Bug 4414
Summary: | Problem: USB bus hangs after 'lsusb -vv' with HP psc1210 printer/scanner | ||
---|---|---|---|
Product: | Drivers | Reporter: | Joel (jd02390) |
Component: | USB | Assignee: | Greg Kroah-Hartman (greg) |
Status: | REJECTED INVALID | ||
Severity: | normal | CC: | stern |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.11.5 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: |
Output from dmesg before running Andrew's script
Output from cat (debugfs)/uhci/*7.2 Output from dmesg after running Andrew's script Followup to last comment |
Description
Joel
2005-03-27 22:17:25 UTC
Please make things hang again and then do dmesg -c echo t > /proc/sysrq-trigger dmesg -s 1000000 > foo and attach foo to this bug. Thanks. Here's some more debugging you can try. Turn on USB verbose debugging in the kernel configuration (CONFIG_USB_DEBUG) and rebuild the USB drivers. Since you appear to have your USB host controller drivers built into the kernel, add the "uhci-hcd.debug=3" parameter to the kernel boot line. If instead you build uhci-hcd as a module, use the "debug=3" parameter on the modprobe line. Then mount a debugfs filesystem (say on /sys/kernel/debug), and after the hang occurs make a copy of /sys/kernel/debug/uhci/*7.2 and post it. Also post the output from dmesg after the hang but _before_ running the script that Andrew suggested. Created attachment 4814 [details]
Output from dmesg before running Andrew's script
Created attachment 4815 [details]
Output from cat (debugfs)/uhci/*7.2
Created attachment 4816 [details]
Output from dmesg after running Andrew's script
Did you notice the last two lines in your first dmesg log: uhci_hcd 0000:00:07.2: host system error, PCI problems? uhci_hcd 0000:00:07.2: host controller halted, very bad! They indicate a hardware-related problem of some sort. Maybe something wrong with the USB controller itself, or maybe something wrong with the motherboard. You may be able to clear the error by doing "rmmod uhci-hcd" followed by "modprobe uhci-hcd". But that's just a temporary solution, since the problem will most likely recur. In the end you may have no choice but to replace the motherboard. Created attachment 4843 [details]
Followup to last comment
The fact that it works with Windows doesn't mean much. Linux uses the hardware in different ways. In case you're interested, here's what the hardware manual has to say about that error: Host System Error. The Host Controller sets this bit to 1 when a serious error occurs during a host system access involving the Host Controller module. In a PCI system, conditions that set this bit to 1 include PCI Parity error, PCI Master Abort, and PCI Target Abort. Just how serious those three possibilities are is a matter of opinion. You can always try buying an add-on PCI USB controller card, in case the problem is in the onboard controller. Does 2.6.12-rc5 still exhibit this problem? If so, are we screwed? 2.6.12-rc5 still exhibits the problem -- the behavior is exactly the same. I guess we're screwed. Oh well, I took Alan's advice and bought a $10 add-on usb card, and it works fine. |