Tuesday 4 October 2016

MuQSS - The Multiple Queue Skiplist Scheduler v0.105

I spent the last few days fighting with various lock debugging techniques and the numerous bug reports and am pleased to announce a new version of MuQSS, version 0.105

There are versions and incrementals available for linux-4.7:
 http://ck.kolivas.org/patches/muqss/4.0/4.7/
and linux-4.8:
http://ck.kolivas.org/patches/muqss/4.0/4.8

If you've been waiting for me to say it's stable enough to try, then now's your chance for I've addressed all known bugs at this time and it's working well for me.

Most of the issues were to do with races and unstable handling of cross-cpu task movement. No effort went into improving performance from 104 though this version should address many of the crashes and hangs that have been reported with earlier versions.

Additionally there is a pending patch being uploaded for BFS512 which, as per usual, had some last minute issues that only just showed up. If enough users complain loudly enough or more issues show up I might just release another bfs and -ck since it should be stable, especially being one of the last BFS releases.

http://ck.kolivas.org/patches/bfs/4.0/4.8/Pending/

Keep the feedback and bug reports coming. Next I need to put more care into the non-interactive mode of muqss for your enjoyment.

Enjoy!
お楽しみ下さい
-ck

12 comments:

  1. Hey,

    is it also possible to integrate the latest fixes for BFS512 to linux 4.7?

    Thank You

    ReplyDelete
    Replies
    1. Sure, I've added to the pending/ directory for 4.7. There are now 5 patches for 4.7 bfs502.

      Delete
    2. He/she most likely wants the actual fixes on top of BFS512.
      Maybe, its worth to try:
      http://ck.kolivas.org/patches/bfs/4.0/4.8/Pending/bfs512-fixes.patch
      even if they're announced as kernel 4.8 related. So, YMMV applies.
      My suggestion: Try MuQSS instead!

      BR, Manuel Krause

      Delete
  2. Con, thank you for your bug-fixing efforts!
    The 105 version, still kernel 4.7.6, solves two issues I've had with 104 +first 2 pending patches: failing to reboot properly and failing to suspend to disk.
    Both are solved now and overall behaviour is in a very good shape. :-)
    Heavy and very good work, "chapeau!" :-)))

    BR, Manuel Krause

    ReplyDelete
  3. I must not be applying the patch to 4.8 correctly:
    https://gist.github.com/anonymous/ede811fe36320eb1bf3f603b721a6143

    ReplyDelete
    Replies
    1. Although I trust your workflow control, can it be that you used an incremental BFS512-to-MuQSS patch without the prerequisite? I cannot verify from here, what may be wrong (no 4.8 on here yet).
      Maybe you'd try the full patch instead:
      http://ck.kolivas.org/patches/muqss/4.0/4.8/4.8-sched-MuQSS_105.patch

      BR, Manuel Krause

      Delete
    2. I probably just diffed them wrong. In any case all it is is a diff from bfs512 to muqss 105. Nothing is left from the bfs.

      Delete
    3. 4.8-sched-MuQSS_105.patch applies cleanly but bfs512-muqss105.patch does not apply to either the 4.8 code or the 4.8-sched-MuQSS_105 patched code.

      Delete
    4. I've reuploaded a fresh diff from bfs512 to muqss 105.

      Delete
    5. Does the newly uploaded bfs512-muqss105.patch apply to the 4.8 for you? I am still experiencing errors.

      Delete
  4. Thanks for the fixes Con.
    However I still have errors related to xfs processes in the log with linux 4.8 + bfs512 + bfs512-fixes.
    http://pastebin.com/5jfSK9qY
    It seems to happen randomly when I stress the cpu. It doesn't happen on cfs, but maybe it's because the test duration is only 2-3hours.
    I'm testing muqss105 now.

    Pedro

    ReplyDelete
    Replies
    1. Ok I've finished testing muqss105. The results are here as usual:
      https://docs.google.com/spreadsheets/d/1ZfXUfcP2fBpQA6LLb-DP6xyDgPdFYZMwJdE0SQ6y3Xg/edit?usp=sharing

      Nothing new to see, it was more to put some stress on the system to see if it triggers errors.
      And indeed I got an error related to an xfs process:
      http://pastebin.com/mQpC1dMx
      I can't reproduce it though, and I'm not shure it's muqss related. I will wait for linux 4.8.1 and try linux 4.7 + muqss106 in the meantime.

      Pedro

      Delete