OSDN Git Service
Dave Airlie [Wed, 5 Aug 2009 23:42:06 +0000 (09:42 +1000)]
Merge branch 'master' into vga-arbiter
Dave Airlie [Wed, 5 Aug 2009 23:39:01 +0000 (09:39 +1000)]
pciaccess: make linux rom reading fallback optional
On kernels which have ROM files, the boot VGA rom file
will contain the 0xc0000 file anyways. If another card
is missing a rom file there is no need for this.
This is fallback is disabled by default, on the premise
that distro shipping this new a pciaccess will be running
on kernels with working rom support which is been upstream
for a long time
Dave Airlie [Wed, 5 Aug 2009 23:39:01 +0000 (09:39 +1000)]
pciaccess: make linux rom reading fallback optional
On kernels which have ROM files, the boot VGA rom file
will contain the 0xc0000 file anyways. If another card
is missing a rom file there is no need for this.
This is fallback is disabled by default, on the premise
that distro shipping this new a pciaccess will be running
on kernels with working rom support which is been upstream
for a long time
Dave Airlie [Mon, 3 Aug 2009 02:43:26 +0000 (12:43 +1000)]
vgaarb: add vga count + resource accessor
Dave Airlie [Sun, 2 Aug 2009 23:39:22 +0000 (09:39 +1000)]
vgaarb: more api changes + cleanup
add a comma for ease of parsing,
use hex on the interface in both directions instead of hex on one
and decimal on the other
Dave Airlie [Fri, 31 Jul 2009 06:15:55 +0000 (16:15 +1000)]
vgaarb: change API to target taking a device + lock/unlock not taking one
working on the target device - a target device of NULL
reselects the target that existed when we started.
this is mainly to allow a udev posting tool to return to the
state of what happened before it ran.
Dave Airlie [Fri, 31 Jul 2009 05:34:32 +0000 (15:34 +1000)]
vgaarb: add support for targetting default device
Dave Airlie [Fri, 31 Jul 2009 01:17:21 +0000 (11:17 +1000)]
vgaarb: clean up debug output
Dave Airlie [Fri, 31 Jul 2009 01:16:16 +0000 (11:16 +1000)]
Merge remote branch 'main/master' into vga-arbiter
Dave Airlie [Mon, 20 Jul 2009 03:26:41 +0000 (13:26 +1000)]
vgaarb: fixup api for new count reporting + decodes interface was wrong.
decodes is a information setting interface so it takes new decodes
and sends them to the kernel
Dave Airlie [Thu, 16 Jul 2009 05:36:30 +0000 (15:36 +1000)]
add support for finding if something has a kernel driver
Jesse Barnes [Tue, 30 Jun 2009 22:24:45 +0000 (15:24 -0700)]
Don't try to use bogus bridge data
If, for whatever reason, we weren't able to read bridge data (therefore
leaving bridge.pci or bridge.pcmcia empty), we shouldn't try to look at
the bus data inside the pci/pcmcia struct.
Dave Airlie [Thu, 4 Jun 2009 23:32:10 +0000 (09:32 +1000)]
vgaarb: fixup api and decode rsrc.
This fixes up the API and stores the vga arb fd in the sys_pci structure,
instead of hiding it in a random dev struct.
It also reads back after setting the target and works out the decodes.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Tiago Vignatti [Thu, 14 May 2009 20:30:17 +0000 (17:30 -0300)]
Comment typo.
Tiago Vignatti [Thu, 14 May 2009 20:28:05 +0000 (17:28 -0300)]
Reorganize the code.
Tiago Vignatti [Thu, 14 May 2009 06:42:56 +0000 (03:42 -0300)]
Fix typo in vgaarb stubs.
Tiago Vignatti [Thu, 14 May 2009 06:40:25 +0000 (03:40 -0300)]
Delete vga_arb_rec.
Tiago Vignatti [Wed, 13 May 2009 21:31:09 +0000 (18:31 -0300)]
vgaarb: a bunch of clean ups to fit in libpciaccess style.
Tiago Vignatti [Wed, 13 May 2009 20:55:31 +0000 (17:55 -0300)]
vgaarb: make it compile.
