[PATCH] drm/radeon: evergreen/cayman indirect draw support

2014-11-09 Thread Marek Olšák
Hi Glenn, First of all, evergreen_cs_track_check only has to be called for the packets that execute a draw call. It's useless to call it for INDEX_BASE, INDEX_BUFFER_SIZE, and SET_BASE. There has never been index buffer bounds checking in the CS checker and this patch doesn't change that, so no a

[PATCH] drm/radeon: evergreen/cayman indirect draw support

2014-11-09 Thread Glenn Kennard
On Sun, 09 Nov 2014 00:15:49 +0100, Jerome Glisse wrote: > On Sat, Nov 08, 2014 at 11:51:30PM +0100, Glenn Kennard wrote: >> Signed-off-by: Glenn Kennard > > NAK insecure. This is missing any kind of boundary checking for the > indirect buffer and thus can be abuse. The indirect command buffe

[PATCH] drm/radeon: evergreen/cayman indirect draw support

2014-11-08 Thread Glenn Kennard
Signed-off-by: Glenn Kennard --- See patch sent to mesa-dev at lists.freedesktop.org for userspace usage. drivers/gpu/drm/radeon/evergreen_cs.c | 76 +++ drivers/gpu/drm/radeon/evergreend.h | 1 + drivers/gpu/drm/radeon/radeon_drv.c | 3 +- 3 files changed,

[PATCH] drm/radeon: evergreen/cayman indirect draw support

2014-11-08 Thread Jerome Glisse
On Sat, Nov 08, 2014 at 11:51:30PM +0100, Glenn Kennard wrote: > Signed-off-by: Glenn Kennard NAK insecure. This is missing any kind of boundary checking for the indirect buffer and thus can be abuse. > --- > See patch sent to mesa-dev at lists.freedesktop.org for userspace usage. > > drivers/