After suspending on a Dell (Latitude) X200 laptop with
kernel 2.6.26, the display is not properly restored on
resume.
quirk-vbe-post and quirk-vbemode-restore are needed (even if
running in terminal mode).
There are two issues:
1) /usr/lib/pm-utils/sleep.d/98smart-kernel-video silently
disables all quirk settings (even explicit command-line
arguments!) if i915 is present on kernel >= 2.6.26.
My workaround was to create
/etc/pm/sleep.d/98smart-kernel-video to override it.
2) The FDI file isn't matching the rules for X200 for my
laptop. There is a rule that would match
system.hardware.product "Latitude X200", but my
system.hardware.product is just "X200".
lshal output:
system.hardware.primary_video.product = 13687 (0x3577) (int)
system.hardware.primary_video.vendor = 32902 (0x8086) (int)
system.hardware.product = 'X200' (string)
system.hardware.vendor = 'Dell Computer Corporation' (string)
system.hardware.version = 'A07' (string)
--- 20-video-quirk-pm-dell.fdi.orig 2008-09-18 11:43:24.000000000 -0700
+++ 20-video-quirk-pm-dell.fdi 2008-09-18 11:45:51.000000000 -0700
@@ -40,6 +40,11 @@
<merge key="power_management.quirk.none" type="bool">true</merge>
</match>
</match>
+ <!-- Dell Latitude X200 -->
+ <match key="system.hardware.product" prefix="X200">
+ <merge key="power_management.quirk.vbemode_restore" type="bool">true</merge>
+ <merge key="power_management.quirk.vbe_post" type="bool">true</merge>
+ </match>
<!-- Dell Inspiron -->
<match key="system.hardware.product" prefix="Inspiron">
Karl Chen wrote: Which graphics driver to you use, which version? A missing fdi rule should be reported against the hal-info package. Just curious, does it also work with --quirk-s3-mode and/or quirk-s3-bios? Those quirks would not be filtered by 98-smart-kernel-video and should be preferred over vbe-* quirks. Cheers, Michael
Michael> Which graphics driver to you use, which version?
It's the i915 driver from the 2.6.26-1-686 kernel.
If you're referring to the X driver, it's "intel" driver version
2.3.2 (xserver-xorg-video-intel 2:2.3.2-2+lenny2). But, I didn't
think the X driver was relevant since the same thing happens in a
text console.
>> 2) The FDI file isn't matching the rules for X200 for my
>> laptop. There is a rule that would match
>> system.hardware.product "Latitude X200", but my
>> system.hardware.product is just "X200".
>>
Michael> A missing fdi rule should be reported against the
Michael> hal-info package.
Michael> Just curious, does it also work with --quirk-s3-mode
Michael> and/or quirk-s3-bios?
Nope, I did try those - it puts the machine into an unrecoverable
state and requires a hard reboot.
For a long time I thought *none* of the quirk options worked since
I had no indication that they were being ignored, when I was
testing with pm-suspend --quirk-.... It was only when exploring
the hook directories that I accidentally discovered the
98smart-kernel-video file. At the very least, it would help if
this behavior were documented.
Thanks,
Karl
Dear Tim, Karl, Jesse, This is to confirm the suspend problems that Karl had on a Dell Latitude X200. Even under 2.6.26 (debian testing) it needs the quirks --quirk-vbe-post --quirk-vbemode-restore, which are now filtered out. But I write less about that (may be fixed in versions newer than used by the frozen debian testing), but more since I noticed that this report led to a long thread on the pm-utils mailing list with in the last entry http://lists.freedesktop.org/archives/pm-utils/2008-October/001747.html Jesse asking for a dump of the video rom. Since this list was likely not read by Karl, I doubt the dump ever arrived. So, I retrieved it on my machine, and loaded it at http://www.astro.utoronto.ca/~mhvk/rom.bin (happy to send this as an attachment, if needed). For completeness, the Dell Latitude X200 uses a i830M chip (not i915 as mentioned in the pm-utils thread), and suspend has been tricky for a long time. See http://sourceforge.net/mailarchive/message.php?msg_name=E1HW1ER-0008GW-00%40swan.astro.utoronto.ca (yes, in a thread on an ASUS, handy...) In particular, as far as I know, suspend-to-ram does not work for a regular vga=0 console (i.e., the simply init=/bin/bash one is always asked to try first), but needs vga=792. Another curious -- possibly completely unrelated -- thing is that the brightness controls normally pop up a little icon in the upper left with a "pie chart" showing the current level. This does not work once agpgart is enabled (one sees something garbled). I've always wondered if that image is just written somewhere in main memory, perhaps overwriting something important, and whether this could explain the random, non-reproduceable crashes I get after changing brightness after a suspend-to-ram. Please ask if you need more details, Thanks, Marten van Kerkwijk
On pre-855 chips, I think you'll still need to POST in order to suspend properly. This is because 830 chips use DVO attached DACs to drive LVDS and other outputs, which the kernel does not yet save/restore properly (this will change once DRM based mode setting is merged). Great, thanks. I'll pull it down and add it to my collection. I definitely need more 830 class ROMs to finish the DRM based mode setting bits (even the current xf86-video-intel doesn't get the info it needs about DVO connections to work everywhere). Hm this sounds like it may be a separate bug. You should probably file it at bugs.freedesktop.org under xf86-video-intel.
Hi, This bug was ubmitted 3 years ago. Apart from tweaking the quirks, I see that xorg developers were trying to fix this issue. Did they succeed? Is this issue still open? Thanks,
Hi, This bug was ubmitted 3 years ago. Apart from tweaking the quirks, I see that xorg developers were trying to fix this issue. Did they succeed? Is this issue still open? Thanks,
There is no tweaking of the quirks necessary. As said they have become obsolete with KMS (kernel mode setting), which the big three (radeon, noveau, intel) support. Quirks are only needed if you disable KMS (which I wouldn't advise to do). Nowadays such bugs such be filed against the kernel.
Dear submitter, as the package pm-utils has just been removed from the Debian archive unstable we hereby close the associated bug reports. We are sorry that we couldn't deal with your issue properly. For details on the removal, please see https://bugs.debian.org/1058701 The version of this package that was in Debian prior to this removal can still be found using https://snapshot.debian.org/. Please note that the changes have been done on the master archive and will not propagate to any mirrors until the next dinstall run at the earliest. This message was generated automatically; if you believe that there is a problem with it please contact the archive administrators by mailing ftpmaster@ftp-master.debian.org. Debian distribution maintenance software pp. Thorsten Alteholz (the ftpmaster behind the curtain)