@@ -349,8 +349,13 @@ int radeon_sa_bo_new(struct radeon_device *rdev,
/* see if we can skip over some allocations */
} while (radeon_sa_bo_next_hole(sa_manager, fences, tries));
+ for (i = 0; i < RADEON_NUM_RINGS; ++i)
+ radeon_fence_ref(fences[i]);
+
spin_unlock(&sa_manager->wq.lock);
r = radeon_fence_wait_any(rdev, fences, false);
+ for (i = 0; i < RADEON_NUM_RINGS; ++i)
+ radeon_fence_unref(&fences[i]);
spin_lock(&sa_manager->wq.lock);
/* if we have nothing to wait for block */
if (r == -ENOENT) {
This is a note to let you know that I have just added a patch titled drm/radeon: hold reference to fences in radeon_sa_bo_new to the linux-3.16.y-queue branch of the 3.16.y-ckt extended stable tree which can be found at: http://kernel.ubuntu.com/git/ubuntu/linux.git/log/?h=linux-3.16.y-queue This patch is scheduled to be released in version 3.16.7-ckt25. If you, or anyone else, feels it should not be added to this tree, please reply to this email. For more information about the 3.16.y-ckt tree, see https://wiki.ubuntu.com/Kernel/Dev/ExtendedStable Thanks. -Luis ---8<------------------------------------------------------------ From 73187980dfefe5198aadcfdf0a377e461eed2bfa Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Nicolai=20H=C3=A4hnle?= <nicolai.haehnle@amd.com> Date: Fri, 5 Feb 2016 14:35:53 -0500 Subject: drm/radeon: hold reference to fences in radeon_sa_bo_new MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit commit f6ff4f67cdf8455d0a4226eeeaf5af17c37d05eb upstream. An arbitrary amount of time can pass between spin_unlock and radeon_fence_wait_any, so we need to ensure that nobody frees the fences from under us. Based on the analogous fix for amdgpu. Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Luis Henriques <luis.henriques@canonical.com> --- drivers/gpu/drm/radeon/radeon_sa.c | 5 +++++ 1 file changed, 5 insertions(+)