Re: Re: ATI Radeon PCI 9200SE no DRI

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, 19 Jan 2004, Fred Welland wrote:

>Yep I tried the "ForcePCIMode" "true" option - and various and sundry other combinations of options in my device section either X works with no hardware acceleation or X hangs as described before.  

Please set your mailer to wrap long lines at 70 columns or so, so 
it is readable.  Your text goes way off my screen, so I can only 
respond to what I can read, and make assumptions from the rest.

It'd be nice if you'd trim excessive list headers/footers and 
other non-essentials in quoted replies also.  ;o)

>So ForcePCIMode  yeilds the same results as the driver the mharris pointed to in the aforementioned link.  And this is the case w/o ForcePCIMode and w/o mharris's tweaked driver.

ForcePCIMode, and the AGP autodetection logic in 4.3.0, and also
in current 4.4.0 are totally broken crap.  I have written a
Radeon driver specific quick hack to for once and for all
*properly* autodetect AGP/PCI hardware by walking the PCI
capabilities list in PCI config space.  The patch is in my 4.3.0
src.rpm in fedora development.  Anyone using stock 4.3.0 or
XFree86 CVS, or any Red Hat builds prior to my patch (check the
spec file changelog for what version/date it went in on), will
have a possibility of misdetection.  In some cases, PCI devices
will be treated as AGP, in other cases AGP devices will be
treated like PCI.  While the latter should work, due to PCI
support not being 100% stable, it can result in AGP Radeon
hardware being totally unuseable.

>Also, noticed the 'ForcePCIMode' bug in XFree's buggzilla and the BusType is the way to go - 'cept that must be a XFree 4.4 thing, because the log file just mentions that it is ignoring it as unknown..

That override might go away once I submit my patch to fix AGP 
autodetection.


>I was hoping to tweak/tune/config this without having to recompile anything....that doesn't look promising...

Use rawhide XFree86 and it should "just work".  If not, submit a 
bug report and I'll investigate, but so far all known problem 
cases have went away on x86, AMD64, ppc, etc..

Hope this helps.


-- 
Mike A. Harris


_______________________________________________
xfree86-list mailing list
xfree86-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/xfree86-list
IRC: #xfree86 on irc.redhat.com

[Red Hat General]     [Red Hat Watch]     [Red Hat Development]     [Kernel Development]     [Yosemite Camping]

  Powered by Linux