Android porting

Copy include/io.h from /usr/include/sys/io.h.

Copy hwdata/pci.ids from git://git.fedorahosted.org/hwdata.git
commit 5732b82c9d6fdbede9095cf11fe72852c37b8e9d.

linux_sysfs: include <limits.h> for PATH_MAX

Signed-off-by: Felix Janda <felix.janda@posteo.de>
Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>

autogen.sh: pass --force to autoreconf, quote string variables

By passing --force autoreconf will update all the aux files, which would
otherwise be ignored if one updates autoconf/automake.

Quote the $srcdir, $ORIGDIR and $0 variables to prevent fall-outs, when
it contains space.

v2: Also handle $srcdir and $0.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>

chmod a-x README.cygwin

Signed-off-by: Adam Jackson <ajax@redhat.com>

linux_sysfs.c: Include <limits.h> for PATH_MAX

Fixes compilation with musl libc.

Tested-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Felix Janda <felix.janda@posteo.de>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Fix quoting issue.

m4 has '[]' as quoting characters, so if we want '[]' to
end up in the configure script, we need to quote them again.

Reported by Greg Troxel <gdt@ir.bbn.com>.

Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Include config.h before anything else in *.c

Debian bug#749008 <https://bugs.debian.org/749008>

Reported-by: Michael Tautschnig <mt@debian.org>
Signed-off-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>

libpciaccess 0.13.4

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Bug 89151: Fix build on NetBSD/alpha and NetBSD/sparc64.

This patch is required for building on NetBSD/alpha (needs platform
specific -lalpha) and NetBSD/sparc64 (no platform specific library,
just -lpci). The patch also generalizes to support all NetBSD
platforms with pci bus.

https://bugs.freedesktop.org/show_bug.cgi?id=89151

Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Thomas Klausner <wiz@NetBSD.org>

Check for __linux__ instead of linux during compilation

"__linux__" is the POSIX define for checking for Linux OS, "linux" is
deprecated and apparently not supported by Android.

Besides correcting the define, patch adds error for the case OS isn't
recognized.

Signed-off-by: Eero Tamminen <eero.t.tamminen@intel.com>
v2: Drop double negative from commit msg. Spotted by Mateusz Jończyk.
[Emil Velikov: Remove irrelevant Android details from commit msg.]
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Include config.h before any other headers.

The former has a series of defines which in some cases are crusial to be
set before including any system headers.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Add missing guards around config.h inclusion

In some cases the header may not exist, leading to compilation issues.
Add the ifdef HAVE_CONFIG_H, as it is those in other parts of the
project.

[Emil Velikov: Split out from a larger commit, rework commit message]
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Include the POSIX fcntl.h instead of sys/fcntl.h

The former is part of the POSIX standard, and (unlike the latter) is
more widely available.

v2: Drop gracious d in header name. Spotted by Alan.

Cc: Adam Jackson <ajax@redhat.com>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

libpciaccess 0.13.3

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Use PCIOCREADMASK on OpenBSD.

If the machdep.allowaperture sysctl(8) variable is set to 0, writing to PCI
config space is not allowed.  So instead of writing 0xffffffff to the BARs
in order to determine their size, use the PCIOCREADMASK ioctl(2) which
returns the mask of changeable bits that was saved by the kernel when the
devices was initially probed.

Reviewed-by: Matthieu Herrb <matthieu@herbb.eu>
Signed-off-by: Mark Kettenis <kettenis@openbsd.org>

Implement the kernel_has_driver() method for NetBSD.

This has the benefit of stopping the "vesa" driver from loading on
hardware that has been claimed by a kernel driver and thus shouldn't
be using "vesa".

Reviewed-by: Matthieu Herrb <matthieu@herrb.eu>
Signed-off-by: Thomas Klausner <wiz@NetBSD.org>

Solaris: Fix fd leak in pci_device_solx_devfs_map_range()

Caching fd's for reuse is most effective when you actually stick the
newly opened fd in the cache, instead of letting it leak at the end
of the function.

Regression introduced by yours truly in commit 9f2d95e61896f41adb.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Randy Fishel <randy.fishel@oracle.com>

Enable use of __attribute__((deprecated)) with Solaris Studio 12.4 compiler

Support for this attribute is added in the 12.4 beta release.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Fix IO access functions on linux+sysfs.

The offsets on the resourceX files are relative to BAR base - don't add
the base address ourselves.

