Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
/*
|
|
|
|
* Copyright (c) 1997, Stefan Esser <se@freebsd.org>
|
|
|
|
* 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 unmodified, 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 ``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.
|
|
|
|
*
|
1999-04-11 02:46:20 +00:00
|
|
|
* $Id: pci_compat.c,v 1.20 1999/01/19 23:29:19 se Exp $
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "pci.h"
|
|
|
|
#if NPCI > 0
|
|
|
|
|
|
|
|
/* for compatibility to FreeBSD-2.2 version of PCI code */
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/malloc.h>
|
1999-01-14 06:22:10 +00:00
|
|
|
#include <sys/linker_set.h>
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
|
|
|
|
#include <vm/vm.h>
|
|
|
|
#include <vm/pmap.h>
|
|
|
|
#include <sys/interrupt.h>
|
|
|
|
|
|
|
|
#include <pci/pcireg.h>
|
|
|
|
#include <pci/pcivar.h>
|
|
|
|
|
|
|
|
#ifdef RESOURCE_CHECK
|
|
|
|
#include <sys/drvresource.h>
|
|
|
|
#endif
|
|
|
|
|
1998-11-09 08:03:01 +00:00
|
|
|
#ifdef APIC_IO
|
|
|
|
#include <machine/smp.h>
|
|
|
|
#endif
|
|
|
|
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
#ifdef PCI_COMPAT
|
|
|
|
|
|
|
|
/* ------------------------------------------------------------------------- */
|
|
|
|
|
|
|
|
static int
|
|
|
|
pci_mapno(pcicfgregs *cfg, int reg)
|
|
|
|
{
|
1999-01-19 23:29:20 +00:00
|
|
|
int i, nummaps;
|
|
|
|
pcimap *map;
|
|
|
|
|
|
|
|
nummaps = cfg->nummaps;
|
|
|
|
map = cfg->map;
|
|
|
|
|
|
|
|
for (i = 0; i < nummaps; i++)
|
|
|
|
if (map[i].reg == reg)
|
|
|
|
return (i);
|
|
|
|
return (-1);
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
}
|
|
|
|
|
1997-05-28 10:10:02 +00:00
|
|
|
static int
|
|
|
|
pci_porten(pcicfgregs *cfg)
|
|
|
|
{
|
|
|
|
return ((cfg->cmdreg & PCIM_CMD_PORTEN) != 0);
|
|
|
|
}
|
|
|
|
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
static int
|
|
|
|
pci_isportmap(pcicfgregs *cfg, int map)
|
1997-05-28 10:10:02 +00:00
|
|
|
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
{
|
|
|
|
return ((unsigned)map < cfg->nummaps
|
|
|
|
&& (cfg->map[map].type & PCI_MAPPORT) != 0);
|
|
|
|
}
|
|
|
|
|
1997-05-28 10:10:02 +00:00
|
|
|
static int
|
|
|
|
pci_memen(pcicfgregs *cfg)
|
|
|
|
{
|
|
|
|
return ((cfg->cmdreg & PCIM_CMD_MEMEN) != 0);
|
|
|
|
}
|
|
|
|
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
static int
|
|
|
|
pci_ismemmap(pcicfgregs *cfg, int map)
|
|
|
|
{
|
|
|
|
return ((unsigned)map < cfg->nummaps
|
|
|
|
&& (cfg->map[map].type & PCI_MAPMEM) != 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ------------------------------------------------------------------------- */
|
|
|
|
|
|
|
|
u_long
|
|
|
|
pci_conf_read(pcici_t tag, u_long reg)
|
|
|
|
{
|
|
|
|
return (pci_cfgread(tag, reg, 4));
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
pci_conf_write(pcici_t tag, u_long reg, u_long data)
|
|
|
|
{
|
|
|
|
pci_cfgwrite(tag, reg, data, 4);
|
|
|
|
}
|
|
|
|
|
1998-07-22 08:39:08 +00:00
|
|
|
int pci_map_port(pcici_t cfg, u_long reg, pci_port_t* pa)
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
{
|
|
|
|
int map;
|
|
|
|
|
|
|
|
map = pci_mapno(cfg, reg);
|
1997-05-28 10:10:02 +00:00
|
|
|
if (pci_porten(cfg) && pci_isportmap(cfg, map)) {
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
u_int32_t iobase;
|
|
|
|
u_int32_t iosize;
|
|
|
|
|
|
|
|
iobase = cfg->map[map].base;
|
|
|
|
iosize = 1 << cfg->map[map].ln2size;
|
|
|
|
#ifdef RESOURCE_CHECK
|
|
|
|
if (resource_claim(cfg, REST_PORT, RESF_NONE,
|
|
|
|
iobase, iobase + iosize -1) == 0)
|
|
|
|
#endif /* RESOURCE_CHECK */
|
|
|
|
{
|
|
|
|
*pa = iobase;
|
|
|
|
return (1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
int pci_map_mem(pcici_t cfg, u_long reg, vm_offset_t* va, vm_offset_t* pa)
|
|
|
|
{
|
|
|
|
int map;
|
|
|
|
|
|
|
|
map = pci_mapno(cfg, reg);
|
1997-05-28 10:10:02 +00:00
|
|
|
if (pci_memen(cfg) && pci_ismemmap(cfg, map)) {
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
u_int32_t paddr;
|
|
|
|
u_int32_t psize;
|
|
|
|
|
|
|
|
paddr = cfg->map[map].base;
|
|
|
|
psize = 1 << cfg->map[map].ln2size;
|
|
|
|
#ifdef RESOURCE_CHECK
|
|
|
|
if (resource_claim(cfg, REST_MEM, RESF_NONE,
|
|
|
|
paddr, paddr + psize -1) == 0)
|
|
|
|
#endif /* RESOURCE_CHECK */
|
|
|
|
{
|
|
|
|
u_int32_t poffs;
|
|
|
|
vm_offset_t vaddr;
|
|
|
|
|
|
|
|
poffs = paddr - trunc_page(paddr);
|
1998-07-22 08:39:08 +00:00
|
|
|
#ifdef __i386__
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
vaddr = (vm_offset_t)pmap_mapdev(paddr-poffs, psize+poffs);
|
1998-07-22 08:39:08 +00:00
|
|
|
#endif
|
|
|
|
#ifdef __alpha__
|
1998-08-07 08:20:36 +00:00
|
|
|
vaddr = paddr-poffs;
|
1998-07-22 08:39:08 +00:00
|
|
|
#endif
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
if (vaddr != NULL) {
|
|
|
|
vaddr += poffs;
|
|
|
|
*va = vaddr;
|
|
|
|
*pa = paddr;
|
|
|
|
return (1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
1998-10-06 14:18:40 +00:00
|
|
|
int
|
|
|
|
pci_map_dense(pcici_t cfg, u_long reg, vm_offset_t* va, vm_offset_t* pa)
|
|
|
|
{
|
|
|
|
if(pci_map_mem(cfg, reg, va, pa)){
|
|
|
|
#ifdef __alpha__
|
1998-12-07 21:58:50 +00:00
|
|
|
vm_offset_t dense;
|
|
|
|
|
1998-10-06 14:18:40 +00:00
|
|
|
if(dense = pci_cvt_to_dense(*pa)){
|
|
|
|
*pa = dense;
|
|
|
|
*va = ALPHA_PHYS_TO_K0SEG(*pa);
|
|
|
|
return (1);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
#ifdef __i386__
|
|
|
|
return(1);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
pci_map_bwx(pcici_t cfg, u_long reg, vm_offset_t* va, vm_offset_t* pa)
|
|
|
|
{
|
|
|
|
if(pci_map_mem(cfg, reg, va, pa)){
|
|
|
|
#ifdef __alpha__
|
1998-12-07 21:58:50 +00:00
|
|
|
vm_offset_t bwx;
|
|
|
|
|
1998-10-06 14:18:40 +00:00
|
|
|
if(bwx = pci_cvt_to_bwx(*pa)){
|
|
|
|
*pa = bwx;
|
|
|
|
*va = ALPHA_PHYS_TO_K0SEG(*pa);
|
|
|
|
return (1);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
#ifdef __i386__
|
|
|
|
return(1);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
1999-01-13 04:59:19 +00:00
|
|
|
int
|
|
|
|
pci_map_int(pcici_t cfg, pci_inthand_t *handler, void *arg, intrmask_t *maskptr)
|
|
|
|
{
|
|
|
|
return (pci_map_int_right(cfg, handler, arg, maskptr, 0));
|
|
|
|
}
|
1998-10-06 14:18:40 +00:00
|
|
|
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
int
|
1999-01-13 04:59:19 +00:00
|
|
|
pci_map_int_right(pcici_t cfg, pci_inthand_t *handler, void *arg,
|
|
|
|
intrmask_t *maskptr, u_int flags)
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
