Debian Bug report logs - #426548
BIOS timing issues cause hard freeze with i855 chips

Toggle useless messages

Message #3 received at submit@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: acpi events cause hard freeze on X40 with l-m-t installed

Date: Tue, 29 May 2007 16:09:01 +0200

[Message part 1 (text/plain, inline)]

Package: laptop-mode-tools
Version: 1.32-1
Severity: important

With l-m-t installed, I can almost reliably reproduce hard freezes
of my IBM Thinkpad X40 (2371-H9G) when I close the display or
unplug/plug the power supply. I've also seen freezes when the
display was blanked or during ususpend/resume.

If I purge l-m-t, the freezes seem unreproducible. If I reinstall
l-m-t, they return, even if I do not make any changes to the
configuration. In fact, I never touched the configuration yet.

Thanks to Reinhard Tartler for the tip. I am still not 100% sure of
the cause but wanted to document this.

-- System Information:
Debian Release: lenny/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.20-1-amd64 (SMP w/1 CPU core)
Locale: LANG=en_GB, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[signature.asc (application/pgp-signature, inline)]

Message #8 received at 426548@bugs.debian.org (full text, mbox, reply):

From: Reinhard Tartler <siretart@tauware.de>

To: martin f krafft <madduck@debian.org>

Cc: 426548@bugs.debian.org

Subject: Re: Bug#426548: acpi events cause hard freeze on X40 with l-m-t installed

Date: Tue, 29 May 2007 17:10:44 +0200

[Message part 1 (text/plain, inline)]

martin f krafft <madduck@debian.org> writes:

> With l-m-t installed, I can almost reliably reproduce hard freezes
> of my IBM Thinkpad X40 (2371-H9G) when I close the display or
> unplug/plug the power supply. I've also seen freezes when the
> display was blanked or during ususpend/resume.
>
> If I purge l-m-t, the freezes seem unreproducible. If I reinstall
> l-m-t, they return, even if I do not make any changes to the
> configuration. In fact, I never touched the configuration yet.
>
> Thanks to Reinhard Tartler for the tip. I am still not 100% sure of
> the cause but wanted to document this.

I think this is known in ubuntu as
https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.15/+bug/12483

-- 
Gruesse/greetings,
Reinhard Tartler, KeyID 945348A4

[Message part 2 (application/pgp-signature, inline)]

Message #11 received at 426548@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: 426548@bugs.debian.org

Subject: Re: Bug#426548: acpi events cause hard freeze on X40

Date: Tue, 29 May 2007 17:38:00 +0200

[Message part 1 (text/plain, inline)]

reassign 426548 kernel
retitle 426548 acpi events cause hard freeze on X40
thanks

also sprach Reinhard Tartler <siretart@tauware.de> [2007.05.29.1710 +0200]:
> I think this is known in ubuntu as
> https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.15/+bug/12483

I have just experienced a freeze without l-m-t installed or
laptop_mode enabled.

I also did not have madwifi drivers loaded nor was I running on
battery.

Thus I think this is a kernel bug.

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[signature.asc (application/pgp-signature, inline)]

Message #20 received at 426548@bugs.debian.org (full text, mbox, reply):

From: Reinhard Tartler <siretart@tauware.de>

To: martin f krafft <madduck@debian.org>

Cc: 426548@bugs.debian.org

Subject: Re: Bug#426548: acpi events cause hard freeze on X40

Date: Wed, 30 May 2007 09:57:37 +0200

[Message part 1 (text/plain, inline)]

martin f krafft <madduck@debian.org> writes:

> reassign 426548 kernel
> retitle 426548 acpi events cause hard freeze on X40
> thanks

What makes you think this was because of ACPI events? I fail to see any
evidence that this problem is related to ACPI at all.

It could be as well because of a broken implementation of the IDE
driver. But this is speculative as well.