Tiago Vignatti [Wed, 13 May 2009 20:44:46 +0000 (17:44 -0300)]
Import libvgaaccess (VGA Arbiter) implementation from C3SL repository:
http://git.c3sl.ufpr.br/
Aaron Plattner [Fri, 1 May 2009 22:04:19 +0000 (15:04 -0700)]
Don't try to write boot_vga since that fails with EACCES
Dave Airlie [Wed, 4 Mar 2009 05:55:11 +0000 (15:55 +1000)]
pciaccess: provide a method to detect if a device is boot VGA
When the linux kernel exposes this information, we can use this interface
in the X server to detect whether the kernel believes the device we
are looking at is the boot VGA device.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Alan Coopersmith [Fri, 24 Apr 2009 23:32:12 +0000 (16:32 -0700)]
Version bump: 0.10.6
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Alan Coopersmith [Fri, 24 Apr 2009 23:19:34 +0000 (16:19 -0700)]
Resync COPYING file with code copyright notices
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Alan Coopersmith [Sat, 18 Apr 2009 05:05:03 +0000 (22:05 -0700)]
Fix unused variable warnings in Solaris code
solx_devfs.c: In function `find_target_node':
solx_devfs.c:672: warning: unused variable `prop'
solx_devfs.c:673: warning: unused variable `i'
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Alan Coopersmith [Sat, 18 Apr 2009 04:59:25 +0000 (21:59 -0700)]
Fix $CWARNFLAGS addition
Original addition to configure.ac had typo (missing N) - moving to
Makefile.am allows easier override at build time
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Alan Coopersmith [Sat, 18 Apr 2009 04:56:31 +0000 (21:56 -0700)]
Move included m4 to a subdir so we can more easily update from upstream
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Owain Ainsworth [Sat, 18 Apr 2009 04:38:46 +0000 (06:38 +0200)]
openbsd_pci.c: use the correct size when mapping the legacy vga rom.
fixes errors (and probably bugs) on intel hardware (at the least).
Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Alan Coopersmith [Wed, 15 Apr 2009 17:06:49 +0000 (10:06 -0700)]
Sun bug
6811468: pci_device_solx_devfs_probe accesses freed memory
di_fini() is being called in pci_device_solx_devfs_probe()
The di_fini (3DEVINFO) man page says "All handles associated with this
snapshot become invalid after the call to di_fini()". But after that,
eight lines down, the subroutine was calling di_prop_lookup_ints with
a handle args.node which was stored from walking the device tree, and
then using the pointers that returned even further down.
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Alan Coopersmith [Wed, 15 Apr 2009 00:29:10 +0000 (17:29 -0700)]
Fix const mismatch compiler warnings in Solaris backend
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
edward shu [Wed, 15 Apr 2009 00:19:09 +0000 (17:19 -0700)]
Sun bug
6785726: libpciaccess needs to use the correct BAR index on solaris
<http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=
6785726>
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Mark Kettenis [Tue, 14 Apr 2009 19:53:26 +0000 (21:53 +0200)]
Make libpciaccess on OpenBSD domain-aware.
Makes X capable of discovering PCI devices
in other domains (on macppc or sparc64 for example).
Darren Smith [Tue, 14 Apr 2009 14:46:26 +0000 (10:46 -0400)]
Be more paranoid about reading bridge info (#20786)
Robert Noland [Fri, 27 Feb 2009 06:34:28 +0000 (00:34 -0600)]
FreeBSD: Incorporate several fixes that have accumulated.
-Don't frob the BARs while they are enabled
-Find proper pci bios address / length
-Use the new PCIOCGETBAR ioctl if it exists
rather than frob the BARs ourself
-Write must also be a power of two
Alan Coopersmith [Tue, 3 Feb 2009 04:34:38 +0000 (20:34 -0800)]
Add README with pointers to mailing list, bugzilla & git repos
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Paulo Cesar Pereira de Andrade [Wed, 28 Jan 2009 20:03:21 +0000 (18:03 -0200)]
Janitor: Correct make distcheck and compiler warnings.
Alan Coopersmith [Fri, 23 Jan 2009 00:14:22 +0000 (16:14 -0800)]
Solaris: Use bus-range properties to limit busses scanned on each node
Based on code provided by Dan.Mick@sun.com
Matthieu Herrb [Sun, 7 Dec 2008 18:18:29 +0000 (19:18 +0100)]
Fix a logic error in pci_device_netbsd_write()
This is the same error as in pci_device_openbsd_write() that actually
prevented it to write anything.
