Bug 28911 - [Arrandale] Booting machine with external monitor attached to laptop make machine crash
Summary: [Arrandale] Booting machine with external monitor attached to laptop make mac...
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Wang Zhenyu
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-07-05 00:42 UTC by Keyzer Suze
Modified: 2010-08-15 19:56 UTC (History)
4 users (show)

See Also:
i915 platform:
i915 features:


Attachments
excerpt from syslog (7.92 KB, text/plain)
2010-07-05 00:43 UTC, Keyzer Suze
no flags Details
Experimental patch (1.45 KB, patch)
2010-07-12 03:42 UTC, Chris Wilson
no flags Details | Splinter Review
dmesg output (123.28 KB, text/plain)
2010-08-05 07:13 UTC, Corona
no flags Details
intel_dp.c downloaed yesterday and edited (no luck) (44.32 KB, text/plain)
2010-08-06 04:18 UTC, lipido
no flags Details
intel_dp.c downloaed 5 aug and edited again (no luck) (44.33 KB, text/plain)
2010-08-07 02:17 UTC, lipido
no flags Details

Description Keyzer Suze 2010-07-05 00:42:39 UTC
Hi

I have a HP 2540p laptop, I have installed debian amd64 (sid).  I am running 2.6.34-1-amd64 (experimental.2).

I have rebuild the i915 kernal module using the patches from #28070 - to allow me to boot up my machine and have a graphics display.

My problem is now when I have a second monitor attach the machine crashes... magic key doesn't work.

Hi

I tried this, against a debian 2.6.34-1 experimental.2 build, works as well,
but if I start X with an external monitor attached the machine hangs....

I can attach after X has started.

So  (with i915 blacklisted)

1.a boot
1.b modprobe i915 modeset=1
1.c startx
1.d attached ext display
1.e xrandr add second display
works

2.a boot
2.b modprobe i915 modeset=1
2.c attach ext monitor
2.d startx
dead have to hold the  power button for a while

I have attached the relevant section from syslog


Alex
Comment 1 Keyzer Suze 2010-07-05 00:43:09 UTC
Created attachment 36758 [details]
excerpt from syslog
Comment 2 Chris Wilson 2010-07-05 10:36:47 UTC
The prime suspect would appear to be:

while ((I915_READ(transconf_reg) & TRANS_STATE_ENABLE) == 0)
    ;
Comment 3 lipido 2010-07-11 10:53:41 UTC
I can confirm. I have a Sony Vaio VPCZ1 on kubuntu 10.04. I updated the kernel to 2.6.34.

To reproduce (almost exactly the same as reported before):

1. Attach the external display before boot (I tried with two different displays).
2. Boot with kernel options (#goes to single text-based mode, KMS disabled): ro single i915.modeset=0 
3. rmmod i915
4. modprobe i915 modeset=1
FREEZE (BEFORE startx)

The freeze seems not to be a kernel panic (no blinking keyboard lights, no messages on netconsole). But it is a total freeze.
Comment 4 Chris Wilson 2010-07-12 03:42:47 UTC
Created attachment 36958 [details] [review]
Experimental patch
Comment 5 Keyzer Suze 2010-07-13 02:08:10 UTC
patch doesn't match up with my source from 2.6.34
Comment 6 Sami Lehtinen 2010-07-15 01:38:44 UTC
I confirm that this bug affets also HP 8440p with Intel GMA HD.

X startup without external monitor connected --> OK
X startup with external monitor connected --> Freeze
Comment 7 Corona 2010-08-05 05:02:21 UTC
For me, this issue was solved in the latest drm-intel-next kernel. I had to
revert a patch which caused a regression giving me a blank screen on boot (see
https://bugzilla.kernel.org/show_bug.cgi?id=16496), but now the VGA output
works without problems.
Comment 8 Chris Wilson 2010-08-05 05:33:07 UTC
Corona, mind attaching a dmesg from boot with drm.debug=0xe just to confirm that we are not simply erroring out?
Comment 9 Corona 2010-08-05 07:13:24 UTC
Created attachment 37587 [details]
dmesg output
Comment 10 Chris Wilson 2010-08-05 08:32:10 UTC
Thanks Corona, it's definitely behaving itself from dmesg.

Keyzer can you confirm that your issue is resolved on the upstream drm-core-next + the brown paper bag patch?
Comment 11 Keyzer Suze 2010-08-06 03:00:20 UTC
(In reply to comment #10)
> Thanks Corona, it's definitely behaving itself from dmesg.
> 
> Keyzer can you confirm that your issue is resolved on the upstream
> drm-core-next + the brown paper bag patch?

Hi

Not sure which patch to apply to what.

I tried the patch in this thread but it didn't apply to my kernel tree.

Keyzer
Comment 12 lipido 2010-08-06 03:10:11 UTC
I use kubuntu. I installed several (daily compiled) ubuntu kernels from http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/, using the "drm-intel-next", but I had no luck.

With the kernel 2.6.35 built on 2010-08-02 I can't attach the external monitor on boot (the problem of this bug). Without the external monitor, it works.

