On Thu, 13 Sep 2012, Don Zickus wrote: > Hi Alan, > > I adapted your patch to our 2.6.32 tree and the customer tested it without > success. The output panic is attached below. I will work on getting a > machine with the latest kernel to reproduce this problem so I don't waste > your time chasing something that might be fixed upstream. > > But if you could take a quick glance at the panic below to see if anything > comes to mind I would appreciate it. > The test failed on the first surprise removal of PCI devices. Last gasp > from the console is posted below. I would guess the faulting process was > reading a file under /proc/bus/usb. I should have a dump if more info is > needed. > ehci_hcd 0000:2c:00.0: HC died; cleaning up > ehci_hcd 0000:2c:00.0: force halt; handhake ffffc90000654024 00004000 > 00004000 -> -19 > ehci_hcd 0000:2c:00.0: HC died; cleaning up > ehci_hcd 0000:2c:00.0: remove, state 0 > usb usb1: USB disconnect, device number 1 > usb 1-1: USB disconnect, device number 2 > usb 1-1.1: USB disconnect, device number 3 > hub 4-1:1.0: unable to enumerate USB device on port 3 > usb 1-1.3: USB disconnect, device number 4 > usb 1-1.6: USB disconnect, device number 5 > usb 1-1.6.1: USB disconnect, device number 6 > ehci_hcd 0000:2c:00.0: USB bus 1 deregistered > hub 4-1:1.0: unable to enumerate USB device on port 3 > general protection fault: 0000 [#1] SMP > last sysfs file: > /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/0000:02:01.0/0000:3d:00.0/0000:3e:01.0/0000:66:00.0/usb3/3-1/3-1.6/3-1.6.1/3-1.6.1:1.2/input/input12/event12/uevent > CPU 0 > Modules linked in: autofs4 sunrpc configfs cachefiles fscache(T) bonding > 8021q garp stp llc vhost_net macvtap macvlan tun uinput ipmi_devintf > ftmod(P)(U) ipmi_msghandler sg matroxfb(U) fosil(U) ext4 mbcache jbd2 > raid1 sr_mod cdrom sd_mod(U) crc_t10dif usb_storage mpt2sas(U) > scsi_hbas(U) scsi_transport_sas raid_class igb(U) dca dm_mirror > dm_region_hash dm_log dm_mod ipv6 cxgb4 cxgb3 mdio libiscsi_tcp libiscsi > scsi_transport_iscsi [last unloaded: scsi_wait_scan] > > Pid: 32752, comm: cat Tainted: P --------------- T > 2.6.32-279.el6.bz849188.test01.x86_64 #1 Stratus ftServer 2700/G7LAY > RIP: 0010:[<ffffffff813b8167>] [<ffffffff813b8167>] usb_device_dump+0x87/0xa70 It would help to know what source statement that memory address corresponds to. Offhand I can't see any remaining races between usb_device_dump() and usb_disconnect(). Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
