On Tue, 15 Aug 2006 [EMAIL PROTECTED] wrote: > Hi, > I think there may be a usb locking problem recently introduced by this > patch: > http://lkml.org/lkml/2006/1/4/458
I don't think that patch is wrong. Probably it just seems that way because some driver is holding a lock and not releasing it. > I got a report from a user with an wireless USB keyboard with 3 > interfaces: What version of the kernel was the user running? > T: Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#= 2 Spd=1.5 MxCh= 0 > D: Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs= 1 > P: Vendor=04f2 ProdID=0200 Rev= 0.03 > S: Manufacturer=Chicony > S: Product=USB Wireless HID Receiver > C:* #Ifs= 3 Cfg#= 1 Atr=a0 MxPwr=100mA > I: If#= 0 Alt= 0 #EPs= 1 Cls=03(HID ) Sub=01 Prot=01 Driver=usbhid > E: Ad=81(I) Atr=03(Int.) MxPS= 8 Ivl=10ms > I: If#= 1 Alt= 0 #EPs= 1 Cls=03(HID ) Sub=00 Prot=00 Driver=usbhid > E: Ad=82(I) Atr=03(Int.) MxPS= 8 Ivl=10ms > I: If#= 2 Alt= 0 #EPs= 1 Cls=03(HID ) Sub=01 Prot=02 Driver=usbhid > E: Ad=83(I) Atr=03(Int.) MxPS= 8 Ivl=10ms > > On boot, with this keyboard plugged the system hangs. What happens if the keyboard is plugged in after booting? > A trace shows > modprobes and a grep hanging in "D state" in usbcore: > > modprobe D C02E8A05 0 540 537 (NOTLB) > c1581d74 53051340 000f41fe c02e8a05 c1581d84 53051340 000f41fe 00000000 > c1580000 c1580000 c156268c c1562560 53051340 000f41fe c1580000 02cd29c0 > 00000000 c1580000 df7cad1c 00000286 df7cad24 c02ea205 c1562560 00000001 > Call Trace: > [<c02e8a05>] schedule+0x285/0x530 > [<c02ea205>] __down+0xd5/0x114 > [<c0119e60>] default_wake_function+0x0/0x20 > [<c0249100>] __driver_attach+0x0/0x70 > [<c02e8737>] __down_failed+0x7/0xc > [<c02492ad>] .text.lock.dd+0x27/0xca > [<c0249100>] __driver_attach+0x0/0x70 > [<c02482ab>] next_device+0xb/0x20 > [<c024831d>] bus_for_each_dev+0x5d/0x80 > [<c0248e65>] driver_attach+0x25/0x30 > [<c0249100>] __driver_attach+0x0/0x70 > [<c02486ec>] bus_add_driver+0x8c/0x180 > [<c02495bb>] driver_register+0x4b/0x90 > [<e0c60560>] usb_register_driver+0x60/0xf0 [usbcore] > [<c01aa61a>] sysfs_create_group+0x7a/0x130 > [<e084b019>] usb_kbd_init+0x19/0x3b [usbkbd] > [<c0139019>] sys_init_module+0x189/0x1a30 > [<e0c57680>] usb_buffer_alloc+0x0/0x50 [usbcore] > [<c0132700>] autoremove_wake_function+0x0/0x60 > [<c011760a>] do_page_fault+0x13a/0x650 > [<c0102fb7>] sysenter_past_esp+0x54/0x75 > > another modprobe just like this from usbmouse instead of usbkbd, and: > > grep D C15DBF50 0 740 739 (NOTLB) > c15dbef8 0bbc1000 000f41ff c15dbf50 c15dbf70 00000002 00000000 ded80000 > c15da000 c15da000 c15f168c c15f1560 0bbc1000 000f41ff c15da000 001e8480 > 00000000 c15da000 df7e0d1c 00000286 df7e0d24 c02ea205 c15f1560 00000001 > Call Trace: > [<c02ea205>] __down+0xd5/0x114 > [<c0119e60>] default_wake_function+0x0/0x20 > [<c02e8737>] __down_failed+0x7/0xc > [<e0c68c6e>] .text.lock.devices+0x4d/0x6f [usbcore] > [<c01635fe>] vfs_read+0xde/0x1b0 > [<c01642ab>] sys_read+0x4b/0x80 > [<c0102fb7>] sysenter_past_esp+0x54/0x75 What you need to do is find the task that is stuck holding the lock. There's a good chance it is khubd. So far all you know is that these tasks are trying to acquire the lock and can't. The question is: Why not? > It does not seem to occur with HZ=100, but with HZ=1000 it occurs > 9 out of 10 times. > > The grep is because the initscripts are doing: > needusbstorage=`LC_ALL=C grep -e "^I.*Cls=08" /proc/bus/usb/devices > 2>/dev/null` > > Not grepping makes booting proceed normally. > > Deleting the lines coming from above patch that aquire a lock on the parent > of the device > in linux/drivers/base/dd.c completely solves the problem. > > Unfortunately, I do not really know how this can lead to cyclic locking. It > may have something todo with the fact that this keyboard has 3 interfaces. > And we > somehow managed to aquire one lock causing devices.c to wait on us, but at > the same time we are trying to aquire a lock on something that devices.c > already has > locked. > > I hope that someone can provide some insight here. Otherwise, how safe is it > to remove the patch mentioned above from 2.6.17 to get a bit more stable > distro-kernel? The patch is necessary; it should not be removed. Alan Stern ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
