Announcement

Collapse
No announcement yet.

Problem with Catalyst 10.5 + Kernel 2.6.34

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

  • Problem with Catalyst 10.5 + Kernel 2.6.34

    Hi guys,

    It's the first time that I've build a custom kernel.

    The installation went fine, but in the end, I'm trying to install the current Catalyst Driver for my HD 5770

    The steps I did:
    - downloading catalyst 10.5 from official site
    - extracting
    - patching with the two patches from http://aur.archlinux.org/packages.php?ID=29111
    - sh ati-install.sh 10.5 --install
    - automatic install

    well .. First Time, I got the following problem:
    The UTS Release version in include/linux/version.h
    ""
    does not match current version:
    "2.6.17"
    After some time googleing I found the solution --> #define UTS_RELEAS "2.6.34"

    And now I'm stuck at the next problem

    :/lib/modules/fglrx# ./make_install.sh
    - recreating module dependency list
    - trying a sample load of the kernel modules
    FATAL: Error inserting fglrx (/lib/modules/2.6.34/kernel/drivers/char/drm/fglrx.ko): Invalid module format
    failed.


    modprobe fglrx shows me the same problem

    modinfo fglrx shows the following lines:
    filename: /lib/modules/2.6.34/kernel/drivers/char/drm/fglrx.ko
    license: Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY
    description: ATI Fire GL
    author: Fire GL - ATI Research GmbH, Germany
    alias: pci:v00001002d000095CFsv*sd*bc*sc*i* [...] <-- a lot of different "alias" lines
    depends:
    vermagic: 2.6.34 SMP mod_unload
    parm: firegl:charp
    Can somebody help me ? ..

    I also tried the installer http://kanotix.com/files/install-fglrx-debian.sh

    But all the time, he downloaded the "drms" stuff from the kanotix site and after it, he wants to download the linux-header 2.6.22-2 with apt

    So I gave up and hope to get some more tips from you

    thanks a lot

    greets,
    Phil

  • #2
    My script is correct just you installed 2.6.34 the completely wrong (manual) way. When you use Debian/Ubuntu use at least kernel-package -> make-kpkg but NEVER without a deb package.

    Comment


    • #3
      mhm .. ok thx 4 info ..

      i installed the kernel in two steps

      make bzImage modules modules_install install ARCH=x86_64
      make header_install

      thought it's enough

      Comment


      • #4
        I used to compile my own kernels months ago, but now i always use the distro kernels. Maybe it is not the best way but this is what i was using.

        1) download the kernel source , uncompress it and go to the directory
        2) su
        3) chmod -R a-s $PWD
        4) CONCURRENCY_LEVEL="your number of cores" make-kpkg --append-to-version=-custom kernel_image kernel_headers --initrd binary
        5) cd..
        6) dpkg -i linux-image-2.6.2*.deb linux-headers-2.6.2*.deb
        7) update-initramfs -u

        Comment


        • #5
          thx jimbo for the manual
          thx kano for the script and help

          With your help, I solved the prob and the fglrx mod is running

          But there is a strange behavior left:

          Since I installed the fglrx driver, after every boot, there is a "kernel oops" message telling me that there is a kernel failure ..

          I searched in the syslog and found some call trace after loading fglrx:
          May 29 15:12:04 WorkDebian kernel: [ 15.223225] [fglrx] IRQ 16 Enabled
          May 29 15:12:04 WorkDebian kernel: [ 15.272109] caller is kas_spin_lock+0x15/0xbe [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272115] Pid: 2698, comm: Xorg Tainted: P 2.6.34-custom #1
          May 29 15:12:04 WorkDebian kernel: [ 15.272118] Call Trace:
          May 29 15:12:04 WorkDebian kernel: [ 15.272127] [<ffffffff81257ba5>] ? debug_smp_processor_id+0xb9/0xd0
          May 29 15:12:04 WorkDebian kernel: [ 15.272168] [<ffffffffa00fa89a>] ? kas_spin_lock+0x15/0xbe [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272209] [<ffffffffa00fac5e>] ? KAS_InterlockedList_RemoveAtHead+0x58/0x15c [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272217] [<ffffffff810c3d7c>] ? __inc_zone_state+0x16/0x7d
          May 29 15:12:04 WorkDebian kernel: [ 15.272271] [<ffffffffa011319d>] ? MCIL_InterlockedList+0x10d/0x2a0 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272342] [<ffffffffa01bbbe0>] ? FIFO_Queue_RemoveFromHead+0x40/0x60 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272408] [<ffffffffa01c7371>] ? AllocateCallbackEntry+0x21/0x50 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272479] [<ffffffffa01bcad7>] ? RegisterIRQClient+0xb7/0x200 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272548] [<ffffffffa01bf1c5>] ? IsR6XXDerivatives+0x55/0x70 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272611] [<ffffffffa01c971e>] ? Is_IRQSource_Valid+0x8e/0xb0 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272684] [<ffffffffa01ba00e>] ? IRQMGR_Access+0xee/0x190 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272745] [<ffffffffa012c2d0>] ? fireglAsyncioIntHandler+0x0/0x70 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272807] [<ffffffffa012c105>] ? fireglAsyncioIntEnableMsgHandler+0x105/0x1e0 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.272946] [<ffffffff810c9114>] ? handle_mm_fault+0x4b7/0x91b
          May 29 15:12:04 WorkDebian kernel: [ 15.273007] [<ffffffffa012c2d0>] ? fireglAsyncioIntHandler+0x0/0x70 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.273094] [<ffffffffa015c32b>] ? getNextMessage+0xab/0x140 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.273178] [<ffffffffa015c224>] ? asyncIOHandler+0x74/0xd0 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.273243] [<ffffffffa012b428>] ? firegl_asyncio_write+0x198/0x260 [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.273290] [<ffffffffa00fe16c>] ? ip_firegl_write+0x4b/0x7d [fglrx]
          May 29 15:12:04 WorkDebian kernel: [ 15.273300] [<ffffffff810e8d74>] ? vfs_write+0xad/0x169
          May 29 15:12:04 WorkDebian kernel: [ 15.273309] [<ffffffff810e8eec>] ? sys_write+0x45/0x6e
          May 29 15:12:04 WorkDebian kernel: [ 15.273319] [<ffffffff81009a82>] ? system_call_fastpath+0x16/0x1b
          Is this a bug or is this a pebkac (Problem exists between keyboard and chair) ?

          greetz

          Comment

          Working...
          X