OSDN Git Service

android-x86/external-libdrm.git
15 years ago[FreeBSD] Add drm_drawable_free_all()
Robert Noland [Sat, 23 Aug 2008 01:45:41 +0000 (21:45 -0400)]
[FreeBSD] Add drm_drawable_free_all()

15 years ago[FreeBSD] Fix long standing memory leak in drm_remove_magic.
Robert Noland [Sat, 23 Aug 2008 01:34:25 +0000 (21:34 -0400)]
[FreeBSD] Fix long standing memory leak in drm_remove_magic.

We shuffled all the links around to disconnect the entry, but
never free it.  We would incorrectly free the last entry in the
hash chain if nothing matched.

15 years agoi915: Convert vblank on disabled pipe DRM_ERROR to DRM_DEBUG.
Robert Noland [Sat, 23 Aug 2008 01:32:37 +0000 (21:32 -0400)]
i915: Convert vblank on disabled pipe DRM_ERROR to DRM_DEBUG.

15 years agoi915: Clear sarea_priv during lastclose.
Robert Noland [Wed, 20 Aug 2008 21:20:31 +0000 (17:20 -0400)]
i915: Clear sarea_priv during lastclose.

sarea_priv needs to be NULL before i915_initialized is called to
properly reset it.  The stale value produces a panic any time something
opens/closes drm without calling initialize.  i.e. version checking

15 years agoi915: Free dev->dev_private on unload.
Robert Noland [Wed, 20 Aug 2008 21:15:46 +0000 (17:15 -0400)]
i915: Free dev->dev_private on unload.

15 years ago[FreeBSD] Move vblank bits into their own structure.
vehemens [Wed, 13 Aug 2008 21:19:35 +0000 (17:19 -0400)]
[FreeBSD] Move vblank bits into their own structure.

Signed-off-by: Robert Noland <rnoland@2hip.net>
15 years ago[FreeBSD] Fix lock leak.
vehemens [Wed, 13 Aug 2008 21:17:05 +0000 (17:17 -0400)]
[FreeBSD] Fix lock leak.

Signed-off-by: Robert Noland <rnoland@2hip.net>
15 years agoi915: Move spinlock init / destroy to load / unload time.
Robert Noland [Wed, 13 Aug 2008 21:05:59 +0000 (17:05 -0400)]
i915: Move spinlock init / destroy to load / unload time.

This resolves the panic on FreeBSD during VT switch, without attempting
any of the more lofty goals for the time being.

15 years agolibdrm: add udev support.
Dave Airlie [Sun, 24 Aug 2008 06:54:43 +0000 (16:54 +1000)]
libdrm: add udev support.

This patch allows you to --enable-udev, and will avoid having libdrm
make device nodes. If you are using udev, you should really --enable-udev
your libdrm.

15 years agonouveau: fifo and graphics engine suspend and resume for nv04-nv4x
root [Tue, 6 May 2008 22:04:55 +0000 (23:04 +0100)]
nouveau: fifo and graphics engine suspend and resume for nv04-nv4x

Corresponding DDX patch at http://people.freedesktop.org/~stuart/nv0x-nv4x_suspend/

15 years agoChange prototype of drmIoctl to unsigned long request.
Coleman Kane [Mon, 18 Aug 2008 21:08:21 +0000 (17:08 -0400)]
Change prototype of drmIoctl to unsigned long request.

This resolves and issue on amd64 FreeBSD and it looks like the
linux ioctl syscall should be unsigned long as well.

Signed-off-by: Robert Noland <rnoland@2hip.net>
15 years agonouveau: make it compile under 2.6.27
Maarten Maathuis [Sat, 16 Aug 2008 22:41:50 +0000 (00:41 +0200)]
nouveau: make it compile under 2.6.27

15 years agolinux: on_each_cpu has 3 args on 2.6.27
Maarten Maathuis [Thu, 14 Aug 2008 17:35:55 +0000 (19:35 +0200)]
linux: on_each_cpu has 3 args on 2.6.27

15 years agoDon't call the vblank tasklet with irqs disabled.
Thomas Hellstrom [Wed, 13 Aug 2008 08:04:21 +0000 (10:04 +0200)]
Don't call the vblank tasklet with irqs disabled.
If a specific tasklet shares data with irq context,
it needs to take a private irq-blocking spinlock within
the tasklet itself.

