Re: [Bug 11898] mke2fs hang on AIC79 device.

[James Bottomley <James.Bottomley@HansenPartnership.com>]

On Tue, 2008-11-04 at 19:55 -0600, Mike Christie wrote:
> bugme-daemon@bugzilla.kernel.org wrote:
> > http://bugzilla.kernel.org/show_bug.cgi?id=11898
> > 
> > 
> > 
> > 
> > 
> > ------- Comment #9 from alex.shi@intel.com  2008-11-04 17:32 -------
> > I did bisect for this bug and find the following commit cause this problem.
> > 
> > commit f0c0a376d0fcd4c5579ecf5e95f88387cba85211
> > Author: Mike Christie <michaelc@cs.wisc.edu>
> 
> Ok, I am looking into this.

I have a theoretical case that might trigger an endless loop:

Supposing we actually get a two deep starved list.  On first pass, we
process the first entry, but set starved_head to be it.  On second pass,
we get a host_busy.  Now we will loop forever (because we have a single
entry sdev and it's not starved_head) under the host lock, which means
we can never unbusy the host.

A fix would be to set starved_head to NULL after we delete sdev from the
list.

James