d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
45271d26ef
freebsd-dev/sys/pci/pci.c

621 lines
14 KiB
C
Raw Normal View History

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/**************************************************************************
**
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
** $Id: pci.c,v 1.2 1994/09/01 02:01:34 se Exp $
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
**
** General subroutines for the PCI bus on 80*86 systems.
** pci_configure ()
**
** 386bsd / FreeBSD
**
**-------------------------------------------------------------------------
**
** Copyright (c) 1994 Wolfgang Stanglmeier. All rights reserved.
**
** Redistribution and use in source and binary forms, with or without
** modification, are permitted provided that the following conditions
** are met:
** 1. Redistributions of source code must retain the above copyright
** notice, this list of conditions and the following disclaimer.
** 2. Redistributions in binary form must reproduce the above copyright
** notice, this list of conditions and the following disclaimer in the
** documentation and/or other materials provided with the distribution.
** 3. The name of the author may not be used to endorse or promote products
** derived from this software without specific prior written permission.
**
** THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
** IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
** OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
** IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
** INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
** NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
** DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
** THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
** (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
** THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
**
**-------------------------------------------------------------------------
**
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
** $Log: pci.c,v $
* Revision 1.2 1994/09/01 02:01:34 se
* Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
* Merged in changes required for NetBSD support (by mycroft@gnu.ai.mit.edu)
* and support for multiple NCR chips.
*
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
** Revision 2.0.0.8 94/08/21 19:57:39 wolf
** Unneeded declarations removed (FreeBSD2.0)
**
** Revision 2.0.0.7 94/08/21 19:25:54 wolf
** pci_intr simplified.
** new not_supported() function.
** Vendor and device ids moved into tables.
**
** Revision 2.0.0.6 94/08/18 22:58:23 wolf
** Symbolic names for pci configuration space registers.
** last_device: from configuration mode
** last_bus: from pcibios.
** PCI_MAX_DPI: changed to 4 (settable by config)
** interrupt configuration by line or pin
**
** Revision 2.0.0.5 94/08/11 19:04:10 wolf
** display of interrupt line configuration register.
**
** Revision 2.0.0.4 94/08/01 20:36:28 wolf
** Tiny clean up.
**
** Revision 2.0.0.3 94/08/01 18:52:33 wolf
** New vendor entry: S3.
** Scan pci busses #0..#255 as default.
** Number of scanned busses and devices settable as option.
** Show these numbers before starting the scan.
**
** Revision 2.0.0.2 94/07/27 09:27:19 wolf
** New option PCI_QUIET: suppress log messages.
**
** Revision 2.0.0.1 94/07/19 19:06:44 wolf
** New vendor entry: MATROX
**
** Revision 2.0 94/07/10 15:53:29 wolf
** FreeBSD release.
**
** Revision 1.0 94/06/07 20:02:19 wolf
** Beta release.
**
***************************************************************************
*/
#include <pci.h>
#if NPCI > 0
/*========================================================
**
** Configuration
**
**========================================================
*/
/*
** maximum number of devices which share one interrupt line
*/
#ifndef PCI_MAX_DPI
#define PCI_MAX_DPI (4)
#endif /*PCI_MAX_DPI*/
/*========================================================
**
** #includes and declarations
**
**========================================================
*/
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/errno.h>
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <i386/isa/icu.h>
#include <i386/isa/isa.h>
#include <i386/isa/isa_device.h>
#include <i386/pci/pci.h>
#include <i386/pci/pci_device.h>
#include <i386/pci/pcibios.h>
char ident_pci_c[] =
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
"\n$Id: pci.c,v 1.2 1994/09/01 02:01:34 se Exp $\n"
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
"Copyright (c) 1994, Wolfgang Stanglmeier\n";
/*
** Function prototypes missing in system headers
*/
#if ! (__FreeBSD__ >= 2)
extern pmap_t pmap_kernel(void);
