USB Mass storage devices don't work in linux 2.6.[9,11.x]

_private · 04-08-2005, 01:58 AM

Here's a problem I couldn't figure out myself: I've got a couple of USB mass storage devices, such as an iPod and a digital camera. They used to work as SCSI emulated devices back when I was using kernel version 2.6.7, but since I have tried several times upgrading to 2.6.9 and then to 2.6.11.6, they stopped functioning...

The configuration I used was copied from the fully working 2.6.7 kernel.

If before I didn't need the newer kernel, now I do, since I upgraded my motherboard to a newest one, based on nVidia nForce3 chipset (and its sensor chips are only 'officially' supported by the latest kernels)

More precisely, they are not anymore detected as USB mass storage devices. Although they are detected as USB devices.

lsmod output:

Code:

Module                  Size  Used by
nls_utf8                1888  - 
sr_mod                 16708  - 
i2c_isa                 1792  - 
i2c_algo_pcf            7012  - 
i2c_algo_pca            5956  - 
i2c_algo_bit            9704  - 
lm90                   12420  - 
i2c_sensor              3456  - 
i2c_nforce2             6400  - 
nvnet                  67268  - 
parport_pc             40612  - 
lp                     10916  - 
parport                35784  - 
binfmt_misc            11208  - 
nvidia               3464188  - 
ipv6                  249504  - 
ip_conntrack_ftp       72656  - 
ip_conntrack           43608  - 
snd_ens1371            22180  - 
snd_rawmidi            23552  - 
snd_ac97_codec         76888  - 
snd_pcm_oss            52032  - 
snd_mixer_oss          19296  - 
snd_pcm                91144  - 
snd_page_alloc          9764  - 
usblp                  12832  - 
usb_storage            39168  -

Besides that, scsi_mod, sg, sd_mod and uhci_hcd are compiled into the kernel.

Here is what dmesg says on topic:

Code:

usb 1-7: new high speed USB device using ehci_hcd and address 4
uba: device 4 capacity nsec 0 bsize 512
usb 1-7: USB disconnect, address 4
usb 1-7: new high speed USB device using ehci_hcd and address 5
ub: cmd #0 cmd status (-104)
ub: cmd #1 cmd status (-104)
uba: device 5 capacity nsec 0 bsize 512
usb 1-7: USB disconnect, address 5
usb 1-7: new high speed USB device using ehci_hcd and address 6
ub: cmd #0 cmd status (-104)
ub: cmd #1 cmd status (-104)
uba: device 6 capacity nsec 0 bsize 512
usb 1-7: USB disconnect, address 6
usb 1-7: new high speed USB device using ehci_hcd and address 7
ub: cmd #0 cmd status (-104)
ub: cmd #1 cmd status (-104)
uba: device 7 capacity nsec 0 bsize 512

# this is with a kernel version with multiple SCSI LUN support:
usb 1-7: new high speed USB device using ehci_hcd and address 3
uba: device 3 capacity nsec 0 bsize 512
usb 1-7: USB disconnect, address 3

kern.log used to be like that with linux v2.6.7:

Code:

Mar 31 00:21:46 tomato kernel: usb 3-1: new full speed USB device using address 2
Mar 31 00:21:46 tomato kernel: scsi1 : SCSI emulation for USB Mass Storage devices
Mar 31 00:21:46 tomato kernel:   Vendor: DSC       Model:                   Rev:     
Mar 31 00:21:46 tomato kernel:   Type:   Direct-Access                      ANSI SCSI revision: 02
Mar 31 00:21:46 tomato kernel: SCSI device sda: 32768 512-byte hdwr sectors (17 MB)
Mar 31 00:21:46 tomato kernel: sda: assuming Write Enabled
Mar 31 00:21:46 tomato kernel: sda: assuming drive cache: write through
Mar 31 00:21:46 tomato kernel:  sda: sda1
Mar 31 00:21:46 tomato kernel: Attached scsi removable disk sda at scsi1, channel 0, id 0, lun 0
Mar 31 00:21:46 tomato kernel: Attached scsi generic sg1 at scsi1, channel 0, id 0, lun 0,  type 0
Mar 31 00:21:46 tomato kernel: USB Mass Storage device found at 2
Mar 31 01:00:50 tomato kernel: usb 3-1: USB disconnect, address 2

I googled a lot yesterday, but couldn't find any good solution to a problem, nor even any other similar problem case. In fact, nothing such has ever been seen. Which tends me to believe I'm missing something very obvious... Or maybe not.

I tried installing and setting up udev, hotplug, but, well, obviously, the problem is in the kernel, not in userspace... :P It didn't help, since no SCSI block device means no /dev/sd* device node for you!

I have been recommended looking into [ linux-usb.org FAQ.html #ts10 ] (sorry, couldn't post a URL). If you could confirm that, since I haven't yet managed to try and maybe it will indeed help.

I'd appreciate any other possible solutions and measures to eliminate the problem... =)

_private · 04-08-2005, 11:57 AM

The problem was fixed by switching off the CONFIG_BLK_DEV_UB option in kernel config.

drrdf · 10-27-2013, 05:10 PM

Hi,

I as a number of others without doubt have had the same problem. It seems that many of these problems have been caused by indiscriminate changes to the Linux kernel, not considering the difficulties which would be caused to existing hardware and other parameters by these changes, and this is due to exactly that failure? I now have various USB devices which worked completely properly before the kernel changes and these have also caused floppy disks drives not to work any longer, and some software drivers not to work any longer, such as those for Windoze wireless card drivers via the NDISwrapper. There are many other posts on the Internet stating these things. In other words there is no adequate Change Control in Linux generally, and that is a sad condemnation.

I do not understand how you switch off the CONFIG_BLK_DEV_UB option in kernel config. Could someone please explain that to me as that is clearly the solution to my problems also. Thanks.

evo2 · 10-27-2013, 08:49 PM

Hi,

Quote:

Originally Posted by drrdf

I do not understand how you switch off the CONFIG_BLK_DEV_UB option in kernel config. Could someone please explain that to me

That is a compilation option: you'd have to be actually building the kernel from source to do this.

Quote:

as that is clearly the solution to my problems also.

I think that is highly unlikely: check the date of the thread.

If you want help with this please start a new thread, clearly stating a specific problem including all relevant information including distro and version, kernel versions, etc.

Evo2.

utanja · 10-29-2013, 07:15 AM

I have a similar problem:

My 150Gig USB portable HD does not work with the 3.1 and 3.2 kernels

utanja · 11-05-2013, 08:18 AM

no answers?

widget · 11-05-2013, 01:06 PM

Quote:

Originally Posted by utanja

no answers?

See post number 4.

utanja · 11-05-2013, 10:10 PM

Quote:

Originally Posted by widget

See post number 4.

Yes I saw that, but could not find it in the makemenu config file prior to the compile.

evo2 · 11-05-2013, 10:20 PM

Hi,

the last two sentences of post #4 are the important ones.

Evo2.

utanja · 11-06-2013, 06:46 AM

Quote:

Originally Posted by evo2

Hi,

the last two sentences of post #4 are the important ones.

Evo2.

OK thanks I will start a new thread