[PATCH] fix root cause of NAND trouble

Andy Green andy at openmoko.com
Sat Nov 1 04:39:09 CET 2008

Hash: SHA1

Somebody in the thread at some point said:
| Andy Green wrote:
|> Not sure this fix about memcpy from u32 is endian-clean (you need a
|> conversion macro to le32?), but good work!
| I wondered about this for a moment, too. If the bus was set up to
| switch the endianness of words, that would indeed be a problem.
| But then, I suppose the readsl would be wrong as well, wouldn't
| it ? So my readl should be at least consistent :)

I never used big-endian ARM, but it seems neither of us are quite sure
what it'd do.  In the datasheet they define the bits in the peripheral
registers with absolute bit numbers, not relative to endianness.  Can
you find anywhere it specifies what it does in the datasheet, then at
least we learn something.

It's the kind of thing you'll get jumped on for by the Lictors of Upstream.

- -Andy

Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org


More information about the openmoko-kernel mailing list