15 years ago[intel_bufmgr_gem] Fix flink buffer name caching (#17085, #17092).
Kristian Høgsberg [Wed, 13 Aug 2008 02:22:15 +0000 (22:22 -0400)]
[intel_bufmgr_gem] Fix flink buffer name caching (#17085, #17092).

Store the global name in global_name, don't overwrite the gem_handle.

15 years agolibdrm: Allow build outside of source tree.
Matthieu Herrb [Mon, 11 Aug 2008 17:44:39 +0000 (10:44 -0700)]
libdrm: Allow build outside of source tree.

15 years agodrm_vm: fix for build on 2.6.22
Dave Airlie [Mon, 11 Aug 2008 00:58:41 +0000 (10:58 +1000)]
drm_vm: fix for build on 2.6.22

thanks to malc0 for pointing it out

15 years agodrm: add OS_HAS_GEM option.
Dave Airlie [Mon, 11 Aug 2008 00:47:00 +0000 (10:47 +1000)]
drm: add OS_HAS_GEM option.

To build i915 driver pass OS_HAS_GEM=1 to make for now

15 years agodrm: TRUE/true
Dave Airlie [Sat, 9 Aug 2008 07:19:16 +0000 (17:19 +1000)]
drm: TRUE/true

15 years agoDrop TTM interfaces from the userland library.
Eric Anholt [Fri, 8 Aug 2008 22:55:34 +0000 (15:55 -0700)]
Drop TTM interfaces from the userland library.

15 years agoFix compile warning from check_aperture change.
Eric Anholt [Fri, 8 Aug 2008 21:08:43 +0000 (14:08 -0700)]
Fix compile warning from check_aperture change.

15 years agoMerge branch 'drm-gem'
Eric Anholt [Fri, 8 Aug 2008 21:05:01 +0000 (14:05 -0700)]
Merge branch 'drm-gem'

Conflicts:

shared-core/i915_dma.c

This brings in kernel support and userland interface for intel GEM.

15 years agoReplace the check_aperture API with one we can make thread-safe.
Eric Anholt [Fri, 8 Aug 2008 20:13:46 +0000 (13:13 -0700)]
Replace the check_aperture API with one we can make thread-safe.

While the bufmgr isn't thread-safe at the moment, we need it to be for shared
objects between contexts.

15 years agonouveau: wait for pgraph idle after loading or saving a context
Stuart Bennett [Fri, 8 Aug 2008 15:20:52 +0000 (16:20 +0100)]
nouveau: wait for pgraph idle after loading or saving a context

the nvidia driver does this, and it stops the error message appearing on nv40

15 years agoMake flink save the kernel-assigned name and return it instead of creating another...
Keith Packard [Thu, 7 Aug 2008 22:26:30 +0000 (15:26 -0700)]
Make flink save the kernel-assigned name and return it instead of creating another name

15 years agoSwitch from shmem_getpage to read_mapping_page
Keith Packard [Tue, 5 Aug 2008 06:33:03 +0000 (23:33 -0700)]
Switch from shmem_getpage to read_mapping_page

15 years ago[gem-intel] Retiring flush requests should clear flushed write_domains
Keith Packard [Tue, 5 Aug 2008 23:06:40 +0000 (16:06 -0700)]
[gem-intel] Retiring flush requests should clear flushed write_domains

When i915_gem_retire_request has a flush which matches an object write
domain, clear the write domain. This will move the object to the inactive
list rather than the flushing list, avoiding trouble with objects left stuck
on the flushing list.

15 years ago[gem-intel] Don't clear write_domain until flush completes
Keith Packard [Tue, 5 Aug 2008 21:44:53 +0000 (14:44 -0700)]
[gem-intel] Don't clear write_domain until flush completes

In i915_gem_object_wait_rendering, if the object write domain is being
written by the GPU, the appropriate flushing commands are written to the
device and an additional request queued to mark that flush. Finally, the
function blocks on that new request.

The bug was that the write_domain in the object was cleared before the
function blocked.

