[ANNOUNCE] [PATCH] [MTD] Flex-OneNAND MTD Driver available.

apgmoorthy moorthy.apg at samsung.com
Mon Sep 29 05:28:02 EDT 2008


Hi David,

   Please do comment on this patch. 

   "Not to expose whole OOB" suggestion by Kyungmin should be fine , but
looking for your suggestion.

With Regards
Gangheyamoorthy.A.P

-----Original Message-----
From: Artem Bityutskiy [mailto:dedekind at infradead.org] 
Sent: Friday, September 26, 2008 1:50 PM
To: Amit Kumar Sharma
Cc: Kyungmin Park; apgmoorthy; linux-mtd at lists.infradead.org; Kyungmin Park;
lkml; David Woodhouse
Subject: Re: [ANNOUNCE] [PATCH] [MTD] Flex-OneNAND MTD Driver available.


On Fri, 2008-09-26 at 13:31 +0530, Amit Kumar Sharma wrote:
> Hi Artem
> 
> we have following erro message  if we maintain 64 bytes 
> eccpos.
> 
> /flash_eraseall -j /dev/mtd4
> ./flash_eraseall: /dev/mtd4: unable to get NAND oobinfo
> 
> ./nandwrite -j /dev/mtd2 /jffs2.img
> MEMSETOOBSEL: Inappropriate ioctl for device
> 
> but i think as OneNAND has hw ecc controller so Jffs2 will 
> not face any problem
> still i think mtd should consider for 128 bytes eccpos for 
> 4K page size.

Well, anyway, in Linux ABI is a holy cow, you cannot change it.
That is a tough rule. Old user-space binaries have to always
works.

Thus you may do one of:
1. Invent a new ioctl for 4KiB page NANDs
2. Add sysfs support
3. Just do not expose whole OOB as Kyungmin suggested

But I think assume dwmw2 would need to comment on this and
tell which approach would agree on.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)




More information about the linux-mtd mailing list