kernel/linux.git/drivers/gpu/drm/vc4, branch v4.9.79

drm/vc4: Fix OOPSes from trying to cache a partially constructed BO.

2017-06-24T05:11:14+00:00

commit ca39b449f6d03e8235969f12f5dd25b8eb4304d6 upstream. If a CMA allocation failed, the partially constructed BO would be unreferenced through the normal path, and we might choose to put it in the BO cache. If we then reused it before it expired from the cache, the kernel would OOPS. Signed-off-by: Eric Anholt Fixes: c826a6e10644 ("drm/vc4: Add a BO cache.") Reviewed-by: Boris Brezillon Link: http://patchwork.freedesktop.org/patch/msgid/20170301185602.6873-2-eric@anholt.net Signed-off-by: Greg Kroah-Hartman

drm/vc4: Allocate the right amount of space for boot-time CRTC state.

2017-04-08T07:30:35+00:00

commit 6d6e500391875cc372336c88e9a8af377be19c36 upstream. Without this, the first modeset would dereference past the allocation when trying to free the mm node. Signed-off-by: Eric Anholt Tested-by: Stefan Wahren Link: http://patchwork.freedesktop.org/patch/msgid/20170328201343.4884-1-eric@anholt.net Fixes: d8dbf44f13b9 ("drm/vc4: Make the CRTCs cooperate on allocating display lists.") Reviewed-by: Daniel Vetter Signed-off-by: Greg Kroah-Hartman

drm/vc4: Use runtime autosuspend to avoid thrashing V3D power state.

2017-03-26T11:05:56+00:00

commit 3a62234680d86efa0239665ed8a0e908f1aef147 upstream. The pm_runtime_put() we were using immediately released power on the device, which meant that we were generally turning the device off and on once per frame. In many profiles I've looked at, that added up to about 1% of CPU time, but this could get worse in the case of frequent rendering and readback (as may happen in X rendering). By keeping the device on until we've been idle for a couple of frames, we drop the overhead of runtime PM down to sub-.1%. Signed-off-by: Eric Anholt Cc: Amit Pundir Signed-off-by: Greg Kroah-Hartman

drm/vc4: Fix termination of the initial scan for branch targets.

2017-03-26T11:05:56+00:00

commit 457e67a728696c4f8e6423c64e93def50530db9a upstream. The loop is scanning until the original max_ip (size of the BO), but we want to not examine any code after the PROG_END's delay slots. There was a block trying to do that, except that we had some early continue statements if the signal wasn't a PROG_END or a BRANCH. The failure mode would be that a valid shader is rejected because some undefined memory after the PROG_END slots is parsed as a branch and the rest of its setup is illegal. I haven't seen this in the wild, but valgrind was complaining when about this up in the userland simulator mode. Signed-off-by: Eric Anholt Cc: Amit Pundir Signed-off-by: Greg Kroah-Hartman

drm/vc4: Fix ->clock_select setting for the VEC encoder

2017-03-22T11:43:39+00:00

commit ab8df60e3a3b68420d0d4477c5f07c00fbfb078b upstream. PV_CONTROL_CLK_SELECT_VEC is actually 2 and not 0. Fix the definition and rework the vc4_set_crtc_possible_masks() to cover the full range of the PV_CONTROL_CLK_SELECT field. Signed-off-by: Boris Brezillon Signed-off-by: Eric Anholt Cc: Amit Pundir Signed-off-by: Greg Kroah-Hartman

drm/vc4: Fix race between page flip completion event and clean-up

2017-03-22T11:43:39+00:00

commit 26fc78f6fef39b9d7a15def5e7e9826ff68303f4 upstream. There was a small window where a userspace program could submit a pageflip after receiving a pageflip completion event yet still receive EBUSY. Signed-off-by: Derek Foreman Signed-off-by: Eric Anholt Reviewed-by: Eric Anholt Reviewed-by: Daniel Stone Cc: Amit Pundir Signed-off-by: Greg Kroah-Hartman

drm/vc4: fix a bounds check

2017-02-01T07:33:03+00:00

commit 21ccc32496b2f63228f5232b3ac0e426e8fb3c31 upstream. We accidentally return success even if vc4_full_res_bounds_check() fails. Fixes: d5b1a78a772f ("drm/vc4: Add support for drawing 3D frames.") Signed-off-by: Dan Carpenter Reviewed-by: Eric Engestrom Reviewed-by: Eric Anholt Signed-off-by: Greg Kroah-Hartman

drm/vc4: Return -EINVAL on the overflow checks failing.

2017-02-01T07:33:03+00:00

commit 6b8ac63847bc2f958dd93c09edc941a0118992d9 upstream. By failing to set the errno, we'd continue on to trying to set up the RCL, and then oops on trying to dereference the tile_bo that binning validation should have set up. Reported-by: Ingo Molnar Signed-off-by: Eric Anholt Fixes: d5b1a78a772f ("drm/vc4: Add support for drawing 3D frames.") Signed-off-by: Greg Kroah-Hartman

drm/vc4: Fix an integer overflow in temporary allocation layout.

2017-02-01T07:33:03+00:00

commit 0f2ff82e11c86c05d051cae32b58226392d33bbf upstream. We copy the unvalidated ioctl arguments from the user into kernel temporary memory to run the validation from, to avoid a race where the user updates the unvalidate contents in between validating them and copying them into the validated BO. However, in setting up the layout of the kernel side, we failed to check one of the additions (the roundup() for shader_rec_offset) against integer overflow, allowing a nearly MAX_UINT value of bin_cl_size to cause us to under-allocate the temporary space that we then copy_from_user into. Reported-by: Murray McAllister Signed-off-by: Eric Anholt Fixes: d5b1a78a772f ("drm/vc4: Add support for drawing 3D frames.") Signed-off-by: Greg Kroah-Hartman

drm/vc4: Fix memory leak of the CRTC state.

2017-02-01T07:33:03+00:00

commit 7622b25543665567d8830a63210385b7d705924b upstream. The underscores variant frees the pointers inside, while the no-underscores variant calls underscores and then frees the struct. Signed-off-by: Eric Anholt Fixes: d8dbf44f13b9 ("drm/vc4: Make the CRTCs cooperate on allocating display lists.") Signed-off-by: Greg Kroah-Hartman