If the wait is interrupted by a signal, the flushing commands may still be
pending. With the current write_domain information lost, the restarted
syscall will drop right through the write_domain test as that value was
lost, and so the function will not block at all. Oops.

Fixed by simply moving the write_domain clear until after the wait_request
succeeds. Note that the restarted system call will generate an additional
flush sequence and request, but that should be 'harmless', aside from a
slight performance impact.

Someday we'll track flushing more accurately and clear write_domains more
efficiently, but for now, this should suffice.

This bug was discovered in the 2d gem development by running x11perf
-copypixwin500 and noticing that the window got cleared accidentally.

15 years agoExpose pin/unpin/set_tiling/flink APIs
Keith Packard [Mon, 4 Aug 2008 07:34:08 +0000 (00:34 -0700)]
Expose pin/unpin/set_tiling/flink APIs

15 years agoRevert "i915: Move all of the irq install/uninstall to load time."
Dave Airlie [Thu, 31 Jul 2008 21:43:58 +0000 (07:43 +1000)]
Revert "i915: Move all of the irq install/uninstall to load time."

This reverts commit 965a72202b439068e62ac341990f51953457b202.

Please re-do over properly

15 years agodrm: add fault handler support so as to be more like possible upstream
Dave Airlie [Thu, 31 Jul 2008 03:12:36 +0000 (13:12 +1000)]
drm: add fault handler support so as to be more like possible upstream

15 years agointel-gem: Replace version bump signalling GEM with I915_PARAM_HAS_GEM.
Eric Anholt [Wed, 30 Jul 2008 21:10:36 +0000 (14:10 -0700)]
intel-gem: Replace version bump signalling GEM with I915_PARAM_HAS_GEM.

15 years agointel-gem: Remove dead field from reverted changes.
Eric Anholt [Wed, 30 Jul 2008 18:29:32 +0000 (11:29 -0700)]
intel-gem: Remove dead field from reverted changes.

15 years agoRevert "Rename drm_mm.c and its fuctions to drm_memrange."
Eric Anholt [Wed, 30 Jul 2008 18:22:48 +0000 (11:22 -0700)]
Revert "Rename drm_mm.c and its fuctions to drm_memrange."

This reverts commit 3ad8db2071d30c198403e605f2726fc5c3e46bfd.

We ended up not needing that namespace, and I'd rather not have the churn
for producing diffs.

15 years agoi915: more version checks
Dave Airlie [Wed, 30 Jul 2008 06:52:13 +0000 (16:52 +1000)]
i915: more version checks

15 years agoi915: add version checks for opregion on old kernels
Dave Airlie [Wed, 30 Jul 2008 06:26:51 +0000 (16:26 +1000)]
i915: add version checks for opregion on old kernels

15 years agointel-gem: actually add i915_gem_debug.c.
Eric Anholt [Wed, 30 Jul 2008 03:26:38 +0000 (20:26 -0700)]
intel-gem: actually add i915_gem_debug.c.

15 years agointel-gem: Check return value of pci_read_base.
Eric Anholt [Tue, 29 Jul 2008 18:10:47 +0000 (11:10 -0700)]
intel-gem: Check return value of pci_read_base.

15 years agor300: Fix cliprect emit
Nicolai Haehnle [Tue, 29 Jul 2008 17:56:51 +0000 (19:56 +0200)]
r300: Fix cliprect emit

This makes our handling of cliprects sane. drm_clip_rect always has exclusive
bottom-right corners, but the hardware expects inclusive bottom-right corners,
so we adjust this here.

This complements Michel Daenzer's commit 57aea290e1e0a26d1e74df6cff777eb9f038f1f8
to Mesa. See also http://bugs.freedesktop.org/show_bug.cgi?id=16123 .

15 years agointel: Fix typo in unused register definition name.
Eric Anholt [Tue, 29 Jul 2008 06:14:11 +0000 (23:14 -0700)]
intel: Fix typo in unused register definition name.

15 years agointel-gem: checkpatch.pl on drm_proc.c new contents.
Eric Anholt [Mon, 28 Jul 2008 23:33:50 +0000 (16:33 -0700)]
intel-gem: checkpatch.pl on drm_proc.c new contents.