{
|
1998-04-01 21:07:37 +00:00
|
|
|
int error;
|
|
|
|
#ifdef APIC_IO
|
|
|
|
int nextpin, muxcnt;
|
|
|
|
#endif
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
if (cfg->intpin != 0) {
|
|
|
|
int irq = cfg->intline;
|
|
|
|
void *dev_instance = (void *)-1; /* XXX use cfg->devdata */
|
1997-05-30 20:58:04 +00:00
|
|
|
void *idesc;
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
|
1999-01-13 04:59:19 +00:00
|
|
|
idesc = intr_create(dev_instance, irq, handler, arg, maskptr,
|
|
|
|
flags);
|
1998-04-01 21:07:37 +00:00
|
|
|
error = intr_connect(idesc);
|
|
|
|
if (error != 0)
|
|
|
|
return 0;
|
|
|
|
#ifdef APIC_IO
|
1998-09-06 22:41:42 +00:00
|
|
|
nextpin = next_apic_irq(irq);
|
1998-04-01 21:07:37 +00:00
|
|
|
|
|
|
|
if (nextpin < 0)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Attempt handling of some broken mp tables.
|
|
|
|
*
|
|
|
|
* It's OK to yell (since the mp tables are broken).
|
|
|
|
*
|
|
|
|
* Hanging in the boot is not OK
|
|
|
|
*/
|
|
|
|
|
|
|
|
muxcnt = 2;
|
1998-09-06 22:41:42 +00:00
|
|
|
nextpin = next_apic_irq(nextpin);
|
1998-04-01 21:07:37 +00:00
|
|
|
while (muxcnt < 5 && nextpin >= 0) {
|
|
|
|
muxcnt++;
|
1998-09-06 22:41:42 +00:00
|
|
|
nextpin = next_apic_irq(nextpin);
|
1998-04-01 21:07:37 +00:00
|
|
|
}
|
|
|
|
if (muxcnt >= 5) {
|
|
|
|
printf("bogus MP table, more than 4 IO APIC pins connected to the same PCI device or ISA/EISA interrupt\n");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
printf("bogus MP table, %d IO APIC pins connected to the same PCI device or ISA/EISA interrupt\n", muxcnt);
|
|
|
|
|
1998-09-06 22:41:42 +00:00
|
|
|
nextpin = next_apic_irq(irq);
|
1998-04-01 21:07:37 +00:00
|
|
|
while (nextpin >= 0) {
|
1999-01-13 04:59:19 +00:00
|
|
|
idesc = intr_create(dev_instance, nextpin, handler,
|
|
|
|
arg, maskptr, flags);
|
1998-04-01 21:07:37 +00:00
|
|
|
error = intr_connect(idesc);
|
|
|
|
if (error != 0)
|
|
|
|
return 0;
|
|
|
|
printf("Registered extra interrupt handler for int %d (in addition to int %d)\n", nextpin, irq);
|
1998-09-06 22:41:42 +00:00
|
|
|
nextpin = next_apic_irq(nextpin);
|
1998-04-01 21:07:37 +00:00
|
|
|
}
|
|
|
|
#endif
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
}
|
|
|
|
return (1);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
pci_unmap_int(pcici_t cfg)
|
|
|
|
{
|
|
|
|
return (0); /* not supported, yet, since cfg doesn't know about idesc */
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ------------------------------------------------------------------------- */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Preliminary support for "wired" PCI devices.
|
|
|
|
* This code supports currently only devices on PCI bus 0, since the
|
|
|
|
* mapping from PCI BIOS bus numbers to configuration file bus numbers
|
|
|
|
* is not yet maintained, whenever a PCI to PCI bridge is found.
|
|
|
|
* The "bus" field of "pciwirecfg" correlates an PCI bus with the bridge
|
|
|
|
* it is attached to. The "biosbus" field is to be updated for each bus,
|
|
|
|
* whose bridge is probed. An entry with bus != 0 and biosbus == 0 is
|
|
|
|
* invalid and will be skipped in the search for a wired unit, but not
|
|
|
|
* in the test for a free unit number.
