platform/upstream/libdrm.git
14 years agotests: Add nouveau to list of supported modules.
Marcin Kościelnicki [Sat, 27 Feb 2010 15:04:41 +0000 (15:04 +0000)]
tests: Add nouveau to list of supported modules.

14 years agolibkms/nouveau: Add support
Marcin Kościelnicki [Sat, 27 Feb 2010 16:02:25 +0000 (16:02 +0000)]
libkms/nouveau: Add support

14 years agolibkms/intel: Throw out unused intel_bo fields.
Marcin Kościelnicki [Sat, 27 Feb 2010 15:04:38 +0000 (15:04 +0000)]
libkms/intel: Throw out unused intel_bo fields.

Acked-by: Jakob Bornecrantz <jakob@vmware.com>
14 years agoAdd config.h macro HAVE_NOUVEAU
Marcin Kościelnicki [Sat, 27 Feb 2010 15:04:37 +0000 (15:04 +0000)]
Add config.h macro HAVE_NOUVEAU

14 years agointel: Add initial support for Sandybridge, and clean up the #defines.
Eric Anholt [Thu, 22 Oct 2009 23:37:56 +0000 (16:37 -0700)]
intel: Add initial support for Sandybridge, and clean up the #defines.

14 years agonouveau: Update nouveau_class.h.
Francisco Jerez [Thu, 25 Feb 2010 17:28:18 +0000 (18:28 +0100)]
nouveau: Update nouveau_class.h.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
14 years agonv30: update for 8 texture units
Patrice Mandin [Sat, 20 Feb 2010 10:57:29 +0000 (11:57 +0100)]
nv30: update for 8 texture units

14 years agonv30: update for front/back stencil inversion
Patrice Mandin [Sat, 20 Feb 2010 10:50:37 +0000 (11:50 +0100)]
nv30: update for front/back stencil inversion

Signed-off-by: Patrice Mandin <patmandin@gmail.com>
14 years agolibkms: Use a standard version number
Jakob Bornecrantz [Thu, 18 Feb 2010 15:29:36 +0000 (16:29 +0100)]
libkms: Use a standard version number

14 years agolibkms: Enable by default
Jakob Bornecrantz [Thu, 18 Feb 2010 12:27:29 +0000 (13:27 +0100)]
libkms: Enable by default

14 years agolibkms: Add missing include file to libkms source
Jakob Bornecrantz [Thu, 18 Feb 2010 12:14:07 +0000 (13:14 +0100)]
libkms: Add missing include file to libkms source

14 years agovmwgfx: Update kernel header
Jakob Bornecrantz [Thu, 18 Feb 2010 12:27:59 +0000 (13:27 +0100)]
vmwgfx: Update kernel header

14 years agoradeon: add square-tiling flag
Marek Olšák [Sun, 14 Feb 2010 16:57:19 +0000 (17:57 +0100)]
radeon: add square-tiling flag

14 years agoIncrement version to 2.4.18 for release. 2.4.18
Carl Worth [Tue, 16 Feb 2010 20:12:41 +0000 (12:12 -0800)]
Increment version to 2.4.18 for release.

The primary motivation of the release is the bug fix in commit
4f0f871730b76730ca58209181d16725b0c40184

14 years agonouveau: bump MAX_PUSH to 512
Ben Skeggs [Tue, 16 Feb 2010 05:42:00 +0000 (15:42 +1000)]
nouveau: bump MAX_PUSH to 512

14 years agonouveau: interface changes for 0.0.16 DRM
Luca Barbieri [Fri, 29 Jan 2010 08:53:24 +0000 (09:53 +0100)]
nouveau: interface changes for 0.0.16 DRM

This commit encompasses the changes necessary to run on top of the 0.0.16
nouveau interface, additional APIs to support the new features of the
interface, as well as code from Luca Barbieri to improve the pushbuf
interface, which just happens to break nouveau's libdrm ABI so was delayed
until now.

