sched: Fix lockup by limiting load-balance retries on lock-break
authorPeter Zijlstra <a.p.zijlstra@chello.nl>
Wed, 11 Jan 2012 12:11:12 +0000 (13:11 +0100)
committerIngo Molnar <mingo@elte.hu>
Wed, 11 Jan 2012 16:15:12 +0000 (17:15 +0100)
commitbced76aeaca03b45e3b4bdb868cada328e497847
treee2965b6cfecdc257a01b89ec6780b8de3b2e6d39
parent6db9dc150eabce7053c8df2a2146aa0d6748ec42
sched: Fix lockup by limiting load-balance retries on lock-break

Eric and David reported dead machines and traced it to commit
a195f004 ("sched: Fix load-balance lock-breaking"), it turns out
there's still a scenario where we can end up re-trying forever.

Since there is no strict forward progress guarantee in the
load-balance iteration we can get stuck re-retrying the same
task-set over and over.

Creating a forward progress guarantee with the existing
structure is somewhat non-trivial, for now simply terminate the
retry loop after a few tries.

Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Eric Dumazet <eric.dumazet@gmail.com>
Reported-by: David Ahern <dsahern@gmail.com>
[ logic cleanup as suggested by Eric ]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1326297936.2442.157.camel@twins
Signed-off-by: Ingo Molnar <mingo@elte.hu>
kernel/sched/fair.c