From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.3) Gecko/20030312 Description of problem: I am using version 1.2 beta 1 of the Speedtouch USB driver from http://speedtouch.sourceforge.net with a Speedtouch USB modem (http://www.speedtouch.com/pdf/datasheetusb.pdf). This was working without any problems with kernel 2.4.20-13.7. I have just upgraded to the latest errata kernel 2.4.20-18.7 and I now get a kernel oops when starting the driver (starting the modem_run program which uploads the firmware to the device). I have attached the relevant portion from /var/log/messages to this bug. Version-Release number of selected component (if applicable): kernel-2.4.20-18.7 How reproducible: Always Steps to Reproduce: 1. Install Speedtouch 1.2 beta 1 drivers. 2. Upgrade to errata kernel 2.4.20-18.7. 3. Start modem_run. Expected Results: I've attached the messages that usually get logged when starting the Speedtouch modem with kernel 2.4.20-13.7. Additional info: My USB card is a PCI card. The following is reported in /proc/pci (when running kernel 2.4.20-13.7): Bus 0, device 15, function 0: USB Controller: OPTi Inc. 82C861 (rev 16). IRQ 11. Master Capable. Latency=64. Non-prefetchable 32 bit memory at 0xffbea000 [0xffbeafff]. The natsemi module referenced is the driver for eth0.
Created attachment 92243 [details] Kernel oops on starting modem_run
Created attachment 92244 [details] Messages logged when starting the Speedtouch modem with kernel 2.4.20-13.7
You should report speedtouch problems to the speedtouch maintainers not to Red Hat. Thanks
I have reported this to the Speedtouch maintainers too. I thought it could be a problem with the latest errata kernel because the RHSA-2003:187-25 advisory lists 'Driver bug fixes are included... for the USB ohci driver'. This is a kernel oops that involves the USB ohci driver, so I thought something may have been broken between 2.4.20-13.7 and 2.4.20-18.7.
Just to clarify - the Speedtouch USB driver I am using is a userspace driver (the ATM connection is terminated in userspace). This crash occurs in a clean (no patches applied) Red Hat kernel.
No problem, reopening. It is possible that the ohci fix is at fault. Unfortunately, it was a very important fix. I cannot back it out.
I've also seen this problem (but haven't tested on the 2.4.20-13.7 kernel). When using usb-uhci there is no kernel oops on my system here. The system I have seen the oops on was either using usb-ohci or ehci-hcd. I'm pretty sure it was the ohci usb controller. Unfortuately it's a remote box so I can't easily check now.
I faced the same problem using RedHat 9.0 after upgrading to from 2.4.20-8 to 2.4.20-18.9 and I have usb-ohci Regards, Emad
This is still a problem with the new errata kernel 2.4.20-19.7 (https://rhn.redhat.com/errata/RHSA-2003-238.html).
Thanks, I know. I didn't fix it yet.
Created attachment 93158 [details] Test fix #1 (ohci6)
Original entry for the ohci disaster was in bug 84814, in case anyone is wondering why we cannot just back out ohci changes from 2.4.20-13.
CVS 2.4.20-19+