15 years agoMerge commit 'origin/master' into drm-gem
Eric Anholt [Mon, 28 Jul 2008 22:17:21 +0000 (15:17 -0700)]
Merge commit 'origin/master' into drm-gem

Conflicts:

linux-core/Makefile.kernel
shared-core/i915_dma.c
shared-core/i915_drv.h
shared-core/i915_irq.c

15 years agonouveau: fix nv04 fifo context save to save reg contents, not reg offset
Stuart Bennett [Tue, 29 Jul 2008 01:32:13 +0000 (02:32 +0100)]
nouveau: fix nv04 fifo context save to save reg contents, not reg offset

clearly the function had never been used :)

15 years agonouveau: fix bad rename from 5a072f32c8f941d1ef301811881e8c89c8d8a5f1
Stuart Bennett [Tue, 29 Jul 2008 01:26:34 +0000 (02:26 +0100)]
nouveau: fix bad rename from 5a072f32c8f941d1ef301811881e8c89c8d8a5f1

15 years agointel-gem: Another checkpatch.pl pass.
Eric Anholt [Mon, 28 Jul 2008 18:45:22 +0000 (11:45 -0700)]
intel-gem: Another checkpatch.pl pass.

15 years agointel-gem: Fix regression tests.
Eric Anholt [Mon, 28 Jul 2008 18:24:00 +0000 (11:24 -0700)]
intel-gem: Fix regression tests.

Main fix is an oops that was triggered by the gtt pwrite path when we don't
have the gtt initialized.  Also, settle on -EBADF for "bad object handle",
and -EINVAL for "reading/writing beyond object boundary".

15 years agointel-gem: Speed up tiled readpixels by tracking which pages have been flushed.
Eric Anholt [Sun, 27 Jul 2008 02:28:14 +0000 (19:28 -0700)]
intel-gem: Speed up tiled readpixels by tracking which pages have been flushed.

This is around 3x or so speedup, since we would read wide rows at a time, and
clflush each tile 8 times as a result.  We'll want code related to this anyway
when we do fault-based per-page clflushing for sw fallbacks.

15 years agointel-gem: Move debug-only functions to a separate file.
Eric Anholt [Sun, 27 Jul 2008 02:51:58 +0000 (19:51 -0700)]
intel-gem: Move debug-only functions to a separate file.

15 years agoradeon: Add suspend/resume hooks for saving/clearing/restoring interrupts.
Stefan Becker [Sat, 26 Jul 2008 14:49:14 +0000 (16:49 +0200)]
radeon: Add suspend/resume hooks for saving/clearing/restoring interrupts.

Fixes http://bugs.freedesktop.org/show_bug.cgi?id=16799 .

15 years agor300_cmdbuf: Always emit INDX_BUFFER immediately after DRAW_INDEX
Nicolai Haehnle [Sat, 26 Jul 2008 09:39:10 +0000 (11:39 +0200)]
r300_cmdbuf: Always emit INDX_BUFFER immediately after DRAW_INDEX

DRAW_INDEX writes a vertex count to VAP_VF_CNTL. Docs say that behaviour
is undefined (i.e. lockups happen) when this write is not followed by the
right number of vertex indices.

Thus we used to do the wrong thing when drawing across many cliprects was
necessary, because we emitted a sequence
 DRAW_INDEX, DRAW_INDEX, INDX_BUFFER, INDX_BUFFER
instead of
 DRAW_INDEX, INDX_BUFFER, DRAW_INDEX, INDX_BUFFER
The latter is what we're doing now and which ought to be correct.

15 years ago[FreeBSD] Duh, we need to actually define the drm_modeset_ctl...
Robert Noland [Fri, 25 Jul 2008 17:46:28 +0000 (13:46 -0400)]
[FreeBSD] Duh, we need to actually define the drm_modeset_ctl...

15 years agodrm: don't set the signal blocker on the master process.
Dave Airlie [Fri, 25 Jul 2008 03:30:08 +0000 (13:30 +1000)]
drm: don't set the signal blocker on the master process.

this lets us debug the X server through xkb startup.

