From mboxrd@z Thu Jan  1 00:00:00 1970
From: Amit Virdi <amit.virdi@st.com>
Date: Thu, 17 May 2012 16:51:09 +0530
Subject: [U-Boot] [PATCH V2 2/4] mtd/NAND: Add FSMC driver support
In-Reply-To: <4FB3CF90.1000202@freescale.com>
References: <cover.1336374946.git.amit.virdi@st.com>
	<9e312b0349ace0f0284be499a21ef1d7eb31e8a5.1336374946.git.amit.virdi@st.com>
	<4FB2C71A.3000208@freescale.com> <4FB38454.3080803@st.com>
	<4FB3CF90.1000202@freescale.com>
Message-ID: <4FB4DF25.90900@st.com>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

>>>> +    while (num_err--) {
>>>> +        change_bit(0,&err_idx[i]);
>>>> +        change_bit(1,&err_idx[i]);
>>>> +
>>>> +        if (err_idx[i]<   512 * 8) {
>>>> +            change_bit(err_idx[i], dat);
>>>> +            i++;
>>>> +        }
>>>> +    }
>>>
>>> Is it normal to not count bit flips in the ECC itself?
>>>
>>
>> Correcting bit flip in ECC isn't of any use, so we skipped it.
>
> I think it is useful to report that the bit flip happened, so that the
> upper layers know when to scrub the page.
>

Ok. I'll add the number of bit flips of OOB while reporting the 
corrected bits.

>> Otherwise also, the ECC and the data area are not contiguous here as
>> they are at different parts of RAM so we couldn't have done
>>      if (err_idx[i]<   (512+13) * 8) {
>> here.
>
> You could still increment the error count.
>

Yes.

Thanks
Amit Virdi