Announcement

Collapse
No announcement yet.

NVIDIA Updates Legacy Drivers With X.Org Server 1.19 Support

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

  • #11
    Originally posted by Amarildo View Post
    dungeon I do believe agdf5's Kernels are very stable too. I did a 'git diff' between a current grsec patch (at the time) and agdf5's 4.7-staging Kernel, and I was surprised: despite being labled "4.7", agdf5's Kernel is somewhat comparable to linux-git: it was not as "old" as 4.7, it had the latest drivers, AND had a ton of security and other patches backported.

    IMO agdf5's Kernels should be an official branch at Kernel.org.
    Ha, ha, and you are also wrong, that won't happen... even if that is true for you at random point at time, no way that should be recommended

    These are so called internal branches, or post internal or pre pushed to drm to Dave or... these just happen to be open well because it is opensource,

    You might try them of course in case you have some bug in current or if you like to test things early... but shouldn't be recommended to average Joe for no particular reason of course
    Last edited by dungeon; 15 February 2017, 04:00 PM.

    Comment


    • #12
      I'm pretty sure the staging branch is for development purposes only, so they won't be advertising something that could break at any time. Instead, they advertise regular kernel.org branches.

      Comment


      • #13
        Originally posted by smitty3268 View Post

        Only if you ignore the Mesa drivers.
        Problem is some GCN cards, especifically the ones with hardvested hardware , doesn't work with Mesa drivers. They also don't work with AMDGPU-PRO and have flickering and microstuttering problems with fglrx, so even staying in an old XOrg with fglrx is not a solution.

        I really appreciate that AMD is putting so much effort in the open source drivers, but if bugs that blocks entire cards for working never get fixed and the closed source drivers come in "doesn't work at all" and "works bad and only with old packages", you can figure out that the users suffering this problems can see NVidia closed source drivers like something to envy, a warranty your hardware will work in Linux.

        Comment


        • #14
          Originally posted by debianxfce View Post

          When you search with the word amdgpu in the kernel buzilla, almost all problems are solved. Use the latest adg5f wip kernel and make a bug report when you find a new bug. Use the latest Mesa git too.
          The bug was opened more than 4 years ago. In this time multiple users have done all tests asked by AMD developers on mailing list and IRC channel to try to debug the problem, with no success.

          You should humbly ask for more info or do some search yourself before assuming you are the only user in the world that know what is the proper procedure to report a bug.

          Comment


          • #15
            Originally posted by debianxfce View Post

            When you search with the word amdgpu in the kernel buzilla, almost all problems are solved. Use the latest adg5f wip kernel and make a bug report when you find a new bug. Use the latest Mesa git too.
            No adg5f wip kernel nor latest mesa need to be used in this case , madmalkav already participated in a Tahiti LE not working bug opened since 4 years ago, as you can see there that is still not fixed:



            Comment


            • #16
              In case I've been not clear enough in my previous message or the mailing list:

              Lot of hours of work where done in debugging this problem during all 2016 in the IRC channel, and we don't email every single failed test to the mailing list, but only the ones that can provide some interesting output.

              Occasional tests are still done, even when agd5f among others AMD developers have said repeatedly that the problem is not in the kernel side of the driver and no work have been done anywhere that can be near to fix the problem.

              PS: this is not a reply for debianxfce, this info is for whatever people can find useful. I have been reading this forum for years and I was too sleepy when I wrote my reply to realize I was replying to the biggest troll in this forums.

              Comment


              • #17
                Originally posted by debianxfce View Post
                And wrong place for kernel bugs.
                fdo bugzilla is not wrong but recommended place, read this:



                "Modified: [TD]2016-12-14 00:13 UTC
                dmesg for linux 4.9 amdgpu driver"

                Modified last year and old kernel used, not latest 4.11-wip.
                But you are free to read from your ass of course and randomly recommending "cutting edge" kernel as "stable"
                Last edited by dungeon; 18 February 2017, 08:37 AM.

                Comment


                • #18
                  Originally posted by debianxfce View Post

                  The bug that you posted is for RadeonSI. If you have problems, first make the kernel to boot without desktop, then you have a working os. Those people in your bug report do not use latest mesa either. They have either no clue have to make a custom kernel. Amd let people living in the dark and forcing them to buy nvidia cards.
                  You are reading all things wrong and recommending all wrong things further at a wrong time and at the wrong places

                  Comment


                  • #19
                    Moderation ate my last message, probably because how explicitly I stated my feelings about debianxfce and how his crusade to be the smartpants of this forum have pulled me off registering until now. So here we go again...

                    The guy can believe he is smarter than all AMD developers and users that have been making tests. Let him be happy. Meanwhile, for the level-headed users here that can find the information useful:

                    - For the last year, a lot of tests have been conducted in the IRC. Only the ones that reported some interested info where documented in the bugtracker. So you can assume that any until this point every driver available have been tested.
                    - Framebuffer works OK, problems are caused when 3D hardware is initialized. This is clearly documented by mareko in the bugtracker.
                    - agd5f, mareko, and other AMD developers have stated once and again that the bug is not in the kernel side of the driver. Even so, some testing is done every now and them just in case. As I already explained in the first point, as this haven't show any useful info, no update to the bugtracker was done.

                    Comment


                    • #20
                      Originally posted by debianxfce View Post

                      Instead of crying here, nobody of you have tested SI cards with amdgpu driver and adg5f 4.11-wip kernel.
                      What part of "until this point every driver available have been tested" or "agd5f, mareko, and other AMD developers have stated once and again that the bug is not in the kernel side of the driver. Even so, some testing is done every now and them just in case" you have problems to understand?

                      Comment

                      Working...
                      X