PROBLEM: USB drive errors with newer kernel versions

DISCUSSION Alan Stern 1 participant 1 message May 27, 2015 2 messages not in the archive SCSI

Share & tools

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): 201505261938.11869.gdalmadge-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

Alan Stern

author

2015-05-27 15:06 UTC #2 lore raw

Quoted 1 line · Gary Dalmadge

Please see below. Please let me know if you need additional information.

Here's the problem, from right near the start of your trace:

Quoted 1 line

ffff8800cf87c000 2445467361 S Bo:1:004:2 -115 31 = 55534243 d2000000 00000000 00000a35 00000000 00000000 00000000 000000

0x35 is a SYNCHRONIZE CACHE command, telling the drive to write the contents of its internal cache out to the physical medium.

Quoted 9 lines

ffff8800cf87c000 2445467440 C Bo:1:004:2 0 31 >
ffff8800cf87c000 2445467445 S Bi:1:004:8 -115 13 <
ffff8800cf87c000 2445467659 C Bi:1:004:8 0 13 = 55534253 d2000000 00000000 01
ffff8800cf87c000 2445467682 S Bo:1:004:2 -115 31 = 55534243 d3000000 12000000 80000603 00000012 00000000 00000000 000000 ffff8800cf87c000 2445467811 C Bo:1:004:2 0 31 >
ffff8800bd45c6c0 2445468603 S Bi:1:004:8 -115 18 <
ffff8800bd45c6c0 2445468688 C Bi:1:004:8 0 18 = 70000500 0000000a 00000000 20000000 0000 ffff8800cf87c000 2445468695 S Bi:1:004:8 -115 13 <
ffff8800cf87c000 2445468811 C Bi:1:004:8 0 13 = 55534253 d3000000 00000000 00

The drive replied with an error status, saying that this was an illegal request with an invalid operation code. Obviously it wasn't, and just as obviously, your drive does not handle the SYNCHRONIZE CACHE command properly.

It makes sense that this would trigger the error-detection code added in commit 89fb4cd1f717 (scsi: handle flush errors properly). After all, it really was an error in the data-flushing path.

You wrote that the error appears with several different brands and types of USB drives. Almost certainly these drivers share a bug in the component that bridges the USB bus to the internal SATA bus.

The real question is what to do about it.

James, do you have any suggestions?

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html