Not sure what the correct answer is, probably X needs to drop
the lock when execing stuff, with input hotplug it can get
xkb stuff at any time I believe.

15 years ago[FreeBSD] Catch up to linux on vblank-rework
Robert Noland [Thu, 24 Jul 2008 04:21:00 +0000 (00:21 -0400)]
[FreeBSD] Catch up to linux on vblank-rework

15 years agoi915: Move all of the irq install/uninstall to load time.
Robert Noland [Wed, 23 Jul 2008 23:55:06 +0000 (19:55 -0400)]
i915: Move all of the irq install/uninstall to load time.

This resolves a panic on FreeBSD which was caused by trying
to re-initialize the swap lock.  It's just much easier to
initialize all of the locks at load time.  It should also
ensure that the vblank structures are available earlier.

15 years agointel-gem: Fix pread math and logic errors.
Eric Anholt [Wed, 23 Jul 2008 18:49:25 +0000 (11:49 -0700)]
intel-gem: Fix pread math and logic errors.

Fixes an oops in fbotexture from walking off the end of the page list.

15 years agointel-gem: Add a quick hack to reduce clflushing on pread.
Eric Anholt [Wed, 23 Jul 2008 17:07:16 +0000 (10:07 -0700)]
intel-gem: Add a quick hack to reduce clflushing on pread.

This increases overhead for the large-readpixels case due to the repeated
page cache accessing, but greatly reduces overhead for the small-readpixels
case.

15 years agointel-gem: Don't do the GTT-pwrite shortcut on tiled buffers.
Eric Anholt [Wed, 23 Jul 2008 15:36:54 +0000 (08:36 -0700)]
intel-gem: Don't do the GTT-pwrite shortcut on tiled buffers.

These will be covered by the fence, while pread/pwrite are supposed to be
CPU-perspective writes, with manual detiling done by the client.

15 years agointel-gem: Move /proc debugging to a separate file.
Eric Anholt [Mon, 21 Jul 2008 21:33:55 +0000 (14:33 -0700)]
intel-gem: Move /proc debugging to a separate file.

15 years agoradeon: fix typo with a better typo
Dave Airlie [Tue, 22 Jul 2008 08:47:27 +0000 (18:47 +1000)]
radeon: fix typo with a better typo

15 years agoradeon: fix type DST vs Z cache flush
Dave Airlie [Tue, 22 Jul 2008 08:10:03 +0000 (18:10 +1000)]
radeon: fix type DST vs Z cache flush

15 years ago[FreeBSD] Improve upper_32_bits define.
Robert Noland [Mon, 21 Jul 2008 22:04:02 +0000 (18:04 -0400)]
[FreeBSD] Improve upper_32_bits define.

Thanks to airlied.

15 years agointel-gem: Remove recently added pci_read_base prototype.
Eric Anholt [Mon, 21 Jul 2008 21:21:16 +0000 (14:21 -0700)]
intel-gem: Remove recently added pci_read_base prototype.

This is in pci.h in the fixed patch to the kernel.

15 years agointel-gem: Set up HWS when it needs a vaddr during GEM init.
Eric Anholt [Tue, 1 Jul 2008 19:31:37 +0000 (12:31 -0700)]
intel-gem: Set up HWS when it needs a vaddr during GEM init.

This requires an updated 2D driver to not try to set it up as well.

15 years agoRemove accidental leftover tests.
Michel Dänzer [Mon, 21 Jul 2008 11:43:12 +0000 (13:43 +0200)]
Remove accidental leftover tests.

Thanks to Nicolai Haehnle for pointing this out on IRC.

15 years agoReinstate dev->vblank_enabled[].
Michel Dänzer [Mon, 21 Jul 2008 09:48:04 +0000 (11:48 +0200)]
Reinstate dev->vblank_enabled[].

I incorrectly thought it was obsolete.

15 years agoDrop workaround for driver vblank counter going backwards.
Michel Dänzer [Mon, 21 Jul 2008 06:16:59 +0000 (08:16 +0200)]
Drop workaround for driver vblank counter going backwards.

The driver code that caused this is no longer necessary and has been dropped.

