Re: cfq build breakage

[Andrew Morton <akpm@osdl.org>]

Jens Axboe <axboe@suse.de> wrote:
>
> On Mon, Jun 27 2005, Andrew Morton wrote:
> > Jeff Garzik <jgarzik@pobox.com> wrote:
> > >
> > > 
> > > In latest git tree...
> > > 
> > >    CC [M]  drivers/block/cfq-iosched.o
> > > drivers/block/cfq-iosched.c: In function `cfq_put_queue':
> > > drivers/block/cfq-iosched.c:303: sorry, unimplemented: inlining failed 
> > > in call to 'cfq_pending_requests': function body not available
> > > drivers/block/cfq-iosched.c:1080: sorry, unimplemented: called from here
> > > drivers/block/cfq-iosched.c: In function `__cfq_may_queue':
> > > drivers/block/cfq-iosched.c:1955: warning: the address of 
> > > `cfq_cfqq_must_alloc_slice', will always evaluate as `true'
> > > make[2]: *** [drivers/block/cfq-iosched.o] Error 1
> > > make[1]: *** [drivers/block] Error 2
> > > make: *** [drivers] Error 2
> > 
> > hm.  The inline thing is trivial, but the misuse of
> > cfq_cfqq_must_alloc_slice() means that we now wander into untested
> > territory.
> 
> Indeed, which compiler errors on that?

4.0 and later, I guess.

> > @@ -1969,7 +1968,7 @@ __cfq_may_queue(struct cfq_data *cfqd, s
> >  		 * only allow 1 ELV_MQUEUE_MUST per slice, otherwise we
> >  		 * can quickly flood the queue with writes from a single task
> >  		 */
> > -		if (rw == READ || !cfq_cfqq_must_alloc_slice) {
> > +		if (rw == READ || !cfq_cfqq_must_alloc_slice(cfqq)) {
> >  			cfq_mark_cfqq_must_alloc_slice(cfqq);
> >  			return ELV_MQUEUE_MUST;
> >  		}
> 
> thanks, clearly a typo but inside if 0.

But the other instance was inside `#if 1'.  This fixup will change behaviour.