#1905 BLOC Trial-2: Field Return: flash corruption - OpenFirmware complaining of 'unknown node type 2006'.

[Zarro Boogs per Child]

#1905: Field Return: flash corruption - OpenFirmware complaining of 'unknown node
type 2006'.
-----------------------+----------------------------------------------------
  Reporter:  dwmw2     |       Owner:  wad     
      Type:  defect    |      Status:  assigned
  Priority:  blocker   |   Milestone:  Trial-2 
 Component:  hardware  |     Version:          
Resolution:            |    Keywords:          
  Verified:  0         |  
-----------------------+----------------------------------------------------
Comment (by dwmw2):

 I've added a CONFIG_JFFS2_FS_WBUF_VERIFY option to the kernel, which will
 read back the page after each time we write the page-buffer (pages which
 are directly written completely because a single node is large enough to
 completely fill the page won't get the same treatment). It will attempt to
 recover by writing the affected nodes out again elsewhere.

 This will catch errors introduced in DMA transfers, but not if the error
 is just corruption in the RAM itself. But if we were seeing RAM corruption
 I'd have expected it to have manifested itself in other ways too --
 including the occasional node with a CRC32 matching the corrupted version
 of the contents instead of the original.

-- 
Ticket URL: <http://dev.laptop.org/ticket/1905#comment:23>
One Laptop Per Child <http://laptop.org/>