API changes as a result of 0.0.16 DRM interface:

1. No more bo_pin()/bo_unpin(), these were only there for UMS and we no
   longer support it.

2. Any random nouveau_bo can be submitted to the GPU as a push buffer.

3. Relocations can be applied on any nouveau_bo

This patch changes the pushbuffer ABI to:

1. No longer use/expose nouveau_pushbuffer. Everything is directly
   in nouveau_channel. This saves the extra "pushbuf" pointer dereference.

2. Use cur/end pointers instead of tracking the remaining size.
   Pushing data now only needs to alter cur and not both cur and remaining.

The goal is to make the *_RING macros faster and make the interface simpler
and cleaner in the process.

The *_RING APIs are unchanged, but those are inlined and the ABI is changed.

Also, anything accessing pushbuf->remaining instead of using AVAIL_RING
will need to be fixed.

14 years agodrm: a some new connector types from the kernel
Alex Deucher [Thu, 11 Feb 2010 00:39:46 +0000 (19:39 -0500)]
drm: a some new connector types from the kernel

Add eDP (embedded displayport) and generic TV

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
14 years agointel: Handle resetting of input params after EINTR during SET_TILING
Chris Wilson [Wed, 10 Feb 2010 09:45:13 +0000 (09:45 +0000)]
intel: Handle resetting of input params after EINTR during SET_TILING

The SET_TILING is pernicious in that it overwrites the input arguments
following an error in order to report the current tiling state of the
buffer. This caught us by surprise as we then fed those arguments back
into to the ioctl unmodified following an EINTR and so the kernel then
reported success for the no-op. We interpreted this success as meaning
that the tiling on the buffer had changed so updated our state and
started using the buffer incorrectly in the new tiled/untiled manner.
This lead to all sorts of random corruption and GPU hangs, even though
the batch buffers would look sane (when the GPU had not wandered off
into forbidden territory).

References:

  Bug 25475 - [i915] Xorg crash / Execbuf while wedged
  http://bugs.freedesktop.org/show_bug.cgi?id=25475

  Bug 25554 - i830_uxa_prepare_access: gtt bo map failed: Input/output error
  http://bugs.freedesktop.org/show_bug.cgi?id=25554

(And probably every other weird bug in the last few months.)

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Account for potential pinned buffers hogging fences
Chris Wilson [Tue, 9 Feb 2010 08:32:54 +0000 (08:32 +0000)]
intel: Account for potential pinned buffers hogging fences

As the kernel reports the total number of fences, we must guess how many
fences are likely to be pinned. In the typical system these will be only
used by the scanout buffers, of which there may be one per pipe, and any
number of manually pinned fenced buffers. So take a conservative guess
and reserve two fences for use by the system.

Note this reduces the number of fences to 3 for i915 and prior.

Reference:
  http://bugs.freedesktop.org/show_bug.cgi?id=25911
  The latest intel driver 2.10.0 causes kernel oops and system hangs

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoTidy up compile warnings by cleaning up types.
Matthew W. S. Bell [Sat, 30 Jan 2010 02:14:44 +0000 (02:14 +0000)]
Tidy up compile warnings by cleaning up types.

14 years agolibdrm/radeon: Fix section size mismatch to reset the section.
Pauli Nieminen [Mon, 1 Feb 2010 18:19:33 +0000 (20:19 +0200)]
libdrm/radeon: Fix section size mismatch to reset the section.

If there is section size mismatch reusing the section object
makes section start fail.
Reseting the object before doing error checking prevents the
possible flood of errors.

