tracing/ring_buffer: Try harder to allocate
authorJoel Fernandes <joelaf@google.com>
Thu, 13 Jul 2017 02:14:16 +0000 (19:14 -0700)
committerSteven Rostedt (VMware) <rostedt@goodmis.org>
Wed, 19 Jul 2017 12:22:12 +0000 (08:22 -0400)
commit848618857d2535176037bdc085f8d012d907071f
tree2d251e7c70404119716aa9a999d60239b78f6d3a
parent5771a8c08880cdca3bfb4a3fc6d309d6bba20877
tracing/ring_buffer: Try harder to allocate

ftrace can fail to allocate per-CPU ring buffer on systems with a large
number of CPUs coupled while large amounts of cache happening in the
page cache. Currently the ring buffer allocation doesn't retry in the VM
implementation even if direct-reclaim made some progress but still
wasn't able to find a free page. On retrying I see that the allocations
almost always succeed. The retry doesn't happen because __GFP_NORETRY is
used in the tracer to prevent the case where we might OOM, however if we
drop __GFP_NORETRY, we risk destabilizing the system if OOM killer is
triggered. To prevent this situation, use the __GFP_RETRY_MAYFAIL flag
introduced recently [1].

Tested the following still succeeds without destabilizing a system with
1GB memory.
echo 300000 > /sys/kernel/debug/tracing/buffer_size_kb

[1] https://marc.info/?l=linux-mm&m=149820805124906&w=2

Link: http://lkml.kernel.org/r/20170713021416.8897-1-joelaf@google.com
Cc: Tim Murray <timmurray@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Joel Fernandes <joelaf@google.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
kernel/trace/ring_buffer.c