]> git.baikalelectronics.ru Git - kernel.git/commit
drm/i915: Reorder context-close to avoid calling i915_vma_close() under RCU
authorChris Wilson <chris@chris-wilson.co.uk>
Thu, 9 Nov 2017 08:55:40 +0000 (08:55 +0000)
committerJani Nikula <jani.nikula@intel.com>
Thu, 9 Nov 2017 14:18:37 +0000 (16:18 +0200)
commit3b77f5d6171dc959eda1dc506eda2dc53c02ae0c
tree4245d55b3fc04709d3aa5144b8957498fc23aa6b
parent015b3d4537481e773d08b9dfe5c0a6783ebde9c1
drm/i915: Reorder context-close to avoid calling i915_vma_close() under RCU

When we close the VMA, we unbind it from the ppgtt and tear down the
page directory pointing at it. That may trigger us to return WC pages
back to the system, requiring conversion back to WB which itself may
sleep. That makes i915_vma_close() unsuitable for use inside the RCU
read lock, which we need to hold to iterate the radixtree.

The fix is quite simple, we can close all the VMA as we close the ppgtt,
we only need to do that instead of closing them during destruction of
the LUT.

v2: Order between closing the LUT and the ppgtt is important; we use the
vma inside the LUT as a means of retrieving the object, and so we must
clear the LUT before freeing the VMA when closing the ppgtt.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103638
Fixes: 547da76b5777 ("drm/i915: Hold rcu_read_lock when iterating over the radixtree (vma idr)")
Fixes: 1b5dde0153b2 ("drm/i915: Replace execbuf vma ht with an idr")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171109085540.32264-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
(cherry picked from commit 94dec87159af6f3dcc0b78d3f909aefa9e29c01a)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
drivers/gpu/drm/i915/i915_gem_context.c