]> git.baikalelectronics.ru Git - kernel.git/commitdiff
drm/i915/userptr: Try to acquire the page lock around set_page_dirty()
authorChris Wilson <chris@chris-wilson.co.uk>
Mon, 11 Nov 2019 13:32:03 +0000 (13:32 +0000)
committerJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Tue, 12 Nov 2019 08:22:58 +0000 (10:22 +0200)
set_page_dirty says:

For pages with a mapping this should be done under the page lock
for the benefit of asynchronous memory errors who prefer a
consistent dirty state. This rule can be broken in some special
cases, but should be better not to.

Under those rules, it is only safe for us to use the plain set_page_dirty
calls for shmemfs/anonymous memory. Userptr may be used with real
mappings and so needs to use the locked version (set_page_dirty_lock).

However, following a try_to_unmap() we may want to remove the userptr and
so call put_pages(). However, try_to_unmap() acquires the page lock and
so we must avoid recursively locking the pages ourselves -- which means
that we cannot safely acquire the lock around set_page_dirty(). Since we
can't be sure of the lock, we have to risk skip dirtying the page, or
else risk calling set_page_dirty() without a lock and so risk fs
corruption.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=203317
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112012
Fixes: 5cc9ed4b9a7a ("drm/i915: Introduce mapping of user pages into video memory (userptr) ioctl")
References: cb6d7c7dc7ff ("drm/i915/userptr: Acquire the page lock around set_page_dirty()")
References: 505a8ec7e11a ("Revert "drm/i915/userptr: Acquire the page lock around set_page_dirty()"")
References: 6dcc693bc57f ("ext4: warn when page is dirtied without buffers")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: stable@vger.kernel.org
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191111133205.11590-1-chris@chris-wilson.co.uk
(cherry picked from commit 0d4bbe3d407f79438dc4f87943db21f7134cfc65)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
drivers/gpu/drm/i915/gem/i915_gem_userptr.c

index 1e045c3370444021de0dd88b94e99cf009bc2145..4c72d74d65764f6f726eca9db8e6e4aa9e9e5c43 100644 (file)
@@ -646,8 +646,28 @@ i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj,
                obj->mm.dirty = false;
 
        for_each_sgt_page(page, sgt_iter, pages) {
-               if (obj->mm.dirty)
+               if (obj->mm.dirty && trylock_page(page)) {
+                       /*
+                        * As this may not be anonymous memory (e.g. shmem)
+                        * but exist on a real mapping, we have to lock
+                        * the page in order to dirty it -- holding
+                        * the page reference is not sufficient to
+                        * prevent the inode from being truncated.
+                        * Play safe and take the lock.
+                        *
+                        * However...!
+                        *
+                        * The mmu-notifier can be invalidated for a
+                        * migrate_page, that is alreadying holding the lock
+                        * on the page. Such a try_to_unmap() will result
+                        * in us calling put_pages() and so recursively try
+                        * to lock the page. We avoid that deadlock with
+                        * a trylock_page() and in exchange we risk missing
+                        * some page dirtying.
+                        */
                        set_page_dirty(page);
+                       unlock_page(page);
+               }
 
                mark_page_accessed(page);
                put_page(page);