Re: MTRR problem, maybe FB related

From: Dave Jones (davej_at_redhat.com)
Date: 08/06/04

  • Next message: William Lee Irwin III: "Re: 2.6.8-rc3-mm1: PROC_FS=n link errors"
    Date:	Fri, 6 Aug 2004 21:34:36 +0100
    To: "admin@wodkahexe.de" <admin@wodkahexe.de>
    
    

    On Fri, Aug 06, 2004 at 07:47:22PM +0200, admin@wodkahexe.de wrote:

    > vesafb: framebuffer at 0xb0000000, mapped to 0xdf80d000, size 6144k
    > vesafb: mode is 1024x768x32, linelength=4096, pages=4

    vesafb's mtrr usage is borken. Instead of creating an MTRR the size
    of video RAM, it creates one the size of the display.

    > mtrr: 0xb0000000,0x8000000 overlaps existing 0xb0000000,0x400000
    > [drm] Initialized i830 1.3.2 20021108 on minor 0: Intel Corp. 82852/855GM Integrated Graphics Device
    > mtrr: 0xb0000000,0x8000000 overlaps existing 0xb0000000,0x400000
    > [drm] Initialized i830 1.3.2 20021108 on minor 1: Intel Corp. 82852/855GM Integrated Graphics Device (#2)

    Then X comes along, sizes the video ram, and tries to create an MTRR
    of the correct size, but the framebuffer got there first and bodged it.

    I used to see this happening also on my Matrox g550, but it seems
    to have 'gone away' in recent times. I haven't checked out why,
    but I'm suspecting X now detects this case, and deletes the crap
    entry, and puts the proper values in its place.

    > when starting X i'm getting the following in dmesg:
    >
    > mtrr: base(0xb0020000) is not aligned on a size(0x180000) boundary

    This one I can't explain however.

    > mtrr: 0xb0000000,0x8000000 overlaps existing 0xb0000000,0x400000
    >
    > is there any way to get both working together? (fb + mtrr)

    Disable MTRR for vesafb. iirc, there's a boot command line option to do it.

                    Dave

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/


  • Next message: William Lee Irwin III: "Re: 2.6.8-rc3-mm1: PROC_FS=n link errors"

    Relevant Pages

    • Re: MTRR problem, maybe FB related
      ... > vesafb's mtrr usage is borken. ... Instead of creating an MTRR the size ... > Then X comes along, sizes the video ram, and tries to create an MTRR ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: [PATCH] add AMD Geode processor support
      ... > What about some of the other ones like MTRR and IOAPIC? ... > I was kinda passing this along from someone I thought knew ... ...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl! ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • mtrr question
      ... /proc/mtrr is empty. ... Does/is setting up mtrr per the old 1999 Docs/mtrr.txt still relevant ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: 4GB memory and Intel Dual-Core system
      ... Subject: x86_64/i386: Compute correct MTRR mask on early Noconas ... Detail info can be found at: ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: Fixing MTRR smp breakage and suspending sysdevs.
      ... I understand that they should be set with IRQ ... I extended it yesterday to do the MTRR settings as well, ... Christian Reformed Church of Tuggeranong ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)