On 3/10/06, Adam K Kirchhoff <adamk@...> wrote:
> Aapo Tahkola wrote:
> >As long as it gets solved by initializing the card with fglrx, I don't s=
ee any point in investigating.
> At the risk of becoming quite unpopular around here, I have to question
> that reasoning. Saying that there's no point in investigating this bug
> due to previous failed attempts to track it down and a lack of
> documentation from ATI is one thing. However, it essentially sounds
> like you're saying "Why investigate a known bug in our drivers if
> there's an ugly hack involving binary only drivers that works around the
> bug?" What will you tell a user for whom fglrx is not an option?
I don't think Aapo is saying that we shouldn't care but that we have
to find what does fglrx and so investigating this bug by tracing and
looking at what does r300 isn't usefull.
I have spend couple of weeks few month ago looking at that.
I give up because i have been kept busy by my work and i
lost one hd in the process... Now i going back to look at that.
This isn't funny at all and i guess i am not too good at it too.
Anyway if anyone want to help or to try tracking down this too,
he is more than welcome.
In the meantime, if you use lastet Benh patch and set agp
speed to 1 and no fastwrite or other fency options you should
have something a bit more stable at least which doesn't lockup
in the minute...
Btw i have also a ppc thus fglrx is not an option for me and
for other ppc user...