On Wed, Mar 12, 2008 at 02:51:23PM +0100, Bartlomiej Zolnierkiewicz wrote:
quoted text
> 
> Hi,
> 
> On Wednesday 12 March 2008, Borislav Petkov wrote:
> 
> [...]
> 
> > > > the read request's buffer directly thru idetape_queue_rw_tail()
> > > > 
> > > > 3. Do next request buffer allocation (tape->merge_stage)
> > > 
> > > Isn't idetape_init_read() taking care of 3.?
> > 
> > i wanted to have the whole handling at one place and let _init_read() only
> > prepare the read. Now we don't allocate any new tape->merge_stage anymore,
> > which is wrong. Originally, this happened in _init_read(), however, if we do
> > idetape_queue_rw_tail(), we should alloc the new stage _after_ queueing the
> 
> The original driver doesn't do this - it just calls idetape_queue_rw_tail(),
> could it be a bug in the original driver?

Damn, i see it now, idetape_queue_rw_tail() queues the request and then simply
_reuses_ the tape->merge_stage buffer by doing

        if (tape->merge_stage)
		idetape_init_merge_stage(tape);

so no need for reallocation. Whew! :)

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