Re: [REGRESSION] (>= v4.12) IO w/dmcrypt causing audio underruns

[vcaputo]

On Fri, Jan 19, 2018 at 11:57:32AM +0100, Enric Balletbo Serra wrote:
> Hi Vito,
> 
> 2018-01-17 23:48 GMT+01:00  <vcaputo@xxxxxxxxxxx>:
> > On Mon, Dec 18, 2017 at 10:25:33AM +0100, Enric Balletbo Serra wrote:
> >> Hi Vito,
> >>
> >> 2017-12-01 22:33 GMT+01:00  <vcaputo@xxxxxxxxxxx>:
> >> > On Wed, Nov 29, 2017 at 10:39:19AM -0800, vcaputo@xxxxxxxxxxx wrote:
> >> >> Hello,
> >> >>
> >> >> Recently I noticed substantial audio dropouts when listening to MP3s in
> >> >> `cmus` while doing big and churny `git checkout` commands in my linux git
> >> >> tree.
> >> >>
> >> >> It's not something I've done much of over the last couple months so I
> >> >> hadn't noticed until yesterday, but didn't remember this being a problem in
> >> >> recent history.
> >> >>
> >> >> As there's quite an accumulation of similarly configured and built kernels
> >> >> in my grub menu, it was trivial to determine approximately when this began:
> >> >>
> >> >> 4.11.0: no dropouts
> >> >> 4.12.0-rc7: dropouts
> >> >> 4.14.0-rc6: dropouts (seem more substantial as well, didn't investigate)
> >> >>
> >> >> Watching top while this is going on in the various kernel versions, it's
> >> >> apparent that the kworker behavior changed.  Both the priority and quantity
> >> >> of running kworker threads is elevated in kernels experiencing dropouts.
> >> >>
> >> >> Searching through the commit history for v4.11..v4.12 uncovered:
> >> >>
> >> >> commit a1b89132dc4f61071bdeaab92ea958e0953380a1
> >> >> Author: Tim Murray <timmurray@xxxxxxxxxx>
> >> >> Date:   Fri Apr 21 11:11:36 2017 +0200
> >> >>
> >> >>     dm crypt: use WQ_HIGHPRI for the IO and crypt workqueues
> >> >>
> >> >>     Running dm-crypt with workqueues at the standard priority results in IO
> >> >>     competing for CPU time with standard user apps, which can lead to
> >> >>     pipeline bubbles and seriously degraded performance.  Move to using
> >> >>     WQ_HIGHPRI workqueues to protect against that.
> >> >>
> >> >>     Signed-off-by: Tim Murray <timmurray@xxxxxxxxxx>
> >> >>     Signed-off-by: Enric Balletbo i Serra <enric.balletbo@xxxxxxxxxxxxx>
> >> >>     Signed-off-by: Mike Snitzer <snitzer@xxxxxxxxxx>
> >> >>
> >> >> ---
> >> >>
> >> >> Reverting a1b8913 from 4.14.0-rc6, my current kernel, eliminates the
> >> >> problem completely.
> >> >>
> >> >> Looking at the diff in that commit, it looks like the commit message isn't
> >> >> even accurate; not only is the priority of the dmcrypt workqueues being
> >> >> changed - they're also being made "CPU intensive" workqueues as well.
> >> >>
> >> >> This combination appears to result in both elevated scheduling priority and
> >> >> greater quantity of participant worker threads effectively starving any
> >> >> normal priority user task under periods of heavy IO on dmcrypt volumes.
> >> >>
> >> >> I don't know what the right solution is here.  It seems to me we're lacking
> >> >> the appropriate mechanism for charging CPU resources consumed on behalf of
> >> >> user processes in kworker threads to the work-causing process.
> >> >>
> >> >> What effectively happens is my normal `git` user process is able to
> >> >> greatly amplify what share of CPU it takes from the system by generating IO
> >> >> on what happens to be a high-priority CPU-intensive storage volume.
> >> >>
> >> >> It looks potentially complicated to fix properly, but I suspect at its core
> >> >> this may be a fairly longstanding shortcoming of the page cache and its
> >> >> asynchronous design.  Something that has been exacerbated substantially by
> >> >> the introduction of CPU-intensive storage subsystems like dmcrypt.
> >> >>
> >> >> If we imagine the whole stack simplified, where all the IO was being done
> >> >> synchronously in-band, and the dmcrypt kernel code simply ran in the
> >> >> IO-causing process context, it would be getting charged to the calling
> >> >> process and scheduled accordingly.  The resource accounting and scheduling
> >> >> problems all emerge with the page cache, buffered IO, and async background
> >> >> writeback in a pool of unrelated worker threads, etc.  That's how it
> >> >> appears to me anyways...
> >> >>
> >> >> The system used is a X61s Thinkpad 1.8Ghz with 840 EVO SSD, lvm on dmcrypt.
> >> >> The kernel .config is attached in case it's of interest.
> >> >>
> >> >> Thanks,
> >> >> Vito Caputo
> >> >
> >> >
> >> >
> >> > Ping...
> >> >
> >> > Could somebody please at least ACK receiving this so I'm not left wondering
> >> > if my mails to lkml are somehow winding up flagged as spam, thanks!
> >>
> >> Sorry I did not notice your email before you ping me directly. It's
> >> interesting that issue, though we didn't notice this problem. It's a
> >> bit far since I tested this patch but I'll setup the environment again
> >> and do more tests to understand better what is happening.
> >>
> >
> > Any update on this?
> >
> 
> I did not reproduce the issue for now. Can you try what happens if you
> remove the WQ_CPU_INTENSIVE in the kcryptd_io workqueue?
> 
> - cc->io_queue = alloc_workqueue("kcryptd_io", WQ_HIGHPRI |
> WQ_CPU_INTENSIVE | WQ_MEM_RECLAIM, 1);
> cc->io_queue = alloc_workqueue("kcryptd_io", WQ_HIGHPRI | WQ_MEM_RECLAIM, 1);
> 

FYI I also tried just removing WQ_HIGHPRI, retaining WQ_CPU_INTENSIVE,
also bad results.

So far just reverting a1b8913 has been the best solution.

I haven't studied the dmcrypt code, is there reason to observe the
effects of these changes on both the workqueues touched by a1b8913
instead of just kcryptd_io?

Regards,
Vito Caputo