Re: [PATCH] ide-disk.c rev 1.13 killed CONFIG_IDEDISK_STROKE

[Erik Andersen <andersen@codepoet.org>]

On Wed Aug 06, 2003 at 08:32:23PM +0200, Bartlomiej Zolnierkiewicz wrote:
> 
> On Wed, 6 Aug 2003, Erik Andersen wrote:
> 
> > > Can you please resend me the patch once its accepted in 2.6 ?
> > >
> > > Maybe we want to test it a while in -ac, also?
> >
> > Ok.  Though it is being incorporated as part of a much larger
> > patch in 2.6.  I suppose that will filter back into 2.4 when
> > it is ready,
> 
> Hi Erik,
> 
> I made init_idedisk_capacity() cleanup.
> Then I ported your patch and reworked it a bit.
> 
> Could you take a look?

I like your improvements.  The only concern I have is you
retained my use of do_div() intact.  That needs to change since
it turns out that do_div() directly modifies the numerator
(violating the principle of least surprise).  See the recent
thread on "do_div considered harmful".  In that thread, Andries
mentions he has made a generic sectors_to_MB() function to
consolidate such things.  which I think is a very good idea since
then this only has to be correct once.  That would also let you
eliminate the somewhat ugly and not particularly obvious division
plus rounding (X - X/625 + 974)/1950 sequence from
idedisk_check_hpa_lba28(),

 -Erik

--
Erik B. Andersen             http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--