* fglrx + radeonfb leads to crashes?
@ 2004-03-17 20:18 David Mosberger
2004-03-18 0:26 ` Benjamin Herrenschmidt
0 siblings, 1 reply; 6+ messages in thread
From: David Mosberger @ 2004-03-17 20:18 UTC (permalink / raw)
To: ajoshi; +Cc: linux-fbdev-devel
Hi Ani,
I have a machine here with an ATI FireGL Z1 card. It works fine with
the ATI fglrx binary-only driver on 2.6.5-rc1. However, if radeonfb
is enabled and I try to stop the X server or switch away from it to
another virtual console, then the screen display gets all messed up
and the machine gets stuck. When this hapens, I periodically see this
kind of message:
radeonfb: FIFO Timeout !
Now, this is on ia64 machine so at this point I'm just wondering
whether this problem also shows up on other machines. If it's an
ia64-specific problem, I can look into it. Otherwise, I guess I'll
just disable radeonfp for now.
Thanks,
--david
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: fglrx + radeonfb leads to crashes?
2004-03-17 20:18 fglrx + radeonfb leads to crashes? David Mosberger
@ 2004-03-18 0:26 ` Benjamin Herrenschmidt
2004-03-18 0:47 ` David Mosberger
2004-03-18 1:23 ` Michel Dänzer
0 siblings, 2 replies; 6+ messages in thread
From: Benjamin Herrenschmidt @ 2004-03-18 0:26 UTC (permalink / raw)
To: davidm; +Cc: ajoshi, Linux Fbdev development list
On Thu, 2004-03-18 at 07:18, David Mosberger wrote:
> Hi Ani,
>
> I have a machine here with an ATI FireGL Z1 card. It works fine with
> the ATI fglrx binary-only driver on 2.6.5-rc1. However, if radeonfb
> is enabled and I try to stop the X server or switch away from it to
> another virtual console, then the screen display gets all messed up
> and the machine gets stuck. When this hapens, I periodically see this
> kind of message:
>
> radeonfb: FIFO Timeout !
>
> Now, this is on ia64 machine so at this point I'm just wondering
> whether this problem also shows up on other machines. If it's an
> ia64-specific problem, I can look into it. Otherwise, I guess I'll
> just disable radeonfp for now.
This is a known problem. I'm not sure at this point what ATI binary
driver is doing, I think it messes up the card internal memory map,
among others, and it definitely doesn't restore things properly when
beeing switched out.
I only have PPC machines so I can't experiment with it myself though.
The "fix" might be as simple as restoring MC_FB_LOCATION,
MC_AGP_LOCATION and the various offset when re-initing the mode (the
latest fbdev will trigger a mode restore & engine reset when switching
out from a KD_GRAPHICS console).
Ben.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: fglrx + radeonfb leads to crashes?
2004-03-18 0:26 ` Benjamin Herrenschmidt
@ 2004-03-18 0:47 ` David Mosberger
2004-03-18 0:52 ` Benjamin Herrenschmidt
2004-03-18 1:23 ` Michel Dänzer
1 sibling, 1 reply; 6+ messages in thread
From: David Mosberger @ 2004-03-18 0:47 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: davidm, ajoshi, Linux Fbdev development list
>>>>> On Thu, 18 Mar 2004 11:26:31 +1100, Benjamin Herrenschmidt <benh@kernel.crashing.org> said:
Benjamin> On Thu, 2004-03-18 at 07:18, David Mosberger wrote:
>> I have a machine here with an ATI FireGL Z1 card. It works fine
>> with the ATI fglrx binary-only driver on 2.6.5-rc1. However, if
>> radeonfb is enabled and I try to stop the X server or switch away
>> from it to another virtual console, then the screen display gets
>> all messed up and the machine gets stuck.
Benjamin> This is a known problem.
OK, that's good to know.
Benjamin> I only have PPC machines so I can't experiment with it
Benjamin> myself though. The "fix" might be as simple as restoring
Benjamin> MC_FB_LOCATION, MC_AGP_LOCATION and the various offset
Benjamin> when re-initing the mode (the latest fbdev will trigger a
Benjamin> mode restore & engine reset when switching out from a
Benjamin> KD_GRAPHICS console).