Matthieu Herrb [Sun, 7 Dec 2008 18:00:22 +0000 (19:00 +0100)]
Sync with OpenBSD code.
- implement pci_read_rom()
- only set MTRR on x86 cpus
- failure to set non-cacheable attributes is not fatal
- fix a logic error in pci_write()
Matthieu Herrb [Tue, 25 Nov 2008 11:42:15 +0000 (12:42 +0100)]
Fix a logic error in pci_device_openbsd_write().
Alan Coopersmith [Wed, 19 Nov 2008 17:37:49 +0000 (09:37 -0800)]
Add AC_SYS_LARGEFILE to configure.ac
Alan Coopersmith [Wed, 19 Nov 2008 16:22:22 +0000 (08:22 -0800)]
Don't open/gzopen pci.ids if we're not going to read it
Alan Coopersmith [Wed, 19 Nov 2008 03:54:42 +0000 (19:54 -0800)]
Fill in byte swapping routines for big-endian Solaris machines
Julien Cristau [Fri, 31 Oct 2008 17:07:52 +0000 (18:07 +0100)]
Bump to 0.10.5
Julien Cristau [Fri, 31 Oct 2008 16:33:50 +0000 (17:33 +0100)]
Make --without-zlib work
Robert Noland [Thu, 30 Oct 2008 02:20:35 +0000 (22:20 -0400)]
Fix FreeBSD systems which support pci domains.
Support for FreeBSD based systems which support pci domains was broken
in the commit to support kFreeBSD. Include config.h so that things are
happy again.
Alan Coopersmith [Fri, 17 Oct 2008 21:09:52 +0000 (14:09 -0700)]
Version 0.10.4
Alan Coopersmith [Fri, 17 Oct 2008 21:09:43 +0000 (14:09 -0700)]
Add scanpci.man to EXTRA_DIST
Alan Coopersmith [Fri, 17 Oct 2008 20:51:28 +0000 (13:51 -0700)]
Update COPYING with the rest of the copyright/license notices from the code
Alan Coopersmith [Fri, 17 Oct 2008 20:48:46 +0000 (13:48 -0700)]
Add scanpci man page from Xorg
Alan Coopersmith [Fri, 17 Oct 2008 04:17:24 +0000 (21:17 -0700)]
More minor Solaris cleanups
Alan Coopersmith [Fri, 17 Oct 2008 01:18:09 +0000 (18:18 -0700)]
Clean up formatting of solx_devfs.c
Alan Coopersmith [Thu, 16 Oct 2008 23:55:27 +0000 (16:55 -0700)]
Fix bus probing on Solaris/SPARC
Alan Coopersmith [Thu, 16 Oct 2008 20:33:01 +0000 (13:33 -0700)]
Correct Sun license notice
Alan Coopersmith [Thu, 16 Oct 2008 01:23:03 +0000 (18:23 -0700)]
scanpci: add -v flag to enable verbose mode like old scanpci
Alan Coopersmith [Wed, 15 Oct 2008 23:14:18 +0000 (16:14 -0700)]
Correct comment about devices used on Solaris
Alan Coopersmith [Wed, 15 Oct 2008 22:35:25 +0000 (15:35 -0700)]
Fix various typos (mostly in comments)
Juan RP [Sat, 11 Oct 2008 18:28:03 +0000 (20:28 +0200)]
Add NetBSD support
It's based in the OpenBSD code, but with differences because on NetBSD,
to set MTRRs we have to use i386/x86_64_set_mtrr() and link to libi386
or libx86_64.
X.Org bug#17931 <http://bugs.freedesktop.org/show_bug.cgi?id=17931>
Petr Salinger [Sat, 11 Oct 2008 18:22:28 +0000 (20:22 +0200)]
Add support for GNU/kFreeBSD
We need to initialize the FreeBSD backend on GNU/kFreeBSD and detect
whether pci_io.pi_sel.pc_domain member exists.
