This message has been processed by Brightmail(r) AntiVirus using
Symantec's AntiVirus Technology.
Unknown00000000.data was not scanned for viruses because too many nested levels of files were found.
For more information on antivirus tips and technology, visit
http://www.brightmail.com/antivirus .
Len Brown <[email protected]> wrote:
>
> Hi Linus, please pull from the release branch here:
>
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6.git release
There are a few bugs which I'd identified as arising from the acpi tree
while it was in -mm. Is this patch likely to drag them into mainline?
They include:
http://bugzilla.kernel.org/show_bug.cgi?id=4977
Summary: ACPI 20050708 fails on HP RX2600 platform
http://bugzilla.kernel.org/show_bug.cgi?id=4867
Summary: bug in ACPI crashes machine when reading
/proc/acpi/thermal_zone/THRM/temperature
http://bugzilla.kernel.org/show_bug.cgi?id=4980
Summary: krash on entering mem sleep
Plus we have all the battery monitor woes, but they're in 2.6.13 already.
Thanks.
>There are a few bugs which I'd identified as arising from the acpi tree
>while it was in -mm. Is this patch likely to drag them into mainline?
>
>They include:
>
>
>http://bugzilla.kernel.org/show_bug.cgi?id=4977
> Summary: ACPI 20050708 fails on HP RX2600 platform
This was filed against ACPICA 20050708
which had a known problem with module-level code.
The ACPI patch now contains ACPICA 20050902 which
fixes that issue, and this system needs to be
re-tested with the latest patch.
>http://bugzilla.kernel.org/show_bug.cgi?id=4867
> Summary: bug in ACPI crashes machine when reading
> /proc/acpi/thermal_zone/THRM/temperature
UNREPRODUCIBLE.
test system died and is no longer available.
>http://bugzilla.kernel.org/show_bug.cgi?id=4980
> Summary: krash on entering mem sleep
The submitter confirmed that suspend to memory now works on this box.
The remaining issue on this box is related to the EC and battery,
and we're getting contradictory feedback on it.
Frankly, I think we need broader testing, and
pushing the latest ec code into 2.6.14 now is the best
way to get that. If it turns out to be a mistake
we can always turn back time and revert
drivers/acpi/ec.c to the one that shipped in 2.6.12 --
but the one in the latest patch has proven to be
superior to 2.6.13 on other systems.
>Plus we have all the battery monitor woes, but they're in
>2.6.13 already.
Re: 2.6.13 regressions vs 2.6.12, I'm aware of these:
http://bugzilla.kernel.org/show_bug.cgi?id=5165
smp c-states on Pentium 4 with hyperthreading causes big slow-down
http://bugzilla.kernel.org/show_bug.cgi?id=5171
2.6.13 SMP kernel crash on boot at pm_idle_save()
I saw lots of transient battery issues from 2.6.13-rc3
until 2.6.13-rc6, but the ones I followed went away
as of 2.6.13 final. Do you have your eye on others
besides 4980?
thanks,
-Len
"Brown, Len" <[email protected]> wrote:
>
> I saw lots of transient battery issues from 2.6.13-rc3
> until 2.6.13-rc6, but the ones I followed went away
> as of 2.6.13 final. Do you have your eye on others
> besides 4980?
Not specifically, but then ACPI bugs are the one sort which I don't track.
a) because there are so many and b) because the ACPI team use bugzilla
well.
Sticking "battery" into the bugzilla Summary field turns up a few.
<vague>There seem to have been four or five reports in recent weeks.
>"Brown, Len" <[email protected]> wrote:
>>
>> I saw lots of transient battery issues from 2.6.13-rc3
>> until 2.6.13-rc6, but the ones I followed went away
>> as of 2.6.13 final. Do you have your eye on others
>> besides 4980?
>
>Not specifically, but then ACPI bugs are the one sort which I
>don't track.
>a) because there are so many and b) because the ACPI team use bugzilla
>well.
In the last 5 weeks we've reduced our unresolved bug count
to 160 from 196 -- even as 50 new sightings rolled in:
9/8/05: 739 Resolved 160 Unresolved 899 Total
8/31/05: 733 Resolved 161 Unresolved 894 Total
8/24/05: 721 Resolved 166 Unresolved 887 Total
8/17/05: 694 Resolved 181 Unresolved 875 Total
8/10/05: 666 Resolved 194 Unresolved 860 Total
8/3/05: 653 Resolved 196 Unresolved 849 Total
>Sticking "battery" into the bugzilla Summary field turns up a few.
><vague>There seem to have been four or five reports in recent weeks.
<specific:->
We have a bugzilla category for battery issues.
There are 12 open, 4 of them resolved -- 3 of those
fixes are included in the proposed patch:
http://bugzilla.kernel.org/show_bug.cgi?id=4892
through the ec_burst=1 fix is still optional:
http://bugzilla.kernel.org/show_bug.cgi?id=3851
http://bugzilla.kernel.org/show_bug.cgi?id=3974
I think it makes sense to proceed to get broader
testing on the latest code. We're not getting
new failure reports from -mm.
Indeed, I'd like to try enabling the new ec_burst=1
by default in -mm. It is not perfect, but it works
for me, so it should do much better than the 2.6.13-rc3 attempt.
We'll keep ec_burst=0 as the default in Linus' tree
for now. The systems that require ec_burst=1 will
have to supply it manually for now, which is better
than not having it all per 2.6.12.
thanks,
-Len
On Thursday 08 September 2005 1:34 am, Andrew Morton wrote:
> Len Brown <[email protected]> wrote:
> >
> > Hi Linus, please pull from the release branch here:
> >
> > rsync://rsync.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6.git release
>
> There are a few bugs which I'd identified as arising from the acpi tree
> while it was in -mm. Is this patch likely to drag them into mainline?
>
> They include:
>
>
> http://bugzilla.kernel.org/show_bug.cgi?id=4977
> Summary: ACPI 20050708 fails on HP RX2600 platform
I tested 2.6.13 + the acpi-20050902 patch, and it works fine, so
I think this bug can be closed. I pinged the submitter to do so.