Xorg 7.4 and umass devices - a bad combination?
Torfinn Ingolfsen
torfinn.ingolfsen at broadpark.no
Tue Mar 17 14:55:03 PDT 2009
On Tue, 17 Mar 2009 08:57:06 +0100
Hans Petter Selasky <hselasky at c2i.net> wrote:
> It's maybe an interrupt problem. Can you do some vmstat'ing before
> and after ?
After starting Xorg (with startx this time), but before doing an 'ls'
on one of the usb drives:
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 0 0 136M 2736M 512 4 11 0 438 0 0 0 198 1604 570 1 1 98
After issuing the 'ls' command (and still waiting for it to complete):
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 474 4 10 0 405 0 0 0 205 1484 573 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 470 4 10 0 402 0 0 0 206 1473 574 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 469 4 10 0 401 0 0 0 206 1469 574 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 468 4 10 0 400 0 0 0 206 1464 574 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 466 4 10 0 399 0 0 0 206 1459 574 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 467 4 10 0 399 0 0 0 207 1460 576 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 466 4 10 0 398 0 0 0 207 1456 576 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 464 4 10 0 397 0 0 0 208 1451 576 1 1 98
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
1 1 0 136M 2736M 463 4 9 0 396 0 0 0 208 1447 576 1 1 98
I can't see anything odd from these. Can you?
Now it is several minutes later,. the 'ls' command still hasn't completed, and I try vmstat once more:
root at kg-quiet# vmstat
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad0 ad1 in sy cs us sy id
0 1 0 136M 2736M 214 2 4 0 183 0 0 0 252 676 606 0 1 99
I now have several timeout messages in /var/log/messages:
Mar 17 22:48:56 kg-quiet kernel: umass1: BBB reset failed, TIMEOUT
Mar 17 22:48:56 kg-quiet kernel: umass3: BBB reset failed, TIMEOUT
Mar 17 22:48:56 kg-quiet kernel: umass2: BBB reset failed, TIMEOUT
Mar 17 22:48:56 kg-quiet kernel: umass0: BBB reset failed, TIMEOUT
Mar 17 22:50:01 kg-quiet kernel: umass1: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:50:01 kg-quiet kernel: umass3: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:50:01 kg-quiet kernel: umass2: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:50:01 kg-quiet kernel: umass0: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:51:06 kg-quiet kernel: umass1: BBB bulk-out clear stall failed, TIMEOUT
Mar 17 22:51:06 kg-quiet kernel: umass3: BBB bulk-out clear stall failed, TIMEOUT
Mar 17 22:51:06 kg-quiet kernel: umass2: BBB bulk-out clear stall failed, TIMEOUT
Mar 17 22:51:06 kg-quiet kernel: umass0: BBB bulk-out clear stall failed, TIMEOUT
Mar 17 22:53:16 kg-quiet kernel: umass1: BBB reset failed, TIMEOUT
Mar 17 22:53:16 kg-quiet kernel: umass3: BBB reset failed, TIMEOUT
Mar 17 22:53:16 kg-quiet kernel: umass2: BBB reset failed, TIMEOUT
Mar 17 22:53:16 kg-quiet kernel: umass0: BBB reset failed, TIMEOUT
Mar 17 22:54:21 kg-quiet kernel: umass1: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:54:21 kg-quiet kernel: umass3: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:54:21 kg-quiet kernel: umass2: BBB bulk-in clear stall failed, TIMEOUT
Mar 17 22:54:21 kg-quiet kernel: umass0: BBB bulk-in clear stall failed, TIMEOUT
> Also try to upgrade the kernel to the lastest sources. There has been
> an EHCI performance quirk added recently.
Is this update in RELENG_6 as well? I am running:
root at kg-quiet# uname -a
FreeBSD kg-quiet.kg4.no 6.4-STABLE FreeBSD 6.4-STABLE #27: Sun Mar 15
19:42:19 CET 2009 root at kg-quiet.kg4.no:/usr/obj/usr/src/sys/QUIET
amd64
which is quite recent (but RELENG_6).
HTH
--
Regards,
Torfinn Ingolfsen
More information about the freebsd-usb
mailing list