#endif
/*========================================================
**
** Autoconfiguration (of isa bus)
**
**========================================================
*/
/*
** per device (interrupt) data structure.
*/
static struct {
u_short number;
u_short isanum;
struct {
int (*proc)(int dev);
dev_t unit;
} vector[PCI_MAX_DPI];
} pcidata [NPCI];
/*
** check device ready
*/
static int pciprobe (struct isa_device *dev)
{
if (dev->id_unit >= NPCI)
return (0);
if (!pci_conf_mode())
return (0);
return (-1);
}
/*
** initialize the driver structure
*/
static int pciattach (struct isa_device *isdp)
{
pcidata[isdp->id_unit].number = 0;
pcidata[isdp->id_unit].isanum = ffs(isdp->id_irq)-1;
return (1);
}
/*
** ISA driver structure
*/
struct isa_driver pcidriver = {
pciprobe,
pciattach,
"pci"
};
/*========================================================
**
** Interrupt forward from isa to pci devices.
**
**========================================================
*/
void pciintr (int unit)
{
u_short i;
if (unit >= NPCI) return;
for (i=0; i<pcidata[unit].number; i++) {
(void)(*pcidata[unit].vector[i].proc)(pcidata[unit].vector[i].unit);
};
}
/*========================================================
**
** Autoconfiguration of pci devices.
**
** This is reverse to the isa configuration.
** (1) find a pci device.
** (2) look for a driver.
**
**========================================================
*/
/*--------------------------------------------------------
**
** The pci devices can be mapped to any address.
** As default we start at the last gigabyte.
**
**--------------------------------------------------------
*/
#ifndef PCI_PMEM_START
#define PCI_PMEM_START 0xc0000000
#endif
static vm_offset_t pci_paddr = PCI_PMEM_START;
/*---------------------------------------------------------
**
** pci_configure ()
**
**---------------------------------------------------------
*/
static void not_supported (pcici_t tag, u_long type);
void pci_configure()
{
u_char device,last_device;
u_short bus,last_bus;
pcici_t tag;
pcidi_t type;
u_long data;
int unit;
int intpin;
int isanum;
int pci_mode;
struct pci_driver *drp;
struct pci_device *dvp;
/*
** check pci bus present
*/
pci_mode = pci_conf_mode ();
if (!pci_mode) return;
last_bus = pci_last_bus ();
last_device = pci_mode==1 ? 31 : 15;
/*
** hello world ..
*/
#ifndef PCI_QUIET
printf ("PCI configuration mode %d.\n", pci_mode);
printf ("Scanning device 0..%d on pci bus 0..%d "
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
"($Revision: 1.2 $)\n",
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
last_device, last_bus);
#endif
for (bus=0;bus<=last_bus; bus++)
for (device=0; device<=last_device; device ++) {
tag = pcitag (bus, device, 0);
type = pci_conf_read (tag, PCI_ID_REG);
if ((!type) || (type==0xfffffffful)) continue;
/*
** lookup device in ioconfiguration:
*/
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
for (dvp = pci_devtab; dvp->pd_device_id; dvp++) {
if (dvp->pd_device_id == type) break;
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
};
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
drp = dvp->pd_driver;
if (!dvp->pd_device_id) {
int data = pci_conf_read(tag, PCI_CLASS_REV_REG);
enum pci_majclass class = PCI_MAJCLASS_OF(data);
vm_offset_t va;
vm_offset_t pa;
int reg;
switch(class) {
case PCI_MJC_OLD:
if(PCI_MINCLASS_OF(data) != PCI_MIN_OVGA)
break;
case PCI_MJC_DISPLAY:
for (reg = PCI_MAP_REG_START;
reg < PCI_MAP_REG_END;
reg += 4) {
data = pci_map_mem(tag, reg, &va, &pa);
if(data == 0)
printf(
"pci%d:%d: mapped VGA-like device at physaddr %lx\n",
bus, device, (u_long)pa);
}
continue;
default:
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
#ifndef PCI_QUIET
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf("pci%d:%d: ", bus, device);
not_supported (tag, type);
#endif
}
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
if (!drp) {
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
if(dvp->pd_flags & PDF_LOADABLE) {
printf("%s: loadable device on pci%d:%d\n",
dvp->pd_name, bus, device);
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
continue;
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** found it.
