Re: Fwd: Please revert wrong kernel OOPS patch in -69 !

["Nicolai Hähnle" <nicolai.haehnle@amd.com>]

[-- Attachment #1: Type: text/plain, Size: 1253 bytes --]

Hmm, I was under the impression from an email I got that this was 
already reverted from 3.16. Clearly, communication is hard.

Let me re-iterate, to be sure: kernels 3.17 and older

1. need to revert the backport of upstream commit 
f6ff4f67cdf8455d0a4226eeeaf5af17c37d05eb (I think most have done that 
already) and

2. should apply the corrected backport which I am attaching yet again 
(seems like when I sent this patch around two weeks ago, it wasn't 
picked up).

Thanks,
Nicolai

On 07.04.2016 04:15, Taras Prokopenko wrote:
>
> ---------- Forwarded message ----------
> From: *Taras Prokopenko* <taras.prokopenko@gmail.com
> <mailto:taras.prokopenko@gmail.com>>
> Date: 2016-04-07 12:04 GMT+03:00
> Subject: Please revert wrong kernel OOPS patch in -69 !
> To: luis.henriques@canonical.com <mailto:luis.henriques@canonical.com>
> Cc: jslaby@suse.cz <mailto:jslaby@suse.cz>
>
>
> Please test your changes! Lost 1/2 a day to figure out whats causing my
> pc to hang.
>
> OopsText:
>   BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
>   IP: [<ffffffffc03b83ed>] radeon_fence_ref+0xd/0x50 [radeon]
>   PGD 0
>   Oops: 0002 [#1] SMP
>
>
> https://lists.ubuntu.com/archives/kernel-team/2016-February/072148.html
>
> Damn
>

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-drm-radeon-hold-reference-to-fences-in-radeon_sa_bo_.patch --]
[-- Type: text/x-patch; name="0001-drm-radeon-hold-reference-to-fences-in-radeon_sa_bo_.patch", Size: 1944 bytes --]

>From ad94965f69c2681832f64473d28c23ae71b6e52f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Nicolai=20H=C3=A4hnle?= <nicolai.haehnle@amd.com>
Date: Tue, 15 Mar 2016 12:56:45 -0500
Subject: [PATCH] drm/radeon: hold reference to fences in radeon_sa_bo_new
 (3.17 and older)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

[Backport of upstream commit f6ff4f67cdf8455d0a4226eeeaf5af17c37d05eb, with
 an additional NULL pointer guard that is required for kernels 3.17 and older.

 To be precise, any kernel that does *not* have commit 954605ca3 "drm/radeon:
 use common fence implementation for fences, v4" requires this additional
 NULL pointer guard.]

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> (v1 + fix)
Tested-by: Lutz Euler <lutz.euler@freenet.de>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/radeon/radeon_sa.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
index f0bac68..8962411 100644
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ b/drivers/gpu/drm/radeon/radeon_sa.c
@@ -349,8 +349,15 @@ 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) {
+			if (fences[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 && block) {
-- 
2.5.0