14 years agoradeon: enable by default now that kms is out of staging
Dave Airlie [Tue, 2 Feb 2010 00:58:50 +0000 (10:58 +1000)]
radeon: enable by default now that kms is out of staging

14 years agointel: check return value for calloc
Dave Airlie [Tue, 2 Feb 2010 00:57:12 +0000 (10:57 +1000)]
intel: check return value for calloc

14 years agonouveau: Regenerate nouveau_class.h.
Francisco Jerez [Sun, 31 Jan 2010 16:47:50 +0000 (17:47 +0100)]
nouveau: Regenerate nouveau_class.h.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
14 years agonouveau: add nouveau_resource_destroy
Marcin Slusarz [Tue, 12 Jan 2010 14:31:28 +0000 (15:31 +0100)]
nouveau: add nouveau_resource_destroy

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
14 years agolibkms: Rework interface to not duplicate fields from kms and make formats explicit
Jakob Bornecrantz [Sat, 23 Jan 2010 01:41:49 +0000 (01:41 +0000)]
libkms: Rework interface to not duplicate fields from kms and make formats explicit

List of changes:
Fixes the cursor size to 64x64, you still need ti supply width and height
Explicitly make the cursor format A8R8G8B8
Explicitly make the scanout format X8R8G8B8

14 years agonouveau: Update nouveau_drm.h
Marcin Kościelnicki [Tue, 26 Jan 2010 18:39:21 +0000 (18:39 +0000)]
nouveau: Update nouveau_drm.h

14 years agolibkms: Fix return value in vmwgfx_bo_create
Jakob Bornecrantz [Sat, 23 Jan 2010 01:38:18 +0000 (01:38 +0000)]
libkms: Fix return value in vmwgfx_bo_create

14 years agonouveau: fail channel creation if pushbuf init fails
Ben Skeggs [Thu, 21 Jan 2010 05:17:26 +0000 (15:17 +1000)]
nouveau: fail channel creation if pushbuf init fails

14 years agointel: Add pkg-config dependency on libdrm.so using 'Requires'
Chris Wilson [Wed, 20 Jan 2010 10:42:46 +0000 (10:42 +0000)]
intel: Add pkg-config dependency on libdrm.so using 'Requires'

14 years agoSupport gcc's __FUNCTION__ for people using other compilers
Alan Coopersmith [Sun, 17 Jan 2010 03:34:13 +0000 (19:34 -0800)]
Support gcc's __FUNCTION__ for people using other compilers

Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
14 years agoAdd support for Solaris libc atomic operations
Alan Coopersmith [Sun, 17 Jan 2010 03:28:50 +0000 (19:28 -0800)]
Add support for Solaris libc atomic operations

Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
14 years agoInclude alloca.h in tests/drmstat.c when configure detects it
Alan Coopersmith [Wed, 16 Dec 2009 23:03:58 +0000 (15:03 -0800)]
Include alloca.h in tests/drmstat.c when configure detects it

Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
14 years agonouveau: disable flush_notify on channel_free
Marcin Slusarz [Tue, 12 Jan 2010 14:38:29 +0000 (15:38 +0100)]
nouveau: disable flush_notify on channel_free

We don't want do call flush_notify when we will FIRE the RING
a couple of lines later, because grobj bound to this channel
might be already freed.

14 years agonouveau: fix memory leak in nouveau_channel_free
Marcin Slusarz [Tue, 12 Jan 2010 14:32:39 +0000 (15:32 +0100)]
nouveau: fix memory leak in nouveau_channel_free

14 years agolibkms: Fix multiple map unmap in vmwgfx and add comment in intel
Jakob Bornecrantz [Fri, 15 Jan 2010 20:13:20 +0000 (20:13 +0000)]
libkms: Fix multiple map unmap in vmwgfx and add comment in intel

14 years agoradeon: get device id from the kernel, use it in cs_print
Jerome Glisse [Thu, 14 Jan 2010 19:01:55 +0000 (20:01 +0100)]
radeon: get device id from the kernel, use it in cs_print

This allow external tools to know for which asics a cs
is destinated to.

14 years agoradeon: simpler cs print function
Jerome Glisse [Thu, 14 Jan 2010 11:28:20 +0000 (12:28 +0100)]
radeon: simpler cs print function

We don't intend libdrm-radeon to become clever enough to
decode cs for all GPU we support. Better to let an external
tool do the job. This will print raw cs in an easy to parse
way.

