| Age | Commit message (Collapse) | Author |
|---|
|
The PCI domain may be larger than 16 bits on Microsoft Azure and other
virtual environments. PCI busses reported by ACPI are limited to 16
bits, but in Azure the domain value for pass through devices is
intentionally larger than 16 bits to avoid clashing with local devices.
This is needed to support pass through of GPU devices.
v3: (ajax)
Update FreeBSD and Solaris backends to preserve the full 32-bit domain
number, since on those OSes it stands a chance of working already.
Update NetBSD and OpenBSD backends to initialize domain_16 compatibly
with older libpciaccess; neither backend appears to support more than a
handful of domains to begin with though. Trivially update the generic
x86 backend for source compatibility, though it still only supports one
domain and will never be better.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=101744
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Support for this attribute is added in the 12.4 beta release.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
|
|
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>
|
|
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>
|
|
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>
|
|
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>
|
|
Acked-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Acked-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
Copied from linuxPci.c in the X server.
Acked-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
|
|
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.
|
|
decodes is a information setting interface so it takes new decodes
and sends them to the kernel
|
|
|
|
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>
|
|
|
|
|
|
|
|
http://git.c3sl.ufpr.br/
|
|
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>
|
|
|
|
|
|
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.
|
|
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.
|
|
|
|
Added new functions pci_device_map_range and pci_device_unmap_range to
handle mapping of PCI device BARs. These new interfaces allow the
possiblity of MTRRs on platforms that support them.
These additional APIs necessitated changing some internal interfaces. The
code for FreeBSD and Solaris has been updated but has not been compiled or
tested.
Old interfaces are marked deprecated and will eventually be removed.
|
|
|
|
At anholt's request, the pci_device_write_u* functions have been
converted to take the value to be written as a parameter instead of a
pointer to the value.
Bump the version to 0.8.0.
|
|
Add pci_device_map_memory_range and pci_device_unmap_memory_range to
map and unmap specific memory ranges. The unmap bit is still a bit
hinkey (unmaps the whole BAR). Works so far for initial conversion of
VESA driver. Will need to be revisited.
|
|
|
|
commit. REALLY add support for querying bridge information. Bump to
version 0.5.0.
|
|
Add multiple-inclusion protection.
Add new function to write masked bits to PCI config space. This mirrors
functionality currently available in X.org that is slated to be
removed.
Gut old regex based search mechanism with a new mechanism that is modeled
after the Linux kernel. In addition to searching for devices by device
/ vendor ID, it is possible to search for devices by their domain / bus
/ slot / function.
Fix serious bus in the reading of ROMs and in the unmapping of regions. The
main point is that the map routine depens on the pci_mem_region::memory
pointer being non-NULL only when the region is mapped. Therefore, the
unmap routine should set it to NULL after unmapping.
Update to use new search API.
|
|
Bump version to 0.3.0.
Replace pci_get_name with pci_get_strings. This function matches the
functionality provided by the Xorg scanpci module almost identically.
|
|
|
