Re: block devices don't work without plugging in 2.4.3

From: Jens Axboe (axboe@suse.de)
Date: Thu Apr 19 2001 - 08:24:43 EST


On Thu, Apr 19 2001, Peter T. Breuer wrote:
> "Jens Axboe wrote:"
> > Examine _why_ you don't want plugging. In 2.2, you would have to edit
> > the kernel manually to disable it for your device.
>
> True. Except that I borrowed a major which already got that special
> treatment.

Ok

> > For 2.4, as long as
> > there has been blk_queue_pluggable, there has also been the
> > disable-merge function mentioned. Why are you disabling plugging??
>
> Fundamentally, to disable merging, as you suggest. I had merging
> working fine in 2.0.*. Then I never could figure out what had to be
> done in 2.2.*, so I disabled it. In 2.4, things work nicely - I don't
> have to do anything and it all happens magically.

Great

> Nevertheless, I am left with baggage that I have to maintain -
> certainly the driver has to work in 2.2 as well as in 2.4. Removing
> the blah_plugging function now in 2.4 after having started off 2.4
> with it around gives me one more #ifdef kernel_version in my code.

On the contrary, you are now given an exceptional opportunity to clean
up your code and get rid of blk_queue_pluggable and your noop plugging
function.

> I don't think that's good for my code, and in general I don't think one
> should remove this function half way through a stable series. Leave it
> there, mark it as deprecated in big letters, and make it do nothing,
> but leave it there, no?

Because most people are using it for the wrong reason anyway, so I'd
consider it a not-so-subtle hint. There should be no need for extra
ifdef's, on the contrary.

-- 
Jens Axboe

- 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/



This archive was generated by hypermail 2b29 : Mon Apr 23 2001 - 21:00:31 EST