Ames Research Center.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the NetBSD
* Foundation, Inc. and its contributors.
* 4. Neither the name of The NetBSD Foundation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
/*-
* Copyright (c) 1992, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Ralph Campbell and Rick Macklem.
*
* 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. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
*
* @(#)if_le.c 8.2 (Berkeley) 11/16/93
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/bus.h>
#include <sys/endian.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/module.h>
#include <sys/mutex.h>
#include <sys/resource.h>
#include <sys/rman.h>
#include <sys/socket.h>
#include <net/ethernet.h>
#include <net/if.h>
#include <net/if_media.h>
#include <machine/bus.h>
#include <machine/resource.h>
#include <dev/pci/pcireg.h>
#include <dev/pci/pcivar.h>
#include <dev/le/lancereg.h>
#include <dev/le/lancevar.h>
#include <dev/le/am79900var.h>
#define AMD_VENDOR 0x1022
#define AMD_PCNET_PCI 0x2000
#define AMD_PCNET_HOME 0x2001
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
#define PCNET_MEMSIZE (32*1024)
#define PCNET_PCI_RDP 0x10
#define PCNET_PCI_RAP 0x12
#define PCNET_PCI_BDP 0x16
struct le_pci_softc {
struct am79900_softc sc_am79900; /* glue to MI code */
struct resource *sc_rres;
struct resource *sc_ires;
void *sc_ih;
bus_dma_tag_t sc_pdmat;
bus_dma_tag_t sc_dmat;
bus_dmamap_t sc_dmam;
};
static device_probe_t le_pci_probe;
static device_attach_t le_pci_attach;
static device_detach_t le_pci_detach;
static device_resume_t le_pci_resume;
static device_suspend_t le_pci_suspend;
static device_method_t le_pci_methods[] = {
/* Device interface */
DEVMETHOD(device_probe, le_pci_probe),
DEVMETHOD(device_attach, le_pci_attach),
DEVMETHOD(device_detach, le_pci_detach),
/* We can just use the suspend method here. */
DEVMETHOD(device_shutdown, le_pci_suspend),
DEVMETHOD(device_suspend, le_pci_suspend),
DEVMETHOD(device_resume, le_pci_resume),
{ 0, 0 }
};
DEFINE_CLASS_0(le, le_pci_driver, le_pci_methods, sizeof(struct le_pci_softc));
DRIVER_MODULE(le, pci, le_pci_driver, le_devclass, 0, 0);
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
MODULE_DEPEND(le, ether, 1, 1, 1);
static const int le_home_supmedia[] = {
IFM_MAKEWORD(IFM_ETHER, IFM_HPNA_1, 0, 0)
};
static const int le_pci_supmedia[] = {
IFM_MAKEWORD(IFM_ETHER, IFM_AUTO, 0, 0),
IFM_MAKEWORD(IFM_ETHER, IFM_AUTO, IFM_FDX, 0),
IFM_MAKEWORD(IFM_ETHER, IFM_10_T, 0, 0),
IFM_MAKEWORD(IFM_ETHER, IFM_10_T, IFM_FDX, 0),
IFM_MAKEWORD(IFM_ETHER, IFM_10_5, 0, 0),
IFM_MAKEWORD(IFM_ETHER, IFM_10_5, IFM_FDX, 0)
};
static void le_pci_wrbcr(struct lance_softc *, uint16_t, uint16_t);
static uint16_t le_pci_rdbcr(struct lance_softc *, uint16_t);
static void le_pci_wrcsr(struct lance_softc *, uint16_t, uint16_t);
static uint16_t le_pci_rdcsr(struct lance_softc *, uint16_t);
