OOPS handling proposal

OOPS handling proposal

[Samium Gromoff]

        Hello folks!
    People most likely needs first oops most of the time, so
  it would be great to make a kernel boot option to stop after
  first oops, so in the situations when its neede life can be made easy.
    Actually afaics this is being done each time by hand, so 
  panic() in BUG() as Rik proposed would be nice.
    What do you folks think?

cheers, Sam

Re: OOPS handling proposal

[Andreas Dilger]

On Sep 13, 2001  06:11 +0000, Samium Gromoff wrote:
>         Hello folks!
>     People most likely needs first oops most of the time, so
>   it would be great to make a kernel boot option to stop after
>   first oops, so in the situations when its neede life can be made easy.
>     Actually afaics this is being done each time by hand, so 
>   panic() in BUG() as Rik proposed would be nice.
>     What do you folks think?

Absolutely NOT.  If you panic() in BUG() then you DO have to copy the oops
by hand each time, whereas the normal situation is that you can keep
running after an oops - to decode it, see what the symptoms are, check
memory, etc.

Cheers, Andreas
-- 
Andreas Dilger  \ "If a man ate a pound of pasta and a pound of antipasto,
                 \  would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/               -- Dogbert

Re: OOPS handling proposal

[Samium Gromoff]

"  Andreas Dilger wrote:"
> 
> On Sep 13, 2001  06:11 +0000, Samium Gromoff wrote:
> >         Hello folks!
> >     People most likely needs first oops most of the time, so
> >   it would be great to make a kernel boot option to stop after
> >   first oops, so in the situations when its neede life can be made easy.
> >     Actually afaics this is being done each time by hand, so 
> >   panic() in BUG() as Rik proposed would be nice.
> >     What do you folks think?
> 
> Absolutely NOT.  If you panic() in BUG() then you DO have to copy the oops
> by hand each time, whereas the normal situation is that you can keep
> running after an oops - to decode it, see what the symptoms are, check
> memory, etc.
   Hmm... I propose to make it a boot option.
   Else sometimes you wont be able to retrieve info at all, as like
 in the situation i have right now: oops in kupdated which i even
 cant post and it remains there all 2.4.8 and 2.4.9.

Cheers, Sam