15 years agoradeon: Post-vblank-rework-rework cleanups.
Michel Dänzer [Mon, 21 Jul 2008 06:16:59 +0000 (08:16 +0200)]
radeon: Post-vblank-rework-rework cleanups.

Thanks to the reworked vblank-rework, we can just use the hardware frame
counter directly, and make the RADEON_PARAM_VBLANK_CRTC getparam just return
what was set by the corresponding setparam.

15 years agovblank-rework rework cleanups.
Michel Dänzer [Mon, 21 Jul 2008 06:16:55 +0000 (08:16 +0200)]
vblank-rework rework cleanups.

Remove some dead/obsolete code and make drm_update_vblank_count() static.

15 years agoRemove obsolete dev->vblank_suspend[crtc] tests.
Michel Dänzer [Mon, 21 Jul 2008 06:13:45 +0000 (08:13 +0200)]
Remove obsolete dev->vblank_suspend[crtc] tests.

Caused drm_update_vblank_count() not to do its thing when called from
drm_modeset_ctl() -> drm_vblank_get().

The vblank functionality no longer needs to be suspended during a modeset, so
rename the field to vblank_inmodeset.

15 years agoi915: convert to using drm_vblank_get/put around vblank counter usage
Jesse Barnes [Sat, 19 Jul 2008 17:18:02 +0000 (13:18 -0400)]
i915: convert to using drm_vblank_get/put around vblank counter usage

All interrupt off vblank count updates are done in drm_vblank_get/put
now, so convert users of the vblank counter over to that interface.

15 years agoUpdate vblank code to new API
Jesse Barnes [Sat, 19 Jul 2008 17:15:23 +0000 (13:15 -0400)]
Update vblank code to new API

In my last push I forgot to convert users of drm_update_vblank_count
over to drm_vblank_get/put, since that's where any interrupt off->on
update accounting is done now.  Since the modeset ioctl did something
similar (an open coded update of the counter) convert it over to using
get/put too, which saves us from having to deal with every combination
of interrupt off & on between calls.

15 years agodrm-gem: Fix build
Ian Romanick [Fri, 18 Jul 2008 19:42:43 +0000 (12:42 -0700)]
drm-gem: Fix build

On some distros missing prototypes cause kernel builds to fail.  These
are hack to make the code build.

15 years agointel-gem: Leave 8xx tiling on until we find any issues.
Eric Anholt [Thu, 17 Jul 2008 23:56:42 +0000 (16:56 -0700)]
intel-gem: Leave 8xx tiling on until we find any issues.

15 years agoradeon: remove microcode version
Dave Airlie [Fri, 18 Jul 2008 04:30:57 +0000 (14:30 +1000)]
radeon: remove microcode version

15 years agodrm/radeon: fixup 0 vs NULL
Dave Airlie [Thu, 19 Jun 2008 01:38:43 +0000 (11:38 +1000)]
drm/radeon: fixup 0 vs NULL

15 years ago[FreeBSD] drm_irq.c updates for vblank fixes.
Robert Noland [Thu, 17 Jul 2008 18:01:04 +0000 (14:01 -0400)]
[FreeBSD] drm_irq.c updates for vblank fixes.

15 years agoi915: remove old broken vblank code
Jesse Barnes [Thu, 17 Jul 2008 17:51:24 +0000 (13:51 -0400)]
i915: remove old broken vblank code

Remove the unused (and broken) "in vblank" code now that the core has
been fixed to use a counter while interrupts are enabled.  Also make the
vblank pipe get/set ioctls into dumb stub functions, since with the new
code we can no longer let userspace control whether vblank interrupts
are enabled, or the core code will misbehave.

15 years agoAvoid incorrect vblank wakeups
Jesse Barnes [Thu, 17 Jul 2008 17:48:14 +0000 (13:48 -0400)]
Avoid incorrect vblank wakeups

The current code uses the hw vblank counter exclusively, which can lead
to wakeups during the active period rather than during the vblank period
if the hw counter counts displayed frames rather than vblank periods.

This change coverts the code over to using the counter while interrupts
are enabled, fixing that issue.  It also includes a couple of related
changes:  one to not enable the new enable/disable behavior until the
modeset ioctl is called (to preserve old client behavior) and another to
account for lost events due to mode setting with the new counter scheme.