Reviewed-by: Adam Jackson <ajax@redhat.com>

Fix a compilation error on GNU Hurd platforms.

This fixes a typo (unAMp instead of unMAp) introduced in the Windows/Cygwin
support patch.

Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>

Windows/Cygwin: Add support through the WinIo library

V2:
- Add support for unmapping
- Add a README.cygwin

Reviewed-by: Martin Peres <martin.peres@free.fr>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>

Improve NetBSD i386 detection.

Sometimes, other values are reported for the CPU part, like
"i486--netbsdelf").

From: Mark Davies <markd@NetBSD.org>
Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-By: Matt Dew <marcoz@osource.org>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Fix zlib handling on NetBSD.

Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

libpciaccess 0.13.2

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Fix Sun bug #7035791: scanpci should report 64 bit registers

On a system which has allocated 64-bit device addresses scanpci -v
will show entries such as:
[...]
  BASE0     0xce000000 SIZE 16777216  MEM
  BASE1     0x00000fc1f8000000 SIZE 134217728  MEM64 PREFETCHABLE
  BASE3     0x00000fc1f6000000 SIZE 33554432  MEM64 PREFETCHABLE
[...]
instead of:
[...]
  BASE0     0xce000000 SIZE 16777216  MEM
  BASE1     0xf8000000 SIZE 134217728  MEM PREFETCHABLE
  BASE3     0xf6000000 SIZE 33554432  MEM PREFETCHABLE
[...]

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Resync COPYING file with copyrights/licenses from code

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

pci_device_solx_devfs_probe: implicit conversion changes signedness warnings

solx_devfs.c:615:32: warning: implicit conversion changes signedness:
 'unsigned long' to 'int' [-Wsign-conversion]
    for (i = 0; i < len; i = i + CELL_NUMS_1275) {
                           ~ ~~^~~~~~~~~~~~~~~~
solx_devfs.c:615:30: warning: implicit conversion changes signedness:
 'int' to 'unsigned long' [-Wsign-conversion]
    for (i = 0; i < len; i = i + CELL_NUMS_1275) {
                             ^ ~

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Solaris: Remove #defines for Ultra 45 southbridge device ids

Workaround for these chipsets was removed in commit d76fb36d9c28be0f7c
with the switch of probing methods from the PCI address poking that upset
them to using the already collected data from the kernel via devinfo,
but the #define to identify them was left behind unused - remove it too.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

pci_device_solx_devfs_read: fix sign conversion/comparison warnings

Fixes:
solx_devfs.c:824:19: warning: comparison of integers of different signs: 'int' and 'pciaddr_t' (aka 'unsigned long') [-Wsign-compare]
    for (i = 0; i < size; i += PCITOOL_ACC_ATTR_SIZE(PCITOOL_ACC_ATTR_SIZE_1))
                ~ ^ ~~~~
solx_devfs.c:826:28: warning: implicit conversion changes signedness: 'int' to 'unsigned long' [-Wsign-conversion]
        cfg_prg.offset = offset + i;
                                ~ ^
solx_devfs.c:844:19: warning: implicit conversion changes signedness: 'int' to 'pciaddr_t' (aka 'unsigned long') [-Wsign-conversion]
    *bytes_read = i;
                ~ ^

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Support more pci functions on NetBSD.

Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Fix build on NetBSD-4.

Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Enable boot_vga support only if WSDISPLAYIO_GET_BUSID is defined.

Fixes build on NetBSD-5.

Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

pci_id_file_open: move variable inside the #ifndef that uses it

Clears up gcc warning on Solaris:
common_device_name.c: In function 'pci_id_file_open':
common_device_name.c:83:17: warning: unused variable 'result' [-Wunused-variable]

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Define PCI_MATCH_ANY as an unsigned int

Clears a bunch of clang warnings of the forms:

common_device_name.c:345:23: warning: comparison of integers of different signs:
 'const uint32_t' (aka 'const unsigned int') and 'int' [-Wsign-compare]
    if ( m->vendor_id == PCI_MATCH_ANY ) {
         ~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~

common_device_name.c:498:19: warning: implicit conversion changes signedness:
 'int' to 'uint32_t' (aka 'unsigned int') [-Wsign-conversion]
    m.device_id = PCI_MATCH_ANY;
                ~ ^~~~~~~~~~~~~

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Update NetBSD support.

Signed-off-by: Michael Lorenz <macallan@NetBSD.org>
Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Move 'const' to correct place, so it has meaning.

Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Protect config.h like usual.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Bug 63583 - add legacy IO routines for FreeBSD

https://bugs.freedesktop.org/show_bug.cgi?id=63583

Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
Signed-off-by: Jung-uk Kim <jkim@FreeBSD.org>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Solaris: probe improvement

Remove pcitool dependency in probing phase. Use the data
collected from devinfo tree instead in creating pci file
system.

Signed-off-by: Henry Zhao <henzhao@glory3.(none)>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Solaris: Add domain support for sparc platform

As a result the code of finding nexus node for a device in sparc is
simplified and made the same as x86.

Signed-off-by: Henry Zhao <henry.zhao@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

autogen.sh: Implement GNOME Build API

http://people.gnome.org/~walters/docs/build-api.txt

Signed-off-by: Adam Jackson <ajax@redhat.com>

configure: Remove AM_MAINTAINER_MODE

Signed-off-by: Adam Jackson <ajax@redhat.com>

PCI: set correct mask value when matching for bridges.

The mask must not be zero otherwise the matching condition will never
be true: ((val & mask) == set).

Signed-off-by: Egbert Eich <eich@freedesktop.org>
Reviewed-by: Guillem Jover <guillem@hadrons.org>

scanpci: print meaningful info on BASEROM

Signed-off-by: Henry Zhao <henry.zhao@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

libpciaccess: sparc rom read fix

Get rom address and size from device's "assigned-addresses" property
if they exist.

Signed-off-by: Henry Zhao <henry.zhao@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

libpciaccess: implementation of boot_vga in Solaris

Signed-off-by: Henry Zhao <henry.zhao@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

libpciaccess: fix memory usage errors

This patch fixes memory usage errors that only occur on large SPARC
machines with more than 256 PCI devices. In this case, memory is being
used after it has been freed by calls to both free() and realloc().
This error was introduced by a previous patch:
2011-03-30 Solaris support for multiple PCI segments (domains)

Signed-off-by: Mark Logan <mark.logan@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Implement legacy io & map for x86 backend

Add the legacy io and map methods for the x86 backend, using ioperm,
in/out, and the existing mmap method.

Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>

Solaris: refactor pci_device_solx_devfs_map_range to reduce code duplication

The sparc & x86 cases were doing essentially the same things with
different paths, so make the path setup be inside the platform
specific #ifdefs, make the open, mmap, & error handling common code.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Solaris: Implement map_legacy & legacy_io functions

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jay Cotton <jay.cotton@oracle.com>

Bump to 0.13.1

Signed-off-by: Julien Cristau <jcristau@debian.org>

Fix pci_io_handle allocation routines.

The previous version of new_io_handle()  would invalidate all previous
allocations when realloc moves the base pointer of the ios array.
Since I cannot figure out where this array is useful, just get rid of
it, providing sound stable memory allocation.

Fixes vgaHWSaveFonts() in 1.12 xserver on OpenBSD, but other
sub-systems relaying on pci_io could be affected too.

Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Tested-by: Tormod Volden <debian.tormod@gmail.com>

Fix openbsd_pci.c:pci_device_vgaarb_set_target();

A return (0) got lost, add it back.

Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Reviewed-by: Mark Kettenis <mark.kettenis@xs4all.nl>

linux: Fix out[bwl] macros on non-I/O architectures

Signed-off-by: Adam Jackson <ajax@redhat.com>

Merge branch 'master' of ssh://git.freedesktop.org/git/xorg/lib/libpciaccess

configure.ac: Bump to 0.13

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

Use O_CLOEXEC for internal file descriptors

Well, almost all of them.  zlib doesn't support "e" in the mode string
in gzopen() though it will silently accept and ignore it, and Solaris appears
not to support "e" in the mode string at all.

Signed-off-by: Adam Jackson <ajax@redhat.com>

linux: Don't use /dev/port

Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>

Use correct type for pci_id file accessor functions

This fixes the compiler warning of using the wrong type for gzgets() and
gzclose() as they want a gzFile argument, not a pointer to gzFile. The
abstraction layer pci_id_file should just abstract the full type.