14 years agoradeon: indentation + trailing space cleanup
Jerome Glisse [Thu, 14 Jan 2010 10:28:25 +0000 (11:28 +0100)]
radeon: indentation + trailing space cleanup

14 years agoradeon: indentation & trailing space cleanup
Jerome Glisse [Thu, 14 Jan 2010 10:24:16 +0000 (11:24 +0100)]
radeon: indentation & trailing space cleanup

14 years agoradeon: indentation + trailing space cleanup
Jerome Glisse [Thu, 14 Jan 2010 10:08:43 +0000 (11:08 +0100)]
radeon: indentation + trailing space cleanup

14 years agotests: Update kmstest a bit
Jakob Bornecrantz [Tue, 12 Jan 2010 21:35:28 +0000 (22:35 +0100)]
tests: Update kmstest a bit

14 years agolibkms: Use sysfs instead of udev to find driver
Jakob Bornecrantz [Tue, 12 Jan 2010 17:53:49 +0000 (17:53 +0000)]
libkms: Use sysfs instead of udev to find driver

Udev code is still there just commented out.

14 years agotests: Add a very small libkms test
Jakob Bornecrantz [Fri, 8 Jan 2010 15:34:44 +0000 (15:34 +0000)]
tests: Add a very small libkms test

14 years agolibkms/intel: Don't fail to create bo if we fail to tile
Jakob Bornecrantz [Fri, 8 Jan 2010 15:19:44 +0000 (15:19 +0000)]
libkms/intel: Don't fail to create bo if we fail to tile

14 years agonouveau: Update nouveau_class.h.
Francisco Jerez [Fri, 8 Jan 2010 03:34:39 +0000 (04:34 +0100)]
nouveau: Update nouveau_class.h.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
14 years agolibkms: Add intel backend
Jakob Bornecrantz [Fri, 8 Jan 2010 03:00:56 +0000 (03:00 +0000)]
libkms: Add intel backend

14 years agolibkms: Pick driver backend via pci id that we get from udev
Jakob Bornecrantz [Fri, 8 Jan 2010 02:51:04 +0000 (02:51 +0000)]
libkms: Pick driver backend via pci id that we get from udev

14 years agomodes: Retry GETCONNECTOR if a hotplug event occurs between the two ioctls
Peter Clifton [Wed, 6 Jan 2010 20:44:11 +0000 (20:44 +0000)]
modes: Retry GETCONNECTOR if a hotplug event occurs between the two ioctls

If the available modes changes between the two GETCONNECTOR ioctls, that
caused the kernel to skip filling one array and led to a crash (as the size
of the allocated and initialised block of memory differed from the reported
size, and might be NULL if no modes were present at first).

This bug manifest its self on my machine due to spurious false positive
detections of a connected TV-out.

Fixes: http://bugs.freedesktop.org/show_bug.cgi?id=25912
       Crash whilst probing modes

Based upon the similar fixes for the GETRESOURCES ioctls by Chris Wilson,
in the following commits:

    commit e6c136ca7a4c54457b48be1aec2be024b3e4a28d
    commit 85fb3e55fdb7af9b5f59c1ec0f15d1950e601b05
    commit d1308f4fe7f94aae51ca9f70947aea8e09597f37

Signed-off-by: Peter Clifton <pcjc2@cam.ac.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agomodes: Fix previous commit for potential legal NULLs
Chris Wilson [Wed, 6 Jan 2010 16:53:33 +0000 (16:53 +0000)]
modes: Fix previous commit for potential legal NULLs

If the count is 0, then the malloc is permitted to return NULL, so don't
throw an error in that case.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agomodes: Free local resources after allocation failure in GETRESOURCES
Chris Wilson [Wed, 6 Jan 2010 15:39:49 +0000 (15:39 +0000)]
modes: Free local resources after allocation failure in GETRESOURCES

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agomodes: Retry GETRESOURCES if a hotplug event occurs between the two ioctls
Chris Wilson [Wed, 6 Jan 2010 15:19:25 +0000 (15:19 +0000)]
modes: Retry GETRESOURCES if a hotplug event occurs between the two ioctls

Peter Clifton hit an issue whereby he had a spurious TV hotplug event
that occurred between the two GETRESOURCES ioctls that caused the kernel
to skip filling one array and led to a crash (as the size of the
allocated and initialised block of memory differed from the reported
size).