-- 
Gruesse/greetings,
Reinhard Tartler, KeyID 945348A4

[Message part 2 (application/pgp-signature, inline)]

Message #23 received at 426548@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Reinhard Tartler <siretart@tauware.de>

Cc: 426548@bugs.debian.org

Subject: Re: Bug#426548: acpi events cause hard freeze on X40

Date: Wed, 30 May 2007 11:23:49 +0200

[Message part 1 (text/plain, inline)]

also sprach Reinhard Tartler <siretart@tauware.de> [2007.05.30.0957 +0200]:
> What makes you think this was because of ACPI events? I fail to see any
> evidence that this problem is related to ACPI at all.
> 
> It could be as well because of a broken implementation of the IDE
> driver. But this is speculative as well.

I see it happen whenever I close the display or unplug/plug the
power adapter, sometimes when I switch video display to the VGA
port, and even sometimes just read()ing /proc/acpi/ibm/video or when
DPMS blanks the screen. These are all ACPI-related and have nothing
to do with the IDE driver.

In fact, I was about to get Lenovo to replace the graphics
chip/display combo because that could also be a source. It's
certainly video-related here.

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[signature.asc (application/pgp-signature, inline)]

Message #26 received at 426548@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: 426548@bugs.debian.org

Subject: bug in i810 X driver, fixed upstream

Date: Wed, 27 Jun 2007 11:58:59 +0200

[Message part 1 (text/plain, inline)]

reassign 426548 xserver-xorg-video-intel
retitle 426548 BIOS timing issues cause hard freeze with i855 chips
tags 426548 fixed-upstream patch
thanks

I believe this issue had to do with a timing issue, which Keith
Packard fixed as follows:

diff --git a/src/i830_display.c b/src/i830_display.c
index 6965337..ce20864 100644
--- a/src/i830_display.c
+++ b/src/i830_display.c
@@ -361,8 +361,10 @@ i830FindBestPLL(xf86CrtcPtr crtc, int target, int refclk, intel_clock_t *best_cl
 void
 i830WaitForVblank(ScrnInfoPtr pScreen)
 {
-    /* Wait for 20ms, i.e. one cycle at 50hz. */
-    usleep(20000);
+    /* Wait for 30ms, i.e. one cycle at slightly less than 50hz. 
+     * THIS PATCHED BY KEITH
+     */
+    usleep(30000);
 }
 
 void


Also, commit fbbb41bc5e03478cb46ee8f64ef68b23ff3fc14b may be related:

Author: Keith Packard <keithp@neko.keithp.com>
Date:   Sun Jun 17 14:59:24 2007 +0100

    Let DPMS functions enable plane/pipe/output on 8xx hardware.
    
    On 855, letting crtc_mode_set enable the plane and pipe will occasionally
    hang the chip. Instead, wait for crtc_enable to light things up. For 9xx,
    leave things alone.

diff --git a/src/i830_display.c b/src/i830_display.c
index adc7479..6965337 100644
--- a/src/i830_display.c
+++ b/src/i830_display.c
@@ -958,11 +958,17 @@ i830_crtc_mode_set(xf86CrtcPtr crtc, DisplayModePtr mode,
  else
      pipeconf &= ~PIPEACONF_DOUBLE_WIDE;
     }
-#if 1
-    dspcntr |= DISPLAY_PLANE_ENABLE;
-    pipeconf |= PIPEACONF_ENABLE;
-    dpll |= DPLL_VCO_ENABLE;
-#endif
+    /*
+     * This "shouldn't" be needed as the dpms on code
+     * will be run after the mode is set. On 9xx, it helps.
+     * On 855, it can lock up the chip (and the entire machine)
+     */
+    if (IS_I9XX (pI830))
+    {
+ dspcntr |= DISPLAY_PLANE_ENABLE;
+ pipeconf |= PIPEACONF_ENABLE;
+ dpll |= DPLL_VCO_ENABLE;
+    }
     
     /* Disable the panel fitter if it was on our pipe */
     if (i830_panel_fitter_pipe (pI830) == pipe)


-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[signature.asc (application/pgp-signature, inline)]

Message #37 received at 426548@bugs.debian.org (full text, mbox, reply):

From: Raúl Sánchez Siles <rasasi78@gmail.com>

To: 426548@bugs.debian.org, 426548-submitter@bugs.debian.org

Subject: Possibly still present upstream

Date: Wed, 4 Jul 2007 00:55:05 +0200

[Message part 1 (text/plain, inline)]

  I've read the bug report and it seems that it corresponds to the upstream 
bug https://bugs.freedesktop.org/show_bug.cgi?id=11432 which is still open.

