Page MenuHomeFreeBSD

uma: Avoid excessive per-CPU draining
ClosedPublic

Authored by des on Fri, Mar 14, 12:56 AM.
Tags
None
Referenced Files
F112540193: D49349.diff
Wed, Mar 19, 2:04 PM
F112483933: D49349.id152240.diff
Tue, Mar 18, 4:31 PM
F112483206: D49349.id152356.diff
Tue, Mar 18, 4:13 PM
Unknown Object (File)
Tue, Mar 18, 12:10 PM
Unknown Object (File)
Tue, Mar 18, 9:25 AM
Unknown Object (File)
Mon, Mar 17, 7:14 PM
Unknown Object (File)
Sun, Mar 16, 4:45 AM

Details

Summary

After commit 389a3fa693ef, uma_reclaim_domain(UMA_RECLAIM_DRAIN_CPU)
calls uma_zone_reclaim_domain(UMA_RECLAIM_DRAIN_CPU) twice on each zone
in addition to globally draining per-CPU caches. This was unintended
and is unnecessarily slow; in particular, draining per-CPU caches
requires binding to each CPU.

Stop draining per-CPU caches when visiting each zone, just do it once in
pcpu_cache_drain_safe() to minimize the amount of expensive sched_bind()
calls.

Fixes: 389a3fa693ef ("uma: Add UMA_ZONE_UNMANAGED")

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

This revision is now accepted and ready to land.Fri, Mar 14, 9:03 AM

FWIW, internal testing with an equivalent diff shows the intermittent (~8%) OOM panics seen with 389a3fa693ef are now resolved. Thanks!

des edited reviewers, added: markj; removed: des.

commandeering since @markj is afk

This revision was automatically updated to reflect the committed changes.