Fixes: http://bugs.freedesktop.org/show_bug.cgi?id=25912
       Crash whilst probing modes

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reported-by: Peter Clifton <pcjc2@cam.ac.uk>
14 years agolibkms: Make vmwgfx optional
Jakob Bornecrantz [Tue, 5 Jan 2010 19:55:02 +0000 (20:55 +0100)]
libkms: Make vmwgfx optional

14 years agolibkms: Remove duped vmwgfx_drm.h header
Jakob Bornecrantz [Tue, 5 Jan 2010 20:04:57 +0000 (21:04 +0100)]
libkms: Remove duped vmwgfx_drm.h header

14 years agoMerge branch 'master' into libkms-master
Jakob Bornecrantz [Tue, 5 Jan 2010 19:59:24 +0000 (20:59 +0100)]
Merge branch 'master' into libkms-master

Conflicts:
configure.ac

14 years agolibkms: Disable by default
Jakob Bornecrantz [Tue, 5 Jan 2010 19:58:37 +0000 (20:58 +0100)]
libkms: Disable by default

14 years agovmwgfx: Add experimental vmwgfx kernel header
Jakob Bornecrantz [Tue, 5 Jan 2010 18:53:10 +0000 (19:53 +0100)]
vmwgfx: Add experimental vmwgfx kernel header

14 years agonouveau: add BEGIN_RING_NI
Marcin Slusarz [Wed, 30 Dec 2009 20:32:40 +0000 (21:32 +0100)]
nouveau: add BEGIN_RING_NI

14 years agonouveau: Unreference pushbuf objects on channel destruction.
Younes Manton [Mon, 28 Dec 2009 23:00:09 +0000 (18:00 -0500)]
nouveau: Unreference pushbuf objects on channel destruction.

- unreference pushbuf objects on channel destruction

Based on Krzysztof Smiechowicz's patch.

14 years agonouveau: Update some object definitions from renouveau.xml.
Francisco Jerez [Mon, 28 Dec 2009 17:01:42 +0000 (18:01 +0100)]
nouveau: Update some object definitions from renouveau.xml.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
14 years agonouveau: Update nouveau_class.h with new renouveau.xml definitions.
Marcin Kościelnicki [Tue, 24 Nov 2009 17:23:54 +0000 (17:23 +0000)]
nouveau: Update nouveau_class.h with new renouveau.xml definitions.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
14 years agolibdrm 2.4.17 2.4.17
Dave Airlie [Mon, 21 Dec 2009 05:03:31 +0000 (15:03 +1000)]
libdrm 2.4.17

14 years agoradeon: fix BO null check, should be in higher level fn
Dave Airlie [Mon, 21 Dec 2009 04:59:48 +0000 (14:59 +1000)]
radeon: fix BO null check, should be in higher level fn

14 years agoradeon: straighten out the API insanity.
Dave Airlie [Thu, 17 Dec 2009 04:11:55 +0000 (14:11 +1000)]
radeon: straighten out the API insanity.

as Michel pointed out we are exposing too much info for these object
for this to be maintainable going forward.

This patch set minimises the exposed parts of the radeon_bo and
radeon_cs objects to the piece necessary for ddx/mesa to operate
at a decent speed.

