diff options
author | Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 2015-05-04 22:34:58 +0200 |
---|---|---|
committer | Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 2015-05-04 22:34:58 +0200 |
commit | 20f34165a92427d16112c4420e0c22fa0e40b3bd (patch) | |
tree | ba9774a85cae1252b3ed609b239697c9913c4eb4 /drivers/pci | |
parent | Linux 4.1-rc2 (diff) | |
parent | GHES: Make NMI handler have a single reader (diff) | |
download | linux-20f34165a92427d16112c4420e0c22fa0e40b3bd.tar.xz linux-20f34165a92427d16112c4420e0c22fa0e40b3bd.zip |
Merge tag 'ras_for_4.2' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras into acpi-apei
Pull GHES changes for 4.2 from Borislav Petkov:
"GHES: Seriously speedup and cleanup NMI handler (Jiri Kosina and Borislav Petkov)
This is the result of us seeing this during boot
[ 24.332560] INFO: NMI handler (ghes_notify_nmi) took too long to run: 3.265 msecs
[ 24.332567] INFO: NMI handler (ghes_notify_nmi) took too long to run: 5.946 msecs
[ 24.332568] INFO: NMI handler (ghes_notify_nmi) took too long to run: 5.948 msecs
and a report of people running perf stat and the machine softlocking.
"hest_disable" was helping in this case, which made us look at that
particular NMI handler. It was grabbing a lock each time it is run and
on each CPU. But this is not needed as the GHES sources are global and
they need only a single reader.
This patchset does that and cleans up the handler in the process."
* tag 'ras_for_4.2' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras:
GHES: Make NMI handler have a single reader
GHES: Elliminate double-loop in the NMI handler
GHES: Panic right after detection
GHES: Carve out the panic functionality
GHES: Carve out error queueing in a separate function
Diffstat (limited to 'drivers/pci')
0 files changed, 0 insertions, 0 deletions