On Fri, Jul 09, 2004 at 08:32:57PM -0600, Xander Hsia wrote: > Anybody else got one of these and notice it takes a really long time > from the time you issue a mount command and the device is actually > mounted? I have the verbose option turned on in the kernel and the log > shows it takes a minute and 10 seconds to complete the mount from the > time it is issued. The logs from turning on verbose usb-storage seem to > indicate the flash drive goes thru several soft resets before finally > mounting. This can't be normal behavior, right?
Normal, no. But explainable, given your logs.
> Jul 9 16:01:48 asengard usb-storage: *** thread awakened.
> Jul 9 16:01:48 asengard usb-storage: Command TEST_UNIT_READY (6 bytes)
> Jul 9 16:01:48 asengard usb-storage: 00 00 00 00 00 00
> Jul 9 16:01:48 asengard usb-storage: Bulk Command S 0x43425355 T 0x433
> L 0 F 0 Trg 0 LUN 0 CL 6
> Jul 9 16:01:48 asengard usb-storage: usb_stor_bulk_transfer_buf: xfer
> 31 bytes
> Jul 9 16:01:48 asengard usb-storage: Status code 0; transferred 31/31
> Jul 9 16:01:48 asengard usb-storage: -- transfer complete
> Jul 9 16:01:48 asengard usb-storage: Bulk command transfer result=0
> Jul 9 16:01:48 asengard usb-storage: Attempting to get CSW...
> Jul 9 16:01:48 asengard usb-storage: usb_stor_bulk_transfer_buf: xfer
> 13 bytes
> Jul 9 16:01:48 asengard usb-storage: Status code 0; transferred 13/13
> Jul 9 16:01:48 asengard usb-storage: -- transfer complete
> Jul 9 16:01:48 asengard usb-storage: Bulk status result = 0
> Jul 9 16:01:48 asengard usb-storage: Bulk Status S 0x53425355 T 0x433 R
> 0 Stat 0x0
> Jul 9 16:01:48 asengard usb-storage: scsi cmd done, result=0x0
> Jul 9 16:01:48 asengard usb-storage: *** thread sleeping.
This section of log is a very good sign. It shows that we're communicating
with the device properly -- that is, we're exchanging meaningful and valid
data encapsulated properly.
But, sometime later, we get:
> Jul 9 16:03:31 asengard usb-storage: Bulk Status S 0x53550000 T
> 0x4405342 R 0 Stat 0x0
> Jul 9 16:03:31 asengard usb-storage: Bulk logical error
And this is not good. In fact, it looks like the USB packet has been very
badly corrupted. The first value there should be 0x53425355 -- to me, it
looks like several bytes in the packet have been swapped around (or perhaps
just pushed around out of alignment?). The tag is scrambled also (it
should be 0x440). Given this input, the driver declares a logical error
and follows the required reset procedure.
The good news: the reset procedure worked (eventually) and your system
didn't crash.
The bad news: this should have never happened in the first place.
My top two guesses:
(1) HCD bug
(2) oddly broken hardware
Matt
--
Matthew Dharm Home: [EMAIL PROTECTED]
Maintainer, Linux USB Mass Storage Driver
God, root, what is difference?
-- Pitr
User Friendly, 11/11/1999
pgpSDjjL2vL60.pgp
Description: PGP signature