X.Org bug#17882 <http://bugs.freedesktop.org/show_bug.cgi?id=17882>
John Tapsell [Tue, 23 Sep 2008 14:26:09 +0000 (17:26 +0300)]
configure.ac: Make mtrr.h test more portable
Use the standard AC_CHECK_HEADERS, which works for cross-compiling as
well as in various other locations.
Matthieu Herrb [Sat, 20 Sep 2008 16:28:59 +0000 (18:28 +0200)]
OpenBSD: allow 2 successives calls to pci_system_init().
And fix pci_system_cleanup() to make it possible to call pci_system_init()
again.
ok kettenis at openbsd.
Kel Modderman [Tue, 8 Jul 2008 11:09:52 +0000 (13:09 +0200)]
Handle compressed pci.ids
Add an option to build with zlib support so we can find
vendor/device information if the pci.ids file is gzipped.
Signed-off-by: Julien Cristau <jcristau@debian.org>
Julien Cristau [Wed, 12 Mar 2008 14:00:26 +0000 (15:00 +0100)]
Hide one more private symbol
Jesse Barnes [Mon, 23 Jun 2008 18:24:04 +0000 (11:24 -0700)]
Support write combine resource files in Linux sysfs
Starting with version 2.6.26, Linux will support resourceN_wc files which
export write combining mappings of PCI resource ranges, so support them if
present in libpciaccess.
Stefan Dirsch [Mon, 23 Jun 2008 08:50:30 +0000 (10:50 +0200)]
Fixed typo in output ("performace" --> "performance").
Adam Jackson [Tue, 10 Jun 2008 19:24:56 +0000 (15:24 -0400)]
libpciaccess 0.10.3
Adam Jackson [Tue, 10 Jun 2008 19:23:06 +0000 (15:23 -0400)]
libpciaccess 0.10.2
Eric Anholt [Thu, 5 Jun 2008 18:39:06 +0000 (11:39 -0700)]
Catch and recover from yet another linux kernel bug in mprotect.
Adam Jackson [Wed, 21 May 2008 17:44:38 +0000 (13:44 -0400)]
Linux: Fail gracefully on machines without PCI.
Dave Airlie [Wed, 21 May 2008 06:10:37 +0000 (16:10 +1000)]
pciaccess: bump to 0.10.1
Dave Airlie [Wed, 21 May 2008 06:10:24 +0000 (16:10 +1000)]
linux: add pci_device_enable entrypoint and sysfs support for it
Alan Coopersmith [Fri, 9 May 2008 22:15:42 +0000 (15:15 -0700)]
Fix lint warnings in solx_devfs.c
Alan Coopersmith [Fri, 9 May 2008 22:05:46 +0000 (15:05 -0700)]
Add pci_system_solx_devfs_create prototype to pciaccess_private.h
Alan Coopersmith [Fri, 9 May 2008 21:49:32 +0000 (14:49 -0700)]
Initialize err to 0 in pci_device_solx_devfs_map_range
Prevents returning errors when mapping actually succeeds
Keith Packard [Wed, 16 Apr 2008 19:10:52 +0000 (12:10 -0700)]
Kludge around linux bug and turn off write-through and cache-disable bits
When mmaping the PCI device, the kernel turns on the write-through and
cache-disable bits in the allocated PTEs. This disables write-combining mode
and dramatically reduces write bandwidth to the frame buffer. While that
should be fixed in the kernel, we'll kludge around it here by using mprotect
to rewrite the PTEs and get those bits turned off.
Hasso Tepper [Mon, 7 Apr 2008 12:28:44 +0000 (15:28 +0300)]
Add DragonFly BSD support
DragonFly behaves exactly like FreeBSD, so no problem here.
Danny van Dyk [Sat, 29 Mar 2008 00:01:29 +0000 (17:01 -0700)]
Fix function prototypes for C++
Stuart Bennett [Fri, 14 Mar 2008 15:58:33 +0000 (11:58 -0400)]
Bug #13988: Fix reads from "0"-sized ROMs.
Mark Kettenis [Wed, 12 Mar 2008 20:29:58 +0000 (21:29 +0100)]
OpenBSD support for libpciaccess.
xserver and libpciaccess both need to open /dev/xf86, which can only
be opened once. I implemented pci_system_init_dev_mem() like Ian
suggested. This requires some minor changes to the BSD-specific
os-support code. Since pci_system_init_dev_mem() is a no-op on
FreeBSD this should be no problem.