Signed-off-by: Mathias Krause <mathias.krause@secunet.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>

solx_devfs.c: fix gcc warnings about casting away const when reading data

solx_devfs.c: In function `pci_device_solx_devfs_write':
solx_devfs.c:1085: warning: cast discards qualifiers from pointer target type
solx_devfs.c:1089: warning: cast discards qualifiers from pointer target type
solx_devfs.c:1093: warning: cast discards qualifiers from pointer target type
solx_devfs.c:1097: warning: cast discards qualifiers from pointer target type

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Close pci.ids file when bailing out because realloc failed

Error: File Leak
   Leaked File f
      at line 272 of src/common_device_name.c in function 'populate_vendor'.
        f initialized at line 204 with fopen("/usr/share/hwdata/pci.ids", "r").

[ This bug was found by the Parfait 0.5.0 bug checking tool.
  For more information see http://labs.oracle.com/projects/parfait/ ]

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Solaris: reorder functions to remove need for lots of static prototypes

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

OpenBSD: Implement map_legacy and legacy_io

Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
Reviewed-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Matthieu Herrb <matthieu.herrb@laas.fr>

delete_io_handle: fix deletion of last handle

When num_ios goes from 1 to 0, a realloc(ios, 0); call is made.
This is equivalent to free(ios) and NULL is returned.

However, the previous logic in the code incorrectly discards this NULL
return value. When we next call new_io_handle(), realloc(ios, X) is
called with "ios" pointing to freed memory. This causes glibc to abort.

Correct this logic to detect the 1-to-0 case and handle it correctly.
Other cases are unchanged; there is still value in checking the
return value from realloc() as it also returns NULL on error.

Signed-off-by: Daniel Drake <dsd@laptop.org>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Add VGA Arbiter support for OpenBSD.

Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>

OpenBSD: Indicate that devices need access to legacy VGA resources.

Makes sure xserver disables DRI if the VGA arbiter is in use.

Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>

Solaris: improve support for sparc platform

(1) added prom property retrieval on sparc
(2) added multiple domain support on sparc
(3) use kernel device as mapping device
(4) performance improvements by removing redundant ioctl

Signed-off-by: Henry Zhao <henry.zhao@oracle.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

configure.ac: Bump to 0.12.902

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

Fix some -Wformat errors in scanpci

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

libpciaccess: close mtrr fd on pci_cleanup

Since the fd is not closed, calling pci_system_init and
pci_system_cleanup more than 1024 times results in "too many files open"
error.

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>

Solaris: Give better error on realloc failure

commit a18460b385ae03 converted from a fixed maximum number of devices
to dynamically growing the list via realloc, but didn't update the
error message shown on failure.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Dirk Wallenstein <halsmit@t-online.de>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>

linux sysfs: Fix read-write access in map_legacy

O_RDONLY | O_WRONLY != O_RDWR

><

Reported-by: Javier Pello <javier.pello@urjc.es>
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

scanpci: Build fix for systems without <err.h>

https://bugs.freedesktop.org/show_bug.cgi?id=31133

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

configure.ac: Bump version to 0.12.901

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

Update library version to reflect new API

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>

linux: Implement map_legacy

Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>

Add map_legacy interface

This allows platforms to hand back mmaps of the low 1M (ISA) address
space on a per-domain basis.

Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>

Strip trailing whitespace

Performed with: find * -type f | xargs perl -i -p -e 's{[ \t]+$}{}'
git diff -w & git diff -b show no diffs from this change

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

linux: Fix a crash in populate_devices

If scandir returns -1, the 'devices' array won't be initialized, and
attempting to free() it will crash.

Signed-off-by: Adam Jackson <ajax@redhat.com>

linux: Only set errno after an error

errno is only valid after an error, and was being filled with a garbage
value upon eof.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Solaris support for multiple PCI segments (domains)

1. Removed hardcoded maximum size of 256 PCI devices, which is
too small for large systems.  The number of devices is dynamically
resized as needed.

2. pci_device_solx_devfs_probe() no longer walks the device tree
from the very top ("/") but instead starts at the nexus which
owns the bus.  Performance optimization for systems with multiple
bus nodes (including systems with just one segment/domain).

3. Added support for multiple domains/segments.  Code tested
on kernels with and without multiple segment support so it should
be safe to integrate independent of the kernel version.

Signed-off-by: John Martin <John.M.Martin@Oracle.COM>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>

config: comment, minor upgrade, quote and layout configure.ac

Group statements per section as per Autoconf standard layout
Quote statements where appropriate.
Autoconf recommends not using dnl instead of # for comments

Use AC_CONFIG_FILES to replace the deprecated AC_OUTPUT with parameters.
Add AC_CONFIG_SRCDIR([Makefile.am])
Update X.Org util-macros to version 1.8

This helps automated maintenance and release activities.
Details can be found in http://wiki.x.org/wiki/NewModuleGuidelines

Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

libpciaccess 0.12.1

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

config: replace deprecated AC_HELP_STRING with AS_HELP_STRING

This silences an Automake warning.

Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

config: remove AC_PROG_CC as it overrides AC_PROG_C_C99

XORG_STRICT_OPTION from XORG_DEFAULT_OPTIONS calls
AC_PROG_C_C99. This sets gcc with -std=gnu99.
If AC_PROG_CC macro is called afterwards, it resets CC to gcc.

Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

config: replace deprecated AM_CONFIG_HEADER with AC_CONFIG_HEADERS

Regroup AC statements under the Autoconf initialization section.
Regroup AM statements under the Automake initialization section.
Add missing AC_CONFIG_SRCDIR
Replace obsolete argument to AC_OUTPUT with AC_CONFIG_FILES

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Fix detection of multi-function cards

Bit 7 of the pci device header type is an indicator of multi-function
devices, so needs to be masked when examining the bridge type.
Thanks go to Chih-Pin Wu for finding the bug and its fix.

https://bugs.launchpad.net/ubuntu/+source/libpciaccess/+bug/681207

Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Bryce Harrington <bryce@canonical.com>

Revert "Don't try to use bogus bridge data"

This reverts commit 2bda5b733bb12854760750c08138db95e77aea0c.

It actually conflicts with 9ba94caf57e3a8c3e9c6f3f5f068f4a7a7b3ff9d
which came earlier, but I didn't catch it probably because I did a blind
rebase before pushing.

When reading bridge bus info, we may need to read out the current bridge
info if the PCI bridge private is uninitialized, so remove the check at
the top that would always short circuit the case where the priv was
still 0.

References:
https://bugs.launchpad.net/ubuntu/+source/libpciaccess/+bug/681207.

Reported-by: Bryce Harrington <bryce@canonical.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>

Sun's copyrights belong to Oracle now

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

config: use LDADD to obtain appropriate dependencies

In generated scanpci/Makefile, this line is added:
scanpci_DEPENDENCIES = $(top_builddir)/src/libpciaccess.la

If the content of the library changes, the tool will rebuild.
When scanpci was in the same directory as the library, the .la file was used.
This will also ensure the "installed" pciaccess library is not used.

Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

include: add a Makefile to install and distribute header file

Makefiles are simpler when they only handle what is in their
directory.

Reviewed-by: Rémi Cardona <remi@gentoo.org>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

config: replace INCLUDES, which is deprecated, with AM_CPPFLAGS

Fixes an Automake warning.

Reviewed-by: Rémi Cardona <remi@gentoo.org>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

config: provide a makefile for scanpci man pages

Reviewed-by: Rémi Cardona <remi@gentoo.org>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

scanpci.c: remove offending whitespace as reported by git diff

Reviewed-by: Rémi Cardona <remi@gentoo.org>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

config: provide a makefile for scanpci dev tool

No function changes.

Note that automake generates a wrapper script if libpciaccess
is not installed at configuration time, which is the most common case.

If it is installed at configuration time, you get an executable.

The makefile build order ensures the lib is build before scanpci.

Reviewed-by: Rémi Cardona <remi@gentoo.org>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

config: upgrade AC_DEFINE_DIR to AX_DEFINE_DIR

Use the latest version now hosted by GNU.
The renaming protects Autoconf namespace.
There is no functional change in the macro.

Signed-off-by: Gaetan Nadon <memsize@videotron.ca>

libpciaccess 0.12.0 - bump version for release

Fix long standing MTRR bug.

Add an include that makes the file common_vgaarb.c see the same
struct pci_system as the file linux_sysfs.c.

Without this fix, on a MTRR system, the vgaarb_fd member
would overwrite the mtrr_fd member and cause the MTRR_IOC_ADD_ENTRY call
to be issued to the incorrect device causing the infamous
"Inappropriate ioctl for device (25)" error.

This error would cause MTRR setup to fail on all systems relying on it
and is severe enough to warrant a new release.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

pciaccess: fix use after free.

Using the pointer passed to realloc earlier isn't a recipe for success.

Signed-off-by: Dave Airlie <airlied@redhat.com>

Let libtool handle library flags for scanpci on Solaris

Stop hardcoding -ldevinfo since configure/libtool already know to add it
for static linking, and ELF dependencies handle it for dynamic linking

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>