Announcement

Collapse
No announcement yet.

Linux 3.18-rc6 Released, A Worrisome Regression Remains

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Linux 3.18-rc6 Released, A Worrisome Regression Remains

    Phoronix: Linux 3.18-rc6 Released, A Worrisome Regression Remains

    Linus Torvalds put out his usual weekly release candidate to the Linux 3.18 kernel...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Does anyone keeps track on kernel regressions these days?

    Comment


    • #3
      I do occasionally as I run a fairly current kernel and occasionally run into issues on my laptop.

      That said, I followed the mailing list on this regression and it went awol into other discussions... can't really tell if it's fixed or not as they apparently found a probable cause? I can't tell if the code posted up is related to the probable cause or not. I hate mailing lists...

      Comment


      • #4
        Originally posted by computerquip View Post
        I do occasionally as I run a fairly current kernel and occasionally run into issues on my laptop.
        I'm thinking about person who manages public list of known regressions for each kernel version.

        Comment


        • #5
          Originally posted by michal View Post
          I'm thinking about person who manages public list of known regressions for each kernel version.
          Every commit can introduce regression for somebody There are some regression tests, but quite limited...

          Now imagine thousand of hardware and probably millions of hardware/software combinations, that is mission impossibile to track hardware support clearly without having all these hardware in one huuuuge impossible to have hangaroo

          So most easier advise is to just git bisect if YOU have hit some regression .

          As for that mentioned regression in 3.18 i can't reproduce it OP might have hardware problem too, who knows
          Last edited by dungeon; 24 November 2014, 06:27 AM.

          Comment


          • #6
            Originally posted by dungeon View Post
            Every commit can introduce regression for somebody There are some regression tests, but quite limited...

            Now imagine thousand of hardware and probably millions of hardware/software combinations, that is mission impossibile to track hardware support clearly without having all these hardware in one huuuuge impossible to have hangaroo

            So most easier advise is to just git bisect if YOU have hit some regression .

            As for that mentioned regression in 3.18 i can't reproduce it OP might have hardware problem too, who knows
            You told me nothing new - really - I'm familiar with "linux kernel tester's guide"...

            I asked about known regressions list, because regressions was tracked a few years ago.

            Comment


            • #7
              Originally posted by michal View Post
              You told me nothing new - really - I'm familiar with "linux kernel tester's guide"...

              I asked about known regressions list, because regressions was tracked a few years ago.
              I don't see why bug trackers wouldn't be enough for tracking regressions. Just needs a good bug tracker with sane filtering and categorization capabilities

              Comment


              • #8
                Originally posted by michal View Post
                You told me nothing new - really - I'm familiar with "linux kernel tester's guide"...

                I asked about known regressions list, because regressions was tracked a few years ago.
                Ah OK then... there are *always* some regressions unfixed, don't know who wants to maintain list for that sort of thing If it is regresion it is real normal bug, go bisect it and that is it
                Last edited by dungeon; 24 November 2014, 06:44 AM.

                Comment


                • #9
                  Originally posted by nanonyme View Post
                  I don't see why bug trackers wouldn't be enough for tracking regressions. Just needs a good bug tracker with sane filtering and categorization capabilities
                  For the same reasons that it wouldn't be enough few years ago:
                  - some peoples prefer bugzilla
                  - some peoples prefer sending bug reports to lkml
                  - developers don't read all emails. if they get a short list of bugs they start discussing it.

                  Comment


                  • #10
                    Originally posted by dungeon View Post
                    Ah OK then... there are *always* some regressions unfixed, don't know who wants to maintain list for that sort of thing
                    It was maintained for a few years.

                    Comment

                    Working...
                    X