cvs commit: src/sys/kern kern_sig.c

[John Baldwin]

On Friday 20 October 2006 20:58, David Xu wrote:
> On Saturday 21 October 2006 00:19, John Baldwin wrote:
> > jhb         2006-10-20 16:19:21 UTC
> >
> >   FreeBSD src repository
> >
> >   Modified files:
> >     sys/kern             kern_sig.c
> >   Log:
> >   Remove the check that prevented signals from being delivered to exiting
> >   processes.  It was originally added back when support for Linux threads
> >   (and thus shared sigacts objects) was added, but no one knows why.  My
> >   guess is that at some point during the Linux threads patches, the 
sigacts
> >   object was torn down during exit1(), so this check was added to prevent
> >   a panic for that race.  However, the stuff that was actually committed 
to
> >   the tree doesn't teardown sigacts until wait() making the above race
> > moot. Re-allowing signals here lets one interrupt a NFS request during
> > process teardown (such as closing descriptors) on an interruptible mount.
> >
> >   Requested by:   kib (long time ago)
> >   MFC after:      1 week
> >
> >   Revision  Changes    Path
> >   1.333     +1 -3      src/sys/kern/kern_sig.c
> This commit opens a window that may cause memory leak, since we have
> signal queue in -CURRENT now, signal queue uses memory, before this
> change, the P_WEXIT will prevents new signal to be queued, after the
> flag is set in kern_exit, we call signqueue_flush to free memory. I think
> we should move sigqueue_flush down to a safe point where the
> PROC lock is no longer unlocked or move it to a place after p_state is set
> to PRS_ZOMBIE.

Or just free it during wait() when most other things about a process are
freed (sigacts, limits, etc.).

-- 
John Baldwin