[RFC/PATCH 2/3] NAND multiple plane feature

[Thomas Gleixner]

Hi,

Can you please use a sane mail client, which does line breaks around
column 78 ?

On Fri, 6 Jun 2008, Alexey Korolev wrote:
> Hi,
> > On Thu, 5 Jun 2008, Alexey Korolev wrote:
> > > > > -	if (!chip->write_page)
> > > > > -		chip->write_page = nand_write_page;
> > > > > -
> > > > 
> > > > Why is this removed?
> > > >
> > > Oh. May be it would be better to make stand alone patch in order to
> > > clean up NAND subsystem. Using chip->write_page is redundend here. It
> > > is set once it is set for write procedure only. There is no chip read
> > > page call like that. IMHO it is better to use static call here. Please
> > > let me know if you are agree with this. 
> > 
> > No, it's definitely not better to use a static call here. Cared to look
> > at the existing drivers, whether one provides it's own function ?
> > 
> > drivers/mtd/nand/cafe_nand.c:	cafe->nand.write_page = cafe_nand_write_page;
> >
> It make sence to remove because of:
> 1. nand_chip structure looks ugly with that function. It contains
  write_page but it does not contain read_page.

Oh yeah, "looks ugly" is definitely a perfect technical reason.

> 2. existence of such redefinition is redundend. There is definetely
> no need to redefine it.

-ENOPARSE

> The only reasignment of write_page call in whole nand subsystem
> exists in cafe driver.  But the purpose of this reasigment looks
> rather unclear.

> Try to find at least one line difference between
> cafe_nand_write_page and nand_write_page. I wonder how this code get
> included ?

You need to check the history why this is the case. And if it is the
same then a separate patch needs to be provide which removes this.

Thanks,
	tglx