From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux 2.4.3-12 i686; en-US; rv:0.9.1) Gecko/20010622 Description of problem: After applying the kernel upgrade to 2.4.3-12 as described on the Redhat errata page the USB Controller fails to initialize when the system is starting up. How reproducible: Always Steps to Reproduce: 1.Boot the system 2. 3. Actual Results: USB initialization fails with error message Expected Results: No error and working USB Additional info: Lines from /var/log/messages: Jul 18 00:28:37 hoelderlin modprobe: Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters Jul 18 00:28:44 hoelderlin kernel: Winbond Super-IO detection, now testing ports 3F0,370,250,4E,2E ... Jul 18 00:28:37 hoelderlin modprobe: /lib/modules/2.4.3-12/kernel/drivers/usb/usb-ohci.o: init_module: No such device Jul 18 00:28:44 hoelderlin kernel: Winbond chip at EFER=0x3f0 key=0x87 devid=52 devrev=f4 oldid=ff Jul 18 00:28:37 hoelderlin modprobe: /lib/modules/2.4.3-12/kernel/drivers/usb/usb-ohci.o: insmod /lib/modules/2.4.3-12/kernel/drivers/usb/usb-ohci.o failed Jul 18 00:28:44 hoelderlin kernel: Winbond chip type 83977EF / SMSC 97w35x Jul 18 00:28:37 hoelderlin modprobe: /lib/modules/2.4.3-12/kernel/drivers/usb/usb-ohci.o: insmod usb-ohci failed Mainboard: Gigabyte GA-7IXE4 (AMD Irongate Chip-Set) CPU: Duron 800Mhz RAM: 256MB 2 ISA slots used: Modem and AHA 1505 SCSI adaptor Graphics: ATI Expert 98 (Rage Pro) Network: Linksys 100LX Full dmesg at http://na.uni-tuebingen.de/~oliver/dmsg Bigger slice of /var/log/messages: http://na.uni-tuebingen.de/~oliver/messages
"usb-ohci.c: 00:07.4 (Advanced Micro Devices [AMD] AMD-756 [Viper] USB): blacklisted, erratum #4" AMD has released an defectsummar^Werratum that states that your specific chipset has a defective USB chip. It will appear to work but cause LOADS of problems if used. (lost interrupts, infinite interrupts etc etc)
I should remark that before the upgrade I was able to do useful work with USB, now I am not. In fact I didn't even notice that there was something wrong with the USB controller. How about a warning like "don't send vital data via USB", rather than blacklisting it altogether? Also, I specifically upgraded the kernel because it fixes those scary mtrr related warnings with the AMD Irongate Chipset. So it would be nice to finally have a flawless kernel for this Chipset.
My problem is similar and more basic. I have a sony vaio fx370 laptop. Originaly I installed 7.2 and immediately upgraded the kernel to 2.4.9-13. A appeared to work "except" nothing on the USB port. The USB port was recognized and the basic drivers were loaded. However, when I attached any usb device an error was always reported in the log that the USB device wouldn't accept new address (i.e. it was timing out). The /proc/interrupts count wasn't increasing. To insure it wasn't the laptop I loaded XP, verified the USB ports and reloaded 7.2. I discovered with the default kernel 2.4.7-10, low and behold the USB and sound worked perfectly. I could use my USB based printer and cd-writer. So, I upgrade again to 2.4.9-13 (using up2date) and once again everything broke. However, this time I had kept around the 2.4.7-10 kernel. Once again after booting 2.4.7-10 all worked. There is definitely a problem with the updated kernel with USB.
We must not keep bugs NEW for months, assigning to myself. Evidently, 2.4.7-10 used a different table somewhere. I'll work with bug reporters to dump their $PIRQ and MPTABLE as appropritate and see what's up. Well, hopefuly.. :)
I would appreciate if both of you guys did this: 1. boot with ok kernel and do "dmesg > /tmp/dmesg.2.4.7-10.ok" 2. boot with broken kernel and do "dmesg > /tmp/dmesg.2.4.9-xx.bad" 3. Attach both dmesg outputs to the bug (please DO NO DROP INTO COMMENTS BOX) If you are on 2.4.9-31 now, that would be helpful too (it does not fix anything AFAIK, but it would be one less diff session for me)
Marcel Oliver (requestor) reported success with 2.4.9-21.