Stack dump on resume - GTA01 - what now?

Andy Green andy at openmoko.com
Fri Jul 4 20:55:42 CEST 2008


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:
| The patch that added the stack dump on suspended transfers works -- see
| the dmesg excerpt below.
|
| I presume that this was added to facilitate fixing the situation that
| got us into this state to begin with (resume ordering?)... or do I
| misunderstand something?

Yes it's not a panic or OOPS but a stack dump like you say, basically
some pcf50606 work queue decided to talk to the pcf50606 using i2c
before the i2c driver has resumed from suspend.  I added a stack dump in
that case so we can see who the culprit is.

In the end on GTA02/pcf50633 I got good results by disallowing pcf50633
interrupt service during suspend / resume:

http://git.openmoko.org/?p=kernel.git;a=commitdiff;h=eaa55efc0b559abebbcf2acea6ce267cea4c26f2

I guess it just needs porting to 06... if you send it on your own tree
let me know where to steal it from ;-)

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

iEYEARECAAYFAkhuci4ACgkQOjLpvpq7dMpbWACfXhuRD9LwOBi6UbC6y6nkWFOf
RMYAniIMzlNR8lbOiaxOa4D77cPyW/Jq
=7IVv
-----END PGP SIGNATURE-----




More information about the openmoko-kernel mailing list