> On Jun 10, 2014, at 12:01 AM, Jonathan Gray <j...@jsg.id.au> wrote:
>
>> On Mon, Jun 09, 2014 at 04:53:23AM -0500, Stan Gammons wrote:
>>> On 06/09/2014 01:56 AM, Jonathan Gray wrote:
>>>> On Sun, Jun 08, 2014 at 07:01:56AM -0500, Stan Gammons wrote:
>>>> Is there a known problem with the Radeon driver on OpenBSD 5.5 AMD64
>>>> stable? I noticed the text on the console was scrolling very slow
>>>> while src.tar.gz was extracting. I didn't time it, but it seem to take
>>>> 2 to 3 times as long to return to the command line compared to
>>>> extracting the same file on a slower i386 machine. Syslog output is
>>>> below.
>>> Well things are going to be slow when all the acceleration is being
>>> disabled. It isn't clear if the problem you're encountering is
>>> due to something like missing firmware as you didn't include a dmesg.
>>>
>>>
>>
>> It loads or tries to load firmware at boot.
>>
>> Here's the dmesg info.
>
> So it seems something isn't working right with the GART.
> Could you try the following diff to force it to a smaller size?
>
> Diff against -current but will likely apply to 5.5
>
> Index: sys/dev/pci/drm/radeon/rs400.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/pci/drm/radeon/rs400.c,v
> retrieving revision 1.4
> diff -u -p -r1.4 rs400.c
> --- sys/dev/pci/drm/radeon/rs400.c 9 Feb 2014 12:33:44 -0000 1.4
> +++ sys/dev/pci/drm/radeon/rs400.c 10 Jun 2014 04:57:21 -0000
> @@ -54,6 +54,8 @@ void rs400_gart_adjust_size(struct radeo
> rdev->mc.gtt_size = 32 * 1024 * 1024;
> return;
> }
> + DRM_ERROR("Forcing to 32M GART size\n");
> + rdev->mc.gtt_size = 32 * 1024 * 1024;
> }
>
> void rs400_gart_tlb_flush(struct radeon_device *rdev)
I'll try the diff later today.
Do you keep getting duplicate emails on this subject? I keep getting
duplicates on this as well as other emails. Something weird is going on here...
Stan