** probe returns the device unit.
*/
unit = (*drp->probe) (tag);
if (unit<0) {
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf ("%s <%s>: probe failed on pci%d:%d\n",
drp->name, drp->vendor, bus, device);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
continue;
};
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
if (drp->name) {
printf ("%s%d <%s>", drp->name, unit, drp->vendor);
} else {
printf ("pci%d: <%s>", bus, drp->vendor);
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** install interrupts
*/
data = pci_conf_read (tag, PCI_INTERRUPT_REG);
intpin = PCI_INTERRUPT_PIN_EXTRACT(data);
if (intpin) {
int idx=NPCI;
/*
** Usage of int line register (if set by bios)
** Matt Thomas <thomas@lkg.dec.com>
*/
isanum = PCI_INTERRUPT_LINE_EXTRACT(data);
if (isanum) {
printf (" il=%d", isanum);
for (idx = 0; idx < NPCI; idx++) {
if (pcidata[idx].isanum == isanum)
break;
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
}
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** Or believe to the interrupt pin register.
*/
if (idx >= NPCI) idx = intpin-1;
/*
** And install the interrupt.
*/
if (idx<NPCI) {
u_short entry = pcidata[idx].number;
printf (" irq %c", 0x60+intpin);
if (entry < PCI_MAX_DPI) {
pcidata[idx].vector[entry].proc = drp->intr;
pcidata[idx].vector[entry].unit = unit;
entry++;
};
printf (" isa=%d [%d]",pcidata[idx].isanum, entry);
pcidata[idx].number=entry;
} else {
printf (" not installed");
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
}
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** enable memory access
*/
data = pci_conf_read (tag, PCI_COMMAND_STATUS_REG)
& 0xffff | PCI_COMMAND_MEM_ENABLE;
pci_conf_write (tag, (u_char) PCI_COMMAND_STATUS_REG, data);
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf (" on pci%d:%d\n", bus, device);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** attach device
** may produce additional log messages,
** i.e. when installing subdevices.
*/
(void) (*drp->attach) (tag);
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
#ifndef PCI_QUIET
printf ("pci uses physical addresses from %lx to %lx\n",
(u_long)PCI_PMEM_START, (u_long)pci_paddr);
#endif
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
}
/*-----------------------------------------------------------------------
**
** Map device into port space.
**
** PCI-Specification: 6.2.5.1: address maps
**
**-----------------------------------------------------------------------
*/
extern vm_map_t kernel_map;
int pci_map_port (pcici_t tag, u_long reg, u_short* pa)
{
/*
** @MAPIO@ not yet implemented.
*/
return (ENOSYS);
}
/*-----------------------------------------------------------------------
**
** Map device into virtual and physical space
**
** PCI-Specification: 6.2.5.1: address maps
**
**-----------------------------------------------------------------------
*/
int pci_map_mem (pcici_t tag, u_long reg, vm_offset_t* va, vm_offset_t* pa)
{
u_long data, result;
vm_size_t vsize;
vm_offset_t vaddr;
/*
** sanity check
*/
if (reg <= PCI_MAP_REG_START || reg >= PCI_MAP_REG_END || (reg & 3))
return (EINVAL);
/*
** get size and type of memory
**
** type is in the lowest four bits.
** If device requires 2^n bytes, the next
** n-4 bits are read as 0.
*/
pci_conf_write (tag, reg, 0xfffffffful);
data = pci_conf_read (tag, reg);
switch (data & 0x0f) {
case PCI_MAP_MEMORY_TYPE_32BIT: /* 32 bit non cachable */
break;
default: /* unknown */
return (EINVAL);
};
/*
** mask out the type,
** and round up to a page size
*/
vsize = round_page (-(data & PCI_MAP_MEMORY_ADDRESS_MASK));
if (!vsize) return (EINVAL);
/*
** align physical address to virtual size
*/
if (data = pci_paddr % vsize)
pci_paddr += vsize - data;
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
vaddr = pmap_mapdev (pci_paddr, vsize);
#if 0
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** display values.
