Lustre devel - Jan 2007 - [Bug 10717] Crash when forcing client to flush locks

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10717



Was this race fixed in post 1.4.5.8 releases with the addition of the
change below.  Or is the observed failure understood to have been 
caused by a different event?

int ldlm_cancel_lru(struct ldlm_namespace *ns, ldlm_sync_t sync)
{
        ...
        list_for_each_entry_safe(lock, next, &ns->ns_unused_list, l_lru)
{
                LASSERT(!lock->l_readers && !lock->l_writers);

                /* If we have chosen to canecl this lock voluntarily, we better
                   send cancel notification to server, so that it frees
                   appropriate state. This might lead to a race where while
                   we are doing cancel here, server is also silently
                   cancelling this lock. */
>>>             lock->l_flags &= ~LDLM_FL_CANCEL_ON_BLOCK;
                ...
        }
        ...
}

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10717



Lustre: lustre-1.4.8-3chaos

Nevermind we were just able to trigger the same issue in a current version
of lustre.  The change I referenced in my previous comment clearly fixes a
bug, but just not this one.