The major problem is mesa contains a legacy BO/CS managers which we still
need to expose functionality to, and we really cannot change the API
until we can drop the non-KMS codepaths.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agonouveau: remove delayed kernel bo creation
Ben Skeggs [Thu, 17 Dec 2009 03:07:18 +0000 (13:07 +1000)]
nouveau: remove delayed kernel bo creation

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
14 years agoBe less chatty in drmSetMaster/drmDropMaster
Kristian Høgsberg [Wed, 9 Dec 2009 15:37:40 +0000 (10:37 -0500)]
Be less chatty in drmSetMaster/drmDropMaster

14 years agomodetest: Error out if pageflipping is requested but not available
Kristian Høgsberg [Wed, 9 Dec 2009 15:36:53 +0000 (10:36 -0500)]
modetest: Error out if pageflipping is requested but not available

14 years agovmwgfx: Fetch the latest header
Jakob Bornecrantz [Mon, 14 Dec 2009 21:53:10 +0000 (22:53 +0100)]
vmwgfx: Fetch the latest header

14 years agonouveau: Use drmIoctl so we restart ioctl on EINTR or EAGAIN
Ben Skeggs [Wed, 9 Dec 2009 04:50:15 +0000 (14:50 +1000)]
nouveau: Use drmIoctl so we restart ioctl on EINTR or EAGAIN

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
14 years agonouveau: move reloc code down, nothing to see here
Ben Skeggs [Wed, 9 Dec 2009 03:15:07 +0000 (13:15 +1000)]
nouveau: move reloc code down, nothing to see here

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
14 years agointel: Clear virtual after failing to mmap_gtt.
Chris Wilson [Tue, 8 Dec 2009 22:35:24 +0000 (22:35 +0000)]
intel: Clear virtual after failing to mmap_gtt.

Don't store the error return in bo_gem->gtt_virtual or else we will
attempt to use that as a valid pointer in future mappings.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agolibkms: Change destroy api a bit
Jakob Bornecrantz [Fri, 4 Dec 2009 15:06:42 +0000 (16:06 +0100)]
libkms: Change destroy api a bit

14 years agolibkms: Add libkms
Jakob Bornecrantz [Tue, 24 Nov 2009 16:54:10 +0000 (17:54 +0100)]
libkms: Add libkms

14 years agoIgnore config.h.in
Jakob Bornecrantz [Fri, 4 Dec 2009 15:07:51 +0000 (16:07 +0100)]
Ignore config.h.in

14 years agoradeon: Use drmIoctl so we restart ioctl on EINTR or EAGAIN
Jerome Glisse [Mon, 7 Dec 2009 17:30:52 +0000 (18:30 +0100)]
radeon: Use drmIoctl so we restart ioctl on EINTR or EAGAIN

This is needed as change in kernel will lead to ioctl returning
EINTR if they are interrupted.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
14 years agointel: Expect caller to guarantee thread-safety of bo during reloc
Chris Wilson [Tue, 1 Dec 2009 23:01:34 +0000 (23:01 +0000)]
intel: Expect caller to guarantee thread-safety of bo during reloc

This removes the foremost prolific user of mutexes in libdrm_intel.so.
The other uses of the bufmgr_gem->mutex to serial access to individual
bos are currently required by Mesa, and are far less frequent.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[anholt: This chunk looks good...]
Acked-by: Eric Anholt <eric@anholt.net>
14 years agomodetest: fix build error due to page_flip_handler name change
Jesse Barnes [Fri, 4 Dec 2009 17:09:19 +0000 (09:09 -0800)]
modetest: fix build error due to page_flip_handler name change

Forgot to update this when pushing the pageflip bits.

