Dear Xin,
On Fri, 3 Aug 2007, LI Xin wrote:
LX> > Log:
LX> > MFp4 - Refine locking to eliminate some potential race/panics:
LX> >
LX> > - Copy before testing a pointer. This closes a race window.
LX> > - Use msleep with the node interlock instead of tsleep.
LX> > - Do proper locking around access to tn_vpstate.
LX> > - Assert vnode VOP lock for dir_{atta,de}tach to capture
LX> > inconsistent locking.
LX> >
LX> > Suggested by: kib
LX> > Submitted by: delphij
LX> > Reviewed by: Howard Su
LX> > Approved by: re (tmpfs blanket)
LX>
LX> Should have mentioned that this is done in response of pho@'s stress
LX> test. Thanks goes to him for all these hard work!
Should it fix panics I reported to you during last couple of days?
Sincerely,
D.Marck [DM5020, MCK-RIPE, DM3-RIPN]
------------------------------------------------------------------------
*** Dmitry Morozovsky --- D.Marck --- Wild Woozle --- marck@rinet.ru ***
------------------------------------------------------------------------