Re: [NFS] 2.6.17.8 - do_vfs_lock: VFS is out of sync with lock manager!

From: Neil Brown
Date: Thu Feb 01 2007 - 17:40:29 EST


On Wednesday January 31, jesper.juhl@xxxxxxxxx wrote:
> On 29/01/07, Trond Myklebust <trond.myklebust@xxxxxxxxxx> wrote:
>
> Finally getting that in will be sooooo nice :-) Thank you.
>
> Btw: any reason why not to include the
> fl->fl_flags &= ~FL_SLEEP;
> bit as well? As in http://lkml.org/lkml/2006/11/27/41 ??
>

Uhmm... I guess I had forgotten it ....

Looking again, I cannot convince myself that it is needed.
The comment says "If we were signalled ...", and if we were signalled
then do_vfs_lock won't block anyway. I think I was probably being
over-cautious as I didn't know the significance of testing for
-ERESTARTSYS.
However it seems to only get returned if a signal is pending (so why
EINTR isn't returned I still don't know) so a signal must be pending
on that piece of code, so there is no need to clear FL_SLEEP.

Thanks,
NeilBrown
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/