[PATCH v3 1/3] mtd-nand: Add "page" parameter to all read_page/read_page_raw APIs
Singh, Vimal
IMCEAEX-_O=TI_OU=BD_CN=RECIPIENTS_CN=X0094262 at dlee86.itg.ti.com
Fri Jul 17 11:12:16 EDT 2009
On Fri, Jul 17, 2009 at 8:18 PM, Narnakaje, Snehaprabha<nsnehaprabha at ti.com> wrote:
>> -----Original Message-----
>> From: Singh, Vimal
>> Sent: Friday, July 17, 2009 10:45 AM
>> To: Narnakaje, Snehaprabha
>> Cc: linux-mtd at lists.infradead.org; davinci-linux-open-
>> source at linux.davincidsp.com; dwmw2 at infradead.org; tglx at linutronix.de;
>> akpm at linux-foundation.org; Paulraj, Sandeep
>> Subject: RE: [PATCH v3 1/3] mtd-nand: Add "page" parameter to all
>> read_page/read_page_raw APIs
>>
>> > > Why not just adding 'page' or 'cur_page' parameter to 'mtd_info'
>> structure and
>> > > populate the same in 'nand_do_read_ops' and 'nand_do_write_ops' as
>> soon as you
>> > > calculate 'page' for every read and write access respectivily. And
>> then this
>> > > info can be used by any one who has knowledge of 'mtd_info'.
>> > >
>> > Yes, that was one option. This meant, some of the existing handlers for
>> read_oob and write_oob required changes to get rid of "page", if we decide
>> to use the page from mtd_info structure.
>>
>> But that would have taken less changes.
>
> The change in read_oob/write_oob handlers would be minimal.
>
> But the API definitions should change all over the place, similar to the way read_page API had to change in all NAND files (in this patch).
AFAIK 'drivers/mtd/nand/cafe_nand.c' is only driver which over rides read_oob/write_oob APIs.
-vimal
More information about the linux-mtd
mailing list