static int le_pci_mediachange(struct lance_softc *);
static void le_pci_hwreset(struct lance_softc *);
static bus_dmamap_callback_t le_pci_dma_callback;
static void
le_pci_wrbcr(struct lance_softc *sc, uint16_t port, uint16_t val)
{
struct le_pci_softc *lesc = (struct le_pci_softc *)sc;
bus_write_2(lesc->sc_rres, PCNET_PCI_RAP, port);
bus_barrier(lesc->sc_rres, PCNET_PCI_RAP, 2, BUS_SPACE_BARRIER_WRITE);
bus_write_2(lesc->sc_rres, PCNET_PCI_BDP, val);
}
static uint16_t
le_pci_rdbcr(struct lance_softc *sc, uint16_t port)
{
struct le_pci_softc *lesc = (struct le_pci_softc *)sc;
bus_write_2(lesc->sc_rres, PCNET_PCI_RAP, port);
bus_barrier(lesc->sc_rres, PCNET_PCI_RAP, 2, BUS_SPACE_BARRIER_WRITE);
return (bus_read_2(lesc->sc_rres, PCNET_PCI_BDP));
}
static void
le_pci_wrcsr(struct lance_softc *sc, uint16_t port, uint16_t val)
{
struct le_pci_softc *lesc = (struct le_pci_softc *)sc;
bus_write_2(lesc->sc_rres, PCNET_PCI_RAP, port);
bus_barrier(lesc->sc_rres, PCNET_PCI_RAP, 2, BUS_SPACE_BARRIER_WRITE);
bus_write_2(lesc->sc_rres, PCNET_PCI_RDP, val);
}
static uint16_t
le_pci_rdcsr(struct lance_softc *sc, uint16_t port)
{
struct le_pci_softc *lesc = (struct le_pci_softc *)sc;
bus_write_2(lesc->sc_rres, PCNET_PCI_RAP, port);
bus_barrier(lesc->sc_rres, PCNET_PCI_RAP, 2, BUS_SPACE_BARRIER_WRITE);
return (bus_read_2(lesc->sc_rres, PCNET_PCI_RDP));
}
static int
le_pci_mediachange(struct lance_softc *sc)
{
struct ifmedia *ifm = &sc->sc_media;
uint16_t reg;
if (IFM_TYPE(ifm->ifm_media) != IFM_ETHER)
return (EINVAL);
if (IFM_SUBTYPE(ifm->ifm_media) == IFM_HPNA_1)
le_pci_wrbcr(sc, LE_BCR49,
(le_pci_rdbcr(sc, LE_BCR49) & ~LE_B49_PHYSEL) | 0x1);
else if (IFM_SUBTYPE(ifm->ifm_media) == IFM_AUTO)
le_pci_wrbcr(sc, LE_BCR2,
le_pci_rdbcr(sc, LE_BCR2) | LE_B2_ASEL);
else {
le_pci_wrbcr(sc, LE_BCR2,
le_pci_rdbcr(sc, LE_BCR2) & ~LE_B2_ASEL);
reg = le_pci_rdcsr(sc, LE_CSR15);
reg &= ~LE_C15_PORTSEL(LE_PORTSEL_MASK);
if (IFM_SUBTYPE(ifm->ifm_media) == IFM_10_T)
reg |= LE_C15_PORTSEL(LE_PORTSEL_10T);
else
reg |= LE_C15_PORTSEL(LE_PORTSEL_AUI);
le_pci_wrcsr(sc, LE_CSR15, reg);
}
reg = le_pci_rdbcr(sc, LE_BCR9);
if (IFM_OPTIONS(ifm->ifm_media) & IFM_FDX) {
reg |= LE_B9_FDEN;
/*
* Allow FDX on AUI only if explicitly chosen,
* not in autoselect mode.
*/
if (IFM_SUBTYPE(ifm->ifm_media) == IFM_10_5)
reg |= LE_B9_AUIFD;
else
reg &= ~LE_B9_AUIFD;
} else
reg &= ~LE_B9_FDEN;
le_pci_wrbcr(sc, LE_BCR9, reg);
return (0);
}
static void
le_pci_hwreset(struct lance_softc *sc)
{
/*
* Chip is stopped. Set software style to PCnet-PCI (32-bit).
* Actually, am79900.c implements ILACC support (hence its
* name) but unfortunately VMware does not. As far as this
* driver is concerned that should not make a difference
* though, as the settings used have the same meaning for
* both, ILACC and PCnet-PCI (note that there would be a
* difference for the ADD_FCS/NO_FCS bit if used).
