I'm adding a couple lacie external USB drives to my CentOS 5.3 system and running into a few issues... when I connect them, I see the following in dmesg:
usb 1-1: new high speed USB device using ehci_hcd and address 11 usb 1-1: configuration #1 chosen from 1 choice
And that's it. No device assignment information, the usb-storage module does not appear to be loaded and I am unable to access the drive.
I do see the following in lsusb:
Bus 001 Device 011: ID 0451:6250 Texas Instruments, Inc.
I'm assuming the kernel or udev doesn't know what to do with the given device ID and isn't correctly mapping it to the usb-storage module (I have manually loaded usb-storage with modprobe usb-storage prior to attaching the device).
I found a post describing how to manually map a vendor id / device id to usb-storage, but it has to do with the old hotplug system from CentOS 4.
Can anyone advise me on how I can tell my system the above device is a mass storage device? Do I need to write some udev rules?
These drives work fine (automatically) on a Fedora 10 machine. Also, I am using the "plus" kernel on my CentOS machine.
Thanks, Ray
On Mon, May 25, 2009 at 11:17:30AM -0700, Ray Van Dolson wrote:
Can anyone advise me on how I can tell my system the above device is a mass storage device? Do I need to write some udev rules?
Sorry for replying to my own post, but upon further investigation, it appears that perhaps the usb-storage module doesn't have the correct matching pattern in place to detect this as a mass storage device:
udevmonitor reports the device having a module alias of:
usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
And modinfo usb-storage shows:
# modinfo usb-storage | grep v045 alias: usb:v045Ap5210d0101dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0151d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0150d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0451p5416d0100dc*dsc*dp*ic*isc*ip*
Which obviously won't match. So the question is, how do I add a new alias to the module? And will the module even support my device?
Ray
On Mon, 25 May 2009, Ray Van Dolson wrote:
On Mon, May 25, 2009 at 11:17:30AM -0700, Ray Van Dolson wrote:
Can anyone advise me on how I can tell my system the above device is a mass storage device? Do I need to write some udev rules?
Sorry for replying to my own post, but upon further investigation, it appears that perhaps the usb-storage module doesn't have the correct matching pattern in place to detect this as a mass storage device:
udevmonitor reports the device having a module alias of:
usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
And modinfo usb-storage shows:
# modinfo usb-storage | grep v045 alias: usb:v045Ap5210d0101dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0151d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0150d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0451p5416d0100dc*dsc*dp*ic*isc*ip*
Which obviously won't match. So the question is, how do I add a new alias to the module? And will the module even support my device?
Just a wild an crazy idea, how about adding it to /etc/modprobe.conf ?
alias usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00 usb-storage
Not sure if it makes sense, but it's worth trying :)
This BTW is a question I have had before and if we find the answer I think this is food for the upcoming CentOS Newsletter ;-)
On Mon, 25 May 2009, Dag Wieers wrote:
On Mon, 25 May 2009, Ray Van Dolson wrote:
On Mon, May 25, 2009 at 11:17:30AM -0700, Ray Van Dolson wrote:
Can anyone advise me on how I can tell my system the above device is a mass storage device? Do I need to write some udev rules?
Sorry for replying to my own post, but upon further investigation, it appears that perhaps the usb-storage module doesn't have the correct matching pattern in place to detect this as a mass storage device:
udevmonitor reports the device having a module alias of:
usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
And modinfo usb-storage shows:
# modinfo usb-storage | grep v045 alias: usb:v045Ap5210d0101dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0151d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0150d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0451p5416d0100dc*dsc*dp*ic*isc*ip*
Which obviously won't match. So the question is, how do I add a new alias to the module? And will the module even support my device?
Just a wild an crazy idea, how about adding it to /etc/modprobe.conf ?
alias usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00 usb-storage
Not sure if it makes sense, but it's worth trying :)
After doing some research it seems that this is valid, if you perform:
modprobe -c
you get an identical list for all known devices. So adding yours should definitely help.
PS What kind of LaCie disk is it ? I recently bought a 500GB LaCie Little Disk and that one worked fine.
On Mon, May 25, 2009 at 09:24:03PM +0200, Dag Wieers wrote:
On Mon, 25 May 2009, Dag Wieers wrote:
On Mon, 25 May 2009, Ray Van Dolson wrote:
On Mon, May 25, 2009 at 11:17:30AM -0700, Ray Van Dolson wrote:
Can anyone advise me on how I can tell my system the above device is a mass storage device? Do I need to write some udev rules?
Sorry for replying to my own post, but upon further investigation, it appears that perhaps the usb-storage module doesn't have the correct matching pattern in place to detect this as a mass storage device:
udevmonitor reports the device having a module alias of:
usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
And modinfo usb-storage shows:
# modinfo usb-storage | grep v045 alias: usb:v045Ap5210d0101dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0151d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0150d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0451p5416d0100dc*dsc*dp*ic*isc*ip*
Which obviously won't match. So the question is, how do I add a new alias to the module? And will the module even support my device?
Just a wild an crazy idea, how about adding it to /etc/modprobe.conf ?
alias usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00 usb-storage
Not sure if it makes sense, but it's worth trying :)
After doing some research it seems that this is valid, if you perform:
modprobe -c
you get an identical list for all known devices. So adding yours should definitely help.
PS What kind of LaCie disk is it ? I recently bought a 500GB LaCie Little Disk and that one worked fine.
I actually did try adding the above alias line with no luck although the following command:
modprobe -v -n --first-time usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
Now shows a "match". I'm beginning to think the kernel doesn't support this device, or isn't properly exposing it to /sys as it should be.
I added a udev rule:
BUS=="usb",ACTION=="add",SYSFS{idVendor}=="0451",SYSFS{idProduct}=="6250",RUN+="/sbin/modprobe usb-storage",NAME="walter"
And when I plug in my device, I get the following entry under /dev:
crw------- 1 root root 189, 19 May 25 12:02 /dev/walter
Obviously this isn't major type 8 which I would expect to see for a disk... also:
With udevmonitor on my Fedora machine, I see a device created with major type 189, but then later, a SCSI device is detected and a device of major type 8 is set up. This doesn't happen on CentOS, only the 189 device is created.
Also, on my Fedora system the vendor id and Product ID's are different.
Fedora:
059f:0527 LaCie, Ltd
Cent:
0451:6250 Texas Instruments, Inc.
So I'm beginning to think that the kernel is sending "incorrect" or dated information into /sys and the event sent to udev is either misleading or incomplete somehow.
It's a 500GB "Big Disk".
I'll keep poking around... thanks for the reponse. :)
Ray
At Mon, 25 May 2009 12:53:55 -0700 CentOS mailing list centos@centos.org wrote:
On Mon, May 25, 2009 at 09:24:03PM +0200, Dag Wieers wrote:
On Mon, 25 May 2009, Dag Wieers wrote:
On Mon, 25 May 2009, Ray Van Dolson wrote:
On Mon, May 25, 2009 at 11:17:30AM -0700, Ray Van Dolson wrote:
Can anyone advise me on how I can tell my system the above device is a mass storage device? Do I need to write some udev rules?
Sorry for replying to my own post, but upon further investigation, it appears that perhaps the usb-storage module doesn't have the correct matching pattern in place to detect this as a mass storage device:
udevmonitor reports the device having a module alias of:
usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
And modinfo usb-storage shows:
# modinfo usb-storage | grep v045 alias: usb:v045Ap5210d0101dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0151d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0457p0150d0100dc*dsc*dp*ic*isc*ip* alias: usb:v0451p5416d0100dc*dsc*dp*ic*isc*ip*
Which obviously won't match. So the question is, how do I add a new alias to the module? And will the module even support my device?
Just a wild an crazy idea, how about adding it to /etc/modprobe.conf ?
alias usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00 usb-storage
Not sure if it makes sense, but it's worth trying :)
After doing some research it seems that this is valid, if you perform:
modprobe -c
you get an identical list for all known devices. So adding yours should definitely help.
PS What kind of LaCie disk is it ? I recently bought a 500GB LaCie Little Disk and that one worked fine.
I actually did try adding the above alias line with no luck although the following command:
modprobe -v -n --first-time usb:v0451p6250d0300dcFFdsc00dp00icFFisc00ip00
Now shows a "match". I'm beginning to think the kernel doesn't support this device, or isn't properly exposing it to /sys as it should be.
I added a udev rule:
BUS=="usb",ACTION=="add",SYSFS{idVendor}=="0451",SYSFS{idProduct}=="6250",RUN+="/sbin/modprobe usb-storage",NAME="walter"
And when I plug in my device, I get the following entry under /dev:
crw------- 1 root root 189, 19 May 25 12:02 /dev/walter
Obviously this isn't major type 8 which I would expect to see for a disk... also:
With udevmonitor on my Fedora machine, I see a device created with major type 189, but then later, a SCSI device is detected and a device of major type 8 is set up. This doesn't happen on CentOS, only the 189 device is created.
Also, on my Fedora system the vendor id and Product ID's are different.
Fedora:
059f:0527 LaCie, Ltd
Cent:
0451:6250 Texas Instruments, Inc.
What are the kernel versions?
So I'm beginning to think that the kernel is sending "incorrect" or dated information into /sys and the event sent to udev is either misleading or incomplete somehow.
It's a 500GB "Big Disk".
I'll keep poking around... thanks for the reponse. :)
Ray _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
At Mon, 25 May 2009 14:25:18 -0700 CentOS mailing list centos@centos.org wrote:
On Mon, May 25, 2009 at 05:20:25PM -0400, Robert Heller wrote:
What are the kernel versions?
CentOS kernel:
2.6.18-128.1.6.el5.centos.plus
Fedora 10 kernel:
2.6.27.21-170.2.56.fc10.i686
It could just be a kernel bug that RedHat didn't back port (or one that can't be backported easily). Or just a driver update (eg adding a few lines to a driver scan structure) that has not been backported yet.
You could get the kernel sources for both kernels and compare the relevant .c and .h files and possibly patch the centos.plus kernel and rebuild it. Probably not for the faint hearted...
It might also be possible to install the FC10 kernel itself...
Ray _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Mon, May 25, 2009 at 06:05:44PM -0400, Robert Heller wrote:
At Mon, 25 May 2009 14:25:18 -0700 CentOS mailing list centos@centos.org wrote:
On Mon, May 25, 2009 at 05:20:25PM -0400, Robert Heller wrote:
What are the kernel versions?
CentOS kernel:
2.6.18-128.1.6.el5.centos.plus
Fedora 10 kernel:
2.6.27.21-170.2.56.fc10.i686
It could just be a kernel bug that RedHat didn't back port (or one that can't be backported easily). Or just a driver update (eg adding a few lines to a driver scan structure) that has not been backported yet.
You could get the kernel sources for both kernels and compare the relevant .c and .h files and possibly patch the centos.plus kernel and rebuild it. Probably not for the faint hearted...
It might also be possible to install the FC10 kernel itself...
Actually, looks like I got it working. And I was way off on a wild goose chase. Turns out that after performing this[1] procedure, the drive is recognized correctly by CentOS.
Well I learned a little bit more about how udev works at least. :-)
Now my lsusb looks like the following:
Bus 001 Device 024: ID 059f:0527 LaCie, Ltd Bus 001 Device 016: ID 0451:6250 Texas Instruments, Inc.
(One drive I still need to "reset").
Sorry for the noise all.
Ray
On Mon, 25 May 2009, Ray Van Dolson wrote:
On Mon, May 25, 2009 at 06:05:44PM -0400, Robert Heller wrote:
At Mon, 25 May 2009 14:25:18 -0700 CentOS mailing list centos@centos.org wrote:
On Mon, May 25, 2009 at 05:20:25PM -0400, Robert Heller wrote:
What are the kernel versions?
CentOS kernel:
2.6.18-128.1.6.el5.centos.plus
Fedora 10 kernel:
2.6.27.21-170.2.56.fc10.i686
It could just be a kernel bug that RedHat didn't back port (or one that can't be backported easily). Or just a driver update (eg adding a few lines to a driver scan structure) that has not been backported yet.
You could get the kernel sources for both kernels and compare the relevant .c and .h files and possibly patch the centos.plus kernel and rebuild it. Probably not for the faint hearted...
It might also be possible to install the FC10 kernel itself...
Actually, looks like I got it working. And I was way off on a wild goose chase. Turns out that after performing this[1] procedure, the drive is recognized correctly by CentOS.
Well I learned a little bit more about how udev works at least. :-)
Now my lsusb looks like the following:
Bus 001 Device 024: ID 059f:0527 LaCie, Ltd Bus 001 Device 016: ID 0451:6250 Texas Instruments, Inc.
(One drive I still need to "reset").
Sorry for the noise all.
And what was that Texas Instruments device that does not show up on Fedora ?
I am intrigued ;-)
On Tue, May 26, 2009 at 01:15:56AM +0200, Dag Wieers wrote:
On Mon, 25 May 2009, Ray Van Dolson wrote:
On Mon, May 25, 2009 at 06:05:44PM -0400, Robert Heller wrote:
At Mon, 25 May 2009 14:25:18 -0700 CentOS mailing list centos@centos.org wrote:
On Mon, May 25, 2009 at 05:20:25PM -0400, Robert Heller wrote:
What are the kernel versions?
CentOS kernel:
2.6.18-128.1.6.el5.centos.plus
Fedora 10 kernel:
2.6.27.21-170.2.56.fc10.i686
It could just be a kernel bug that RedHat didn't back port (or one that can't be backported easily). Or just a driver update (eg adding a few lines to a driver scan structure) that has not been backported yet.
You could get the kernel sources for both kernels and compare the relevant .c and .h files and possibly patch the centos.plus kernel and rebuild it. Probably not for the faint hearted...
It might also be possible to install the FC10 kernel itself...
Actually, looks like I got it working. And I was way off on a wild goose chase. Turns out that after performing this[1] procedure, the drive is recognized correctly by CentOS.
Well I learned a little bit more about how udev works at least. :-)
Now my lsusb looks like the following:
Bus 001 Device 024: ID 059f:0527 LaCie, Ltd Bus 001 Device 016: ID 0451:6250 Texas Instruments, Inc.
(One drive I still need to "reset").
Sorry for the noise all.
And what was that Texas Instruments device that does not show up on Fedora ?
I am intrigued ;-)
I can only guess it's the device presented to the OS when it hasn't been unmounted properly...
http://focus.ti.com/docs/prod/folders/print/tusb6250.html
Shows it to be a USB/ATA bridge of some sort -- Texas Instruments. Why Fedora seemed more adept at seeing the device may have been purely coincidence. I'd assume I could get the drive in a funky state so that Fedora only sees this TI device as well... though so far I've been unable to reproduce (on Fedora). :)
Ray
-
Dag Wieers -
Ray Van Dolson -
Robert Heller