[PATCH v3 5/9] mtd: pxa3xx_nand: add support for the Marvell Berlin nand controller

Ezequiel Garcia ezequiel.garcia at free-electrons.com
Sun Mar 8 13:22:04 PDT 2015


On 03/08/2015 02:14 PM, Robert Jarzmik wrote:
> Ezequiel Garcia <ezequiel.garcia at free-electrons.com> writes:
> 
>> Hi Antoine,
>>
>> On 03/05/2015 08:31 AM, Antoine Tenart wrote:
>> [..]
>>> +
>>> +static struct pxa3xx_nand_flash berlin_builtin_flash_types[] = {
>>> +{ "4GiB 8-bit",    0xd7ec, 128, 8192,  8,  8, 4096 },
>>> +{ },
>>
>> IMHO, supporting a specific flash shouldn't be part of this patch.
>>
>> In any case, why do you need this? If you can share the details about
>> this device, it would be interesting for me to take a look.
>>
>> This driver's open-coded, legacy-style flash detection is nasty, and the
>> only reason I've kept it is to avoid breaking some wacky user with some
>> old board. In fact, maybe we can just kill it so nobody thinks it's sane.
> 
> I think you'll kill the zylonite board, and I'll nack it if that's the case. At
> least that's what happened when I tried to use onfi default values last time in
> barebox development.
> 
> I can test your changes, but if the specific zylonite nand (ie. nand id 0xba20,
> ie. pxa310 embedded flash) gets broken, I'm against the removal of the legacy
> timings removal.
>

I'm not speaking of any timing params here, but about the flash
identification.

Which flash do you have there?

-- 
Ezequiel García, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com



More information about the linux-arm-kernel mailing list