14 years agoMerge branch 'modesetting-dirty-libdrm'
Jakob Bornecrantz [Wed, 2 Dec 2009 18:40:58 +0000 (19:40 +0100)]
Merge branch 'modesetting-dirty-libdrm'

Conflicts:
include/drm/drm.h

14 years agoAdd RELEASING to document the release process
Kristian Høgsberg [Thu, 3 Dec 2009 22:49:31 +0000 (17:49 -0500)]
Add RELEASING to document the release process

14 years agoBump event context structure version for page flipping
Jesse Barnes [Thu, 3 Dec 2009 22:20:51 +0000 (14:20 -0800)]
Bump event context structure version for page flipping

14 years agoMerge branch 'pageflip' of git://people.freedesktop.org/~jbarnes/drm
Jesse Barnes [Thu, 3 Dec 2009 22:17:26 +0000 (14:17 -0800)]
Merge branch 'pageflip' of git://people.freedesktop.org/~jbarnes/drm

Conflicts:
include/drm/drm.h - RMFB had its signature changed to avoid uint32_t

14 years agoBump to 2.4.16 for release 2.4.16
Kristian Høgsberg [Thu, 3 Dec 2009 16:31:27 +0000 (11:31 -0500)]
Bump to 2.4.16 for release

14 years agoEnable experimental APIs for distcheck
Kristian Høgsberg [Thu, 3 Dec 2009 16:27:16 +0000 (11:27 -0500)]
Enable experimental APIs for distcheck

14 years agoChange the number on the dirty ioctl to match upstream
Jakob Bornecrantz [Wed, 2 Dec 2009 18:39:16 +0000 (19:39 +0100)]
Change the number on the dirty ioctl to match upstream

14 years agointel: Free memory before inserting bo into cache.
Chris Wilson [Wed, 2 Dec 2009 13:36:22 +0000 (13:36 +0000)]
intel: Free memory before inserting bo into cache.

This has the unfortunate behaviour of releasing our malloc cache, but
the alternative is for X to consume a couple of gigabytes of ram and
die during testing. Fortunately the extra mallocs have little impact on
performance whereas avoiding swap and death, lots.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Check and propagate errors from building reloc-tree
Chris Wilson [Wed, 2 Dec 2009 13:12:39 +0000 (13:12 +0000)]
intel: Check and propagate errors from building reloc-tree

Instead of forcing the caller to check after every emit_reloc(), we can
flag the object as being in error, propagating that error upwards through
the relocation tree, and failing the eventual batch buffer execution.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Repeat execbuffer after EINTR
Chris Wilson [Wed, 2 Dec 2009 12:58:00 +0000 (12:58 +0000)]
intel: Repeat execbuffer after EINTR

EAGAIN cannot be raised by the current code, but the system call maybe
interrupted and so return EINTR.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Review use of errno.
Chris Wilson [Wed, 2 Dec 2009 12:40:26 +0000 (12:40 +0000)]
intel: Review use of errno.

Hitting this error lead to a segfault:

  intel_bufmgr_gem.c:919: Error mapping buffer 48607 (pixmap):
                          Cannot allocate memory.

because the errno was reused as the function return value after being
reset by the fprintf(), so caller thought the mapping had succeeded. The
convention established by libdrm is that the return value is the
negative errno and that uses of libdrm cannot trust the value of errno
afterwards, but must use the return code.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Make bo_reference() inline for internal use.
Chris Wilson [Wed, 2 Dec 2009 10:42:51 +0000 (10:42 +0000)]
intel: Make bo_reference() inline for internal use.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Remove the extra reference while validating the reloc tree
Chris Wilson [Wed, 2 Dec 2009 10:41:39 +0000 (10:41 +0000)]
intel: Remove the extra reference while validating the reloc tree

Buffers on the relocation tree are guarded by the reference to the batch
object and so do not need an extra reference whilst constructing the
list of execution buffer objects.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Wrap a few more syscalls with EINTR protection
Chris Wilson [Tue, 1 Dec 2009 13:08:04 +0000 (13:08 +0000)]
intel: Wrap a few more syscalls with EINTR protection

Having been bitten by a missing EINTR check during mmap_gtt(), I thought
it prudent to add some more protection around the ioctls.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Clear bo->used_as_reloc_target flag on destroy
Chris Wilson [Mon, 30 Nov 2009 23:07:19 +0000 (23:07 +0000)]
intel: Clear bo->used_as_reloc_target flag on destroy

