Announcement

Collapse
No announcement yet.

Intel Mesa Gives Problems With KDE's KWin, Again

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

  • Intel Mesa Gives Problems With KDE's KWin, Again

    Phoronix: Intel Mesa Gives Problems With KDE's KWin, Again

    Martin Grlin, the lead KDE developer of the KWin compositing window manager, usually has fairly insightful and technical blog posts. Last week he was talking about possibly moving the KDE screensaver into the KWin compositor for KDE SC 4.8 after writing the KDE view on GNOME's new compositing manager. Today he has written a new post, but this time it's about the open-source Intel Mesa driver breaking (again) for KWin...

    http://www.phoronix.com/vr.php?view=OTM0NA

  • #2
    Instead of talking to driver developers like Wine and Gnome people do and resolving issues together, he prefers writing ranty blog posts where he describes how bloody hacks in KWin are driver developers' fault. Typical coward.

    Comment


    • #3
      From his post:

      Still I could talk to them now instead of writing angry blog posts? Sadly I cannot. First of all we were notified about the change too late to do anything. As just explained we were notified after 4.6.2 and no chance to adjust before Natty. Furthermore I was ill last week, are sitting in a plane to the US right now, so I can write blog posts but cannot use Internet and I will stay more or less disconnected throughout the next week.

      Comment


      • #4
        Originally posted by marek View Post
        Instead of talking to driver developers like Wine and Gnome people do and resolving issues together, he prefers writing ranty blog posts where he describes how bloody hacks in KWin are driver developers' fault. Typical coward.
        Oh right, so now it's the kwin developers' fault that the free graphics driver developers fucked up once again?

        Comment


        • #5
          I'm really on Intel's side here. Both changes are improvements, having the 20090101 date in there forever isn't really informative.

          Why the heck do they parse direct rendering from the name anyway, it would take five minutes to look at how glxinfo checks for it.

          Comment


          • #6
            I think that Marek was less concerned with passing the blame and more about the approach to resolving problems.

            I don't think that checking the renderer string is the correct thing to do, but I also don't know any other way.

            It's also quite amazing that the Intel devs broke one of the most important linux apps out there in a point release without even checking whether it works. The solution is to downgrade the driver and use the old one until the issue is resolved, and that's quite pointless.

            A simple regression test would have given everyone enough time to find the problem and resolve it.

            Comment


            • #7
              For the record, there appears to be a nice function called glXIsDirect, which is what glxinfo uses.

              Comment


              • #8
                Software rasteriser uses direct rendering, so checking that is completely pointless.

                What KWin needs to do is whether GL is hardware-accelerated, and there is no way to check that. Even in forums when people debug driver issues, people look at the renderer string.

                Martin is a volunteer and has my thanks for all his hard work. He also rants a lot, which is sometimes unfortunate. But the sad fact is that Intel released a driver which broke KWin, one of the most important applications in Linux. They either

                - released a driver without even testing whether it works, or
                - released a driver although they knew it would break the desktop for many users, and didn't tell anyone.

                Either way, it's very very unfortunate.

                Comment


                • #9
                  Hm, right. In that case, how about the driver name (as output by xdriinfo)? Parsing that ought to be easier, with only three (AFAIK) software drivers.

                  Comment


                  • #10
                    I don't know how well that works, but KWin needs OpenGL 2 functionality for some things, so even hardware-accelerated drivers using old codepaths don't work -- that's why they're checking for DRI2. Only Intel does not list DRI2 in the renderer string either, so they tested for GEM, which got removed.

                    Grepping the renderer string is an old evil which broke many things in the past, even 10 years ago, but there is simply no good way to test what the driver actually supports today, and that's a problem.

                    With KDE 4.5, KWin made changes which broke things on Mesa drivers and were (rightly) criticised for not testing with open drivers. They said that they trusted the drivers about the functionality they advertised, but didn't implement.

                    This time, Mesa devs made changes which broke KWin, and KWin guys did not even know about it until it hit Ubuntu repositories.

                    Comment


                    • #11
                      intel driver devs fuck up all the time. Just look up Linus' rants whenever they drop a load of crap into the kernel.

                      They could also learn a lot from the kernel: never change behaviour expected by the userspace.

                      Such a change in a point release is more than stupid. Once again intel's devs fucked up.

                      Comment


                      • #12
                        Originally posted by AnonymousCoward View Post
                        Oh right, so now it's the kwin developers' fault that the free graphics driver developers fucked up once again?
                        They didn't fuck up at all. The Intel devs only changed the renderer string. It's a valid change, a normal one, it may happen from time to time and apps shouldn't rely on it. On the other hand, detecting KMS/GEM/DRI2 via some string in OpenGL is clearly a hack. There are better ways of detecting those than some random string that happened to contain "GEM" by luck. (I am surprised too that someone parses that string.)

                        Anyway, the fact the driver devs haven't tested it is, of course, their fault. But the real problem lies in the Kwin code, Martin has even admitted that with the code snippet in his blog post. It's pretty clear that hacks like that would cause breakage sooner or later.

                        Comment


                        • #13
                          Marek, how would you test correctly if a driver provides appropriate functionality?

                          It is known since the last KWin fiasco that you can't trust the advertised functionality, and checking for direct rendering is also a hack which doesn't even work anymore. Blacklists caused their own set of problems, and if there is no check, KWin will crash all the time, like 4.5 did.

                          What is the correct way to check this?

                          Comment


                          • #14
                            Originally posted by marek View Post
                            There are better ways of detecting those than some random string that happened to contain "GEM" by luck.
                            Such as...?

                            Comment


                            • #15
                              Kwin do it again

                              BTW. AMD/ATI is not mentioned because no one from KDE test there Stuff on AMD/ATI and if only on an old Code Base.

                              Comment

                              Working...
                              X