Slow boot due to USB errors

[Alan Stern]

On Mon, 12 May 2014, poma wrote:

> ...
> > "initial 64-byte descriptor request timeout"
> > https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/tree/drivers/usb/core/hub.c#n67
> > https://www.kernel.org/doc/Documentation/kernel-parameters.txt
> > usbcore.initial_descriptor_timeout=
> > & Co.
> > 
> 
> I made several tests concerning the topic and incidentally found one bug. :)

> https://bugzilla.redhat.com/show_bug.cgi?id=1096786
> Unplug the device without unmounting, 
> 
> ...
> usb 2-1: USB disconnect, device number 2
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 4692 at fs/sysfs/group.c:216 sysfs_remove_group+0x93/0xa0()
> sysfs group ffffffff81cb1300 not found for kobject 'target8:0:0'
> CPU: 0 PID: 4692 Comm: umount Not tainted 3.14.3-200.fc20.x86_64 #1
> Call Trace:
>  [<ffffffff816ef242>] dump_stack+0x45/0x56
>  [<ffffffff8108a1bd>] warn_slowpath_common+0x7d/0xa0
>  [<ffffffff8108a23c>] warn_slowpath_fmt+0x5c/0x80
>  [<ffffffff81265598>] ? kernfs_find_and_get_ns+0x48/0x60
>  [<ffffffff81264233>] sysfs_remove_group+0x93/0xa0
>  [<ffffffff8146b543>] dpm_sysfs_remove+0x43/0x50
>  [<ffffffff81460b35>] device_del+0x45/0x1c0
>  [<ffffffff8148deb7>] scsi_target_reap_usercontext+0x27/0x40
>  [<ffffffff810a4e67>] execute_in_process_context+0x67/0x70
>  [<ffffffff8148f334>] scsi_target_reap+0xc4/0xf0
>  [<ffffffff81491366>] scsi_device_dev_release_usercontext+0xe6/0x120
>  [<ffffffff810a4e67>] execute_in_process_context+0x67/0x70
>  [<ffffffff8149127c>] scsi_device_dev_release+0x1c/0x20
>  [<ffffffff814600e2>] device_release+0x32/0xa0
>  [<ffffffff813560a7>] kobject_cleanup+0x77/0x1b0
>  [<ffffffff81355f60>] kobject_put+0x30/0x60
>  [<ffffffff81460407>] put_device+0x17/0x20
>  [<ffffffff8148268b>] scsi_device_put+0x3b/0x50
>  [<ffffffff81495e50>] scsi_disk_put+0x30/0x50
>  [<ffffffff814961b2>] sd_release+0x42/0xd0
>  [<ffffffff8122489c>] __blkdev_put+0x16c/0x1a0
>  [<ffffffff8122483e>] __blkdev_put+0x10e/0x1a0
>  [<ffffffff8122521c>] blkdev_put+0x4c/0x120
>  [<ffffffff811ebd64>] kill_block_super+0x44/0x70
>  [<ffffffff811ec06d>] deactivate_locked_super+0x3d/0x60
>  [<ffffffff811ec636>] deactivate_super+0x46/0x60
>  [<ffffffff81208dac>] mntput_no_expire+0xac/0x140
>  [<ffffffff8120a37d>] SyS_umount+0x9d/0x110
>  [<ffffffff816ff5e9>] system_call_fastpath+0x16/0x1b
> ---[ end trace 6a4cda17930756e9 ]---
> 
> BTW 3.15.0-0.rc5.git0.1.fc21.x86_64 doesn't have this bug.

This is a known bug, as you can tell from the fact that it has already 
been fixed.  (And, as you can tell by looking at the stack dump, the 
bug was in the SCSI layer rather than the USB layer.)

Alan Stern