2004-01-07 17:21:27

by Matthias Hentges

[permalink] [raw]
Subject: 2.6.* + suspend/resume kills USB

Hello all!

Fist of all let me thank the kernel developers for their _great_ work.
Kernel 2.6 is really amazing, especially on my old laptop (Celeron 600)
it worked pure wonders performance-wise. Thank you! :)

I'm running 2.6.0-mm2 right now which fixed some serious problems with
encrypted loop devices on my machine.

Now i've run into a problem with USB after a suspend/resume event on my
laptop. USB won't wakeup after a resume event which is a pain if you
happen to depend on an USB mouse (the build-in touchpad is toast...)

I've googled for more info on this issue and found a thread from
linux-usb-devel:
http://www.mail-archive.com/[email protected]/msg17660.html

The poster (Michael Vogt) had the exact same problem and came up with a
solution:

"When I change drivers/usb/host/uhci-hcd.c:suspend_allowed() to always
return 0 (don't suspend on uhci_suspend the but reset), usb seems to
work after a suspend/resume with apm."

This worked for me, too.

Now i wonder if a fix will be included into the 2.6 line.
The solution works ATM but it's a very "hacky" way to do it IMO.

I'm continuously checking the changelogs of the 2.6 -rc and -mm kernels
but didn't see any mention of the problem (or a fix) there, so i thought
i'd ask here if the bug is known and if someone is working on it or is
planning to do so (if i had any coding skills, i'd try it myself).

If someone is planning to work on it, please feel free to contact me
if you need a tester. I'll do what i can to help resolve this issue.

I have attached a lspci and dmesg dumps in case they are needed.

TIA and HAND
--

Matthias Hentges
Cologne / Germany

[http://www.hentges.net] -> PGP welcome, HTML tolerated
ICQ: 97 26 97 4 -> No files, no URL's

My OS: Debian Woody. Geek by Nature, Linux by Choice


Attachments:
lcpci.txt (795.00 B)
dmesg.txt (11.76 kB)
Download all attachments