|
|
|
|
*/
|
|
|
|
|
|
|
|
typedef struct {
|
|
|
|
char *name;
|
|
|
|
int unit;
|
|
|
|
u_int8_t bus;
|
|
|
|
u_int8_t slot;
|
|
|
|
u_int8_t func;
|
|
|
|
u_int8_t biosbus;
|
|
|
|
} pciwirecfg;
|
|
|
|
|
|
|
|
static pciwirecfg pci_wireddevs[] = {
|
|
|
|
/* driver, unit, bus, slot, func, BIOS bus */
|
1997-05-30 20:58:04 +00:00
|
|
|
#ifdef PCI_DEBUG
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
{ "ncr", 2, 1, 4, 0, 0 },
|
|
|
|
{ "ed", 2, 1, 5, 0, 0 },
|
1997-05-30 20:58:04 +00:00
|
|
|
#endif /* PCI_DEBUG */
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
/* do not delete the end marker that follows this comment !!! */
|
|
|
|
{ NULL }
|
|
|
|
};
|
|
|
|
|
|
|
|
/* return unit number of wired device, or -1 if no match */
|
|
|
|
|
|
|
|
static int
|
|
|
|
pci_wiredunit(pcicfgregs *cfg, char *name)
|
|
|
|
{
|
|
|
|
pciwirecfg *p;
|
|
|
|
|
|
|
|
p = pci_wireddevs;
|
|
|
|
while (p->name != NULL) {
|
|
|
|
if (p->bus == cfg->bus
|
|
|
|
&& p->slot == cfg->slot
|
|
|
|
&& p->func == cfg->func
|
|
|
|
&& strcmp(p->name, name) == 0)
|
|
|
|
return (p->unit);
|
|
|
|
p++;
|
|
|
|
}
|
|
|
|
return (-1);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* return free unit number equal or greater to the one supplied as parameter */
|
|
|
|
|
|
|
|
static int
|
|
|
|
pci_freeunit(pcicfgregs *cfg, char *name, int unit)
|
|
|
|
{
|
|
|
|
pciwirecfg *p;
|
|
|
|
|
|
|
|
p = pci_wireddevs;
|
|
|
|
while (p->name != NULL) {
|
|
|
|
if (p->unit == unit && strcmp(p->name, name) == 0) {
|
|
|
|
p = pci_wireddevs;
|
|
|
|
unit++;
|
|
|
|
} else {
|
|
|
|
p++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return (unit);
|
|
|
|
}
|
|
|
|
|
1998-12-14 05:47:29 +00:00
|
|
|
static const char *drvname;
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
|
1998-12-14 05:47:29 +00:00
|
|
|
static const char*
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
pci_probedrv(pcicfgregs *cfg, struct pci_device *dvp)
|
|
|
|
{
|
|
|
|
if (dvp && dvp->pd_probe) {
|
|
|
|
pcidi_t type = (cfg->device << 16) + cfg->vendor;
|
|
|
|
return (dvp->pd_probe(cfg, type));
|
|
|
|
}
|
|
|
|
return (NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pci_lkm *pci_lkm_head;
|
|
|
|
|
|
|
|
static struct pci_device*
|
|
|
|
pci_finddrv(pcicfgregs *cfg)
|
|
|
|
{
|
|
|
|
struct pci_device **dvpp;
|
|
|
|
struct pci_device *dvp = NULL;
|
|
|
|
struct pci_lkm *lkm;
|
|
|
|
|
|
|
|
drvname = NULL;
|
|
|
|
lkm = pci_lkm_head;
|
|
|
|
while (drvname == NULL && lkm != NULL) {
|
|
|
|
dvp = lkm->dvp;
|
|
|
|
drvname = pci_probedrv(cfg, dvp);
|
|
|
|
lkm = lkm->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
dvpp = (struct pci_device **)pcidevice_set.ls_items;
|
|
|
|
while (drvname == NULL && (dvp = *dvpp++) != NULL)
|
|
|
|
drvname = pci_probedrv(cfg, dvp);
|
|
|
|
return (dvp);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
pci_drvmessage(pcicfgregs *cfg, char *name, int unit)
|
|
|
|
{
|
|
|
|
if (drvname == NULL || *drvname == '\0')
|
|
|
|
return;
|
|
|
|
printf("%s%d: <%s> rev 0x%02x", name, unit, drvname, cfg->revid);
|
|
|
|
if (cfg->intpin != 0)
|
|
|
|
printf(" int %c irq %d", cfg->intpin + 'a' -1, cfg->intline);
|
|
|
|