This allows us to keep the assert added in the previous commit that we do
not modify the tree_reloc_size after inserting the buffer into a relocation
tree, which was being hit here:

  #0  0xb78c2424 in __kernel_vsyscall ()
  #1  0xb74f6401 in raise () from /lib/libc.so.6
  #2  0xb74f7b42 in abort () from /lib/libc.so.6
  #3  0xb74ef5a8 in __assert_fail () from /lib/libc.so.6
  #4  0xb737e78b in drm_intel_bo_gem_set_in_aperture_size (bufmgr_gem=<value optimized out>, bo_gem=0x6) at intel_bufmgr_gem.c:373
  #5  0xb737f519 in drm_intel_gem_bo_set_tiling (bo=0xa1030a0, tiling_mode=0xbff6c85c, stride=0) at intel_bufmgr_gem.c:1386
  #6  0xb737f67f in drm_intel_gem_bo_unreference_final (bo=0xa1030a0, time=<value optimized out>) at intel_bufmgr_gem.c:768
  #7  0xb737f5e3 in drm_intel_gem_bo_unreference_locked_timed (bo=0xa1e50d0, time=<value optimized out>) at intel_bufmgr_gem.c:805
  #8  drm_intel_gem_bo_unreference_final (bo=0xa1e50d0, time=<value optimized out>) at intel_bufmgr_gem.c:756
  #9  0xb737fcbb in drm_intel_gem_bo_unreference (bo=0xa1e50d0) at intel_bufmgr_gem.c:821
  #10 0xb737b4e6 in drm_intel_bo_unreference (bo=0x0) at intel_bufmgr.c:80
  #11 0xb7325625 in intel_batch_flush (scrn=0x9d91f78, flush=1) at i830_batchbuffer.c:200

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Apply pessimistic alignment to in-aperture buffer size
Chris Wilson [Mon, 30 Nov 2009 22:14:30 +0000 (22:14 +0000)]
intel: Apply pessimistic alignment to in-aperture buffer size

For the older chipsets, i.e. pre-i965, which have severe alignment
restrictions for tiled buffers we need to pessimistically assume that we
will waste the size of buffer to meet those alignment constraints.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agointel: Only store a buffer in the cache if it is retained.
Chris Wilson [Mon, 30 Nov 2009 20:02:05 +0000 (20:02 +0000)]
intel: Only store a buffer in the cache if it is retained.

If the kernel immediately frees the backing store for a buffer when
marking it purgeable, then there is not point adding to the cache. Free
it immediately, instead.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoChange the dirty ioctl a bit and comment it
Jakob Bornecrantz [Mon, 30 Nov 2009 00:40:51 +0000 (00:40 +0000)]
Change the dirty ioctl a bit and comment it

14 years agoBring dirty code from old branch
Jakob Bornecrantz [Tue, 24 Nov 2009 17:00:12 +0000 (18:00 +0100)]
Bring dirty code from old branch

14 years agoCorrectly set DRM_MAX_MINOR for all platforms.
Robert Noland [Wed, 25 Nov 2009 21:09:24 +0000 (15:09 -0600)]
Correctly set DRM_MAX_MINOR for all platforms.

DRM_MAJOR is platform specific, but not used outside of xf86drm.c
that I can find.

14 years agoAdd missing DRM_MAX_MINOR define
Alex Deucher [Tue, 24 Nov 2009 20:03:32 +0000 (15:03 -0500)]
Add missing DRM_MAX_MINOR define

lost in 500f5b524000ed5930301f4303744cb4c0a19b75

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
14 years agoMore fixups to allow mesa to build
Robert Noland [Tue, 24 Nov 2009 16:31:41 +0000 (10:31 -0600)]
More fixups to allow mesa to build

14 years agoFinish fixing the build on FreeBSD
Robert Noland [Tue, 24 Nov 2009 15:27:29 +0000 (09:27 -0600)]
Finish fixing the build on FreeBSD