kern/51928: Errors reading a USB pen drive on macppc 7.99.59

[dhgutteridge%sympatico.ca@localhost]

>Number:         51928
>Category:       kern
>Synopsis:       Errors reading a USB pen drive on macppc 7.99.59
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Jan 31 02:05:00 +0000 2017
>Originator:     David H. Gutteridge
>Release:        current as of Jan 30, 2017
>Organization:
>Environment:
NetBSD arcusv.nonus-porta.net 7.99.59 NetBSD 7.99.59 (ARCUSV_DEBUG) #1: Mon Jan 30 18:36:13 EST 2017  disciple%arcus-v3.nonus-porta.net@localhost:/home/disciple/netbsd-current/src/sys/arch/macppc/compile/obj/ARCUSV_DEBUG macppc

>Description:
While testing PR kern/51528, I've encountered another issue. The
original USB pen drive I was testing with now mounts and functions
without errors, but with another USB pen drive (much newer and larger)
I get:

umass0 at uhub0 port 1 configuration 1 interface 0
umass0: Kingston DataTraveler 3.0, rev 2.10/1.10, addr 2
umass0: using SCSI over Bulk-Only
scsibus0 at umass0: 2 targets, 1 lun per target
sd0 at scsibus0 target 0 lun 0: <Kingston, DataTraveler 3.0, PMAP> disk removable
sd0: fabricating a geometry
sd0: 59136 MB, 59136 cyl, 64 head, 32 sec, 512 bytes/sect x 121110528 sectors
sd0: fabricating a geometry
sd0: no NetBSD disk label

At which point I tried to mount it ("mount_msdos -u 1000 -m 0644
-M 0755 -o nodev -o noexec -o nosuid /dev/sd0a /mnt"), which
succeeded, with only a repetition of an earlier kernel message:

sd0: no NetBSD disk label

At which point I tried "mkdir /mnt/test", which resulted in:

sd0a: error writing fsbn 41024 of 41024-41087 (sd0 bn 49088; cn 23 tn 62 sn 0)
sd0a: error writing fsbn 18880 of 18880-19007 (sd0 bn 26944; cn 13 tn 10 sn 0)
sd0a: error writing fsbn 33664 of 33664-33791 (sd0 bn 41728; cn 20 tn 24 sn 0)
sd0a: error writing fsbn 1 (sd0 bn 8065; cn 3 tn 60 sn 1)
sd0a: error reading fsbn 41024 of 41024-41087 (sd0 bn 49088; cn 23 tn 62 sn 0)
sd0a: error reading fsbn 18880 of 18880-19007 (sd0 bn 26944; cn 13 tn 10 sn 0)

At which point I tried "ls -l /mnt", which resulted in:

sd0a: error reading fsbn 41024 of 41024-41087 (sd0 bn 49088; cn 23 tn 62 sn 0)
sd0a: error reading fsbn 41024 of 41024-41087 (sd0 bn 49088; cn 23 tn 62 sn 0)

At which point I tried "umount /mnt", which resulted in:

umass0: BBB bulk-out clear stall failed, TIMEOUT
sd0: cache synchronization failed
sd0: detached
scsibus0: detached
ohci_close_pipe: pipe not empty sed=0x1004db40 hd=0xb0018191 tl=0x100f2840 pipe=0x11b950a8, std=0x0
ohci_close_pipe: pipe still not empty
umass0: detached
umass0: at uhub0 port 1 (addr 2) disconnected

After that, all the data on that drive was lost. This was a drive that
was functioning without any issues on Linux and MacOS X. (I hadn't
tried it on NetBSD/amd64, but suspect it would've worked fine there
too. I'll try that once I reimage it.)

I don't have USB debug output for it yet, if that's relevant, but I've
built a kernel that supports that (and OHCI debug) now, and can
follow up. I didn't see any pre-existing PRs about this, but I may
have missed something, so I thought I'd report it now.

Dave

>How-To-Repeat:
Mount a USB pen drive like the one noted above.
>Fix:
Unknown.