*/
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf (" virtual=0x%lx physical=0x%lx\n", (u_long)vaddr,
(u_long)pci_paddr);
#endif
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
/*
** return them to the driver
*/
*va = vaddr;
*pa = pci_paddr;
/*
** set device address
*/
pci_conf_write (tag, reg, pci_paddr);
return (0);
}
struct vt {
u_short ident;
char* name;
};
struct dt {
u_long ident;
char* name;
};
static struct vt VendorTable[] = {
{0x1002, "ATI TECHNOLOGIES INC"},
{0x1011, "DIGITAL EQUIPMENT COMPANY"},
{0x101A, "NCR"},
{0x102B, "MATROX"},
{0x1045, "OPTI"},
{0x5333, "S3 INC."},
{0x8086, "INTEL CORPORATION"},
{0,0}
};
static struct dt DeviceTable[] = {
{0x04848086, " 82378IB pci-isa bridge"},
{0x04838086, " 82424ZX cache dram controller"},
{0x04828086, " 82375EB pci-eisa bridge"},
{0x04A38086, " 82434LX pci cache memory controller"},
{0,0}
};
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
static const char *const majclasses[] = {
"old", "storage", "network", "display", "multimedia", "memory",
"bridge"
};
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
void not_supported (pcici_t tag, u_long type)
{
u_char reg;
u_long data;
struct vt * vp;
struct dt * dp;
/*
** lookup the names.
*/
for (vp=VendorTable; vp->ident; vp++)
if (vp->ident == (type & 0xffff))
break;
for (dp=DeviceTable; dp->ident; dp++)
if (dp->ident == type)
break;
/*
** and display them.
*/
if (vp->ident) printf (vp->name);
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
else printf ("vendor=%lx", type & 0xffff);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
if (dp->ident) printf (dp->name);
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
else printf (", device=%lx", type >> 16);
data = pci_conf_read(tag, PCI_CLASS_REV_REG);
if (PCI_MAJCLASS_OF(data) < sizeof(majclasses) / sizeof(majclasses[0]))
printf(", class=%s", majclasses[PCI_MAJCLASS_OF(data)]);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
printf (" [not supported]\n");
for (reg=PCI_MAP_REG_START; reg<PCI_MAP_REG_END; reg+=4) {
data = pci_conf_read (tag, reg);
if (!data) continue;
switch (data&7) {
case 1:
case 5:
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf (" map(%lx): io(%lx)\n", reg, data & ~3);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
break;
case 0:
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf (" map(%lx): mem32(%lx)\n", reg, data & ~7);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
break;
case 2:
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf (" map(%lx): mem20(%lx)\n", reg, data & ~7);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
break;
case 4:
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
printf (" map(%lx): mem64(%lx)\n", reg, data & ~7);
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
break;
Added a bit of missing functionality to make this work correctly on a wider variety of systems. Include the deivers from pci_intel.c in pci_config.c (I hope this is what was intended; my system works ok). Use pmap_mapdev(). Automatically map any large linear frame buffers or whatnot in VGA-style devices which ordinarily would not have their own drivers, and don't call not_supported() for them. (This shuts up complaints about my Matrox card.) Include the beginnings of what could eventually become dynamically-loadable PCI devices. Allow for the possibility of PCI devices simply providing a PCI veneer over an existing ISA device, and shut up about them, too. Make autoconfiguration text conform more to the style of other supported buses.
1994-09-14 01:34:51 +00:00
}
}
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + Stefan Esser <se> Directory for PCI autoconfigure and device driver code.
1994-09-01 01:45:19 +00:00
}
#endif
Reference in New Issue Copy Permalink