SRU: Fix system hang when stress S3 on radeon with TTM
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
HWE Next |
Fix Released
|
Undecided
|
Unassigned | ||
linux-oem-5.6 (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
Focal |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
SRU Justification:
[Impact]
System hang when stress S3 more than 90 times.
[Fix]
Upstream kernel is good, after bisecting, bad commit is a Ubuntu sauce
patch: "vfio -- release device lock before userspace requests"
5.8 kernel got this fix.
5.4 kernel got this commit by stable update LP:#1888560
[Test Case]
Verified for 500 times of S3, system runs good.
[Regression Potential]
Low
uptream fix for specific commit, verified with postive result.
=======
[Summary]
System got hang during executing the S3-30-cycle test case.
[Steps to reproduce]
$ sudo checkbox-
[Expected result]
The test script should be finished smoothly.
[Actual result]
System got hang at the 7th or 8th S3 resumed back.
[Failure rate]
100%
[Additional information]
CPU: Intel(R) Core(TM) i5-10600 CPU @ 3.30GHz (12x)
GPU: 00:02.0 Display controller [0380]: Intel Corporation Device [8086:9bc8]
01:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Oland [Radeon HD 8570 / R7 240/340 OEM] [1002:6611] (rev 87)
kernel-version: 5.6.0-1020-oem
CVE References
description: | updated |
summary: |
- System hang when stress S3 on radeon with TTM + SRU: System hang when stress S3 on radeon with TTM |
summary: |
- SRU: System hang when stress S3 on radeon with TTM + SRU: Fix system hang when stress S3 on radeon with TTM |
no longer affects: | linux (Ubuntu Focal) |
no longer affects: | linux (Ubuntu) |
Changed in linux-oem-5.6 (Ubuntu Focal): | |
status: | New → Fix Committed |
tags: | added: verification-done verification-done-focal |
tags: | removed: verification-done verification-done-focal |
Changed in linux-oem-5.6 (Ubuntu): | |
status: | New → Invalid |
Changed in hwe-next: | |
status: | New → Fix Released |
After revert commit in Focal and oem-5.6, this issue can be fixed.
From b1cc82bb294939c 76dc6b61146e960 bd9f810222 Mon Sep 17 00:00:00 2001
From: Aaron Ma <email address hidden>
Date: Fri, 28 Aug 2020 11:03:46 +0800
Subject: [PATCH] Revert "UBUNTU: SAUCE: vfio -- release device lock before
userspace requests"
This reverts part of commit 240766c8b029a44 45f632f822a744c 6ae34c48e3.
Part of changes break stress S3 test on ttm driver with radeon.
Error log:
[TTM] Erroneous page count. Leaking pages.
Signed-off-by: Aaron Ma <email address hidden>
---
drivers/base/dd.c | 7 -------
1 file changed, 7 deletions(-)
diff --git a/drivers/base/dd.c b/drivers/base/dd.c .b25bcab2a26b 100644 release_ driver( struct device *dev, struct device *parent)
dev- >bus->remove( dev);
else if (drv->remove)
drv- >remove( dev);
index b4f9e99f7372.
--- a/drivers/base/dd.c
+++ b/drivers/base/dd.c
@@ -1135,13 +1135,6 @@ static void __device_
- /*
- * A concurrent invocation of the same function might
- * have released the driver successfully while this one
- * was waiting, so check for that.
- */
- if (dev->driver != drv)
- return;
--