[OpenAFS-devel] Linux deadlocks (possibly fixed in IBM-AFS)

[chas williams]

>	srikanth-IY31752-afs3.6-race-condition-in-afs-buffer-cache 1.2

i cannot comment on this.  with the description though it should be 
easy to check/fix/add.

>	srikanth-12885-afs3.6-race.condition.in.linux.event.handling 1.5

i believe i helped srikanth with this issue -- he messaged this list
about the problem.  i developed what i consider a 'correct' fix which
is in the latest devel and stable releases.  its probably similar, although
however, i also eliminated the glock droppping in osi_linux_alloc() since
you could still 'loose' a wakeup when adding the event the first time
through the loop.