Boy, am I having fun. I can't stop.
Anyways, found something new. Before running scanimage, when I cat the usb devices file, there are two entries below the produce 'P:' markde 'S:' . These two lines contain
Manufacturer=EPSON
Produce=Perfection636
I assume these are free-form strings stored in the scanner. The devices text is returned immediately.
During the command 'scanimage -v -d epson', the devices text is returned immediately, but the two 'S:' lines are not there!
During the beginning of the command 'scanimage -v -d epson:/dev/usbscanner0' the devices, with the two 'S:' lines included. Near the end of the command, there is a long delay for the devices text to retrun (~4 seconds) and the two 'S:' lines are not there again.
After running 'scanimage -v -d epson:/dev/usbscanner0', whose exact syntax causes a segmentation fault, I get the output:
[epson] Unknown type 2 or levl /, using B3
[epson] Extended status flag request failed
Segmentation fault
---or---
[epson] ident failed
Segmentation fault
and thereafter, every time I cat the devices file, it takes several seconds to retreive the text, and the two string entries "S:" are not there anymore.
To get everything back working fast, with string entries, I must now 'rmmod scanner' and 'insmod scanner' again. Sometimes I reboot to make sure everything is clean for the insmod's. I can also power off/on the scanner, and everything returns to what seems to be normal operation.
The scanner itself seems to do nothing during all this, but after sitting awhile, the green status light will go red and flashing (necesitating a power off/on).
Anyways, here are a few questions for the knowledgeable:
Does scanimage determine the scanner identity from these 'S:' entries, as seen in the usb/devices file?
What could the access to the scanner driver change, such that the usb susbsystem can no longer identify the scanner?
How, exactly, are scanner identities determined? Inside each individual driver code? By standard mechanism? Are there differences between the USB/SCSI command requests (esp. for the epson)?
Oh well, it's ten pm and I must study for dynamics test. More on this later.
Thanks for any and all help in advance,
Torsten
Torsten Howard wrote:
>
> Ok, More update.
>
> I have finally managed to compile 2.3.99-pre1 kernel. I had to remove some stuff to get it to compile, then had to remove damn near everything else to get it to boot. I had about five stops: just after unpacking, ppp, fd, partition check, daemon startup. Blah.
>
> Well, the good news is I can mount /proc/bus/usb and get the two files, devices and drivers (after insmod usbcore, insmod usb-uhci). I can plug in the Epson, and I can see lines added to the devices files, with the correct manufacturer EPSON and product PERFECTION 636. I can insmod scanner, and the devices files goes from "no driver"--->"usbscanner", so I know the scanner module is claiming the epson.
>
> I am unable to do 'insmod scanner vendor=04b0 product=0101' because insmod complains about a 'b'. When I change 'b' to 'B', it complains about 'B'. What, insmod no read hex?
>
> I also know that the scanimage is doing something to the scanner driver, because when I try to rmmod scanner, it tells me device or resource is busy, but only when scanimage is running (and scanimage takes forever! several minutes)
>
> Anyways, the bad news is, the scanner still won't work. Here's the log of the commands I tried:
>
> vaio:~# scanimage -d epson
>
> vaio:~# scanimage -v -d epson
> [epson] ident failed
> scanimage: open of device epson failed: Invalid argument
> vaio:~# scanimage -v -d /dev/usbscanner0
> scanimage: open of device /dev/usbscanner0 failed: Invaild argument
> //note - this command executed immediate
> vaio:~# scanimage -v -d epson:/dev/usbscanner0
> [epson] ident failed
> Segmentation fault
> vaio:~# scanimage -v -d epson
> [epson] ident failed
> scanimage: open of device epson failed: Invalid argument
> vaio:~# scanimage -v -d epson:/dev/usbscanner0
> [epson] ident failed
> Segmentation fault
> vaio:~#
>
> If anyone has been able to get the epson perfection, or similar model up and running, please let me know which kernel and usb-backport (or development kernel) you are using. Any info is much appreciated.
>
> Thanks in advance,
> Torsten
>
> Torsten Howard wrote:
> >
> > Hello,
> >
> > This is a short addendum. I've noticed that when the scanner green light is on and steady, the scanner carriage light is also on. When I first power on/plug in/replug in the scanner, and all is ready,
> >
> > 'scanimage -d epson' returns
> > [epson] error in receive - status = 62
> > [epson] error in receive - status = 5
> > [epson] ident failed
> > scanimage: open of device epson failed: Invalid argument
> >
> > the second time I run this command, before printing anything out, the green light changed red and flashed quickly, and the scanner carriage light goes out (!).
> >
> > The programs seems to never quit, and the printer indefinitely flashing the red light. I can now control-C out of the scanimage program (normally the scanimage program doesn't respond to control-C) and power off/on the scanner.
> >
> > Torsten
> >
> > Torsten Howard wrote:
> > >
> > > Hello,
> > >
> > > Please help. My epson scanner is more stubborn than Windows' registry. I'll try to
> > > list everything I've tried and done very concisely and explicitly.
> > >
> > > Much thanks for any help,
> > > Torsten Howard
> > >
> > > The Scanner
> > > -----------
> > > Epson Perfection 636U
> > > USB interface
> > > Supported by SANE backend
> > >
> > > The Computer
> > > ------------
> > > Sony VAIO 505TR laptop
> > > 1 USB Port
> > > UHCI (not alternate)
> > > Kernels used: 2.2.13, 2.2.14, 2.3.99pre1
> > > Distribution: Slackware 7.0
> > >
> > > The Problem
> > > -----------
> > > When I plug the scanner in, unplug and then replug it in, power-off then power-on,
> > > 'scanimage --list-devices' returns
> > > [epson] error in receive - status = 62
> > > [epson] error in receive - status = 5
> > > [epson] ident failed
> > >
> > > It takes ~2 min. 0.12 sec to run this command every time I run it.
> > >
> > > A couple of times, it just prints first and third, the second and third, or just the
> > > third lines from above.
> > >
> > > After trying 'scanimage --list-devices' several times, eventually just
> > > [epson] ident failed
> > > is printed, and after a little while longer, and I'm not 100% on the error conditions
> > > here, but it seems like one error is printed, the green ready light on the scanner
> > > turns red and flashes quickly.
> > >
> > > Hereafter, I must power off/power on the scanner. It seems this red light started
> > > flashing once after the second line was printed (but no the third), and all other
> > > times it seems the red light started flashing sometime after the scanimage program
> > > quits, but I'm not sure. Every time I check the scanner right after the program
> > > quits, it's light is green. But often I look back after not having checked the
> > > scanner for awhile, and the light flashes red. Hmmmph.
> > >
> > > What I Know
> > > -----------
> > > the scanner works. I can scan with printshop pro under windows 98.
> > >
> > > when the "plug n' play" is enabled in bios, nothing works. When disabled, I get the
> > > error mentioned above. So it remains disabled. This was done under recommendation
> > > from Karl's homepage usb setup.
> > >
> > > /dev/usbscanner0 was created with the command 'mknod /dev/usbscanner c 180 48'
> > >
> > > I've tried using sane v1.0.2 and save v1.0.2 modified with the epson files (version
> > > 20000306) from www.khk.net (Karl Heinz Kremer's page) - don't worry, I deleted all the
> > > old files before reinstalling the new version to make sure there were no duplicate
> > > files anywhere.
> > >
> > > I have modified the file /usr/local/etc/sane.d/epson.conf to one have one line,
> > > '/dev/usbscanner0'
> > >
> > > the output of /var/log/debug prints, when I connect the scanner
> > > ... kernel: hub.c: port 1 connection change
> > > ... kernel: hub.c: portstatus 101, change 1, High Speed
> > > ... kernel: hub.c: portstatus 103, change 0, High Speed
> > > ... kernel: hub.c: kmalloc IF c696d2c0, numif 1
> > > ... kernel: usb.c: usbscanner driver claimed interface c696d2c0
> > > so I know the scanner is being recognized and initialized, and it appears the scanner
> > > driver is grabbing the interface(?)
> > >
> > > 'scanimage -d epson' command returns
> > > [epson] error in receive - status = 62
> > > [epson] error in receive - status = 5
> > > [epson] ident failed
> > > scanimage: open of device epson failed: Invalid argument
> > >
> > > 'xscanimage epson' returns a dialog box, something like 'failed to open epson' and
> > >
> > > 'scanimage /dev/usbscanner0' returns
> > > [epson] ident failed
> > > scanimage: no SANE devices found
> > >
> > > I have used the kernel 2.2.13 patched with the usb backport (for 2.2.14) and I have
> > > used the 2.2.14 kernel with the usb-backport (for 2.2.14). Both have the exact same
> > > errors and outputs. I have compiled both as modules, and directly in the kernel.
> > > Kernels are compiled with usbcore, usbdevfs, usbscanner, and uhci (not "alternate").
> > > Also, I was sure to make modules and make modules_install because I can insmod usbcore
> > > and insmod usb-uhci and insmod scanner....
> > >
> > > I can mount /proc/bus/usb but it never ever shows any files (devices, etc).
> > >
> > > I have tried the 2.3.99pre1 kernel, but it failed to boot. It printed something like,
> > > 'Uncompressing kernel......... OK, booting" and hung right there. I made very few
> > > configuration changed, and even set 'vga=normal' in lilo.conf (and ran lilo) to make
> > > sure kernel could use the display, but it still failed to boot.
> > >
> > > --
> > > Source code, list archive, and docs: http://www.mostang.com/sane/
> > > To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
> >
> > --
> > Source code, list archive, and docs: http://www.mostang.com/sane/
> > To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
>
> --
> Source code, list archive, and docs: http://www.mostang.com/sane/
> To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Mon Mar 20 2000 - 18:56:27 PST