[Disco][SRU][PATCH 0/2] Fix for kernel machine check events logging

Po-Hsu Lin po-hsu.lin at canonical.com
Tue Dec 31 09:53:24 UTC 2019

== SRU Justification ==
With the 5.0 Disco kernel, the kernel cannot record the mce log while
injecting 1bit ecc error.

== Fix ==
  * 09cbd219 (RAS/CEC: Increment cec_entered under the mutex lock)
  * de0e0624 (RAS/CEC: Check count_threshold unconditionally)

Commit de0e0624 is the real fix for this issue. 09cbd219 is a fix to
avoid race condition, and it can make the latter become a clean

These have been landed on newer kernels.

== Test ==
Test kernel could be found here:

Verified by the bug reporter, fan jinke, the patched kernel can log
the error correctly.

== Regression Potential ==
Low, changes are limited to the RAS Correctable Errors Collector. And
the fix has been verified as working as expected.

Borislav Petkov (1):
  RAS/CEC: Check count_threshold unconditionally

WANG Chao (1):
  RAS/CEC: Increment cec_entered under the mutex lock

 drivers/ras/cec.c | 31 ++++++++++++-------------------
 1 file changed, 12 insertions(+), 19 deletions(-)


More information about the kernel-team mailing list