Re: patch for 2.1.69 buffer.c

[Andrew E. Mileski]

> > Hmm, it seems to me the loop device buffers shouldn't be locked until
> > all the resources have been rounded up. But if this is what's being
> > done, it's easy enough to add a test not to wait on a loop buffer.

The loop driver expects the kernel to be re-entrant, and not all
of it is. Specifically, driver requests, buffers, and interrupts
are a problem.

I've attempted a fix long ago on the first two - it seems to work okay.
I took a swipe at solving the interrupt problem [saving and restoring
the flag instead of assuming it is in a specific state], which also
seems to work, but I doubt if I got every single instance.
   http://fractal.mta.ca/pub/crypto/aem/

--
Andrew E. Mileski   mailto:[e-mail]