APM problem after 2.6.13.5

APM problem after 2.6.13.5

[Alessio Sangalli]

Hi, I've found a problem with kernels after 2.6.13.5.

I have an old (2001) laptop. Any kernel I tried:

2.6.14
2.6.14.7
2.6.15.7
2.6.16.2
2.6.16.9
2.6.16.16
2.6.16.20
2.6.17-rc6

if I enable "APM support" I get a freeze at the very beginning of the
boot, without any explicit erro message, just after the PCI stuff. If
you need a transcript of the messages at boot, let me know, I will have
to write them by hand).
2.6.13.5 is ok. I need APM support to let the "Fn" key and the battery
meter work!

I'm not subscribed to the lkml so please, CC me.

Thank you
Alessio Sangalli

Re: APM problem after 2.6.13.5

[Pekka Enberg]

Hi,

On 6/16/06, Alessio Sangalli <alesan@manoweb.com> wrote:
> if I enable "APM support" I get a freeze at the very beginning of the
> boot, without any explicit erro message, just after the PCI stuff. If
> you need a transcript of the messages at boot, let me know, I will have
> to write them by hand).
> 2.6.13.5 is ok. I need APM support to let the "Fn" key and the battery
> meter work!

There's a lot of changes between 2.6.13 and 2.6.14.  It would be
helpful if you could narrow down the exact changeset that broke your
setup.  Please see the following URL for details:

  http://www.kernel.org/pub/software/scm/git/docs/howto/isolate-bugs-with-bisect.txt

                             Pekka

Re: APM problem after 2.6.13.5

[Alessio Sangalli]

Pekka Enberg wrote:
> Hi,
> 
> On 6/16/06, Alessio Sangalli <alesan@manoweb.com> wrote:
>> if I enable "APM support" I get a freeze at the very beginning of the
>> boot, without any explicit erro message, just after the PCI stuff. If
>> you need a transcript of the messages at boot, let me know, I will have
>> to write them by hand).
>> 2.6.13.5 is ok. I need APM support to let the "Fn" key and the battery
>> meter work!
> 
> There's a lot of changes between 2.6.13 and 2.6.14.  It would be
> helpful if you could narrow down the exact changeset that broke your


done:

4196c3af25d98204216a5d6c37ad2cb303a1f2bf is first bad commit
diff-tree 4196c3af25d98204216a5d6c37ad2cb303a1f2bf (from
9092b20803e4b3b3a480592794a73030f17370b3)
Author: Linus Torvalds <torvalds@g5.osdl.org>
Date:   Sun Oct 23 16:31:16 2005 -0700

    cardbus: limit IO windows to 256 bytes

    That's what we've always historically done, and bigger windows seem to
    confuse some cardbus bridges. Or something.

    Alan reports that this makes the ThinkPad 600x series work properly
    again: the 4kB IO window for some reason made IDE DMA not work, which
    makes IDE painfully slow even if it works after DMA timeouts.

    Signed-off-by: Linus Torvalds <torvalds@osdl.org>

:040000 040000 629d4d303048bffa610017e81e0e744bae08660d
33e154ffe96822d09f37ae2d433de5152216501b M      drivers


let me know any other test I should do to help find a solution to this
problem, thank you!

Alessio Sangalli

Re: APM problem after 2.6.13.5

[Pekka Enberg]

Hi Alessio,

On 6/16/06, Alessio Sangalli <alesan@manoweb.com> wrote:
> > > if I enable "APM support" I get a freeze at the very beginning of the
> > > boot, without any explicit erro message, just after the PCI stuff. If
> > > you need a transcript of the messages at boot, let me know, I will have
> > > to write them by hand).
> > > 2.6.13.5 is ok. I need APM support to let the "Fn" key and the battery
> > > meter work!

