From: Daniel Drake Date: Wed, 4 Apr 2012 21:27:36 +0000 (+0100) Subject: UBI FAQ: document behaviour when lots of blocks go bad X-Git-Url: https://www.infradead.org/git/?a=commitdiff_plain;h=6953b6f41c57c986581350ed5f041d858716d32f;p=mtd-www.git UBI FAQ: document behaviour when lots of blocks go bad Based on an email from Artem. Signed-off-by: Daniel Drake Signed-off-by: Artem Bityutskiy --- diff --git a/faq/ubi.xml b/faq/ubi.xml index c9f1c57..8af461e 100644 --- a/faq/ubi.xml +++ b/faq/ubi.xml @@ -26,6 +26,7 @@
  • Can UBI logical eraseblocks be written randomly?
  • Why UBI does not use OOB area of NAND flashes?
  • Is UBI tolerant of power failures?
    • +
  • What happens when the PEBs reserved for bad block handling run out?
  • May UBI be used on MLC flash?
  • Why does ubiattach on a freshly formatted device fail with "Invalid argument"?
  • What is a sub-page?
    • @@ -456,6 +457,22 @@ must be an issue for UBI as well.

    +

    What happens when the PEBs reserved for bad block handling run out?

    + +By default, 1% of the available PEBs are reserved for handling bad blocks. + +If the number of blocks that turn bad exceeds that allocation, an error +will be presented and UBI will switch to read-only mode. + +To recover from this error you could re-flash the device. The run-time +recovery would require deleting or shrinking one of the UBI volumes. + +So, you need to carefully select the amount of PEBs reserved for bad +blocks handling. For Nokia phones like N900 (with Samsung OneNAND flash, +256MiB in size, 128KiB PEBs) 1% was just fine. + + +

    May UBI be used on MLC flash?

    Yes, it may, as long as the flash is supported by the MTD layer. UBI does