On Thu 11-06-20 09:13:07, Paolo Valente wrote:
Il giorno 5 giu 2020, alle ore 16:16, Jan Kara jack@suse.cz ha scritto:
bfq_setup_cooperator() uses bfqd->in_serv_last_pos so detect whether it makes sense to merge current bfq queue with the in-service queue. However if the in-service queue is freshly scheduled and didn't dispatch any requests yet, bfqd->in_serv_last_pos is stale and contains value from the previously scheduled bfq queue which can thus result in a bogus decision that the two queues should be merged.
Good catch!
This bug can be observed for example with the following fio jobfile:
[global] direct=0 ioengine=sync invalidate=1 size=1g rw=read
[reader] numjobs=4 directory=/mnt
where the 4 processes will end up in the one shared bfq queue although they do IO to physically very distant files (for some reason I was able to observe this only with slice_idle=1ms setting).
Fix the problem by invalidating bfqd->in_serv_last_pos when switching in-service queue.
Apart from the nonexistent problem that even 0 is a valid LBA :)
Yes, I was also thinking about that and decided 0 is "good enough" :). But I just as well just switch to (sector_t)-1 if you think it would be better.
Acked-by: Paolo Valente paolo.valente@linaro.org
Thanks!
Honza