  This upstream bug is reproducible in my system using July 2nd git version. 
My system: Dell inspiron 510m, Intel 855GM and unstable. The problem in case 
it exists should be easily repeatable just pressing the lid close button in 
the middle of an Xorg session. In my case it even happens on the kdm logon 
screen.

  Could you please if this stills happen to you using the experimental version 
of the driver?

  Thanks.

-- 
     Raúl Sánchez Siles
----->Proud Debian user<-----
Linux registered user #416098

[signature.asc (application/pgp-signature, inline)]

Message #43 received at 426548@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Raúl Sánchez Siles <rasasi78@gmail.com>, 426548@bugs.debian.org

Cc: keithp@freedesktop.org

Subject: Re: Bug#426548: Possibly still present upstream

Date: Wed, 4 Jul 2007 13:04:24 +0200

[Message part 1 (text/plain, inline)]

also sprach Raúl Sánchez Siles <rasasi78@gmail.com> [2007.07.04.0055 +0200]:
>   I've read the bug report and it seems that it corresponds to the upstream 
> bug https://bugs.freedesktop.org/show_bug.cgi?id=11432 which is still open.
> 
>   This upstream bug is reproducible in my system using July 2nd git version. 
> My system: Dell inspiron 510m, Intel 855GM and unstable. The problem in case 
> it exists should be easily repeatable just pressing the lid close button in 
> the middle of an Xorg session. In my case it even happens on the kdm logon 
> screen.
> 
>   Could you please if this stills happen to you using the experimental version 
> of the driver?

I found a patch that is not yet upstream. Does this fix it?

From 1442fe79f2add91ba6df2417ad75243ed65952c7 Mon Sep 17 00:00:00 2001
From: martin f. krafft <madduck@lapse.madduck.net>
Date: Mon, 18 Jun 2007 13:35:18 +0100
Subject: [PATCH] only run dpms on code on 9xx

---
 src/i830_display.c |   16 +++++++++++-----
 1 files changed, 11 insertions(+), 5 deletions(-)

diff --git a/src/i830_display.c b/src/i830_display.c
index adc7479..6965337 100644
--- a/src/i830_display.c
+++ b/src/i830_display.c
@@ -958,11 +958,17 @@ i830_crtc_mode_set(xf86CrtcPtr crtc, DisplayModePtr mode,
  else
      pipeconf &= ~PIPEACONF_DOUBLE_WIDE;
     }
-#if 1
-    dspcntr |= DISPLAY_PLANE_ENABLE;
-    pipeconf |= PIPEACONF_ENABLE;
-    dpll |= DPLL_VCO_ENABLE;
-#endif
+    /*
+     * This "shouldn't" be needed as the dpms on code
+     * will be run after the mode is set. On 9xx, it helps.
+     * On 855, it can lock up the chip (and the entire machine)
+     */
+    if (IS_I9XX (pI830))
+    {
+ dspcntr |= DISPLAY_PLANE_ENABLE;
+ pipeconf |= PIPEACONF_ENABLE;
+ dpll |= DPLL_VCO_ENABLE;
+    }
     
     /* Disable the panel fitter if it was on our pipe */
     if (i830_panel_fitter_pipe (pI830) == pipe)
-- 
1.5.2.2

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[signature.asc (application/pgp-signature, inline)]

Message #48 received at 426548@bugs.debian.org (full text, mbox, reply):

From: Raúl Sánchez Siles <rasasi78@gmail.com>

To: 426548@bugs.debian.org

Subject: Patch already applied upstream.

Date: Thu, 5 Jul 2007 01:01:29 +0200

[Message part 1 (text/plain, inline)]

  Hello Martin:

  The patch you mention is already applied upstream as of June 17th, so 
included in my up-to-date local git repo. This means bug is still present in 
my experience.

