RE: spinlock requests (was RE: [Xen-devel] [Patch] don't spin with irq disabled)

[Dan Magenheimer <dan.magenheimer@xxxxxxxxxx>]

> Well the race is still impossible afaics. Unless your
> _raw_rw_is_write_locked() was checking for ==0 rather than 
> <=0.

Finally got a chance to go back and reproduce and figure
this one out.  The crash is still occurring.  My code
definitely checks for <=0.  BUT the declaration of
raw_rwlock_t declares the lock as "volatile UNSIGNED int",
so the compiler blithely generates a check for == 0.

Can I assume the fix is to change the declaration to
int instead of unsigned int, or will that cause problems
elsewhere?

Thanks,
Dan

> -----Original Message-----
> From: Keir Fraser [mailto:keir.fraser@xxxxxxxxxxxxx]
> Sent: Friday, March 27, 2009 12:12 PM
> To: Dan Magenheimer; Jan Beulich; Juergen Gross
> Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
> Subject: Re: spinlock requests (was RE: [Xen-devel] [Patch] don't spin
> with irq disabled)
> 
> 
> On 27/03/2009 18:00, "Dan Magenheimer" 
> <dan.magenheimer@xxxxxxxxxx> wrote:
> 
> >> What try_readlock would that be? There is no such function.
> > 
> > Oops, my memory was faulty.  It is read_lock() itself
> > that decrements the lock and then re-increments it.
> > This appeared to be causing the race.  When I added
> > debug code, the problem went away (which was what led
> > me to the "hack" working version).
> 
> Well the race is still impossible afaics. Unless your
> _raw_rw_is_write_locked() was checking for ==0 rather than 
> <=0. I think
> something fishy was going on in your actual original implementation if
> is_write_locked(). :-)
> 
>  -- Keir
> 
> 
>

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel