usb-storage problem

Message

Ulukay · Post by **Ulukay** » Fri Mar 05, 2004 1:41 pm

hoi, i got a big problem with my external usb 2.0 hdd

i tried already 2 different pcs (athlon xp2800/kt333 and p4 ht/i865)
2 different kernels (2.4.25, 2.6.4-pre mm sources)
and many different cables

the hardware is working perfectly, because i can use the ext. hdd in windows without any problems (i moved already ~200gigs with this drive)

here is what i get in linux

on connect:

Code: Select all

usb 1-2: new high speed USB device using address 3
scsi3 : SCSI emulation for USB Mass Storage devices
  Vendor: ST360021  Model: A                 Rev: 0811
  Type:   Direct-Access                      ANSI SCSI revision: 02
SCSI device sdb: 117231408 512-byte hdwr sectors (60022 MB)
sdb: assuming drive cache: write through
 /dev/scsi/host3/bus0/target0/lun0: p1
Attached scsi disk sdb at scsi3, channel 0, id 0, lun 0
Attached scsi generic sg3 at scsi3, channel 0, id 0, lun 0,  type 0
USB Mass Storage device found at 3

then i mount the drive, and e.g. try to copy something
after a few minutes (or seconds) the copy process "freezes" for about 2mins and then i get an I/O error

here is what kern.log is telling me:

Code: Select all

Mar  5 14:39:34 sunnyserver scsi: Device offlined - not ready after error recovery: host 3 channel 0 id 0 lun 0
Mar  5 14:39:34 sunnyserver SCSI error : <3 0 0 0> return code = 0x50000
Mar  5 14:39:34 sunnyserver end_request: I/O error, dev sdb, sector 358631
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358568
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358569
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358570
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358571
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358572
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358573
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358574
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358575
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358576
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver Buffer I/O error on device sdb1, logical block 358577
Mar  5 14:39:34 sunnyserver lost page write due to I/O error on sdb1
Mar  5 14:39:34 sunnyserver SCSI error : <3 0 0 0> return code = 0x70000
Mar  5 14:39:34 sunnyserver end_request: I/O error, dev sdb, sector 358759
Mar  5 14:39:34 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:34 sunnyserver FAT: bread(block 129) in fat_access failed
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:35 sunnyserver scsi3 (0:0): rejecting I/O to offline device
.
.
.
MANY of them ;)
.
.
Mar  5 14:39:36 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:36 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:43 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 14:39:43 sunnyserver FAT: Directory bread(block 28648) failed

Ulukay · Post by **Ulukay** » Fri Mar 05, 2004 2:35 pm

i tried with ext2:

Code: Select all

Mar  5 15:32:40 sunnyserver scsi: Device offlined - not ready after error recovery: host 3 channel 0 id 0 lun 0
Mar  5 15:32:40 sunnyserver SCSI error : <3 0 0 0> return code = 0x50000
Mar  5 15:32:40 sunnyserver end_request: I/O error, dev sdb, sector 104961479
Mar  5 15:32:40 sunnyserver printk: 425844 messages suppressed.
Mar  5 15:32:40 sunnyserver Buffer I/O error on device sdb1, logical block 13120181
Mar  5 15:32:40 sunnyserver lost page write due to I/O error on sdb1
Mar  5 15:32:40 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 15:32:40 sunnyserver scsi3 (0:0): rejecting I/O to offline device
..
many
..
Mar  5 15:32:40 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 15:32:41 sunnyserver Buffer I/O error on device sdb1, lorejecting I/O to offline device
Mar  5 15:32:41 sunnyserver scsi3 (0:0): rejecting I/O to offline device
..
many
..
Mar  5 15:32:41 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 15:32:41 sunnyserver buffer layer error at fs/buffer.c:1266
Mar  5 15:32:41 sunnyserver Call Trace:
Mar  5 15:32:41 sunnyserver [<c0152ad8>] mark_buffer_dirty+0x48/0x50
Mar  5 15:32:41 sunnyserver [<c01cce0f>] ext2_sync_super+0x3f/0x60
Mar  5 15:32:41 sunnyserver [<c01cbc6c>] ext2_error+0xac/0xe0
Mar  5 15:32:41 sunnyserver [<c0152ed5>] __bread+0x35/0x40
Mar  5 15:32:41 sunnyserver [<c01ca736>] ext2_get_inode+0x116/0x130
Mar  5 15:32:41 sunnyserver [<c01caaf9>] ext2_update_inode+0x49/0x350
Mar  5 15:32:41 sunnyserver [<c01701e6>] write_inode+0x46/0x50
Mar  5 15:32:41 sunnyserver [<c01704a9>] __sync_single_inode+0x2b9/0x320
Mar  5 15:32:41 sunnyserver [<c01c9299>] ext2_discard_prealloc+0x59/0x90
Mar  5 15:32:41 sunnyserver [<c0170752>] sync_sb_inodes+0x192/0x260
Mar  5 15:32:41 sunnyserver [<c0170873>] writeback_inodes+0x53/0xb0
Mar  5 15:32:41 sunnyserver [<c0138f6b>] background_writeout+0x7b/0xc0
Mar  5 15:32:41 sunnyserver [<c013972d>] __pdflush+0xed/0x1d0
Mar  5 15:32:41 sunnyserver [<c0139810>] pdflush+0x0/0x20
Mar  5 15:32:41 sunnyserver [<c013981f>] pdflush+0xf/0x20
Mar  5 15:32:41 sunnyserver [<c0138ef0>] background_writeout+0x0/0xc0
Mar  5 15:32:41 sunnyserver [<c0139810>] pdflush+0x0/0x20
Mar  5 15:32:41 sunnyserver [<c012de05>] kthread+0xa5/0xb0
Mar  5 15:32:41 sunnyserver [<c012dd60>] kthread+0x0/0xb0
Mar  5 15:32:41 sunnyserver [<c0106989>] kernel_thread_helper+0x5/0xc
Mar  5 15:32:41 sunnyserver
Mar  5 15:32:41 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 15:32:41 sunnyserver EXT2-fs error (device sdb1): ext2_get_inode: unable to read inode block - inode=6619138, block=13238274
Mar  5 15:32:41 sunnyserver scsi3 (0:0): rejecting I/O to offline device
..
many
..
Mar  5 15:32:41 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 15:32:41 sunnyserver EXT2-fs error (device sdb1): ext2_free_branches: Read failure, inode=6619138, block=13253646
Mar  5 15:32:41 sunnyserver scsi3 (0:0): rejecting I/O to offline device
..
many
..
Mar  5 15:33:00 sunnyserver scsi3 (0:0): rejecting I/O to offline device
Mar  5 15:33:00 sunnyserver printk: 1178 messages suppressed.
Mar  5 15:33:00 sunnyserver Buffer I/O error on device sdb1, logical block 0
Mar  5 15:33:00 sunnyserver lost page write due to I/O error on sdb1

Ulukay · Post by **Ulukay** » Fri Mar 05, 2004 4:01 pm

lsusb is telling me:

Bus 001 Device 002: ID 05e3:0702 Genesys Logic, Inc.
Device Descriptor:
bLength 18
bDescriptorType 1
bcdUSB 2.00
bDeviceClass 0 Interface
bDeviceSubClass 0
bDeviceProtocol 0
bMaxPacketSize0 64
idVendor 0x05e3 Genesys Logic, Inc.
idProduct 0x0702
bcdDevice 0.02
iManufacturer 2 Genesyslogic
iProduct 3 USB Mass Storage Device
iSerial 0
bNumConfigurations 1
Configuration Descriptor:
bLength 9
bDescriptorType 2
wTotalLength 32
bNumInterfaces 1
bConfigurationValue 1
iConfiguration 0
bmAttributes 0xc0
Self Powered
MaxPower 4mA
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 0
bAlternateSetting 0
bNumEndpoints 2
bInterfaceClass 8 Mass Storage
bInterfaceSubClass 6 SCSI
bInterfaceProtocol 80 Bulk (Zip)
iInterface 0
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x81 EP 1 IN
bmAttributes 2
Transfer Type Bulk
Synch Type none
wMaxPacketSize 512
bInterval 1
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x02 EP 2 OUT
bmAttributes 2
Transfer Type Bulk
Synch Type none
wMaxPacketSize 512
bInterval 1
Language IDs: (length=4)
0409 English(US)

jdoe · Post by **jdoe** » Mon Mar 08, 2004 5:36 pm

are you on nforce2 ?? i've a similare problem with a usb-to-IDE adapter, which I use for moving files from my rec studio to home. On win i've used it and i've no problem, but on linux the transfer freezes and the I/O error, like. I've been told it's a nforce issue... bah...

John

Ulukay · Post by **Ulukay** » Mon Mar 08, 2004 10:18 pm

Ulukay wrote: i tried already 2 different pcs (athlon xp2800/kt333 and p4 ht/i865)

http://www.qbik.ch/usb/devices/showdev.php?id=2383

brenden · Post by **brenden** » Tue Mar 09, 2004 1:36 am

I've been having the same problem. Try disabling usbfs and acpi.

Ulukay · Post by **Ulukay** » Tue Mar 09, 2004 12:22 pm

brenden wrote:I've been having the same problem. Try disabling usbfs and acpi.

no effect

but i found: http://www.qbik.ch/usb/devices/showdev.php?id=2383

elpollodiablo · Post by **elpollodiablo** » Wed Mar 10, 2004 7:47 pm

the same thing happens to me if i use the usb2.0!!! if i connect my hdd to the usb1 port on my hub, the transfer works (but is so damn slow....)

markuslk · Post by **markuslk** » Sun Apr 11, 2004 8:48 pm

Similar problems here, I've got a MAP-H31x Series enclosure for my Samsung 160GB harddisk with USB 2.0 connector.

[little OT]
When I start the disk at boot time and boot into winxp, my hdd is set up right when i log in and I've got no problems while accessing/moving data. When I turn it on while XP's running, my disk sometimes needs 3 times of ON/OFF to get recognized by the os....
[/OT]

With gentoo-2.6.5 (and all the kernels before) I've the same problem that my harddisk begins to "pump" after several minutes or even a few hours... it sounds like a unhealthily spin down/spin up while transferring data.
Then I can't umount the drive properly and most of the times I'd to reboot by pressing the button

because X/KDE freezed.

I thought that it might be a problem with my Samsung Drive (only 1/2 year old).

Which external cases are working fine with linux? I'll try out the cheapest external case which I might find on ebay next week, but your experiences with working cases/chips would help me anyway

Greetings

Markus

RoYzter · Post by **RoYzter** » Mon Apr 12, 2004 12:41 am

glad to see that i am not the only one having problems... i bought an icybox 350u (english page doesn't work...) and a seagate 200gb hdd. the enclosure has a prolific 2507 chip which (according to this page) should work ok.
if i disable usb 2.0 support in the bios of my msi hermes 651 barebone system (SiS 651 chipset), it works without any problems (except for the slow usb 1.1 speed). but if usb 2.0 is enabled, it starts getting weird...
it seems that transfers first get buffered and if the buffer is full the system nearly freezes (no network traffic possible, everything very slow). top shows over 80% of my cpu waiting, but i can't identify which process is eating everything up. it has to be something usb/external hdd related though.

Code: Select all

top - 02:27:34 up 13:17,  1 user,  load average: 2.71, 0.96, 0.54
Tasks:  70 total,   1 running,  69 sleeping,   0 stopped,   0 zombie
Cpu(s): 12.0% us,  4.0% sy,  0.0% ni,  0.0% id, 83.9% wa,  0.0% hi,  0.0% si
Mem:    484156k total,   481180k used,     2976k free,   109168k buffers
Swap:   506036k total,    10616k used,   495420k free,   222884k cached

if i wait long enough, everything gets in order again... just to give you an example what long enough means: copying 50mb takes nearly 3 minutes to finish (even usb 1.1 is supposed to be faster!)