Julien Cristau [Fri, 7 Mar 2008 13:41:47 +0000 (14:41 +0100)]
solaris: add pci_tools.h to OS_SUPPORT
pci_tools.h was missing from the tarball, but is needed by solx_devfs.c
Adam Jackson [Thu, 6 Mar 2008 20:31:51 +0000 (15:31 -0500)]
libpciaccess 0.10
Doug Chapman [Thu, 6 Mar 2008 19:22:17 +0000 (14:22 -0500)]
Bug #14818: Actually return the PCI vendor name, instead of NULL.
Julien Cristau [Tue, 19 Feb 2008 13:45:40 +0000 (14:45 +0100)]
Don't export private symbols.
Julien Cristau [Tue, 19 Feb 2008 13:26:54 +0000 (14:26 +0100)]
Use <byteswap.h> when using glibc, not just on linux
This fixes the build on GNU/kFreeBSD.
James Cloos [Thu, 6 Dec 2007 21:38:51 +0000 (16:38 -0500)]
Replace static ChangeLog with dist-hook to generate from git log
Matthias Hopf [Tue, 23 Oct 2007 13:19:36 +0000 (15:19 +0200)]
Bit-fields have to have type _Bool, signed int, or unsigned int.
ISO/IEC 9899:1999 (E), 6.7.2.1 Structure and union specifiers, (4).
_Bool is only supported for C99 and up, and 1-bit signed types don't make
sense -> unsigned int.
Ian Romanick [Thu, 18 Oct 2007 22:59:14 +0000 (15:59 -0700)]
Update bug reporting link.
Eric Anholt [Wed, 10 Oct 2007 22:20:51 +0000 (15:20 -0700)]
FreeBSD: for 64-bit BARs, skip the resource slot used for the upper 32 bits.
This gets us the same resource numbering as on Linux.
Eric Anholt [Wed, 10 Oct 2007 21:55:45 +0000 (14:55 -0700)]
FreeBSD: Don't try to unset an MTRR if we didn't set it.
Eric Anholt [Wed, 10 Oct 2007 21:32:38 +0000 (14:32 -0700)]
Add domain output to scanpci.
Eric Anholt [Wed, 10 Oct 2007 21:32:09 +0000 (14:32 -0700)]
FreeBSD: Add support for multiple PCI domains.
Eric Anholt [Tue, 9 Oct 2007 19:13:49 +0000 (12:13 -0700)]
FreeBSD: Fix unmap_range to return an error value.
Eric Anholt [Tue, 9 Oct 2007 19:12:34 +0000 (12:12 -0700)]
FreeBSD: don't set the MTRR if it's the default mode (uncacheable).
Alan Coopersmith [Thu, 27 Sep 2007 22:22:51 +0000 (15:22 -0700)]
Need to link with -ldevinfo on Solaris
Alan Coopersmith [Thu, 27 Sep 2007 00:26:47 +0000 (17:26 -0700)]
Fix Solaris build: missing static prototype & typo in variable name
Eric Anholt [Tue, 11 Sep 2007 15:07:55 +0000 (15:07 +0000)]
Add FreeBSD MTRR setting support.
Ian Romanick [Tue, 4 Sep 2007 23:13:24 +0000 (16:13 -0700)]
Add stub version of pci_device_map_memory_range
This stub version of pci_device_map_memory_range allows the vesa driver and
other users of this interface to continue functioning with current
libpciaccess bits. That said, users of this interface should convert over
to pci_device_map_range as soon as possible.
James Cloos [Mon, 3 Sep 2007 10:15:58 +0000 (06:15 -0400)]
Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/lib/libpciaccess
James Cloos [Mon, 3 Sep 2007 09:53:57 +0000 (05:53 -0400)]
Add *~ to .gitignore to skip patch/emacs droppings
Keith Packard [Fri, 31 Aug 2007 20:43:18 +0000 (13:43 -0700)]
Don't add MTRR for uncached regions. Remove MTRR on unmap.
MTRR regions aren't needed for uncached mappings, so don't add them. Also,
when unmapping memory, remove the MTRR entry.