handling bad blocks on flash - Embedded

This is a discussion on handling bad blocks on flash - Embedded ; Hi, I'm interested to know how bad block handling mechanism works on flash devices in Linux. When data found found to be recoverable via ECC check, does it still need to be relocated to a good flash block? When data ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: handling bad blocks on flash

  1. handling bad blocks on flash

    Hi,
    I'm interested to know how bad block handling mechanism works on flash
    devices in Linux.
    When data found found to be recoverable via ECC check, does it still
    need to be relocated to a good flash block?
    When data found found to be unrecoverable via ECC check, should the
    rest of the pages on the block to be relocated?
    Who handles the relocation process, mtd or the file system?

    Thanks.

  2. Re: handling bad blocks on flash

    jackhab wrote:

    > I'm interested to know how bad block handling mechanism works on flash
    > devices in Linux.
    > When data found found to be recoverable via ECC check, does it still
    > need to be relocated to a good flash block?


    Marking each block with a single ECC error as BAD is not a good idea.

    Some sort of statistic or stress test is better.

    > When data found found to be unrecoverable via ECC check, should the
    > rest of the pages on the block to be relocated?


    Yes, of course.

    > Who handles the relocation process, mtd or the file system?


    The file system. Or a middle layer (UBI).
    Have a look at the ubi documentation!

    best regards

    Wolfgang

+ Reply to Thread