[v2,0/6] x86/mce: Various fixes and cleanups for MCE handling

Message

Jan H. Schönherr Jan. 3, 2020, 3:07 p.m. UTC

Hi Boris.

This is the 2nd iteration of a smallish series with some fixes/cleanups
for the handling of MCEs. It should apply cleanly to your ras/core branch.

The first iteration can be found here:
  https://lore.kernel.org/linux-edac/[e-mail]/T/

Changes v1 -> v2:
- dropped patches 3, 5, 6  as you already cherry-picked them into
  ras/core or ras/urgent (this renumbers patch 4 in v1 to patch 3 in v2);
- addressed remaining comments on patches 1-3;
- added patch 5 as per Yazen's comment that the SRAO notifier shall
  not be used on AMD for now;
- added patch 4 as a prerequisite for the given realization of patch 5;
- added patch 6 as an example, what else can be done due to patch 4.

See individual patches 1-3 for more detailed comments on changes.

I'm not yet convinced, that patch 6 is an entirely good idea. I've
still included it for discussion. If we end up not doing something
like it, we can as well rewrite patch 5 to be just another "if"
within srao_decode_notifier()/uc_decode_notifier().

Regards
Jan

Jan H. Schönherr (6):
  x86/mce: Take action on UCNA/Deferred errors again
  x86/mce: Make mce=nobootlog work again
  x86/mce: Fix use of uninitialized MCE message string
  x86/mce: Allow a variable number of internal MCE decode notifiers
  x86/mce: Do not take action on SRAO/Deferred errors on AMD for now
  x86/mce: Dynamically register default MCE handler

 arch/x86/include/asm/mce.h     |   2 +-
 arch/x86/kernel/cpu/mce/core.c | 145 ++++++++++++++++++---------------
 2 files changed, 81 insertions(+), 66 deletions(-)