From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; de-DE; rv:1.7.7) Gecko/20050416 Red Hat/1.0.3-1.4.1 Firefox/1.0.3 Description of problem: When trying to activate my UMTS/3G PCMCIA card, which is known to the system as simple USB serial devices, the usbserial modules are a) failing to load, and b) when loaded manually, the devices are not created properly. Instead of creating the ttyUSB0 ttyUSB1 etc devices, it created <NULL>0 etc devices, with major and minor numbers = 0. Harald had a look into this and identified this as a bug in the usbserial module. System is an IBM X31 (26725KG), running RHEL4DT. messages, when PCMCIA card is plugged in: Apr 27 13:45:45 konstanz kernel: ohci_hcd 0000:03:00.0: OHCI Host Controller Apr 27 13:45:45 konstanz kernel: ohci_hcd 0000:03:00.0: irq 11, pci mem e0840000 Apr 27 13:45:45 konstanz kernel: ohci_hcd 0000:03:00.0: new USB bus registered, assigned bus number 5 Apr 27 13:45:45 konstanz kernel: ohci_hcd 0000:03:00.0: WARNING: OPTi workarounds unavailable Apr 27 13:45:45 konstanz kernel: hub 5-0:1.0: USB hub found Apr 27 13:45:45 konstanz kernel: hub 5-0:1.0: 2 ports detected Apr 27 13:45:55 konstanz kernel: ohci_hcd 0000:03:00.0: wakeup Apr 27 13:45:56 konstanz kernel: usb 5-1: new full speed USB device using address 2 [root@konstanz log]# udevinfo -q all -p /sys/class/tty/\<NULL\>0 P: /sys/class/tty/<NULL>0 N: <NULL>0 T: c M: 020600 S: O: root G: root F: L: 0 U: 1076 R: 0 when i manually called a MAKEDEV ttyUSB, the devices were created properly and the card runs OK. Version-Release number of selected component (if applicable): kernel-2.6.9-5.0.5.EL How reproducible: Always Steps to Reproduce: 1. plug in PCMCIA UMTS/3G card Actual Results: nada. Expected Results: USB Serial devices should be initialized and should be used as modem devices Additional info:
the kernel module usbserial does not provide a proper sysfs "dev" file... the path is scrambled and major and minor are "0"!!
So the '<NULL>'0 file is being created on (un)plug of the pcmcia card? Seems like the usb-serial 'probe' routine isn't creating the /sys/ files correctly. but i'm really not sure at this point. The file drivers/usb/serial/usb-serial.h has a 'dbg' MACRO, that might help a lot in tracking this down.
The '<NULL>'0 file is being created on _plug_ of the pcmcia card after loading usbserial by hand.
Created attachment 113784 [details] var/log/messages with hotplug debug turned on, on 2.6.9-6.39.EL var/log/messages with hotplug debug turned on, on 2.6.9-6.39.EL. devices show up as /sys/class/tty/<NULL>0 etc, not as ttyUSB0 etc as they should.
tried my chances with the U1 kernel from RHEL4-U1-re0421.0, no joy. see attachment above for detailled logs. udevinfo etc remains the same. tried to work around this by doing a MAKEDEV ttyUSB, and when the right /dev/ttyUSB0 device is there, the card is responsive.. the workaround shouldnt be the solution 'though.
I've tracked this down a bit...the /sys/class/tty/ttyUSBN file should be getting created by 'tty_register_device', which is called from drivers/usb/serial/bus.c, line 133. If i had to guess this routine is not being passed valid data for some reason. There is a 'dev_info' printk on the next line that would confirm this suspicion. We can turn that on, i think, but defining DEBUG, in include/linux/device.h. Also, i don't think dbg was properly turned on in comment #4, as i see no output from it. If i had h/w to test this locally, that would help...any suggestions?
Thank you for submitting this issue for consideration in Red Hat Enterprise Linux. The release for which you requested us to review is now End of Life. Please See https://access.redhat.com/support/policy/updates/errata/ If you would like Red Hat to re-consider your feature request for an active release, please re-open the request via appropriate support channels and provide additional supporting details about the importance of this issue.