BSD will require similar changes to its drm_irq.c code, but they should
be straightforward.

15 years agoFreeBSD: Fix radeon build
Robert Noland [Thu, 17 Jul 2008 03:39:25 +0000 (23:39 -0400)]
FreeBSD: Fix radeon build

15 years agoBSD: change drm_locked_task*() to use the same scheme as linux.
Owain Gordon Ainsworth [Mon, 7 Jul 2008 16:23:48 +0000 (17:23 +0100)]
BSD: change drm_locked_task*() to use the same scheme as linux.

The current code can sleep in an interrupt handler, that is bad. So
instead if we can't grab the lock, flag it and run the tasklet on
unlock.

Signed-off-by: Robert Noland <rnoland@2hip.net>
15 years ago[FreeBSD] We aren't allowed to hold locks over bus_dma_tag_create or bus_dmamem_alloc.
Robert Noland [Mon, 9 Jun 2008 12:54:53 +0000 (08:54 -0400)]
[FreeBSD] We aren't allowed to hold locks over bus_dma_tag_create or bus_dmamem_alloc.

15 years agoThis is a modified version of Hong's patch from last month, with a few
Hong Liu [Tue, 15 Jul 2008 17:14:17 +0000 (10:14 -0700)]
This is a modified version of Hong's patch from last month, with a few
modifications to make it work correctly on my test hardware (altered the
backlight write function, made it enable the legacy backlight controller
interrupts on mobile hardware, sorted the interrupt function so we don't
get an excessive number of vblank interrupts). This lets the backlight
keys on my T61 work properly, though there's a 750msec or so delay
between the request and the brightness actually changing - this sounds
awfully like the hardware spinning waiting for a status flag to become
ready, but as far as I can tell they're all set correctly. If anyone can
figure out what's wrong here, it'd be nice to know.

Some of the functions are still stubs and just tell the hardware that
the request was successful. These can be filled in as kernel modesetting
gets integrated. I think it's worth getting this in anyway, since it's
required for backlight control to work properly on some new platforms.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
15 years agodrm: fix missing symbol export
Dave Airlie [Tue, 15 Jul 2008 10:27:14 +0000 (20:27 +1000)]
drm: fix missing symbol export

15 years agodrm: add fix for PAT on radeon with 2.6.26
Dave Airlie [Tue, 15 Jul 2008 06:18:22 +0000 (16:18 +1000)]
drm: add fix for PAT on radeon with 2.6.26

15 years agodrm/pcigart: fix warning
Dave Airlie [Tue, 15 Jul 2008 06:18:04 +0000 (16:18 +1000)]
drm/pcigart: fix warning

15 years agointel-gem: Disable tiling if we get junk from the MCHBAR read.
Eric Anholt [Mon, 14 Jul 2008 16:16:45 +0000 (09:16 -0700)]
intel-gem: Disable tiling if we get junk from the MCHBAR read.

One of our systems has been returning 0xffffffff from all MCHBAR reads, which
means we'll need to figure out why, or add an alternate detection method.

15 years agointel-gem: Add two new ioctls for managing tiling on objects.
Eric Anholt [Fri, 11 Jul 2008 21:47:33 +0000 (14:47 -0700)]
intel-gem: Add two new ioctls for managing tiling on objects.

Various chips have exciting interactions between the CPU and the GPU's
different ways of accessing interleaved memory, so we need some kernel
assistance in determining how it works.

Only fully tested on GM965 so far.