printf(" on pci%d.%d.%d\n", cfg->bus, cfg->slot, cfg->func);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void
|
1998-09-15 08:21:13 +00:00
|
|
|
pci_drvattach(struct pci_devinfo *dinfo)
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
{
|
|
|
|
struct pci_device *dvp;
|
1998-09-15 08:21:13 +00:00
|
|
|
pcicfgregs *cfg;
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
|
1998-09-15 08:21:13 +00:00
|
|
|
cfg = &dinfo->cfg;
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
dvp = pci_finddrv(cfg);
|
|
|
|
if (dvp != NULL) {
|
|
|
|
int unit;
|
|
|
|
|
|
|
|
unit = pci_wiredunit(cfg, dvp->pd_name);
|
|
|
|
if (unit < 0) {
|
|
|
|
unit = pci_freeunit(cfg, dvp->pd_name, *dvp->pd_count);
|
|
|
|
*dvp->pd_count = unit +1;
|
|
|
|
}
|
|
|
|
pci_drvmessage(cfg, dvp->pd_name, unit);
|
|
|
|
if (dvp->pd_attach)
|
|
|
|
dvp->pd_attach(cfg, unit);
|
1998-09-15 08:21:13 +00:00
|
|
|
|
|
|
|
dinfo->device = dvp;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* XXX KDM for some devices, dvp->pd_name winds up NULL.
|
|
|
|
* I haven't investigated enough to figure out why this
|
|
|
|
* would happen.
|
|
|
|
*/
|
|
|
|
if (dvp->pd_name != NULL)
|
|
|
|
strncpy(dinfo->conf.pd_name, dvp->pd_name,
|
|
|
|
sizeof(dinfo->conf.pd_name));
|
|
|
|
else
|
|
|
|
strncpy(dinfo->conf.pd_name, "????",
|
|
|
|
sizeof(dinfo->conf.pd_name));
|
1998-12-04 22:54:57 +00:00
|
|
|
dinfo->conf.pd_name[sizeof(dinfo->conf.pd_name) - 1] = 0;
|
1998-09-15 08:21:13 +00:00
|
|
|
|
|
|
|
dinfo->conf.pd_unit = unit;
|
|
|
|
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ------------------------------------------------------------------------- */
|
|
|
|
|
|
|
|
static void
|
|
|
|
pci_rescan(void)
|
|
|
|
{
|
|
|
|
/* XXX do nothing, currently, soon to come ... */
|
|
|
|
}
|
|
|
|
|
|
|
|
int pci_register_lkm (struct pci_device *dvp, int if_revision)
|
|
|
|
{
|
|
|
|
struct pci_lkm *lkm;
|
|
|
|
|
|
|
|
if (if_revision != 0) {
|
|
|
|
return (-1);
|
|
|
|
}
|
|
|
|
if (dvp == NULL || dvp->pd_probe == NULL || dvp->pd_attach == NULL) {
|
|
|
|
return (-1);
|
|
|
|
}
|
1999-04-11 02:46:20 +00:00
|
|
|
lkm = malloc (sizeof (*lkm), M_DEVBUF, M_NOWAIT);
|
1998-12-09 01:27:29 +00:00
|
|
|
if (lkm == NULL) {
|
Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .
The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...
This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.
A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:
1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.
The following features are currently missing, but will be added back,
soon:
1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets
This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).
1997-05-26 15:08:43 +00:00
|
|
|
return (-1);
|
|
|
|
}
|
|
|
|
|
|
|
|
lkm->dvp = dvp;
|
|
|
|
lkm->next = pci_lkm_head;
|
|
|
|
pci_lkm_head = lkm;
|
|
|
|
pci_rescan();
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
pci_configure(void)
|
|
|
|
{
|
|
|
|
pci_probe(NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* ------------------------------------------------------------------------- */
|
|
|
|
|
|
|
|
#endif /* PCI_COMPAT */
|
|
|
|
#endif /* NPCI > 0 */
|