Subject: kern/31428: umass (BBB?) problem on 3.0_BETA with USB memory stick
To: None <kern-bug-people@netbsd.org, gnats-admin@netbsd.org,>
From: None <jkuittinen@gmx.de>
List: netbsd-bugs
Date: 09/30/2005 14:10:00
>Number: 31428
>Category: kern
>Synopsis: umass (BBB?) problem on 3.0_BETA with USB memory stick
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: kern-bug-people
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Fri Sep 30 14:10:00 +0000 2005
>Originator: Jari Kuittinen
>Release: 3.0_BETA
>Organization:
>Environment:
NetBSD triacetin.boofar.tld 3.0_BETA NetBSD 3.0_BETA (GENERIC) #0: \
Wed Sep 21 19:14:45 UTC 2005 builds@works.netbsd.org:\
/home/builds/ab/netbsd-3/i386/200509210000Z-obj/home/builds/\
ab/netbsd-3/src/sys/arch/i386/compile/GENERIC i386
>Description:
My USB memory stick (Verbatim Store'n'Go 256MB) has some problems with 3.0_BETA. NetBSD detects it fine but all i/o with files larger than few kilobytes causes problems. Sometimes copying of some small text files can work without problems.
Plugged in the stick:
/netbsd: umass0 at uhub2 port 5 configuration 1 interface 0
/netbsd: --cut-- what's this empty line? --cut--
/netbsd: umass0: Verbatim Store 'n' Go, rev 2.00/1.00, addr 2
/netbsd: umass0: using SCSI over Bulk-Only
/netbsd: scsibus0 at umass0: 2 targets, 1 lun per target
/netbsd: sd0 at scsibus0 target 0 lun 0: <VBTM, Store 'n' Go, 1.04> disk removable
/netbsd: sd0: 245 MB, 980 cyl, 16 head, 32 sec, 512 bytes/sect x 501760 sectors
Reading a small (about 5MB file) from the stick (cp /mnt/foo.mp3 /tmp):
the led on the stick starts flashing, so there is some i/o going on.
1 minute passed: nothing happened, the led is still flashing
2 minutes passed: /netbsd: umass0: BBB reset failed, TIMEOUT
3 minutes passed: /netbsd: umass0: BBB bulk-in clear stall failed, TIMEOUT
4 minutes passed: /netbsd: umass0: BBB bulk-out clear stall failed, TIMEOUT
Then it keeps timeouting those two actions forever (tested for couple of hours). Canceling the copy with ctrl+c doesn't work, cp process doesn't die with sighup or sigkill before the stick is physically removed from the system.
/netbsd: umass0: at uhub2 port 5 (addr 2) disconnected
/netbsd: sd0(umass0:0:0:0): generic HBA error
/netbsd: sd0(umass0:0:0:0): generic HBA error
/netbsd: sd0 detached
/netbsd: scsibus0 detached
/netbsd: umass0 detached
Well. Atleast umount doesn't hang when unmounting /mnt after removing the stick :)
I really think this is not hardware related, but here's my dmesg:
http://soul.fiveam.org/~ascizo/triacetin.dmesg
My 3.0_BETA is from 20050921, although older betas did the same. 3.99.7 had the same problem as 3.0_beta has.
>How-To-Repeat:
doing the above
>Fix:
NetBSD 2.1 works fine (tested with 2.1RC5 netbsd-INSTALL kernel). Transfers are fast, everything works :)