OK. If somebody had a patch to do these things, I could try it out
here. The problem isn't really critical for me, so I'll probably just
live without radeonfb for now.
Thanks,
--david
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: fglrx + radeonfb leads to crashes?
2004-03-18 0:47 ` David Mosberger
@ 2004-03-18 0:52 ` Benjamin Herrenschmidt
0 siblings, 0 replies; 6+ messages in thread
From: Benjamin Herrenschmidt @ 2004-03-18 0:52 UTC (permalink / raw)
To: davidm; +Cc: ajoshi, Linux Fbdev development list
> OK. If somebody had a patch to do these things, I could try it out
> here. The problem isn't really critical for me, so I'll probably just
> live without radeonfb for now.
I'm a bit too busy at the moment, I'll try to come up with something
one of these days. Ping me if you don't hear back from me in a couple
of weeks :)
Ben.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: fglrx + radeonfb leads to crashes?
2004-03-18 1:23 ` Michel Dänzer
@ 2004-03-18 1:15 ` Benjamin Herrenschmidt
0 siblings, 0 replies; 6+ messages in thread
From: Benjamin Herrenschmidt @ 2004-03-18 1:15 UTC (permalink / raw)
To: Michel Dänzer; +Cc: davidm, ajoshi, Linux Fbdev development list
> Note that I don't have problems switching between radeonfb and an X
> server from DRI CVS which sets up the card memory layout in the
> recommended way and never restores it. Am I missing something?
I'm not completely sure what ATI binary does...
Ben.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: fglrx + radeonfb leads to crashes?
2004-03-18 0:26 ` Benjamin Herrenschmidt
2004-03-18 0:47 ` David Mosberger
@ 2004-03-18 1:23 ` Michel Dänzer
2004-03-18 1:15 ` Benjamin Herrenschmidt
1 sibling, 1 reply; 6+ messages in thread
From: Michel Dänzer @ 2004-03-18 1:23 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: davidm, ajoshi, Linux Fbdev development list
On Thu, 2004-03-18 at 01:26, Benjamin Herrenschmidt wrote:
> On Thu, 2004-03-18 at 07:18, David Mosberger wrote:
> >
> > I have a machine here with an ATI FireGL Z1 card. It works fine with
> > the ATI fglrx binary-only driver on 2.6.5-rc1. However, if radeonfb
> > is enabled and I try to stop the X server or switch away from it to
> > another virtual console, then the screen display gets all messed up
> > and the machine gets stuck. When this hapens, I periodically see this
> > kind of message:
> >
> > radeonfb: FIFO Timeout !
> >
> > Now, this is on ia64 machine so at this point I'm just wondering
> > whether this problem also shows up on other machines. If it's an
> > ia64-specific problem, I can look into it. Otherwise, I guess I'll
> > just disable radeonfp for now.
>
> This is a known problem. I'm not sure at this point what ATI binary
> driver is doing, I think it messes up the card internal memory map,
> among others, and it definitely doesn't restore things properly when
> beeing switched out.
>
> I only have PPC machines so I can't experiment with it myself though.
> The "fix" might be as simple as restoring MC_FB_LOCATION,
> MC_AGP_LOCATION and the various offset when re-initing the mode (the
> latest fbdev will trigger a mode restore & engine reset when switching
> out from a KD_GRAPHICS console).
Note that I don't have problems switching between radeonfb and an X
server from DRI CVS which sets up the card memory layout in the
recommended way and never restores it. Am I missing something?
--
Earthling Michel Dänzer | Debian (powerpc), X and DRI developer
Libre software enthusiast | http://svcs.affero.net/rm.php?r=daenzer
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-03-18 1:30 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-03-17 20:18 fglrx + radeonfb leads to crashes? David Mosberger
2004-03-18 0:26 ` Benjamin Herrenschmidt
2004-03-18 0:47 ` David Mosberger
2004-03-18 0:52 ` Benjamin Herrenschmidt
2004-03-18 1:23 ` Michel Dänzer
2004-03-18 1:15 ` Benjamin Herrenschmidt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).