2003-03-04 21:06:29

by Rusty Lynch

[permalink] [raw]
Subject: Available watchdog test cases

The test cases that I have been using for testing my watchdog work are
available at
http://www.stinkycat.com/patches/watchdog_stuff/watchdog_test.tar.gz

The test are for both the legacy (well, current) /dev/watchdog interface
and my proposed sysfs based interface.

--rustyl




2003-03-05 16:05:16

by Paul Larson

[permalink] [raw]
Subject: Re: Available watchdog test cases

On Tue, 2003-03-04 at 15:04, Rusty Lynch wrote:
> The test cases that I have been using for testing my watchdog work are
> available at
> http://www.stinkycat.com/patches/watchdog_stuff/watchdog_test.tar.gz
>
> The test are for both the legacy (well, current) /dev/watchdog interface
> and my proposed sysfs based interface.
Thanks a bunch Rusty, I'll take a look at them for LTP.

-Paul Larson


Attachments:
signature.asc (240.00 B)
This is a digitally signed message part

2003-03-09 09:30:23

by Paul Rolland

[permalink] [raw]
Subject: Re: Available watchdog test cases

Just a quick question : is there an easy to force the kernel to Oops,
to make sure that the watchdog will even be working under such
conditions ?

I know people are all trying to avoid Oops... but I think the testplan
should include that too...

Regards,
Paul

> On Tue, 2003-03-04 at 15:04, Rusty Lynch wrote:
> > The test cases that I have been using for testing my
> watchdog work are
> > available at
> > http://www.stinkycat.com/patches/watchdog_stuff/watchdog_test.tar.gz
> >
> > The test are for both the legacy (well, current) /dev/watchdog
> > interface and my proposed sysfs based interface.
> Thanks a bunch Rusty, I'll take a look at them for LTP.

2003-03-10 05:18:01

by Rusty Lynch

[permalink] [raw]
Subject: Re: Available watchdog test cases

On Sun, 2003-03-09 at 01:40, Paul Rolland wrote:
> Just a quick question : is there an easy to force the kernel to Oops,
> to make sure that the watchdog will even be working under such
> conditions ?
>
> I know people are all trying to avoid Oops... but I think the testplan
> should include that too...
>
> Regards,
> Paul

You can write a kernel module that when loaded will disable all
interrupts and sit and spin, or even easier just call panic().

--rustyl

2003-03-10 06:51:38

by Paul Rolland

[permalink] [raw]
Subject: Re: Available watchdog test cases

Hello,

> > I know people are all trying to avoid Oops... but I think
> the testplan
> > should include that too...
>
> You can write a kernel module that when loaded will disable
> all interrupts and sit and spin, or even easier just call panic().
>
Too easy I should have thought about it myself...

Thanks for the tip...

Regards,
Paul

2003-03-11 04:25:02

by Randy.Dunlap

[permalink] [raw]
Subject: Re: Available watchdog test cases

> Hello,
>
>> > I know people are all trying to avoid Oops... but I think
>> the testplan
>> > should include that too...
>>
>> You can write a kernel module that when loaded will disable
>> all interrupts and sit and spin, or even easier just call panic().
>>
> Too easy I should have thought about it myself...

Steve Hemminger of OSDL did one of these. You can get it at
http://www.osdl.org/archive/rddunlap/tools/dump_test.c
if you are interested. (I just put it there and there is a
15-minute rsync delay...)

~Randy