[coreboot] T60: Waking up from suspend to disk (`s2disk`) sometimes fails with `PM: Image mismatch: …`

[Paul Menzel]

Dear coreboot folks,

before reporting this problem to the Linux hibernation folks
(), could you please confirm that you have
seen this problem or if it is likely to be caused by using coreboot?

DMI: LENOVO ThinkPad T60 / T60p, BIOS 4.0-2807-g4807892 10/04/2012

This is basically the Peter’s patchset »lenovo/t60 lenovo/x60: Override
SMBIOS vendor name to all upper case« [1], which is based on the
following commit.

commit 1ee8b45740a2c888742cb5917ca71cd0ed86cec2
        Author: Siyuan Wang 
        Date:   Fri Sep 7 19:20:02 2012 +0800

add tyan s8226: add a new mainboard

Reviewed-on: http://review.coreboot.org/1495

I reported this issue to the Debian Bug Tracking System (BTS) already,
where it is tracked under the number #703787 [2], and I paste my report
below.

Using Debian you can import the message using `bts show --mbox 703787`.
The program `bts` is packaged in `devscripts`.

---- 8< ----- >8 ----
Dear Debian folks,

running Debian Wheezy/testing on a Lenovo T60 with coreboot and SeaBIOS
(not sure about vendor BIOS) after having done `s2disk` three or four
times, sometimes the previous session is not loaded and it boots
normally. Looking at the Linux messages, I see the following

/var/log/kern.log:Mar 23 02:18:42 myhostname kernel: [   17.878056] PM: Image mismatch: memory size

or

/var/log/kern.log.1:Mar 17 10:33:16 myhostname kernel: [   53.905020] PM: Image mismatch: kernel version

written several times to the log files. Strangely some message seem to
be not written to disk as seen by the missing »I« in »Image«.

Mar 22 17:56:07 myhostname kernel: [73413.739108] PM: Basic memory bitmaps created
    Mar 23 02:18:42 myhostname kernel: imklog 5.8.11, log source = /proc/kmsg started.
    Mar 23 02:18:42 myhostname kernel: mage mismatch: memory size
    Mar 23 02:18:42 myhostname kernel: [   17.878056] PM: Image mismatch: memory size

Do you have any idea how that can happen? I am sure that nothing was
changed about the memory. The second one might be due to an Linux kernel
package upgrade.

Aptitude 0.6.8.2: Protokoll
    So, 17. Mär 2013 00:10:11 +0100

WICHTIG: Dieses Protokoll zeigt nur geplante Aktionen an. Aktionen, die wegen
    dpkg-Problemen fehlschlagen, sind vielleicht nicht abgeschlossen.

86 Pakete werden installiert und 2 Pakete entfernt.
    676 kB der Festplatte werden belegt.
    ===============================================================================
    […]
    [AKTUALISIERUNG] linux-image-3.2.0-4-686-pae:i386 3.2.35-2 -> 3.2.39-2
    [AKTUALISIERUNG] linux-libc-dev:i386 3.2.35-2 -> 3.2.39-2
    […]

So might there be two problems or just one?

Though this should not happen according to a comment in a reply on the
pm-utils list [1].

I am attaching the corresponding `kern.log{,1.}.log` files containing
the logs of Linux from the working suspend to disk cycles and the
non-working ones.

Thanks,

Paul

[1] http://lists.freedesktop.org/archives/pm-utils/2010-April/002117.html
---- 8< ----- >8 ----

Thanks,

Paul

[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=703787
[2] http://review.coreboot.org/#/c/1557/