linux-stable.git/drivers/gpu/drm/drm_gem.c, branch linux-3.16.y

drm: Release driver tracking before making the object available again

2017-11-11T13:33:23+00:00

commit fe4600a548f2763dec91b3b27a1245c370ceee2a upstream.

This is the same bug as we fixed in commit f6cd7daecff5 ("drm: Release
driver references to handle before making it available again"), but now
the exposure is via the PRIME lookup tables. If we remove the
object/handle from the PRIME lut, then a new request for the same
object/fd will generate a new handle, thus for a short window that
object is known to userspace by two different handles. Fix this by
releasing the driver tracking before PRIME.

Fixes: 0ff926c7d4f0 ("drm/prime: add exported buffers to current fprivs
imported buffer list (v2)")
Signed-off-by: Chris Wilson 
Cc: David Airlie 
Cc: Daniel Vetter 
Cc: Rob Clark 
Cc: Ville Syrjälä 
Cc: Thierry Reding 
Reviewed-by: Daniel Vetter 
Signed-off-by: Joonas Lahtinen 
Link: https://patchwork.freedesktop.org/patch/msgid/20170819120558.6465-1-chris@chris-wilson.co.uk
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

drm/gem: replace misleading comment

2014-05-27T05:50:57+00:00

shmem supports page-relocations during swapin since quite some time. It
was implemented in:

    commit bde05d1ccd512696b09db9dd2e5f33ad19152605
    Author: Hugh Dickins 
    Date:   Tue May 29 15:06:38 2012 -0700

        shmem: replace page if mapping excludes its zone

The gem-comment about wrongly placed DMA32 pages is no longer valid.
Replace it with a proper comment but keep the BUG_ON() to verify correct
shmem behavior.

Signed-off-by: David Herrmann 
Reviewed-by: Daniel Vetter 
Signed-off-by: Dave Airlie

Merge branch 'drm-docs' of ssh://people.freedesktop.org/~danvet/drm into drm-next

2014-03-18T09:09:10+00:00

Here's my drm documentation update and driver api polish pull request.
Alex reviewed the entire pile, I've applied a little bit of spelling
polish in a few places since then and otherwise the Usual Suspects (David,
Rob, ...) don't seem up to have another look at it (I've poked them on
irc). So I think it's as good as it gets ;-)

Note that I've dropped the final imx breaker patch since that's blocked on
imx getting sane. Once that's landed I'll ping you to pick up that
straggler.

* 'drm-docs' of ssh://people.freedesktop.org/~danvet/drm: (34 commits)
  drm/imx: remove drm_mode_connector_detach_encoder harder
  drm: kerneldoc polish for drm_crtc.c
  drm: kerneldoc polish for drm_crtc_helper.c
  drm: drop error code for drm_helper_resume_force_mode
  drm/crtc-helper: remove LOCKING from kerneldoc
  drm: remove return value from drm_helper_mode_fill_fb_struct
  drm/doc: Fix misplaced 
  drm: remove drm_display_mode->private_size
  drm: polish function kerneldoc for drm_modes.[hc]
  drm/modes: drop maxPitch from drm_mode_validate_size
  drm/modes: drop return value from drm_display_mode_from_videomode
  drm/modes: remove drm_mode_height/width
  drm: extract drm_modes.h for drm_crtc.h functions
  drm: move drm_mode related functions into drm_modes.c
  drm/doc: Repleace LOCKING kerneldoc sections in drm_modes.c
  drm/doc: Integrate drm_modes.c kerneldoc
  drm/kms: rip out drm_mode_connector_detach_encoder
  drm/doc: Add function reference documentation for drm_mm.c
  drm/doc: Overview documentation for drm_mm.c
  drm/mm: Remove MM_UNUSED_TARGET
  ...

drm/gem: dont init "ret" in drm_gem_mmap()

2014-03-16T11:11:01+00:00

There is no need to initialize this variable, so drop it. Otherwise, the
compiler won't warn if we use it unintialized.

Reviewed-by: Daniel Vetter 
Signed-off-by: David Herrmann

drm/gem: free vma-node during object-cleanup

2014-03-16T11:11:01+00:00

All drivers currently need to clean up the vma-node manually. There is no
fancy logic involved so lets just clean it up unconditionally. The
vma-manager correctly catches multiple calls so we are fine.

Signed-off-by: David Herrmann

drm/gem: fix indentation

2014-03-16T11:11:01+00:00

Remove double-whitespace and wrong indentation.

Signed-off-by: David Herrmann

drm/doc: Clean up and integrate kerneldoc for drm_gem.c

2014-03-13T11:48:27+00:00

Fairly incomplete, but at least a start.

Reviewed-by: Alex Deucher 
Signed-off-by: Daniel Vetter

drm/gem: Always initialize the gem object in object_init

2014-01-21T09:19:58+00:00

At least drm/i915 expects that the obj->dev pointer is set even in
failure paths. Specifically when the shmem initialization fails we
call i915_gem_object_free which needs to deref obj->base.dev to get at
the slab pointer in the device private structure. And the shmem
allocation can easily fail when userspace is hitting open file limits.

Doing the structure init even when the shmem file allocation fails
prevents this Oops.

This is a regression from

commit 89c8233f82d9c8af5b20e72e4a185a38a7d3c50b
Author: David Herrmann 
Date:   Thu Jul 11 11:56:32 2013 +0200

    drm/gem: simplify object initialization

v2: Add regression note which Chris supplied.

Testcase: igt/gem_fd_exhaustion
Reported-and-Suggested-by: Linus Torvalds 
Cc: Linus Torvalds 
References: http://lists.freedesktop.org/archives/intel-gfx/2014-January/038433.html
Cc: stable@vger.kernel.org
Reviewed-by: David Herrmann 
Cc: David Herrmann 
Signed-off-by: Daniel Vetter

drm: store the gem vma offset manager in a typed pointer

2014-01-14T02:38:32+00:00

This was hidden in a generic void * dev->mm_private. But only ever
used for gem. But thanks to this fake generic pretension no one
noticed that Rob's drm drivers are now all broken.

So just give the offset manager a type pointer and fix up msm, omapdrm
and tilcdc.

v2: Fixup compile fail.

v3: Fixup rebase fail that David spotted.

Cc: David Herrmann 
Cc: Rob Clark 
Signed-off-by: Daniel Vetter 
Signed-off-by: Dave Airlie

drm: Don't reference objects in the flink name idr

2013-12-18T00:48:17+00:00

There's no reason to keep a reference to objects in the name idr.  Each
handle to an object has a reference to the object and just before we
destroy the last handle we take the object out of the name idr.  Thus,
if an object is in the name idr, there's at least one reference to the
object.

Or to put it another way, the name idr reference will never keep the
object alive.  It just looks like it, which is confusing.

Signed-off-by: Kristian Høgsberg 
Reviewed-by: Daniel Vetter 
Signed-off-by: Dave Airlie