Re: [PATCH] Add Amstrad Delta NAND support.

From: Jörn Engel
Date: Fri May 19 2006 - 06:02:47 EST


On Fri, 19 May 2006 10:01:42 +0100, Jonathan McDowell wrote:
> On Thu, May 18, 2006 at 06:57:28PM +0200, Jörn Engel wrote:
> > On Thu, 18 May 2006 17:09:41 +0100, Jonathan McDowell wrote:
> > > + omap_writew(0, (OMAP_MPUIO_BASE + OMAP_MPUIO_IO_CNTL));
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > Could that be done in a macro?
>
> Is there any benefit to doing so?

Not a big one. A common pattern is to have some defines like

#define FOO_BASE 0x12340000
#define FOO_THIS (FOO_BASE + 0)
#define FOO_THAT (FOO_BASE + 4)
etc.

Your code looks similar to this pattern, so I asked. If it doesn't
make sense, fair.

> > > + udelay(0.04);
> >
> > Floating point in the kernel?
>
> Not quite. udelay is a macro on ARM so this ends up as an integer before
> it ever hits a function call.

Hmm. It does what you say. Whether this is robust and will continue
to do so... I'm not sure.

> In an ideal world I'd use "ndelay(40);"
> but that would result in a delay of over 1µs as ARM doesn't have ndelay
> defined so we hit the generic fallback.

Can you either introduce a proper ndelay or get rmk to officially
bless your use of (constant, I know) floats in the kernel source? A
proper ndelay would obviously be preferred.

> > > + ams_delta_mtd = kmalloc (sizeof(struct mtd_info) +
> > ^
> > > + sizeof (struct nand_chip), GFP_KERNEL);
> >
> > Remove space
> >
> > And please create a structure containing both struct mtd_info and
> > struct nand_chip. Then use sizeof(that structure)...
>
> This format is used throughout the drivers/mtd/nand/ directory. I'd
> suggest it'd be more appropriate to have a separate patch that did this
> for all of them if it's desired, rather than having each driver do its
> own thing.
>
> Agreed on all the spacing comments you raised; hangovers from toto.c
> that I used as a base.

We have suboptimal code in the kernel, true. I still prefer new code
to have slightly higher standards, so the overall quality improves
slowly. Therefore, pointing to the other ugly duckling is not an
excuse for being mud-covered. ;)

But you are right. The other ducklings could use a bath as well.

Jörn

--
There are two ways of constructing a software design: one way is to make
it so simple that there are obviously no deficiencies, and the other is
to make it so complicated that there are no obvious deficiencies.
-- C. A. R. Hoare
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/