Andy Green andy at
Wed Oct 29 10:35:32 CET 2008

Hash: SHA1

Somebody in the thread at some point said:

| Andy, are the two items below (from stable-tracking, commit
| def57aeb4f2a5838e8ab814204af3b0465caeff6) just noise or do
| they indicate that something more sinister is going on there ?
| (Trying to find out why stable-tracking broke my WLAN ...)

There was a problem with a bad timer for a while turning up much later
with a NULL "base" member that caused big trouble... I suspected it was
WLAN but after a day or so I realized the problem was the then failing
WM8753 init / probe; it didn't clean up after itself somehow and much
later the timer created in there fired.  When I fixed the WM8753 init to
use the new i2c method it was all OK.

Actually thanks for pointing out I left that WARN_ON() in there because
I was puzzled why WLAN spewed a couple of warnings now, I'll fix it.

Make sure your tree has "fix-set-wlan-power-mgt-to-default-on.patch";a=commitdiff;h=b027fc1eacf71b3616b0eb77a9a4917093815e2a

that would explain it not putting in an appearance but maybe that is too

BTW you looked at NAND issue on stable-tracking, I looked further at it
today and temporarily replaced the new ECC implementation with the
2.6.24 one as a test.  It stopped the ECC error reports OK but mounting
jffs2 rootfs still failed with a spew of CRC errors at that level.  So
something else goes on.  I diffed the s3c2410.c nand thing against
2.6.24 but nothing jumped out.

- -Andy
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora -


More information about the openmoko-kernel mailing list