Hello, I'm seeing what appears to be a deadlock between khubd disconnecting a malfunctioning USB device (due to EMI likely) and a userspace app querying a sysfs attribute for the device at the same time. This is with Linux 2.6.29.1 - I'm building 2.6.37 to test now but I'm wondering if someone has seen something like this before and can point me at patches. I found: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=e16362a0c8d90e9adbfe477acbe32b021823fb22 which seems relevant, but is already applied to our tree. Here is the SysRq-w (blocked tasks) output: Jan 28 15:08:00 localhost kernel: [53940.956991] SysRq : Show Blocked State Jan 28 15:08:00 localhost kernel: [53940.957027] task PC stack pid father Jan 28 15:08:00 localhost kernel: [53940.957065] khubd D 00002fcf 5348 27 2 Jan 28 15:08:00 localhost kernel: [53940.957111] f6b49d94 00000046 fa11db42 00002fcf 00000000 f6aac3a0 c09525c4 c0a0a7cc Jan 28 15:08:00 localhost kernel: [53940.957177] c0a0d880 c0a0d880 c0a0d880 00000046 f6b466e0 f6b46948 c9d06880 00000000 Jan 28 15:08:00 localhost kernel: [53940.957238] fa15c269 00002fcf f6b49d60 f5da979c f70184c0 00000246 f6b49d78 f6b46948 Jan 28 15:08:00 localhost kernel: [53940.957300] Call Trace: Jan 28 15:08:00 localhost kernel: [53940.957340] [<c07a1166>] schedule_timeout+0x1c/0xc3 Jan 28 15:08:00 localhost kernel: [53940.957371] [<c04262c3>] ? get_parent_ip+0x10/0x36 Jan 28 15:08:00 localhost kernel: [53940.957396] [<c07a5472>] ? sub_preempt_count+0x94/0xa3 Jan 28 15:08:00 localhost kernel: [53940.957422] [<c07a0685>] wait_for_common+0xa1/0xee Jan 28 15:08:00 localhost kernel: [53940.957448] [<c0428ef5>] ? default_wake_function+0x0/0x12 Jan 28 15:08:00 localhost kernel: [53940.957475] [<c07a0765>] wait_for_completion+0x17/0x19 Jan 28 15:08:00 localhost kernel: [53940.957504] [<c04da08d>] sysfs_addrm_finish+0x19d/0x1c9 Jan 28 15:08:00 localhost kernel: [53940.957532] [<c04d8c16>] sysfs_hash_and_remove+0x42/0x55 Jan 28 15:08:00 localhost kernel: [53940.957559] [<c04daf82>] sysfs_remove_group+0x7e/0xad Jan 28 15:08:00 localhost kernel: [53940.957587] [<c0619cf6>] device_remove_attrs+0x3a/0x9a Jan 28 15:08:00 localhost kernel: [53940.957611] [<c0619e49>] device_del+0xf3/0x151 Jan 28 15:08:00 localhost kernel: [53940.957634] [<c069f246>] usb_disable_device+0x72/0xd0 Jan 28 15:08:00 localhost kernel: [53940.957658] [<c069b33c>] usb_disconnect+0x7b/0xe3 Jan 28 15:08:00 localhost kernel: [53940.957679] [<c069c092>] hub_thread+0x55a/0xe08 Jan 28 15:08:00 localhost kernel: [53940.957704] [<c044058a>] ? autoremove_wake_function+0x0/0x38 Jan 28 15:08:00 localhost kernel: [53940.957726] [<c069bb38>] ? hub_thread+0x0/0xe08 Jan 28 15:08:00 localhost kernel: [53940.957747] [<c044026c>] kthread+0x40/0x66 Jan 28 15:08:00 localhost kernel: [53940.957766] [<c044022c>] ? kthread+0x0/0x66 Jan 28 15:08:00 localhost kernel: [53940.957788] [<c0403cd7>] kernel_thread_helper+0x7/0x10 Jan 28 15:08:00 localhost kernel: [53940.957860] qt_gui D 00002fcf 5032 8745 6257 Jan 28 15:08:00 localhost kernel: [53940.957896] f2801e98 00200046 f8e74ff6 00002fcf 00000000 f3e300a0 c0d952af c0a0a7cc Jan 28 15:08:00 localhost kernel: [53940.957949] c0a0d880 c0a0d880 c0a0d880 00200246 f5d2a0a0 f5d2a308 c9d06880 00000000 Jan 28 15:08:00 localhost kernel: [53940.958001] f8f94847 00002fcf f2801e64 c04262c3 00000001 c9aca21c f2801e74 f5d2a308 Jan 28 15:08:00 localhost kernel: [53940.958054] Call Trace: Jan 28 15:08:00 localhost kernel: [53940.958074] [<c04262c3>] ? get_parent_ip+0x10/0x36 Jan 28 15:08:00 localhost kernel: [53940.958095] [<c07a5472>] ? sub_preempt_count+0x94/0xa3 Jan 28 15:08:00 localhost kernel: [53940.958117] [<c041b517>] ? kunmap_atomic+0x52/0x68 Jan 28 15:08:00 localhost kernel: [53940.958138] [<c07a1166>] schedule_timeout+0x1c/0xc3 Jan 28 15:08:00 localhost kernel: [53940.958159] [<c04262c3>] ? get_parent_ip+0x10/0x36 Jan 28 15:08:01 localhost kernel: [53940.958178] [<c07a5472>] ? sub_preempt_count+0x94/0xa3 Jan 28 15:08:01 localhost kernel: [53940.958199] [<c07a1e5e>] __down_common+0x83/0xc0 Jan 28 15:08:01 localhost kernel: [53940.958220] [<c07a1efa>] __down+0x17/0x19 Jan 28 15:08:01 localhost kernel: [53940.958241] [<c04440d9>] down+0x27/0x36 Jan 28 15:08:01 localhost kernel: [53940.958262] [<c06a37b1>] show_supports_autosuspend+0x20/0x57 Jan 28 15:08:01 localhost kernel: [53940.958283] [<c06a3791>] ? show_supports_autosuspend+0x0/0x57 Jan 28 15:08:01 localhost kernel: [53940.958305] [<c0619805>] dev_attr_show+0x1e/0x3b Jan 28 15:08:01 localhost kernel: [53940.958325] [<c04d939a>] sysfs_read_file+0x97/0xf4 Jan 28 15:08:01 localhost kernel: [53940.958346] [<c04d9303>] ? sysfs_read_file+0x0/0xf4 Jan 28 15:08:01 localhost kernel: [53940.958367] [<c049e7c4>] vfs_read+0x7e/0xaf Jan 28 15:08:01 localhost kernel: [53940.958386] [<c049e898>] sys_read+0x40/0x65 Jan 28 15:08:01 localhost kernel: [53940.958406] [<c0403402>] syscall_call+0x7/0xb qt_gui is a QT console application. As I noted, I'm going to rerun with 2.6.37, and gather more info (enable lockdep, etc), but I'm hoping someone has fixed this already (we need to stick with 2.6.29.1 for various reasons) and there are patches available. Thanks, Pete -- 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
