#1905 BLOC Trial-3: 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:  new    
  Priority:  blocker   |   Milestone:  Trial-3
 Component:  hardware  |     Version:         
Resolution:            |    Keywords:         
  Verified:  0         |  
-----------------------+----------------------------------------------------
Comment (by wmb at firmworks.com):

 Replying to [comment:52 dwmw2]:
 > It may be possible to develop a non-destructive test, if we can write to
 and read from internal buffer of the NAND flash device rather than
 actually writing to the flash itself. That would allow units to be tested
 for this problem without using up flash erase cycles.

 I tried for most of the afternoon today to write the internal buffer and
 read it back, with little success.  The "random data output" command
 (05/e0) seems to work reasonably well, delivering data from the NAND
 chip's internal buffer into system memory.

 I didn't have much luck with the "random data input" command (85/-).  I
 tried to use it to put data into the buffer, then to use the "random data
 output" command to read it back.  I didn't manage to get back anything
 approximating what I tried to write.

 I tried using the 85 command after the 10/30 (read into memory) and 10/35
 (read for copy-back, only into the internal buffer) sequences - still no
 luck.  In the latter case, the 85 command changed the data that the
 subsequent 05/e0 sequence returned, but that data did not match what I
 thought I sent.

-- 
Ticket URL: <http://dev.laptop.org/ticket/1905#comment:54>
One Laptop Per Child <http://laptop.org/>