Re: [PATCH] mtd: gpmi: Always report ECC stats and return max_bitflips

[Zach Sadecki <zsadecki@itwatchdogs.com>]

On 12/03/2012 09:34 PM, Huang Shijie wrote:
> 于 2012年12月04日 03:48, Zach Sadecki 写道:
>> Always report corrected and failed ECC stats back up to the MTD 
>> layer. Also
>> return max_bitflips from read_page() as is expected from NAND drivers 
>> now.
>>
>> Signed-off-by: Zach Sadecki <zsadecki@itwatchdogs.com>
>> ---
>> drivers/mtd/nand/gpmi-nand/gpmi-nand.c | 27 ++++++++-------------------
>> 1 file changed, 8 insertions(+), 19 deletions(-)
>>
>> diff --git a/drivers/mtd/nand/gpmi-nand/gpmi-nand.c 
>> b/drivers/mtd/nand/gpmi-nand/gpmi-nand.c
>> index d79696b..fe10c53 100644
>> --- a/drivers/mtd/nand/gpmi-nand/gpmi-nand.c
>> +++ b/drivers/mtd/nand/gpmi-nand/gpmi-nand.c
>> @@ -914,8 +914,7 @@ static int gpmi_ecc_read_page(struct mtd_info 
>> *mtd, struct nand_chip *chip,
>> dma_addr_t auxiliary_phys;
>> unsigned int i;
>> unsigned char *status;
>> - unsigned int failed;
>> - unsigned int corrected;
>> + unsigned int max_bitflips = 0;
>> int ret;
>> pr_debug("page number is : %d\n", page);
>> @@ -939,35 +938,25 @@ static int gpmi_ecc_read_page(struct mtd_info 
>> *mtd, struct nand_chip *chip,
>> payload_virt, payload_phys);
>> if (ret) {
>> pr_err("Error in ECC-based read: %d\n", ret);
>> - goto exit_nfc;
>> + return ret;
>> }
>> /* handle the block mark swapping */
>> block_mark_swapping(this, payload_virt, auxiliary_virt);
>> /* Loop over status bytes, accumulating ECC status. */
>> - failed = 0;
>> - corrected = 0;
>> - status = auxiliary_virt + nfc_geo->auxiliary_status_offset;
>> + status = auxiliary_virt + nfc_geo->auxiliary_status_offset;
>> for (i = 0; i < nfc_geo->ecc_chunk_count; i++, status++) {
>> if ((*status == STATUS_GOOD) || (*status == STATUS_ERASED))
>> continue;
>> if (*status == STATUS_UNCORRECTABLE) {
>> - failed++;
>> + mtd->ecc_stats.failed++;
>> continue;
>> }
>> - corrected += *status;
>> - }
>> -
>> - /*
>> - * Propagate ECC status to the owning MTD only when failed or
>> - * corrected times nearly reaches our ECC correction threshold.
>> - */
>> - if (failed || corrected >= (nfc_geo->ecc_strength - 1)) {
>> - mtd->ecc_stats.failed += failed;
>> - mtd->ecc_stats.corrected += corrected;
>> + mtd->ecc_stats.corrected += *status;
>> + max_bitflips = max_t(unsigned int, max_bitflips, *status);
>
> Does the max_bitflips stand for the whole page's bitfips ? or just the 
> bitfips of one part of the page?
> The gpmi driver splits a page into several parts when it does the ECC.
>
> thanks
> Huang Shijie
>
max_bitflips represents the maximum number of bitflips per ECC corrected 
region in a page.  The MTD layer needs it to compare to 
bitflip_threshold to know when to return -EUCLEAN to upper layers (like 
UBI) so a nearly corrupt page can be copied elsewhere.  This is pretty 
much identical to the way other drivers do this as I copied it from 
those.  Here's where it was added in the other drivers: 
http://git.infradead.org/linux-mtd.git/commit/3f91e94f7f511de74c0d2abe08672ccdbdd1961c 


Zach
>
>> }
>> if (oob_required) {
>> @@ -989,8 +978,8 @@ static int gpmi_ecc_read_page(struct mtd_info 
>> *mtd, struct nand_chip *chip,
>> this->payload_virt, this->payload_phys,
>> nfc_geo->payload_size,
>> payload_virt, payload_phys);
>> -exit_nfc:
>> - return ret;
>> +
>> + return max_bitflips;
>> }
>> static int gpmi_ecc_write_page(struct mtd_info *mtd, struct nand_chip 
>> *chip,
>
>