On Wed, Jul 31, 2002 at 02:51:32PM +0200, Joerg Dorchain wrote:
> On Wed, Jul 31, 2002 at 11:32:20AM +1000, David Gibson wrote:
> >
> > Ok, sorry, it's not quite as bad as I thought. But this highlights
> > part of the point - I missed the extra if statements in your patch
> > because your indentation doesn't match mine.
>
> Sorry, I tried to adapt it. I promise to learn.
> >
> > How similar are the PLX and TMD chips really? Is it worth putting
> > support into the same module, or should we create a new orinoco_tmd.c
> > with support for the TMD chips?
>
> AFAIK they both only provide access to the Prism2 chip. Differences are:
> - The PLX chip has three recourse areas. The addional one provides access
> to the PCMCIA cis, which is only used for checking wether the card has
> a know value. This additional check is not possible for the TMD chip.
> The only place where it is used is the init_one function. As I see it,
> it is redundancy checking.
> - The two remaing recources areas are swapped for their functionality.
>
> These roughly 32 lines of additional code are all that is needed to make
> the driver work. The layout of the prism2 registers is the same, it just
> fits. The rest of the code is shared, so creating a new file with 95%
> identical code makes no sense to me.

Ok, that's reasonable. Send me a patch that's properly formatted, and
I'll probably apply it.

> Maybe it would be an idea to put the common code into one file and then
> have two files of about 30 lines which contain the init_one and
> remove_one functions. IMHO it is overkill. It might be worth a config
> option, though. You are the maintainer, so you will have the final word
> ;)

No, I agree that is overkill.

> For the patch itself, I have two more remarks. I noticed meanwhile that
> the card also initialzes with an mdelay(1). The other is that the address
> for the outb(0x45) corresponds to the corr register, so for better
> readability it could be replaced by appropriate macros.

Send me a patch... :)

-- 
David Gibson			| For every complex problem there is a
davidgibson.dropbear.id.au	| solution which is simple, neat and
				| wrong.
http://www.ozlabs.org/people/dgibson

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]