Hi All. I am new to this forum, so hope this is the right place to post
this question. We have an Android tablet running with Toshiba NAND
TC58TEGDCJTA00. The devices boot up fine, but seem to generate ECC errors
rather rapidly. Within about 60 days of operation we see unrecoverable ECC
errors in the boot logs. Looking at the boot logs, we can see the
following:
[0.203] [NAND] nand driver(b) version: 0x0x00000002, 0x0x00000012, data:
0x20130325
[0.211] *get the good blk ratio from hwscan : 912*
[0.216] NB1 : nand phy init ok
The device starts up, but with ECC errors and partition not aligned
messages (below)
[1.080] nand part 2048 start sec is not align
[1.080] nand part 2048 start sec is not align
[1.084] nand part 2048 start sec is not align
[1.088] nand part 2048 start sec is not align
[1.093] nand part 2048 start sec is not align
[1.098] nand part 2048 start sec is not align
[1.103] nand part 2048 start sec is not align
[1.107] nand part 2048 start sec is not align
[1.112] nand part 2048 start sec is not align
[1.117] nand part 2048 start sec is not align
[1.122] nand part 6144 sec count is not align
>From the source code of the Sun5i NAND.ko driver, we can see the Toshiba
chip entry for the physical information on the chip is below:
*//NAND_CHIP_ID DieCnt SecCnt
PagCnt BlkCnt OpOpt DatBlk Freq EccMode ReadRetry
OperationPar*
{ {0x98, 0xd7, 0x84, 0x93, 0x72, 0x57, 0xff, 0xff }, 1, 32,
256, 1024, 0x0580, 918, 40, 4, 0x100504,
0, &PhysicArchiPara9}, // TC58TEG5DCJTA00
*Here's my question:*
How can the chip think the DatBlk count is *912 from hwscan *when the
driver is programmed to think the *DatBlk count is 918?* Is this a bug in
the driver and would it explain why we are seeing
faster failure rates with these chips and ECC errors that can't be
corrected?
If I am reading the driver code correctly, the good_blk_ratio can be set in
the script.ini [nand_para]. We were going to try overriding the 918 and set
this to 912 at boot and see if it changed the ECC
error correction behavior.
Any information will help!!!!!!!!
thanks.
--
You received this message because you are subscribed to the Google Groups
"linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