Pekka Enberg wrote:
> > There's a lot of changes between 2.6.13 and 2.6.14.  It would be
> > helpful if you could narrow down the exact changeset that broke your

On 6/16/06, Alessio Sangalli <alesan@manoweb.com> wrote:
> done:
>
> 4196c3af25d98204216a5d6c37ad2cb303a1f2bf is first bad commit
> diff-tree 4196c3af25d98204216a5d6c37ad2cb303a1f2bf (from
> 9092b20803e4b3b3a480592794a73030f17370b3)
> Author: Linus Torvalds <torvalds@g5.osdl.org>
> Date:   Sun Oct 23 16:31:16 2005 -0700
>
>     cardbus: limit IO windows to 256 bytes
>
>     That's what we've always historically done, and bigger windows seem to
>     confuse some cardbus bridges. Or something.
>
>     Alan reports that this makes the ThinkPad 600x series work properly
>     again: the 4kB IO window for some reason made IDE DMA not work, which
>     makes IDE painfully slow even if it works after DMA timeouts.
>
>     Signed-off-by: Linus Torvalds <torvalds@osdl.org>
>
> :040000 040000 629d4d303048bffa610017e81e0e744bae08660d
> 33e154ffe96822d09f37ae2d433de5152216501b M      drivers
>
>
> let me know any other test I should do to help find a solution to this
> problem, thank you!

So reverting the above commit from git head makes your box boot again?
Linus, any thoughts?

                                      Pekka

Re: APM problem after 2.6.13.5

[Alessio Sangalli]

Pekka Enberg wrote:
> Hi Alessio,
> 
> On 6/16/06, Alessio Sangalli <alesan@manoweb.com> wrote:
>>>> if I enable "APM support" I get a freeze at the very beginning of the
>>>> boot, without any explicit erro message, just after the PCI stuff. If
>>>> you need a transcript of the messages at boot, let me know, I will have
>>>> to write them by hand).
>>>> 2.6.13.5 is ok. I need APM support to let the "Fn" key and the battery
>>>> meter work!
> 
> Pekka Enberg wrote:
>>> There's a lot of changes between 2.6.13 and 2.6.14.  It would be
>>> helpful if you could narrow down the exact changeset that broke your
> 
> On 6/16/06, Alessio Sangalli <alesan@manoweb.com> wrote:
>> done:
>>
>> 4196c3af25d98204216a5d6c37ad2cb303a1f2bf is first bad commit
>> diff-tree 4196c3af25d98204216a5d6c37ad2cb303a1f2bf (from
>> 9092b20803e4b3b3a480592794a73030f17370b3)
>> Author: Linus Torvalds <torvalds@g5.osdl.org>
>> Date:   Sun Oct 23 16:31:16 2005 -0700
>>
>>     cardbus: limit IO windows to 256 bytes
>>
>>     That's what we've always historically done, and bigger windows seem to
>>     confuse some cardbus bridges. Or something.
>>
>>     Alan reports that this makes the ThinkPad 600x series work properly
>>     again: the 4kB IO window for some reason made IDE DMA not work, which
>>     makes IDE painfully slow even if it works after DMA timeouts.
>>
>>     Signed-off-by: Linus Torvalds <torvalds@osdl.org>
>>
>> :040000 040000 629d4d303048bffa610017e81e0e744bae08660d
>> 33e154ffe96822d09f37ae2d433de5152216501b M      drivers
>>
>>
>> let me know any other test I should do to help find a solution to this
>> problem, thank you!
> 
> So reverting the above commit from git head makes your box boot again?
> Linus, any thoughts?


Yes, exactly. I can run 2.6.17-rc6 with only that commit reverted.

The mahcine is a notebook COMPAQ Presario 800 Model 80XL4
Pentium3@700MHz 256MB ram etc

Thank you
ciao
Alessio Sangalli

Re: APM problem after 2.6.13.5

[Alessio Sangalli]

Alessio Sangalli wrote:

>>> done:
>>>
>>> 4196c3af25d98204216a5d6c37ad2cb303a1f2bf is first bad commit
>>> diff-tree 4196c3af25d98204216a5d6c37ad2cb303a1f2bf (from
>>> 9092b20803e4b3b3a480592794a73030f17370b3)
>>> Author: Linus Torvalds <torvalds@g5.osdl.org>
>>> Date:   Sun Oct 23 16:31:16 2005 -0700
>>>
>>>     cardbus: limit IO windows to 256 bytes

>> So reverting the above commit from git head makes your box boot again?
>> Linus, any thoughts?
> 
> 
> Yes, exactly. I can run 2.6.17-rc6 with only that commit reverted.

Any news about this problem?

Thank you
Alessio Sangalli