Bug 2549 - S3 resume: blank screen - Pentium-3 ATI Radeon M6 LY
Summary: S3 resume: blank screen - Pentium-3 ATI Radeon M6 LY
Status: REJECTED INSUFFICIENT_DATA
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Fan (show other bugs)
Hardware: i386 Linux
: P2 high
Assignee: Venkatesh Pallipadi
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-04-20 03:25 UTC by Stefan de Bruijn
Modified: 2006-12-03 09:30 UTC (History)
4 users (show)

See Also:
Kernel Version: 2.6.4
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
dmesg, dmidecode, acpidmp, lspci -vv, /proc/interrupts (105.18 KB, text/plain)
2004-04-20 03:27 UTC, Stefan de Bruijn
Details
ASL file from dtst (137.30 KB, text/plain)
2004-04-20 03:29 UTC, Stefan de Bruijn
Details

Description Stefan de Bruijn 2004-04-20 03:25:57 UTC
Distribution: Debian, custom dri-trunk-unstable X driver (that's the only one
that works properly with DRI/DRM); other X drivers give the same problems ('ati'
and 'radeon' drivers)
Hardware Environment: Pentium-3 933, 256 MB ram, ATI Radeon M6 LY, intel 82801-ICH3.
Problem Description (1):
CPU Fan doesn't spin up; ACPI doesn't appear to recognise temperature
measurements (acpi -t returns 'no support for device type: thermal').

Problem Description (2):
When pressing a function key (brightness up, brightness down), the system
freezes. Ssh logins etc also fail. Using a fbdev, other X driver, etc doesn't
resolve the issue.

Problem Description (3):
After a suspend (3), the system works fine, but the screen remains blank.

In that order, I think the severity should be read; a malfunctioning suspend is
not as bad as a cpu fan not responding :)
Comment 1 Stefan de Bruijn 2004-04-20 03:27:31 UTC
Created attachment 2635 [details]
dmesg, dmidecode, acpidmp, lspci -vv, /proc/interrupts
Comment 2 Stefan de Bruijn 2004-04-20 03:29:45 UTC
Created attachment 2636 [details]
ASL file from dtst
Comment 3 Luming Yu 2004-04-27 19:12:24 UTC
1. I didn't find any FAN or ThermalZone definition in your DSDT.
So it is impossible to recognise temperature by ACPI.

2.Please try disabling APIC.

3.What is video driver you are using? 

Hopefully, you can open tickets separately.

Thanks,
Luming
Comment 4 Stefan de Bruijn 2004-04-29 01:27:45 UTC
1. I suppose this is the end of story about this subject then.

2. I've just compiled a kernel without APIC. This _does_ appear to resolve the 
issue. Thanks a lot, I suppose this matter is resolved then.

3. The video driver I'm currently using is the dri-trunk-unstable driver, which 
can be found here:
http://people.debian.org/~daenzer/

I already tried other drivers, as well as the linux native ATI and Radeon 
drivers (from Debian testing), but somehow they don't appear to support DRI. 
Apart from that fact, they too fail after suspend.

So about the tickets - there's only 1 problem left at the moment.
Comment 5 Venkatesh Pallipadi 2004-11-03 19:05:35 UTC
Is the video problem still there with latest 2.6 kernel?

for the video problem, 
- Please try the options in Documentation/power/video.txt
- Try the workaround in bug #3670
Comment 6 Stefan de Bruijn 2004-11-06 05:07:51 UTC
Nope, not working.

/var/log/messages now does show a "resuming tasks", but after that the screen 
shows a white bar and the laptop goes numb. Things like "echo 3 
> /proc/acpi/sleep && reboot" don't work either.

I'm wondering... is there any way to let the kernel generate more debug 
information (what it's doing), so it's easier to trace the bug?
Comment 7 Len Brown 2005-01-03 14:29:42 UTC
still a problem with linux-2.6.10 or later?
Comment 8 Stefan de Bruijn 2005-01-17 00:33:38 UTC
Actually yes.

I tried the complete excercise again yesterday... First thing I did was compile
2.6.10 with as little options as possible (with ACPI enabled that is), putting
in the lilo.conf an append="init=/bin/bash" and tried it. After going to suspend
nicely it failed to go back up. That is: total crash.

Second was to add "append="init=/bin/bash acpi_sleep=s3_bios". Same here- it
crashed when trying to get back up.

After that I tried the  bug as described in
http://bugme.osdl.org/show_bug.cgi?id=3609 which involves:
setpci -s 00:1f.0 f2.b=1

Same here - it still hangs after going out of suspend mode. In all cases the
screen went black and no keyboard response (I typed things like 'halt' and tried
num-lock and so on).

In one of the cases the HD did something after the halt. This was the first case.

Last thing I tried was going to S3 with the complete system in X using the
dri-trunk driver (as described earlier). Here it seemed to do something more, as
the complete screen went garbled at the time of the resume.

I've been wondering if there's any way to let sleep produce some debug
information. It would be nice to know *where* it crashes so we can finally solve
this.
Comment 9 Matija Polajnar 2005-06-02 15:05:51 UTC
I have two computers with radeon card. 
I am using kernel 2.6.6 (old, I know), but I discovered (on my desktop PC with 
AMI bios) that it works if I select the an option in the CMOS: "Repost Video 
on S3 Resume" - bios help states that it "determines, whether to invoke VGA 
BIOS post on S3/STR resume". 
On my laptop (also with radeon, but with Phoenix bios) I have no such option 
and I can't get resume to work. 
Comment 10 Stefan de Bruijn 2005-06-19 13:32:17 UTC
If I had that option I would have tried of course :)

Stefan.

--- bugme-daemon@kernel-bugs.osdl.org wrote:

> http://bugzilla.kernel.org/show_bug.cgi?id=2549
> 
> Matija.Polajnar@guest.arnes.si changed:
> 
>            What    |Removed                    
> |Added
>
----------------------------------------------------------------------------
>                  CC|                           
> |Matija.Polajnar@guest.arnes.
>                    |                            |si
> 
> 
> 
> ------- Additional Comments From
> Matija.Polajnar@guest.arnes.si  2005-06-02 15:05
> -------
> I have two computers with radeon card. 
> I am using kernel 2.6.6 (old, I know), but I
> discovered (on my desktop PC with 
> AMI bios) that it works if I select the an option in
> the CMOS: "Repost Video 
> on S3 Resume" - bios help states that it
> "determines, whether to invoke VGA 
> BIOS post on S3/STR resume". 
> On my laptop (also with radeon, but with Phoenix
> bios) I have no such option 
> and I can't get resume to work. 
> 
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Comment 11 Diego Calleja 2006-09-28 09:10:42 UTC
Does it still happens in recent kernels?
Comment 12 Rafael J. Wysocki 2006-09-28 09:55:39 UTC
The problems with video after an S3 resume can be handled by the s2ram tool
(please see http://en.opensuse.org/S2ram).  If that helps, please close this bug.
Comment 13 Adrian Bunk 2006-12-03 09:30:21 UTC
Please reopen this bug if:
- it is still present with kernel 2.6.19 and
- you can provide the requested information.

Note You need to log in before you can comment on or make changes to this bug.