On  8 Apr, Andrew Bennetts wrote:
>  Usually they get dumped to /var/log/messages if the system isn't too far 
>  gone, so that you often don't need to write anything down. 

That should be mentioned in the man page for ksymoops.  Not that it
would have helped me, since I never saw any Oops stuff in my messages
file back in January, when it was really bad.  (Though grepping again
tonight, I see that I had one for a crash on March 10th.)
 
>  > Does anyone know whether something more like Solaris's kind of facility 
>  > is being planned for Linux? 
>   
>  I seem to recall various patches at different times; try searching the 
>  kt.zork.net Kernel Traffic archives.  I think people have implemented 
>  everything from dumping to a file on disk, prompting for a floppy and 
>  dumping to that, and providing live debugging over a serial link... whether 
>  any of these will ever make it into the main kernel or even a 
>  vendor-supplied kernel is a different matter ;) 

Hmm.

>  Linus' view on kernel debugging aids seems to be that anything more than a 
>  register and stack dump is too much hassle for too little gain.  Registers 
>  and stack already give you the bulk of what you want to know (assuming the 
>  symbols have been decoded by ksymoops, of course), which is what the kernel 
>  was doing at the time it died.  Others have disagreed...

Yes, and I can sympathise with them.

> I once tried to 
>  keep up with the linux-kernel list, but I just couldn't keep it up.  These 
>  days I just read kt.zork.net and lwn.net for my kernel news. 

Thanks,

luke

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Reply via email to