On Sun, 2005-04-03 at 15:47, Wolfgang Denk wrote:
In message 1B848D3797CA9841B0A46694F055A67A02FBBA86@MW-AUS02.mtwk.sps.mot.com you wrote:
Here is the patch.
...
--- u-boot-pq3-base/drivers/cfi_flash.c 2004-08-06 13:08:19.000000000 -0500 +++ u-boot-pq3-for-release/drivers/cfi_flash.c 2004-10-27 16:02:34.028749216 -0500 @@ -10,6 +10,11 @@
- Ed Okerson
- Modified to work with little-endian systems.
- Copyright (C) 2004
- York Sun
- Modified to support 8-bit port and 32-bit chip, not verified
- Modified to support polling DQ7 and DQ6 for AMD flash
The patch was corrupted by your mailer; I restored it. Nevetheless, there were a few problems with some boards -- I guess you didn't bother to run the "MAKEALL" script, did you?
True. And I've been bugging folks locally to fix it too... :-)
I tried to fix the problems, but you should have a look at my changes and re-test.
My own tests are not so positive - I got a "Timeout writing to Flash" error on the first board I tried (which was working fine before applying the patch).
Wolfgang,
There was a follow-up bug fix that might be important here. It fixes the problem with the retcode/flag typo here:
@@ -1176,7 +1402,8 @@ if (flag) enable_interrupts ();
- return flash_full_status_check (info, 0, info->write_tout, "write"); + retcode = flash_full_status_check (info, cptr, cword, info->write_tout, "write"); + return flag;
Should return "retcode" now, not "flag".
Thanks, jdl