Code: Select all

# time dd if=/dev/zero of=/dev/sda2 bs=1M count=50
50+0 records in
50+0 records out

real    2m50.744s
user    0m0.001s
sys     0m0.288s

here is a part of /var/log/messages:

Code: Select all

Apr 11 21:32:43 cassiopeia usb 1-1: new high speed USB device using address 3
Apr 11 21:32:43 cassiopeia scsi1 : SCSI emulation for USB Mass Storage devices
Apr 11 21:32:43 cassiopeia Vendor: ST320082  Model: 2A                Rev: 3.01
Apr 11 21:32:43 cassiopeia Type:   Direct-Access                      ANSI SCSI revision: 02
Apr 11 21:32:43 cassiopeia SCSI device sda: 390721969 512-byte hdwr sectors (200050 MB)
Apr 11 21:32:43 cassiopeia sda: assuming drive cache: write through
Apr 11 21:32:44 cassiopeia /dev/scsi/host1/bus0/target0/lun0: p1 p2
Apr 11 21:32:44 cassiopeia Attached scsi disk sda at scsi1, channel 0, id 0, lun 0
Apr 11 21:32:44 cassiopeia Attached scsi generic sg0 at scsi1, channel 0, id 0, lun 0,  type 0
Apr 11 21:32:44 cassiopeia USB Mass Storage device found at 3
Apr 11 21:32:44 cassiopeia scsi.agent[20871]: disk at /devices/pci0000:00/0000:00:03.3/usb1/1-1/1-1:1.0/host1/1:0:0:0
Apr 11 21:32:44 cassiopeia modprobe: FATAL: Module sd_mod already in kernel.

i didn't yet really test if it works ok when using windoze... gonna try this tomorrow
anyway... can somebody help me please?
thx in advance

[edit] corrected "copying 100mb" to "copying 50mb"

so it's even worse! [/edit]

kottlettstanze · Post by **kottlettstanze** » Wed Apr 21, 2004 11:27 pm

hi,
same troubles here as well. disabling usbfs and acpi (and apic, hint from a different thread) didn't solve it either.
In the moment I'm using the 2.6.5 kernel. Using a 2.4.? kernel I could sort of solve the problem by giving the ehci_hcd module the parameter:

Code: Select all

log2_irq_thresh=6

That made it slow (because it waits then), but stable. This doesn't work for the 2.6.5 kernel

any other ideas? and why the hell is it working with m$?

cheers,
Kevin

kurifu · Post by **kurifu** » Tue Apr 27, 2004 4:34 am

This is actually a well known problem... many people appear to be trying to figure out what is actually causing this, and though there was a bug fix in kernel 2.6.4 which managed to make its way past the 0 length messages the hardware was sending (invalidly) to the mass storage driver, there are objviously other issues that need to be ironed out.

You can follow up more closly at this url: http://www.mail-archive.com/linux-usb-d ... 20692.html

Any help anyone can provide with debugging this, well I am sure we would all appreciate it.

thepi · Post by **thepi** » Sun Jun 20, 2004 6:53 pm

i had the same problem, but found no solution. so i returned the disk for refund.

later, i found this.
it's actually for genesys devices, but maybe it helps. in my case the copying would freeze tho, i had to turn off the whole external hdd. as far as i understand it the usb chip on the hdd case had a data overflow (i.e. the data came in too fast).

hth, never mind if it adresses a totally different issue.

pi~

kottlettstanze · Post by **kottlettstanze** » Sun Aug 15, 2004 12:55 pm

This one did the trick for me:
http://www.mail-archive.com/linux-usb-d ... 26639.html

Using gentoo-dev-sources 2.6.8, I just wrote 3.9GB to the usb disk without error. The speed is okay as well (between 10 and 15 MB/s).

The patch tool doesn't like to do the patch (too much differences), but it is easy to do manually.
Would be great if the patch was included in the gentoo-dev-sources

cheers,
Kevin

usb-storage problem

usb-storage problem

Re: usb-storage problem

Genesys USB2->IDE KNOWN problem...

Patch!