Messages in this thread | | | Date | Tue, 26 Jun 2012 13:43:22 -0700 | From | Tejun Heo <> | Subject | Re: multi-second application stall in open() |
| |
Hello,
On Tue, Jun 26, 2012 at 10:18:04AM -0500, Josh Hunt wrote: > Do you have any suggestions on how to debug this? > > I did "perf record -a -e workqueue:*" and grabbed some tracepoint > data, but it's hard to correlate when these events are occurring in > the blktrace logs. Will keep investigating.
Yeah, as Vivek said, it would be better to get both traces through perf instead of using blktrace. queue_work and execute_start TPs print out the function pointer. Only printing out entries which match the unplug_work would reduce a lot of noise - you can grep cfq_kick_queue's address from /proc/kallsyms and set up /sys/kernel/debug/tracing/events/workqueue/EVENT/filter - "echo function==0xFUNCADDR > /sys/kernel/debug.../filter". Not sure how to do that with perf tho.
Ooh, you posted another message. Will continue there.
Thanks.
-- tejun
| |