]> git.baikalelectronics.ru Git - kernel.git/commitdiff
drm/vc4: Fix false positive WARN() backtrace on refcount_inc() usage
authorBoris Brezillon <boris.brezillon@free-electrons.com>
Wed, 22 Nov 2017 20:39:28 +0000 (21:39 +0100)
committerBoris Brezillon <boris.brezillon@free-electrons.com>
Thu, 7 Dec 2017 09:09:43 +0000 (10:09 +0100)
With CONFIG_REFCOUNT_FULL enabled, refcount_inc() complains when it's
passed a refcount object that has its counter set to 0. In this driver,
this is a valid use case since we want to increment ->usecnt only when
the BO object starts to be used by real HW components and this is
definitely not the case when the BO is created.

Fix the problem by using refcount_inc_not_zero() instead of
refcount_inc() and fallback to refcount_set(1) when
refcount_inc_not_zero() returns false. Note that this 2-steps operation
is not racy here because the whole section is protected by a mutex
which guarantees that the counter does not change between the
refcount_inc_not_zero() and refcount_set() calls.

Fixes: 8943349f4512 ("drm/vc4: Add the DRM_IOCTL_VC4_GEM_MADVISE ioctl")
Reported-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20171122203928.28135-1-boris.brezillon@free-electrons.com
drivers/gpu/drm/vc4/vc4_bo.c

index 4ae45d7dac428984fa63081338ac23f283fc5ac8..2decc8e2c79f58aad6ea7626a32a0f26665d23eb 100644 (file)
@@ -637,7 +637,8 @@ int vc4_bo_inc_usecnt(struct vc4_bo *bo)
        mutex_lock(&bo->madv_lock);
        switch (bo->madv) {
        case VC4_MADV_WILLNEED:
-               refcount_inc(&bo->usecnt);
+               if (!refcount_inc_not_zero(&bo->usecnt))
+                       refcount_set(&bo->usecnt, 1);
                ret = 0;
                break;
        case VC4_MADV_DONTNEED: