Bug 3821 - won't start on ppc64 with a Radeon 9600
Summary: won't start on ppc64 with a Radeon 9600
Status: CLOSED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: git
Hardware: PowerPC Linux (All)
: high major
Assignee: Xorg Project Team
QA Contact:
URL: http://bugs.gentoo.org/show_bug.cgi?i...
Whiteboard:
Keywords:
Depends on:
Blocks: 1690
  Show dependency treegraph
 
Reported: 2005-07-20 05:21 UTC by o590694
Modified: 2005-11-20 06:55 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments

Description o590694 2005-07-20 05:21:17 UTC
Hi,

all snapshots after xorg-x11-6.8.99.1 won't start on my G5 running Gentoo/PPC64.
Won't start means that the display turns black, the fans begin to spin and the
system hard locks. This a 23" Apple Cinema Display connected to a Radeon 9600
Mac Edition. I have tracked that prolem down to a patch which was applied to
radeon_driver.c after 6.8.99.1 was released.

The patch added this two lines:

OUTREG(info->DDCReg, INREG(info->DDCReg) &
       ~(RADEON_GPIO_EN_0 | RADEON_GPIO_EN_1));

Removing this two lines solves the startup problem.

Regards,

Markus Rothe
Comment 1 o590694 2005-07-20 05:24:23 UTC
Created attachment 3109 [details]
xorg-6.8.99.14 single works

This patch turns radeon_driver.c into the old state, which makes it useable
again on my G5.
Comment 2 Benjamin Herrenschmidt 2005-07-22 01:58:19 UTC
Hrm... It is definitely necessary to release the DDC lines after probe, I don't
see how doing so would crash the system, that makes little sense at this point.
Without this patch, a whole bunch of monitors will turn off (like most modern
apple cinema displays). Back then, ATI confirmed that this was the right thing
to do.

What might be happening is that the wrong DDC line is used, that this line is
routed to something else than a DDC output GPIOs and we are somewhat crashing
whatever is connected to that i2c bus (TMDS transmitter ?) but that would be
really weird, the chip has other i2c busses for that sort of thing.

Are you 100% certain that just changing those two lines actually causes the
entire machine to lock up ? Make very sure no other change is happening.
Comment 3 o590694 2005-07-25 00:31:46 UTC
I have double checked that this patch works and found out that I am wrong. The
problem is that X starts without error if you use fbdev. I haven't had a config
in /etc/X11/ and it seems like Xorg defaulted to fbdev.

I'm trying to debug this a little bit more, but as I don't get any error and gdb
(over ssh) does not give a hint, too, my only idea to debug this is to recompile
X step by step with the patches from 6.8.99.1 to 6.8.99.2 and try to find the
one, which broke ppc64.

BTW: this problem is not limited to my Cinema Display connected with ADC. I've
also reproduced this problem with an LCD connected to the DVI port of my Radeon.
Comment 4 o590694 2005-08-10 19:52:19 UTC
I was wrong. The breakage didn't happened from .1 to .2. I figured out that .1
was just starting because I haven't had a xorg.conf in /etc/X11/ and the auto
configuration seems to have built a configuration which uses fbdev.

So I have no clue what is going on here. Could someone please give me a hint how
to debug this?
Comment 5 o590694 2005-08-10 19:55:37 UTC
this is a configuration which let's every xorg-x11 6.8.99.x start on ppc64:
http://bugs.gentoo.org/attachment.cgi?id=60196

As you can see it uses Driver "fbdev". using Driver "radeon" makes xorg-x11
crashing.
Comment 6 Michel Dänzer 2005-11-08 06:35:06 UTC
Can you verify whether a 32 bit X server shows the same problem?
Comment 7 Benjamin Herrenschmidt 2005-11-14 09:53:00 UTC
I'm running 32 bits X servers on G5 regulary without problem. Paulus has been
running 64 bits ones too. There was a bug that broke them but I think it was
fixed. I'll ask him the details, it could have been a kernel issue.
Comment 8 Yang Dehua 2005-11-14 16:12:39 UTC
I just tried the xorg-6.8.99.15-r4 in gentoo portage on my G5 running linux 2.6.
14, and it's worked well till now.
Comment 9 o590694 2005-11-21 01:55:38 UTC
As benh said: seems to be fixed. 6.8.99.15 starts as normal. 
Comment 10 o590694 2005-11-21 01:55:50 UTC
closing 


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.