15 years agoxgi: use true/false instead of TRUE/FALSE
martin capitanio [Tue, 8 Jul 2008 07:58:10 +0000 (17:58 +1000)]
xgi: use true/false instead of TRUE/FALSE

15 years agonouveau: interface changes for nv5x 3d
Ben Skeggs [Tue, 24 Jun 2008 18:39:32 +0000 (04:39 +1000)]
nouveau: interface changes for nv5x 3d

15 years ago[intel-gem] typo fix in DRM_ERROR
Eric Anholt [Thu, 26 Jun 2008 16:35:14 +0000 (09:35 -0700)]
[intel-gem] typo fix in DRM_ERROR

15 years agoi915: official name for GM45 chipset
Zhenyu Wang [Wed, 2 Jul 2008 16:49:19 +0000 (00:49 +0800)]
i915: official name for GM45 chipset

Signed-off-by: Zhenyu Wang <zhenyu.z.wang@intel.com>
15 years agoi915: only use tiled blits on 965+
Jesse Barnes [Tue, 1 Jul 2008 23:10:01 +0000 (16:10 -0700)]
i915: only use tiled blits on 965+

When scheduled swaps occur, we need to blit between front & back buffers.  I
the buffers are tiled, we need to set the appropriate XY_SRC_COPY tile bit,
only on 965 chips, since it will cause corruption on pre-965 (e.g. 945).

Bug reported by and fix tested by Tomas Janousek <tomi@nomi.cz>.

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
15 years agoRevert "i915: only use tiled blits on 965+"
Jesse Barnes [Tue, 1 Jul 2008 23:09:02 +0000 (16:09 -0700)]
Revert "i915: only use tiled blits on 965+"

This reverts commit 727d4f1d1667e43b3558bd5f6ed6dc2cd9c29401, somehow git
deleted the symlink and replaced it with the file.

15 years agoi915: only use tiled blits on 965+
Jesse Barnes [Tue, 1 Jul 2008 19:25:16 +0000 (12:25 -0700)]
i915: only use tiled blits on 965+

When scheduled swaps occur, we need to blit between front & back buffers.  If
the buffers are tiled, we need to set the appropriate XY_SRC_COPY tile bit, but
only on 965 chips, since it will cause corruption on pre-965 (e.g. 945).

Bug reported by and fix tested by Tomas Janousek <tomi@nomi.cz>.

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
15 years agoi915: enable bus mastering on i915 at resume time
Jie Luo [Tue, 24 Jun 2008 17:38:31 +0000 (10:38 -0700)]
i915: enable bus mastering on i915 at resume time

On 9xx chips, bus mastering needs to be enabled at resume time for much of the
chip to function.  With this patch, vblank interrupts will work as expected
on resume, along with other chip functions.   Fixes kernel bugzilla #10844.

Signed-off-by: Jie Luo <clotho67@gmail.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
16 years agonv50: when destroying a channel make sure it's not still current on PFIFO
Ben Skeggs [Wed, 25 Jun 2008 06:45:41 +0000 (16:45 +1000)]
nv50: when destroying a channel make sure it's not still current on PFIFO

We won't get a PFIFO context switch when the same channel ID is recreated if
the hw still thinks the channel is already active, which causes fun issues.

Should allow X to be stopped and started without tearing down the entire
card state in lastclose().

16 years ago[intel] Get vblank pipe from irq_mask_reg instead of hardware enable reg
Keith Packard [Tue, 24 Jun 2008 20:39:25 +0000 (13:39 -0700)]
[intel] Get vblank pipe from irq_mask_reg instead of hardware enable reg

With the interrupt enable/disable using only the mask register, it was wrong
to use the enable register to detect which pipes had vblank detection
turned on. Also, as we keep a local copy of the mask register around, and
MSI machines smack the hardware during the interrupt handler, it is more
efficient and more correct to use the local copy.

16 years ago[intel] Create functions to enable/disable interrupts
Keith Packard [Tue, 24 Jun 2008 20:08:04 +0000 (13:08 -0700)]
[intel] Create functions to enable/disable interrupts

This shares common code sequences for managing the interrupt register bits

16 years agoi915: remove unused variable
Jesse Barnes [Tue, 24 Jun 2008 19:57:21 +0000 (12:57 -0700)]
i915: remove unused variable

Leftover dev_priv from the move of the suspend/resume code into shared-core.

16 years agoi915: register definition & header file cleanup
Jesse Barnes [Tue, 24 Jun 2008 19:50:29 +0000 (12:50 -0700)]
i915: register definition & header file cleanup

It would be nice if one day the DRM driver was the canonical source for
register definitions and core macros.  To that end, this patch cleans things up
quite a bit, removing redundant definitions (some with different names
referring to the same register) and generally tidying up the header file.