*/
le_pci_wrbcr(sc, LE_BCR20, LE_B20_SSTYLE_PCNETPCI2);
}
static void
le_pci_dma_callback(void *xsc, bus_dma_segment_t *segs, int nsegs, int error)
{
struct lance_softc *sc = (struct lance_softc *)xsc;
if (error != 0)
return;
KASSERT(nsegs == 1, ("%s: bad DMA segment count", __func__));
sc->sc_addr = segs[0].ds_addr;
}
static int
le_pci_probe(device_t dev)
{
if (pci_get_vendor(dev) != AMD_VENDOR)
return (ENXIO);
switch (pci_get_device(dev)) {
case AMD_PCNET_PCI:
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
device_set_desc(dev, "AMD PCnet-PCI");
/* Let pcn(4) win. */
return (BUS_PROBE_LOW_PRIORITY);
case AMD_PCNET_HOME:
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
device_set_desc(dev, "AMD PCnet-Home");
/* Let pcn(4) win. */
return (BUS_PROBE_LOW_PRIORITY);
default:
return (ENXIO);
}
}
static int
le_pci_attach(device_t dev)
{
struct le_pci_softc *lesc;
struct lance_softc *sc;
int error, i;
lesc = device_get_softc(dev);
sc = &lesc->sc_am79900.lsc;
LE_LOCK_INIT(sc, device_get_nameunit(dev));
pci_enable_busmaster(dev);
i = PCIR_BAR(0);
lesc->sc_rres = bus_alloc_resource_any(dev, SYS_RES_IOPORT,
&i, RF_ACTIVE);
if (lesc->sc_rres == NULL) {
device_printf(dev, "cannot allocate registers\n");
error = ENXIO;
goto fail_mtx;
}
i = 0;
if ((lesc->sc_ires = bus_alloc_resource_any(dev, SYS_RES_IRQ,
&i, RF_SHAREABLE | RF_ACTIVE)) == NULL) {
device_printf(dev, "cannot allocate interrupt\n");
error = ENXIO;
goto fail_rres;
}
error = bus_dma_tag_create(
bus_get_dma_tag(dev), /* parent */
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
1, 0, /* alignment, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */
BUS_SPACE_MAXSIZE_32BIT, /* maxsize */
0, /* nsegments */
BUS_SPACE_MAXSIZE_32BIT, /* maxsegsize */
0, /* flags */
NULL, NULL, /* lockfunc, lockarg */
&lesc->sc_pdmat);
if (error != 0) {
device_printf(dev, "cannot allocate parent DMA tag\n");
goto fail_ires;
}
sc->sc_memsize = PCNET_MEMSIZE;
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
/*
* For Am79C970A, Am79C971 and Am79C978 the init block must be 2-byte
* aligned and the ring descriptors must be 16-byte aligned when using
* a 32-bit software style.
*/
error = bus_dma_tag_create(
lesc->sc_pdmat, /* parent */
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
16, 0, /* alignment, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */
sc->sc_memsize, /* maxsize */
1, /* nsegments */
sc->sc_memsize, /* maxsegsize */
0, /* flags */
NULL, NULL, /* lockfunc, lockarg */
&lesc->sc_dmat);
if (error != 0) {
device_printf(dev, "cannot allocate buffer DMA tag\n");
goto fail_pdtag;
}
error = bus_dmamem_alloc(lesc->sc_dmat, (void **)&sc->sc_mem,
BUS_DMA_WAITOK | BUS_DMA_COHERENT, &lesc->sc_dmam);
if (error != 0) {
device_printf(dev, "cannot allocate DMA buffer memory\n");
goto fail_dtag;
}
sc->sc_addr = 0;
error = bus_dmamap_load(lesc->sc_dmat, lesc->sc_dmam, sc->sc_mem,
sc->sc_memsize, le_pci_dma_callback, sc, 0);
if (error != 0 || sc->sc_addr == 0) {
device_printf(dev, "cannot load DMA buffer map\n");
goto fail_dmem;
}
sc->sc_flags = LE_BSWAP;
sc->sc_conf3 = 0;
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
sc->sc_mediastatus = NULL;
switch (pci_get_device(dev)) {
case AMD_PCNET_HOME:
sc->sc_mediachange = le_pci_mediachange;
sc->sc_supmedia = le_home_supmedia;
sc->sc_nsupmedia = sizeof(le_home_supmedia) / sizeof(int);
sc->sc_defaultmedia = le_home_supmedia[0];
break;
default:
sc->sc_mediachange = le_pci_mediachange;
sc->sc_supmedia = le_pci_supmedia;
sc->sc_nsupmedia = sizeof(le_pci_supmedia) / sizeof(int);
sc->sc_defaultmedia = le_pci_supmedia[0];
}
/*
* Extract the physical MAC address from the ROM.
