Re: [PATCH 2/3] loop: No need to initialize ->queue_lock explicitly before calling blk_cleanup_queue()

[Sergey Senozhatsky <sergey.senozhatsky@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2049 bytes --]

On (02/22/11 09:20), Vivek Goyal wrote:
> On Tue, Feb 22, 2011 at 09:30:32AM +0200, Sergey Senozhatsky wrote:
> > On (02/21/11 22:53), Vivek Goyal wrote:
> > > o Now we initialize ->queue_lock at queue allocation time so driver does
> > >   not have to worry about initializing it before calling blk_cleanup_queue().
> > > 
> > > Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
> > > ---
> > >  drivers/block/loop.c |    3 ---
> > >  1 files changed, 0 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/block/loop.c b/drivers/block/loop.c
> > > index 49e6a54..44e18c0 100644
> > > --- a/drivers/block/loop.c
> > > +++ b/drivers/block/loop.c
> > > @@ -1641,9 +1641,6 @@ out:
> > >  
> > >  static void loop_free(struct loop_device *lo)
> > >  {
> > > -	if (!lo->lo_queue->queue_lock)
> > > -		lo->lo_queue->queue_lock = &lo->lo_queue->__queue_lock;
> > > -
> > >  	blk_cleanup_queue(lo->lo_queue);
> > >  	put_disk(lo->lo_disk);
> > >  	list_del(&lo->lo_list);
> > 
> > Hi,
> > 
> > (just for note) 
> > There is an incremental patch fixing this case in Andrew's mm tree: 
> > https://lkml.org/lkml/2011/2/11/165
> >                                    
> > (block-fix-queue_lock-null-pointer-derefence-in-blk_throtl_exit-v4.patch
> > added to -mm tree).
> 
> Hi Sergey,
> 
> Thinking more about it, initializing queue lock in blk_alloc_queue() seems
> to be even more cleaner to me instead of initializing it to internal lock
> during blk_cleanup_queue(). If others like the idea, then we can either
> ask Andrew to drop the patch or I can generate one on top of it.
> 

Hi Vivek,

Sure, fixing the even probability of NULL queue_lock is better (correct/sane/etc.) 
then fixing NULL queue_lock in `different places'. I'm presonally OK with dropping 
my patch.

By the way, I guess we should Cc stable when we decide which one [fix] will land 
on the mainline (taking in account that we're on -rc6 already).

Side Note: Need to check if .37 is affected (I suspect it is).


	Sergey

[-- Attachment #2: Type: application/pgp-signature, Size: 316 bytes --]