On Wed, 23 May 2001, Eric Seppanen wrote:
> So maybe you could do some neat post-crash debugging, or recover the most
> recent pending filesystem writes, but I have doubts that "restarting" the
> crashed system would ever happen.
I think you're right. I should have been more specific. We are kind of
curious here about whether we could resuscitate a process.
There are lots of ways this could fail. Are there any that can succeed?
That's what we're wondering.
> Hey, I bet people would be impressed if you could just come up with a
> simple kernel panic recovery method, where linuxbios could detect the
> residue (oops data stashed in the log buffer or hidden somewhere else in
> ram), and preserve it for later debugging.
That we definitely want due to Myrinet driver problems.
ron
