lkml.org 
[lkml]   [2010]   [Oct]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: nfsd changes for 2.6.37
On Tue, Oct 26, 2010 at 06:11:56PM -0400, J. Bruce Fields wrote:
> On Tue, Oct 26, 2010 at 05:44:41PM -0400, J. Bruce Fields wrote:
> > On Tue, Oct 26, 2010 at 02:37:26PM -0700, Linus Torvalds wrote:
> > > On Tue, Oct 26, 2010 at 2:24 PM, J. Bruce Fields <bfields@fieldses.org> wrote:
> > > >
> > > > I did a couple connectathon runs just now with no obvious ill effects
> > > > except for some sleep-within-spinlock warnings in the lease code.
> > >
> > > Hmm. Those sleep-within-spinlock warnings are very likely serious
> > > bugs.
> >
> > Yeah, didn't mean to belittle them.
> >
> > > Can you quote the whole warning with stack trace?
> >
> > It's just obvious allocations in setlease:
> >
> > BUG: sleeping function called from invalid context at mm/slab.c:3101
> > in_atomic(): 1, irqs_disabled(): 0, pid: 4345, name: lease_tests
> > 1 lock held by lease_tests/4345:
> > #0: (file_lock_lock){+.+.+.}, at: [<ffffffff81128be5>] lock_flocks+0x15/0x20
> > Pid: 4345, comm: lease_tests Not tainted 2.6.36-05858-gbd5e20b #1028
> > Call Trace:
> > [<ffffffff8103141d>] __might_sleep+0x10d/0x140
> > [<ffffffff810e3ad3>] kmem_cache_alloc+0x1f3/0x230
> > [<ffffffff8112a4d2>] generic_setlease+0x112/0x2c0
> > [<ffffffff8112a6b5>] __vfs_setlease+0x35/0x40
> > [<ffffffff8112acfe>] fcntl_setlease+0xce/0x180
> > [<ffffffff810f7c2e>] sys_fcntl+0x2fe/0x630
> > [<ffffffff81961999>] ? trace_hardirqs_on_thunk+0x3a/0x3f
> > [<ffffffff81002658>] system_call_fastpath+0x16/0x1b
> >
> > I'm testing a patch.
>
> This works for me.
>
> I'm not saying it's correct, but it does at least pass my dumb tests
> without complaining.

I can't think of any more missing locking, though I did notice this on a
quick look.

--b.

commit fc42117585672abd3cbf247dd311869233d1606a
Author: J. Bruce Fields <bfields@redhat.com>
Date: Tue Oct 26 18:25:30 2010 -0400

fix nlmsvc_notify_blocked locking

Signed-off-by: J. Bruce Fields <bfields@redhat.com>

diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index 6f1ef00..c462d34 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -700,14 +700,16 @@ nlmsvc_notify_blocked(struct file_lock *fl)
struct nlm_block *block;

dprintk("lockd: VFS unblock notification for block %p\n", fl);
+ spin_lock(&nlm_blocked_lock);
list_for_each_entry(block, &nlm_blocked, b_list) {
if (nlm_compare_locks(&block->b_call->a_args.lock.fl, fl)) {
- nlmsvc_insert_block(block, 0);
+ nlmsvc_insert_block_locked(block, 0);
+ spin_unlock(&nlm_blocked_lock);
svc_wake_up(block->b_daemon);
return;
}
}
-
+ spin_unlock(&nlm_blocked_lock);
printk(KERN_WARNING "lockd: notification for unknown block!\n");
}


\
 
 \ /
  Last update: 2010-10-27 00:43    [from the cache]
©2003-2014 Jasper Spaans. Advertise on this site