New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 721931 link

Starred by 1 user

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 3
Type: Bug



Sign in to add a comment

Should we set SPLIT_PTLOCK_CPUS=8 for kernel 4.4/arm64?

Project Member Reported by diand...@chromium.org, May 12 2017

Issue description

There's a kernel config option "SPLIT_PTLOCK_CPUS".  As far as I can tell the value itself is ignored except for a comparison to "NR_CPUS".  If the value is >= NR_CPUS then we use SPLIT_PTLOCK.  If not then we don't.

For kernel 4.4 on arm64 we set NR_CPUS to 8, so we don't get SPLIT_PTLOCK_CPUS.

Question is: do we want it?
...is there any type of benchmark that shows good / bad things here?
 
I believe that will split up the page table lock to allow better concurrency of page faults in a process.  For a synthetic test I'd say we'd need a multi-threaded version of balloon.  For a real(ish) test -- probably something similar to what we're doing for general memory pressure on arc++ with lots of tabs + several big android applications.
Project Member

Comment 2 by sheriffbot@chromium.org, May 16 2018

Labels: Hotlist-Recharge-Cold
Status: Untriaged (was: Available)
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Luigi -- we could use your new test with this kernel change and see if it improves things on kevin
Status: Assigned (was: Untriaged)
This bug has an owner, thus, it's been triaged. Changing status to "assigned".

Sign in to add a comment