2005-01-27 18:46:22

by Pavel Machek

[permalink] [raw]
Subject: Applications segfault on evo n620c with 2.6.10

Hi!

It happened for 3rd in a week now...

When problem happens, processes start to segfault, usually right
during startup. Programs that were loaded prior to problem usualy
works, and can be restarted. I also seen sendmail exec failing with
"no such file or directory" when it clearly was there. Reboot corrects
things, and filesystem (ext3) is not damaged.

Unfortunately I do not know how to reproduce it. I tried
parallel-building kernels for few hours and that worked okay. Swsusp
is not involved (but usb, bluetooth, acpi and sound may be).

Does anyone else see something similar?
Pavel
--
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


2005-01-27 22:58:11

by Nigel Cunningham

[permalink] [raw]
Subject: Re: Applications segfault on evo n620c with 2.6.10

Hi.

On Fri, 2005-01-28 at 05:43, Pavel Machek wrote:
> Unfortunately I do not know how to reproduce it. I tried
> parallel-building kernels for few hours and that worked okay. Swsusp
> is not involved (but usb, bluetooth, acpi and sound may be).

I take it you're sure suspending is not involved because it happens
before you've ever suspended? If you hadn't said that, I'd say it sounds
very much like something suspend related.

Regards,

Nigel
--
Nigel Cunningham
Software Engineer
Cyclades Corporation

http://cyclades.com

2005-01-27 23:05:21

by Pavel Machek

[permalink] [raw]
Subject: Re: Applications segfault on evo n620c with 2.6.10

Hi!

> > Unfortunately I do not know how to reproduce it. I tried
> > parallel-building kernels for few hours and that worked okay. Swsusp
> > is not involved (but usb, bluetooth, acpi and sound may be).
>
> I take it you're sure suspending is not involved because it happens
> before you've ever suspended? If you hadn't said that, I'd say it sounds
> very much like something suspend related.

Yes, it happened even in cases when machine was not ever suspended. I
guess I should also add that kernel is "tainted: pavel", (that means I
have my own patches in; but I really believe that my changes are not
responsible).
Pavel
--
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!

2005-01-27 23:17:33

by Nigel Cunningham

[permalink] [raw]
Subject: Re: Applications segfault on evo n620c with 2.6.10

Hi.

On Fri, 2005-01-28 at 10:01, Pavel Machek wrote:
> Yes, it happened even in cases when machine was not ever suspended. I
> guess I should also add that kernel is "tainted: pavel", (that means I
> have my own patches in; but I really believe that my changes are not
> responsible).

I often believe that too ;>

Nigel
--
Nigel Cunningham
Software Engineer
Cyclades Corporation

http://cyclades.com

2005-01-28 02:40:03

by Hu Gang

[permalink] [raw]
Subject: Re: Applications segfault on evo n620c with 2.6.10

On Thu, Jan 27, 2005 at 07:43:34PM +0100, Pavel Machek wrote:
> Hi!
>
> It happened for 3rd in a week now...
>
> When problem happens, processes start to segfault, usually right
> during startup. Programs that were loaded prior to problem usualy
> works, and can be restarted. I also seen sendmail exec failing with
> "no such file or directory" when it clearly was there. Reboot corrects
> things, and filesystem (ext3) is not damaged.
>
> Unfortunately I do not know how to reproduce it. I tried
> parallel-building kernels for few hours and that worked okay. Swsusp
> is not involved (but usb, bluetooth, acpi and sound may be).
>
> Does anyone else see something similar?

I got the same thing in my computer.

Maybe this can reproduce it.
1: add this in boot loader
"init=/bin/sh"
2: after system boot, then active swap space, then do suspend.
3: after system resume, the sh will crash like.
that can 100% reproduce it my in X86, X86_64, PPC32.

The Software suspend2 has not that problem.

--
Hu Gang .-.
/v\
// \\
Linux User /( )\ [204016]
GPG Key ID ^^-^^ http://soulinfo.com/~hugang/hugang.asc

2005-01-28 10:13:27

by Pierre Chifflier

[permalink] [raw]
Subject: Re: Applications segfault on evo n620c with 2.6.10

On Thu, Jan 27, 2005 at 07:43:34PM +0100, Pavel Machek wrote:
> Hi!
>
> It happened for 3rd in a week now...
>
> When problem happens, processes start to segfault, usually right
> during startup. Programs that were loaded prior to problem usualy
> works, and can be restarted. I also seen sendmail exec failing with
> "no such file or directory" when it clearly was there. Reboot corrects
> things, and filesystem (ext3) is not damaged.
>
> Unfortunately I do not know how to reproduce it. I tried
> parallel-building kernels for few hours and that worked okay. Swsusp
> is not involved (but usb, bluetooth, acpi and sound may be).
>
> Does anyone else see something similar?
> Pavel

I have the same laptop and there is no error here.
However, I remember this laptop was affected by a RAM problem, which
could cause these symptoms.

More infos here:
http://www.theregister.co.uk/2004/06/26/hp_ram_recall/

Cheers,

Pierre

2005-01-28 13:04:42

by Pavel Machek

[permalink] [raw]
Subject: Re: Applications segfault on evo n620c with 2.6.10

Hi!

> > It happened for 3rd in a week now...
> >
> > When problem happens, processes start to segfault, usually right
> > during startup. Programs that were loaded prior to problem usualy
> > works, and can be restarted. I also seen sendmail exec failing with
> > "no such file or directory" when it clearly was there. Reboot corrects
> > things, and filesystem (ext3) is not damaged.
> >
> > Unfortunately I do not know how to reproduce it. I tried
> > parallel-building kernels for few hours and that worked okay. Swsusp
> > is not involved (but usb, bluetooth, acpi and sound may be).
> >
> > Does anyone else see something similar?
>
> I have the same laptop and there is no error here.
> However, I remember this laptop was affected by a RAM problem, which
> could cause these symptoms.
>
> More infos here:
> http://www.theregister.co.uk/2004/06/26/hp_ram_recall/

I see... unfortunately this is some strange kind of engineering sample
:-(, and they are no longer replacing the memory.

Does someone still have the application that tests if the flaw is
present? Is there easy way to tell that from markings on the chip?

"korea 253 PC2100S-25330-Z M470L6423DN0-CB0 512MB DDR PC2100CL2.5"

Some sources report that it only happens with C3 -- and I had usb
plugged in, that should result in no C3...

Pavel
--
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!