[Bug 218890] reset SuperSpeed USB device number 2 using xhci_hcd

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=218890

--- Comment #1 from Alan Stern (stern@rowland.harvard.edu) ---
There are at least two very suspicious-looking things in the usbmon trace:

ffff9244d6cf5000 1301812804 S Ci:6:004:0 s 80 06 0301 0409 00ff 255 <
ffff9244d6cf5000 1306956154 C Ci:6:004:0 -2 255 = 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000
ffff9244d6cf5000 1306956185 S Ci:6:004:0 s 80 06 0301 0409 0002 2 <
ffff9244d6cf5000 1310626032 C Ci:6:004:0 0 2 = 2a03

The first two lines show the computer asking the drive to send one of its
string descriptors and the transfer timing out after 5 seconds.  The trace says
that the device sent 255 bytes of zeros, which seems unlikely, but if it is
true it should have caused the transfer to end right away.  The second two
lines show the computer asking for just the first two bytes of the string
descriptor and the response arriving correctly, but not until an additional 3.7
seconds have passed.

ffff9244d6cf5300 1311630806 S Bo:6:004:2 -115 31 = 55534243 05000000 00020000
80000a28 00000000 00000001 00000000 000000
ffff9244d6cf5300 1311630823 C Bo:6:004:2 0 31 >
ffff9244ea669300 1311630835 S Bi:6:004:1 -115 512 <
ffff9244ea669300 1342568184 C Bi:6:004:1 -104 512 = 00000000 00000000 00000000
00000000 00000000 00000000 00000000 00000000

This shows the computer trying to read the first sector of the drive.  The
transfer times out after 30 seconds (again after 512 bytes of zeros are
apparently received), and this error is what causes the reset.

It's possible that the problem is in the device, but more likely the problem is
in the kernel.  You've made a good start by finding that the 6.6.1 kernel works
properly and the 6.7.7 kernel doesn't.  You should try doing a git bisection to
find out which commit in the interval between those kernels caused the problem
to appear.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.