Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

On 01/04/2011 02:40 PM, Jiri Slaby wrote:

On 12/24/2010 01:58 AM, akpm@linux-foundation.org wrote:

quoted

The mm-of-the-moment snapshot 2010-12-23-16-58 has been uploaded to

Hi, this kernel regresses with respect to suspend to ram in comparison
with mmotm 2010-12-16-14-56.

This is OK:
echo devices > /sys/power/pm_test
pm-suspend
This hangs at suspend phase:
echo platform > /sys/power/pm_test
pm-suspend

Note that this kernel is based on next-20101221. Should I try newer (and
clean) -next?

Ok, bisected down to:
16dc39c98a6ca56a27f22f7ac6731d8223237a2e is first bad commit
commit 16dc39c98a6ca56a27f22f7ac6731d8223237a2e
Author: Len Brown [off-list ref]
Date:   Thu Dec 16 23:12:23 2010 -0500

    ACPI: use ioremap_cache()

    Although the temporary boot-time ACPI table mappings
    were set up with CPU caching enabled, the permanent table
    mappings and AML run-time region memory accesses were
    set up with ioremap(), which on x86 is a synonym for
    ioremap_nocache().

    Changing this to ioremap_cache() improves performance as
    seen when accessing the tables via acpidump,
    or /sys/firmware/acpi/tables.  It should also improve
    AML run-time performance.

    No change on ia64.

    Reported-by: Jack Steiner [off-list ref]
    Signed-off-by: Len Brown [off-list ref]

:040000 040000 be35c5e8f214f10f94688c1a27f33ecfb8505220
52581222d0edf190f160f3e5aa5d2c1af8e76988 M      arch
:040000 040000 ccdca0d41938b8312e946cde3c01c59b32d1c17c
96ccf2357f2ac4a31d19cc41f5728d9f87b6cac0 M      drivers

Revert of that patch fixes the problem.

regards,
-- 
js