While working on my system somtimes one of my USB devices suddenly "freeze". Mouse does not move or keyboard doesn't work anymore. When my USB headset freeze I get problems shutting down the computer because it hangs while unloading the usb-audio module. The devices don't freeze together (I can still use the keyboard when the mouse crash), but the problem spreads (switching to a text console with CTRL+Alt+F1 freeze the keyboard when the mouse is already gone). I was unable to find a single message in the systems logs that could tell me what is happening. I had this bug both with gentoo-sources and vanilla-sources. Reproducible: Sometimes Steps to Reproduce: ASUS P5N-E SLI Board with Core2 Duo 6600. Cherry eVolution Stream keyboard Logitech MX400 mouse Logitech USB headset
Created attachment 122714 [details] Output of lspci
Created attachment 122715 [details] Output of dmesg
Created attachment 122717 [details] Output of emerge --info
Created attachment 122720 [details] /usr/src/linux/.config
Can you test with kernel sources 2.6.22-rc5? This is the latest stable prepatch as of this comment. A note on bugzilla states that much USB work was done in 2.6.22-rc4. Do you remember the last working kernel
I will test 2.6.22 over the weekend... I don't remember the last "working" kernel because I accused the OSS interface of Teamspeak for the headset hangup a log time (beginning of the year ?)... but when I got a USB keyboard and mouse I discovered it's a generic USB bug (and much more difficult to ignore).
Okay, I tested 2.6.22-rc5 over the weekend. First I thought the new kernel solved my problem (had no lockup on Saturday), but I lot's of them on Sunday and Monday... Just installed 2.6.22-rc6 to test it too.
Had three USB freeze (2x mouse, 1x keyboard) today. So the problem (I think it is some kind of deadlock) is still present in 2.6.22-rc6.
Can you enable CONFIG_USB_DEBUG on 2.6.22-rc and post dmesg logs from after when one of the devices breaks?
The dmesg above (see attachment) was made after my mouse "crashed"... and I had CONFIG_USB_DEBUG activated this time (kernel 2.6.21). Is there another kernel option I can activate to get more logs ?
Just downloaded a Kubuntu 7.04 CD and did some experiments... had a mouse freeze within minutes, so it's not a special Gentoo problem. I don't have this problem with Windows XP, so I think it's unlikely that it is a hardware defect.
For more logging, you can enable CONFIG_HID_DEBUG. Could you post the output of "lsusb -v", both before and after a device locks up? From your dmesg output, it looks like you're using an external USB hub. If these devices (kbd, mouse, headset) are connected to this external hub, could you please test without it, connecting the devices to internal USB ports?
Created attachment 123710 [details] lsusb -v same output before and after mouse freeze
I don't have an USB hub... I only use the 8 USB ports of my Asus mainboard (4 front, 4 aft). I have done a BIOS update today, maybe it will solve my problem... if I get a freeze again I will test 2.6.22-rc7.
I'm having almost the same problem with kernel 2.6.20-gentoo-r8. It occurs to me when (and only when) I plug in a USB drive, every time but not immediately. The drive mounts okay and works for a while, but after navigating too much the WHOLE USB system, including keyboard and mouse, is dead. Doing updatedb is a sure kill. A PS/2 keyboard/mouse still works, so I can see that there is nothing in /var/log/messages. Also nothing new registers on the usb. Doing lsusb works once, but the second time hangs the shell. I recorded udevmonitor --env through a crash but didn't notice anything special. Just as Henning says shutdown hangs and the only way to restart is with the power button. Should I start a new bug or post the usual data here? It's a new install so I don't have an earlier working kernel, but other machines are working. This time, for the first time, I used genkernel. Maybe there's some conflicting setting?
Yes, please file a separate bug report for this. The original (this) bug is reproducable only sometimes, while for you running updatedb is a "sure kill". Please attach your .config, dmesg output, and output from lspci -v and lsusb -v. Thanks. If you want, you could always compile a kernel manually instead of using genkernel, to see if that works for you.
I nearly killed my motherboard with the bios update (no POST, no boot, ect.) and had to pull out the cmos battery, but now I my system seems to run without any USB freeze. I will keek an eye on the problem but it seems the bios update solved it for my ASUS board. Maybe it is a small bug in the USB controller. It seems windows has a workaround built in but Linux has not...
What model mobo do you have? Do you have a link to the bios update you downloaded? Maybe there's a changelog of some sort that could shed some light on this. If you've gone four days without any freezes then it would seem the bios update did the trick though.
See first post: ASUS P5N-E SLI Board with Core2 Duo 6600. Cherry eVolution Stream keyboard Logitech MX400 mouse Logitech USB headset I just downloaded the bios file directly from the ASUS webpage, it contained no explanation file or change summary (only a binary image of the bios).
Not a single freeze the whole week... I think we can close this bug now.
Great, thanks for all the testing.