nbd for 2.0.* [was Re: MM deadlock]

Peter T. Breuer (ptb@it.uc3m.es)
Mon, 11 Jan 1999 16:55:07 +0100 (MET)

"A month of sundays ago Eric W. Biederman wrote:"
> ST> It means that if our superblock-locked IO has to stall waiting for an
> ST> nbd server process or a raid daemon, then those daemons cannot safely do
> ST> GFP_IO. It's really gross.
>
> Right. And the flag not to do I/O doesn't propogate across processes.
> This sounds like a variation of the priority inheritance problem.
>
> I wonder if this is why there are some known deadlocks with raid?

I've been having deadlock problems with the port to 2.0.36 of NBD. SInce
you seem to be on the same track and as I can't figure it out, I'll post
my current nbd patch made for 2.0.36 and my comments to Pavel. I see the
problems under raid too.

I would appreciate help.

Peter

------