Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Len Brown]

fastboot also causes the S3 regression here:

http://bugzilla.kernel.org/show_bug.cgi?id=12936

Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Linus Torvalds]

On Fri, 10 Apr 2009, Len Brown wrote:
>
> fastboot also causes the S3 regression here:
> 
> http://bugzilla.kernel.org/show_bug.cgi?id=12936

Looks like the same issue. Does my one-liner fix that case too?

		Linus

Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Heinz Diehl]

On 11.04.2009, Linus Torvalds wrote: 

[Len Brown:]
> > fastboot also causes the S3 regression here:
> > http://bugzilla.kernel.org/show_bug.cgi?id=12936
 
> Looks like the same issue. Does my one-liner fix that case too?

http://bugzilla.kernel.org/show_bug.cgi?id=13063

Your one-liner fixes this bug, too. Without the revertion mentioned in the
bugreport applied!

Regards,
Heinz.

Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Arjan van de Ven]

Linus Torvalds wrote:
> 
> On Fri, 10 Apr 2009, Len Brown wrote:
>> fastboot also causes the S3 regression here:
>>
>> http://bugzilla.kernel.org/show_bug.cgi?id=12936
> 
> Looks like the same issue. Does my one-liner fix that case too?
Now that I've been able to scan most of my mail; it looks like the one-liner should just go in.
Restricting module loading from async work to only non-blocking is not a big deal,
and it seems that userspace in various distros is really broken (and breaks on scsi already today),
but that is not something we can really fix quickly.
The sad part is that the userland is unlikely get fixed unless it shows breakage, and we can't break it (obviously).

I don't know of any good other solutions; right now it's mostly the partition scan (which is already mostly async on
scsi since a really long time, just in practice fast enough unless you have a big server).

(The __init thing in modules is separate and easy to fix, but without the other stuff being fixed there's no point)

Re: [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Linus Torvalds]

On Sat, 11 Apr 2009, Arjan van de Ven wrote:
>
> Now that I've been able to scan most of my mail; it looks like the one-liner
> should just go in.

Ok, I committed my one-liner, and then also reverted the ACPI battery 
workaround of removing __init, since the one-liner should fix that too.

Cc'ing the people involved with that commit, just so that they know to 
test to make sure the alternate fix really did fix it for them (I'm pretty 
sure it does, but still a good idea to verify or at least let people know 
that a previous fix got reverted)

			Linus

Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Arjan van de Ven]

Linus Torvalds wrote:
> 
> On Sat, 11 Apr 2009, Arjan van de Ven wrote:
>> Now that I've been able to scan most of my mail; it looks like the one-liner
>> should just go in.
> 
> Ok, I committed my one-liner, and then also reverted the ACPI battery 
> workaround of removing __init, since the one-liner should fix that too.
> 
> Cc'ing the people involved with that commit, just so that they know to 
> test to make sure the alternate fix really did fix it for them (I'm pretty 
> sure it does, but still a good idea to verify or at least let people know 
> that a previous fix got reverted)

longer term I would like to work on two things
1) Fix the __init thing by just only freeing when appropriate, but not having to wait for it
2) Find a way for userland to trigger a sync
    - for all storage probing
    - for just async work
    - "global sync", which includes all driver init
    We can then have insmod/modprobe use this always, unless a flag is set
    (so say, udev, can load a slew of modules, and then sync only on the last one)

and then after a year or two we can maybe remove the sync from the module loader ;(

Re: [linux-pm] [2.6.30-rc1-git2 regressions] Hibernation broken and (minor but annoying) audio problem

[Arkadiusz Miskiewicz]

On Saturday 11 of April 2009, Linus Torvalds wrote:
> On Sat, 11 Apr 2009, Arjan van de Ven wrote:
> > Now that I've been able to scan most of my mail; it looks like the
> > one-liner should just go in.
>
> Ok, I committed my one-liner, and then also reverted the ACPI battery
> workaround of removing __init, since the one-liner should fix that too.

Tested - works here|.

> 			Linus


-- 
Arkadiusz Miśkiewicz        PLD/Linux Team
arekm / maven.pl            http://ftp.pld-linux.org/

--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html