However, with the kernel 2.6.35 built on 2010-08-04 I get a total black screen allways (having or not the external monitor plugged).
Comment 13 lipido 2010-08-06 03:27:38 UTC
Sorry, the dates were wrong. These are correct:
2.6.35 2010-07-09: boot without external monitor
2.6.35 2010-08-02: no boot (with or without external monitor)
2.6.35 2010-08-04: no boot (with or without external monitor)


(In reply to comment #12)
> I use kubuntu. I installed several (daily compiled) ubuntu kernels from
> http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/, using the
> "drm-intel-next", but I had no luck.
> 
> With the kernel 2.6.35 built on 2010-08-02 I can't attach the external monitor
> on boot (the problem of this bug). Without the external monitor, it works.
> 
> However, with the kernel 2.6.35 built on 2010-08-04 I get a total black screen
> allways (having or not the external monitor plugged).
Comment 14 Corona 2010-08-06 03:39:58 UTC
(In reply to comment #11)
> (In reply to comment #10)
> > Thanks Corona, it's definitely behaving itself from dmesg.
> > 
> > Keyzer can you confirm that your issue is resolved on the upstream
> > drm-core-next + the brown paper bag patch?
> 
> Hi
> 
> Not sure which patch to apply to what.
> 
> I tried the patch in this thread but it didn't apply to my kernel tree.
> 
> Keyzer

You need to download the drm-intel-next tree and revert the patch mentioned here: https://bugzilla.kernel.org/show_bug.cgi?id=16496 . It's only one line so it's actually more convenient if you do it by hand. If you don't revert the patch you will get the results lipido is reporting (i.e. black screen on boot with or without external monitor attached).
Comment 15 lipido 2010-08-06 04:18:04 UTC
Created attachment 37628 [details]
intel_dp.c downloaed yesterday and edited (no luck)

Yesterday, I downloaded the drm-intel-next head from git, and edited the intel_dp.c (attached). Corona, is this the change you are saying? (line 866)

I compiled the kernel with:
make oldconfig
make
sudo make modules_install install
mkinitramfs -o /boot/initrd.img-2.6.35 2.6.35

However, I had no luck, total screen black.
Comment 16 Corona 2010-08-06 13:34:21 UTC
No, you need to change line 866 to this:

				ironlake_edp_backlight_off(dev);

So you have the same line twice (865 and 866)

(In reply to comment #15)
> Created an attachment (id=37628) [details]
> intel_dp.c downloaed yesterday and edited (no luck)
> 
> Yesterday, I downloaded the drm-intel-next head from git, and edited the
> intel_dp.c (attached). Corona, is this the change you are saying? (line 866)
> 
> I compiled the kernel with:
> make oldconfig
> make
> sudo make modules_install install
> mkinitramfs -o /boot/initrd.img-2.6.35 2.6.35
> 
> However, I had no luck, total screen black.
Comment 17 lipido 2010-08-07 02:17:14 UTC
Created attachment 37658 [details]
intel_dp.c downloaed 5 aug and edited again (no luck)

No luck again. I think that now I edited the line in the right way, but no luck: black laptop screen (strange!)

I have also attached the new intel_dp.c
Comment 18 Corona 2010-08-07 06:07:14 UTC
(In reply to comment #17)
> Created an attachment (id=37658) [details]
> intel_dp.c downloaed 5 aug and edited again (no luck)
> 
> No luck again. I think that now I edited the line in the right way, but no
> luck: black laptop screen (strange!)
> 
> I have also attached the new intel_dp.c

weird. have you tried to remove the comment? it shouldn't make a difference, but you never know. do you have an intel core i5 cpu? otherwise your problem might be caused by something else. in any case, i am not a specialist in this, so i think you'd better open a separate bug about your issue.
Comment 19 Keyzer Suze 2010-08-07 19:49:53 UTC
> 
> You need to download the drm-intel-next tree and revert the patch mentioned
can you point me to this, I am not familiar with git

Thanks
Comment 20 Keyzer Suze 2010-08-07 20:12:41 UTC
(In reply to comment #19)
> > 
> > You need to download the drm-intel-next tree and revert the patch mentioned
> can you point me to this, I am not familiar with git
> 
> Thanks

Sorry found it, took a snapshot from the git/web interface will try that
Comment 21 lipido 2010-08-09 09:04:46 UTC
In summary (in my case), with the latest kernel next snapshot from the 5 Aug (4 days ago):

1. Laptop screen black, but it boots in background. If I attach an external monitor after boot, I can see everything OK, so I can check dmesg: There is a debug message (booting with drm.debug=0x06) saying "LVDS is not present in VBT". This issue seems to be introduced between 2010-07-09 and 2010-08-02.

2. Booting with external monitor attached -> total freeze, so using the drm-intel-next seems not to fix this issue in my laptop. (vaio VPCZ1).

Any ideas?
Comment 22 Keyzer Suze 2010-08-15 13:59:09 UTC
[snip]
> > 
> > Keyzer
> 
> You need to download the drm-intel-next tree and revert the patch mentioned
> here: https://bugzilla.kernel.org/show_bug.cgi?id=16496 . It's only one line so
> it's actually more convenient if you do it by hand. If you don't revert the
> patch you will get the results lipido is reporting (i.e. black screen on boot
> with or without external monitor attached).

sorry for the delay, I have downloaded the latest 2.6.35 kernel from debian and it seems to have the fix in place, I can boot now to multiple monitors attached.


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.