  Thanks,

-- 
     Raúl Sánchez Siles
----->Proud Debian user<-----
Linux registered user #416098

[signature.asc (application/pgp-signature, inline)]

Message #53 received at 426548@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Brice Goglin <Brice.Goglin@ens-lyon.org>

Cc: 426548@bugs.debian.org, Keith Packard <keithp@freedesktop.org>

Subject: Re: Bug#426548: Possibly still present upstream

Date: Thu, 5 Jul 2007 09:08:22 +0200

[Message part 1 (text/plain, inline)]

also sprach Brice Goglin <Brice.Goglin@ens-lyon.org> [2007.07.05.0033 +0200]:
> I don't know where this patch comes from and whether keithp was aware of
> it before your mail, but if it doesn't get applied upstream soon, we
> should put it in the upstream bug so that keithp and/or anholt see it again.

Keith wrote the patch. I have already written to him to ask him what
to do with it.

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems
 
a qui sait comprendre, peu de mots suffisent.
                                                 -- intelligenti pauca

[signature.asc (application/pgp-signature, inline)]

Message #58 received at 426548@bugs.debian.org (full text, mbox, reply):

From: Keith Packard <keithp@keithp.com>

To: martin f krafft <madduck@debian.org>

Cc: keithp@keithp.com, Brice Goglin <Brice.Goglin@ens-lyon.org>, 426548@bugs.debian.org, Keith Packard <keithp@freedesktop.org>

Subject: Re: Bug#426548: Possibly still present upstream

Date: Thu, 05 Jul 2007 00:50:19 -0700

[Message part 1 (text/plain, inline)]

On Thu, 2007-07-05 at 09:08 +0200, martin f krafft wrote:
> also sprach Brice Goglin <Brice.Goglin@ens-lyon.org> [2007.07.05.0033 +0200]:
> > I don't know where this patch comes from and whether keithp was aware of
> > it before your mail, but if it doesn't get applied upstream soon, we
> > should put it in the upstream bug so that keithp and/or anholt see it again.
> 
> Keith wrote the patch. I have already written to him to ask him what
> to do with it.

Both of those changes (the WaitForVblank and crtc_mode_set portions) are
upstream and should have shipped in driver version 2.1.0

-- 
keith.packard@intel.com

[signature.asc (application/pgp-signature, inline)]

Message #61 received at 426548@bugs.debian.org (full text, mbox, reply):

From: martin f krafft <madduck@debian.org>

To: Keith Packard <keithp@keithp.com>

Cc: Brice Goglin <Brice.Goglin@ens-lyon.org>, 426548@bugs.debian.org

Subject: Re: Bug#426548: Possibly still present upstream

Date: Thu, 5 Jul 2007 09:55:39 +0200

[Message part 1 (text/plain, inline)]

also sprach Keith Packard <keithp@keithp.com> [2007.07.05.0950 +0200]:
> Both of those changes (the WaitForVblank and crtc_mode_set portions) are
> upstream and should have shipped in driver version 2.1.0

Ah, yes. The patch I sent to the bug report yesterday is indeed
upstream:

  http://gitweb.freedesktop.org/?p=xorg/driver/xf86-video-intel.git;a=commitdiff;h=ec236c76b93aea5f2ee1e8b8509cde4625974fcb

I am sorry, I got majorly confused.

-- 
 .''`.   martin f. krafft <madduck@debian.org>
: :'  :  proud Debian developer, author, administrator, and user
`. `'`   http://people.debian.org/~madduck - http://debiansystem.info
  `-  Debian - when you have better things to do than fixing systems

[signature.asc (application/pgp-signature, inline)]

Message #72 received at 426548-done@bugs.debian.org (full text, mbox, reply):

From: Brice Goglin <Brice.Goglin@ens-lyon.org>

To: 426548-done@bugs.debian.org

Subject: Bug#426548: BIOS timing issues cause hard freeze with i855 chips

Date: Wed, 06 Feb 2008 06:56:26 +0100

Version: 2:2.2.0.90-1

This should be fixed in the new 2.2.0.90 in unstable, according to the
upstream bug report status.

Brice

Send a report that this bug log contains spam.