2.4 > 2.6 usb-storage fails to work w/ scanner [SOLVED]

[jmusits]

This is a dupe of a post I made in the Got udev working now what thread

First off I want to sat thanks to Decible and dsd for the guides they have made. Great work guys, made getting udev up on my system painless. But I do have one problem.

Recently I migrated from gentoo-sources-2.4.25 with devfs to gentoo-dev-sources-2.6.8 with udev. My Sony CLIE and my USB thumbdrive both work fine with udev, even wrote some some rules to get persistent names. But now for my problem I have a HP Photo Scanner 1000 which worked with devfs through usb-storage and sd_mod. However it does not work with udev.
Code:
Aug 23 21:27:36 particle usb 1-1: new full speed USB device using address 3
Aug 23 21:27:37 particle usb-storage: probe of 1-1:1.0 failed with error -1

It seems as though udev is getting the info about the device:
Code:
root@particle linux# lsusb -v
Bus 001 Device 003: ID 03f0:1001 Hewlett-Packard
Language IDs: none (invalid length string descriptor 09; len=39)
Device Descriptor:
bLength 18
bDescriptorType 1
bcdUSB 1.00
bDeviceClass 0 Interface
bDeviceSubClass 0
bDeviceProtocol 0
bMaxPacketSize0 8
idVendor 0x03f0 Hewlett-Packard
idProduct 0x1001
bcdDevice 1.00
iManufacturer 0
iProduct 0
iSerial 0
bNumConfigurations 1
Configuration Descriptor:
bLength 9
bDescriptorType 2

wTotalLength 39
bNumInterfaces 1
bConfigurationValue 1
iConfiguration 0
bmAttributes 0x80
MaxPower 500mA
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 0
bAlternateSetting 0
bNumEndpoints 3
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 3
Transfer Type Interrupt
Synch Type none
wMaxPacketSize 2
bInterval 20
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x02 EP 2 OUT
bmAttributes 2
Transfer Type Bulk
Synch Type none
wMaxPacketSize 64
bInterval 0
Endpoint Descriptor:
bLength 7
bDescriptorType 5
bEndpointAddress 0x83 EP 3 IN
bmAttributes 2
Transfer Type Bulk
Synch Type none
wMaxPacketSize 64
bInterval 0
Language IDs: none (invalid length string descriptor 09; len=39)

However /dev/sda1 is never created as it was w/ devfs.

Anyone have any ideas on what I can try? Am I missing something? I have searched the forums and google for the past week and can't find anything related to my problem.

Thanks,

Jason
_________________
If you got the DO RE I got MI.

--Jerry Garcia "Run For The Roses"

[Nate_S]

you were using usb-storage for a scanner? how did that work? I thought you had to use sane (http://www.sane-project.org/)

-Nate

[jmusits]

The scanner is just recognized as a mass storage device ie: a thumbdrive. The scanner basically has a 16MB memory onboard, it stores the scan there, devfs creates /dev/sdX for the divice when I plug it in, not sg like a scanner that works with sane. I mount it and then copy the image to the hard drive. That's the way it worked with the 2.4 series kernel, I doubt anything would change with the change to 2.6. I think it's just a problem with udev recognizing the device properly or binding it to the driver.

Jason
_________________
If you got the DO RE I got MI.

--Jerry Garcia "Run For The Roses"

[Nate_S]

Hmm... well, have you checked just in case it is creating a sg device for it instead?

Also, just in case it's this, do you have 'probe multiple luns' enabled in your kernel config? some devices need that.

A lot of things did change from 2.4 -> 2.6, however, I'd bet it's more likely the udev switch that caused this, or perhaps a kernel config option you missed. If those ideas don't help, you can post your kernel config here.

-Nate

EDIT: also, I noticed in the kernel config, under usb, CONFIG_USB_HPUSBSCSI, you can check if you need that. you also might try enabling all of the usb ata adapters and smartmedia things, in case your scanner uses one of those chips. (unlikely, but your scanner is the first I've heard of that allows direct access to the memory, so who knows?)

[jmusits]

I have a feeling those kernel options won't change anything because they were not need in 2.4 but I will give it a try. As far as creating /dev/sg that doesn't happen because that it not compiled for my kernel, I will throw that in too just in case. I seems to me that udev is recognizeing the device, but not 100% properly, because if usb-storage and sd_mod are not loaded when I plug in the scanner they are autoloaded. Maybe there was a change in the usb-storage module code going from 2.4 > 2.6, I will have to look into that.

Thanks for the ideas.

Jason
_________________
If you got the DO RE I got MI.

--Jerry Garcia "Run For The Roses"

[Nate_S]

AFAIK, very little of the usb storage code has actually changed, however, some of it may have been broken out into seperate modules. Or, perhaps simply the way the kernel handels your scanner has changed. Perhaps your scanner is now supported by sg, instead? Also, go ahead and post your /usr/src/linux/.confg, as well as your dmesg.

-Nate

[jmusits]

Nate:

First off I want to thank you for the time and thought you are giving my problem - it's greatly appreciated. Second...I think I narrowed the probelm down to the usb-storage module. I turned on verbose messageing for that modules and it gives me more info in dmesg. Here are the appropriate lines:

[Nate_S]

Hmm... well if it's the module itself, I may be a bit out of my league here...

You might try a divverent version and/or patchset of the kernel, in case it's something that recently changed. you could try vanilla-2.6.3 and see if the problem persists.

Now that you've narrowed it down, you might rename the topic to something like '2.4 -> 2.6 usb-storage fails to work with scanner [UNSOLVED]' perhaps people with more knowledge of usb-storage can be of more assistance.

also, I saw another thread about recent issues with hp scanners, here's the link in case it's related, though i'm not sure it is
https://forums.gentoo.org/viewtopic.php?t=214500

Good Luck
-Nate

[jmusits]

I got the problem solved. I spent some time over at linux-usb-devel and found that the problem is actually with my scanner. As it turns out when usb-storage probes the device for the LUN count my device returns 0, it should be 1. Back in kernel 2.4.x the usb-storage module was more lax, it checked LUN_count >= 0, in 2.6.8 it checked LUN_count == 1. About 5 days ago they released a patch to go back to the way 2.4.x checked allowing faulty devices to be recogized. Now my scanner works to way it used to, nice and simple, wrote some udev rules so it won't step on my thumbdrive either.

[Nate_S]

Well cool, glad you got that working

-Nate