*/
bus_read_region_1(lesc->sc_rres, 0, sc->sc_enaddr,
sizeof(sc->sc_enaddr));
sc->sc_copytodesc = lance_copytobuf_contig;
sc->sc_copyfromdesc = lance_copyfrombuf_contig;
sc->sc_copytobuf = lance_copytobuf_contig;
sc->sc_copyfrombuf = lance_copyfrombuf_contig;
sc->sc_zerobuf = lance_zerobuf_contig;
sc->sc_rdcsr = le_pci_rdcsr;
sc->sc_wrcsr = le_pci_wrcsr;
sc->sc_hwreset = le_pci_hwreset;
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
sc->sc_hwinit = NULL;
sc->sc_hwintr = NULL;
sc->sc_nocarrier = NULL;
error = am79900_config(&lesc->sc_am79900, device_get_name(dev),
device_get_unit(dev));
if (error != 0) {
- Revert if_le_pci.c rev. 1.2; although lnc(4) is now gone, le_pci_probe() still should return BUS_PROBE_LOW_PRIORITY instead of BUS_PROBE_DEFAULT in order to give pcn(4) a chance to attach in case it probes after le(4). - Rearrange the code related to RX interrupt handling so that ownership of RX descriptors is immediately returned to the NIC after we have copied the data of the hardware, allowing the NIC to already reuse the descriptor while we are processing the data in ifp->if_input(). This results in a small but measurable increase in RX throughput. As a side-effect, this moves the workaround for the LANCE revision C bug to am7900.c (still off by default as I doubt we will actually encounter such an old chip in a machine running FreeBSD) and the workaround for the bug in the VMware PCnet-PCI emulation to am79000.c, which is now also only compiled on i386 (resulting in a small increase in RX throughput on the other platforms). - Change the RX interrupt handlers so that the descriptor error bits are only check once in case there was no error instead of twice (inspired by the NetBSD pcn(4), which additionally predicts the error branch as false). - Fix the debugging output of the RX and TX interrupt handlers; while looping through the descriptors print info about the currently processed one instead of always the previously last used one; remove pointless printing of info about the RX descriptor bits after their values were reset. - Create the DMA tags used to allocate the memory for the init block, descriptors and packet buffers with the alignment the respective NIC actually requires rather than using PAGE_SIZE unconditionally. This might as well fix the alignment of the memory as it seems we do not inherit the alignment constraint from the parent DMA tag. - For the PCI variants double the number of RX descriptors and buffers from 8 to 16 as this minimizes the number of RX overflows im seeing with one NIC-mainboard combination. Nevertheless move reporting of overflows under debugging as they seem unavoidable with some crappy hardware. - Set the software style of the PCI variants to ILACC rather than PCnet-PCI as the former is was am79000.c actually implements. Should not make a difference for this driver though. - Fix the driver name part in the MODULE_DEPEND of the PCI front-end for ether. - Use different device descriptions for PCnet-Home and PCnet-PCI. - Fix some 0/NULL confusion in lance_get(). - Use bus_addr_t for sc_addr and bus_size_t for sc_memsize as these are more appropriate than u_long for these. - Remove the unused LE_DRIVER_NAME macro. - Add a comment describing why we are taking the LE_HTOLE* etc approach instead of using byteorder(9) functions directly. - Improve some comments and fix some wording. MFC after: 2 weeks
2006-05-16 21:04:01 +00:00
device_printf(dev, "cannot attach Am79900\n");
goto fail_dmap;
}
error = bus_setup_intr(dev, lesc->sc_ires, INTR_TYPE_NET | INTR_MPSAFE,
NULL, am79900_intr, sc, &lesc->sc_ih);
if (error != 0) {
device_printf(dev, "cannot set up interrupt\n");
goto fail_am79900;
}
return (0);
fail_am79900:
am79900_detach(&lesc->sc_am79900);
fail_dmap:
bus_dmamap_unload(lesc->sc_dmat, lesc->sc_dmam);
fail_dmem:
bus_dmamem_free(lesc->sc_dmat, sc->sc_mem, lesc->sc_dmam);
fail_dtag:
bus_dma_tag_destroy(lesc->sc_dmat);
fail_pdtag:
bus_dma_tag_destroy(lesc->sc_pdmat);
fail_ires:
bus_release_resource(dev, SYS_RES_IRQ,
rman_get_rid(lesc->sc_ires), lesc->sc_ires);
fail_rres:
bus_release_resource(dev, SYS_RES_IOPORT,
rman_get_rid(lesc->sc_rres), lesc->sc_rres);
fail_mtx:
LE_LOCK_DESTROY(sc);
return (error);
}
static int
le_pci_detach(device_t dev)
{
struct le_pci_softc *lesc;
struct lance_softc *sc;
lesc = device_get_softc(dev);
sc = &lesc->sc_am79900.lsc;
bus_teardown_intr(dev, lesc->sc_ires, lesc->sc_ih);
am79900_detach(&lesc->sc_am79900);
bus_dmamap_unload(lesc->sc_dmat, lesc->sc_dmam);
bus_dmamem_free(lesc->sc_dmat, sc->sc_mem, lesc->sc_dmam);
bus_dma_tag_destroy(lesc->sc_dmat);
bus_dma_tag_destroy(lesc->sc_pdmat);
bus_release_resource(dev, SYS_RES_IRQ,
rman_get_rid(lesc->sc_ires), lesc->sc_ires);
bus_release_resource(dev, SYS_RES_IOPORT,
rman_get_rid(lesc->sc_rres), lesc->sc_rres);
LE_LOCK_DESTROY(sc);
return (0);
}
static int
le_pci_suspend(device_t dev)
{
struct le_pci_softc *lesc;
lesc = device_get_softc(dev);
lance_suspend(&lesc->sc_am79900.lsc);
return (0);
}
static int
le_pci_resume(device_t dev)
{
struct le_pci_softc *lesc;
lesc = device_get_softc(dev);
lance_resume(&lesc->sc_am79900.lsc);
return (0);
}