so much it's borderline not funny anymore.
(tested on a similar patch to libpciaccess 0.10.8, since my X still
refuses to start with libpciaccess 0.11)
from jmcneill
checking if xorg-macros used to generate configure is at least 1.3... configure: error: configure built with too old of a version of xorg-macros.m4 - requires version 1.2.0 or newer
Synced the netbsd support patch from xsrc HEAD, with a minor tweak:
Now we first try to map memory in the newfangled /dev/pciN way.
If that fails, then we fall back to /dev/mem.
We need to do this to deal with netbsd-5 and earlier.
Add NetBSD/sparc64 to the list of supported platforms.
Upstream changes:
0.11.0:
Add pci_device_get_parent_bridge()
I/O port access routines
Fix pci_device_open_io() to actually return something useful
Fix I/O handle array allocator to work for devices past the first
libpciaccess 0.11.0
.gitignore: use common defaults with custom section # 24239
Makefile.am: ChangeLog not required: EXTRA_DIST or *CLEANFILES #24432
Deploy the new XORG_DEFAULT_OPTIONS #24242
INSTALL, NEWS, README or AUTHORS files are missing/incorrect #24206
Makefile.am: add ChangeLog and INSTALL on MAINTAINERCLEANFILES
0.10.9:
vgaarb: check pci_sys exists before initing vga arb
shutup stupid gcc warning, strtok_r ignores the first parameter
Free return value from scandir instead of leaking it.
vgaarb: check for fd before close it
vgaarb: fix newbie typo
Slight brown paper bag release, 0.10.7 fails to build
when the vga arb stubs were required. also whot pointed
out I had no info in 0.10.7 release mail.
libpciacesss 0.10.8 provides support for VGA arbiter on Linux,
along with support to check if a device is the boot VGA and
if a device has a kernel driver associated with it.
Dave Airlie (1):
pciaccess: bump to 0.10.8
Thomas Klausner (1):
pciaccess: fix vga arb stubs
Aaron Plattner (1):
Don't try to write boot_vga since that fails with EACCES
Dave Airlie (15):
pciaccess: provide a method to detect if a device is boot VGA
vgaarb: fixup api and decode rsrc.
add support for finding if something has a kernel driver
vgaarb: fixup api for new count reporting + decodes interface was wrong.
vgaarb: clean up debug output
vgaarb: add support for targetting default device
vgaarb: change API to target taking a device + lock/unlock not taking one
vgaarb: more api changes + cleanup
vgaarb: add vga count + resource accessor
pciaccess: make linux rom reading fallback optional
pciaccess: make linux rom reading fallback optional
pciaccess: make linux rom reading fallback optional
Revert "vgaarb: delete stub functions"
vgaarb: update stub interface
libpciaccess: bump to 0.10.7 for vga arb interface addition
Jesse Barnes (1):
Don't try to use bogus bridge data
Julien Cristau (2):
configure.ac: fix help text for the linux-rom-fallback option
Silence compiler warning about pci_id_file_open definition
Tiago Vignatti (11):
Import libvgaaccess (VGA Arbiter) implementation from C3SL repository:
vgaarb: make it compile.
vgaarb: a bunch of clean ups to fit in libpciaccess style.
Delete vga_arb_rec.
Fix typo in vgaarb stubs.
Reorganize the code.
Comment typo.
vgaarb: delete stub functions
vgaarb: fix indentation and remove trailing white spaces
vgaarb: fix pci_device_vgaarb_get_info() return value.
Revert "vgaarb: fix pci_device_vgaarb_get_info() return value."
good enough to get newer xorg servers running.
(I need to double-check the PCI spec, but it is well possible that we
can't do much better anyway: ISTR the PCI spec allows devices to share
decoding ressources between ROM and runtime logics, which means there
is no safe way to read the original ROM at runtime. Correct me if
I'm wrong.)
Changes since 0.10.5:
Alan Coopersmith (12):
Fill in byte swapping routines for big-endian Solaris machines
Don't open/gzopen pci.ids if we're not going to read it
Add AC_SYS_LARGEFILE to configure.ac
Solaris: Use bus-range properties to limit busses scanned on each node
Add README with pointers to mailing list, bugzilla & git repos
Fix const mismatch compiler warnings in Solaris backend
Sun bug 6811468: pci_device_solx_devfs_probe accesses freed memory
Move included m4 to a subdir so we can more easily update from upstream
Fix $CWARNFLAGS addition
Fix unused variable warnings in Solaris code
Resync COPYING file with code copyright notices
Version bump: 0.10.6
Darren Smith (1):
Be more paranoid about reading bridge info (#20786)
Mark Kettenis (1):
Make libpciaccess on OpenBSD domain-aware.
Matthieu Herrb (3):
Fix a logic error in pci_device_openbsd_write().
Sync with OpenBSD code.
Fix a logic error in pci_device_netbsd_write()
Owain Ainsworth (1):
openbsd_pci.c: use the correct size when mapping the legacy vga rom.
Paulo Cesar Pereira de Andrade (1):
Janitor: Correct make distcheck and compiler warnings.
Robert Noland (1):
FreeBSD: Incorporate several fixes that have accumulated.
edward shu (1):
Sun bug 6785726: libpciaccess needs to use the correct BAR index on solaris
Changes since 0.10.4:
Julien Cristau (2):
Make --without-zlib work
Bump to 0.10.5
Robert Noland (1):
Fix FreeBSD systems which support pci domains.
This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
Tested on NetBSD-current (i386 and amd64); support for
other architectures is in the works.
Currently this should work with:
ONLY_FOR_PLATFORM= FreeBSD-*-i386 NetBSD-*-i386 OpenBSD-*-i386
ONLY_FOR_PLATFORM+= FreeBSD-*-x86_64 NetBSD-*-x86_64 OpenBSD-*-x86_64
ONLY_FOR_PLATFORM+= DragonFly-i386-* DragonFly-x86_64-*
ONLY_FOR_PLATFORM+= Linux-*-* SunOS-*-*
