> I vote for removing this file and add a paragraph in scanner.txt about
> where to find documentation about SANE.

I'd say kill it.

> Documentation/usb/scanner.txt
>
> | REQUIREMENTS
> |
> | A host with a USB port.  Ideally, either a UHCI (Intel) or OHCI
> | (Compaq and others) hardware port should work.  At the time of this
> | writing, there are two UHCI drivers and one OHCI.
>
> I guess it also works with EHCI?

It should.

> | A Linux kernel with USB support enabled or a backported version to
> | linux-2.2.x.  See http://www.linux-usb.org for more information on
> | accomplishing this.
>
> Well, if you read scanner.txt, you alredy have a kernel with USB
> capability.
>
> | 'lspci' which is only needed to determine the type of USB hardware
> | available/installed in your machine.
> |
> | CONFIGURATION
> |
> | Using `lspci -v`, determine the type of USB hardware available/installed.
> |
> |  If you see something like:
> |
> |    USB Controller: ......
> |    Flags: .....
> |    I/O ports at ....
>
> At least my version of lsusb doesn't list anything if the hcd isn't
> loaded. And if it's loaded, you already know which one to use.
> Even if it's loaded, it doesn't show the above mentioned output.
>
> Should this paragraph be just removed or is there an easy way to find
> out if you have ohci or uhci?

Remove it. Just stating that you need working USB should be enough.
You cannot replace a HOWTO with a single paragraph anyway.

> | Beginning with version 0.4 of the driver, up to 16 scanners can be
> | connected/used simultaneously.  For devfs support, see next section.
> | If you intend to use more than one scanner at a time w/o devfs support:
> |
> |   Add a device for the USB scanner:
> |     `mknod /dev/usbscanner0 c 180 48`
> |     `mknod /dev/usbscanner1 c 180 49`
> |                     .
> |                      .
> |     `mknod /dev/usbscanner15 c 180 63`
>
> Shouldn't this be changed to "/dev/usb/scanner0"? At least that's what
> MAKEDEV and Documentation/devices.txt mentions.
>
> Apropos MAKEDEV. I would add a reference to that script here.

Definitely.

[..]
> | There is a small test program (hp_scan.c -- appended below) that can
> | be used to test the scanner device if it's an HP scanner that supports
> | SCL (Scanner Control Language).  Known HP scanner that support SCL are
> | the 4100, 5200, 6200, the 6300 -- note that the 4200 is *not*
> | supported since it does not understand SCL; it's also strongly
> | suspected that the 3300 and the PhotoSmart S20 are not SCL compliant.
> | Hp_scan.c's purpose is to test the driver without having to
> | retrieve/configure SANE.  Hp_scan.c will scan the entire bed and put
> | the output into a file called 'out.dat' in the current directory.  The
> | data in the file is raw data so it's not very useful for imaging.
>
> Should a C program that works with some old HP scanners be really
> included in kernel documentation? I don't think so. If necessary, I
> can put it on a website and add a link here.

How does it hurt having it in the kernel?

> Without getting SANE (or another user space application), the user
> can't do much with his scanner anyway.

But it'll show him whether to use scanner.o or hpusbscsi.o

> | funky result -- Most of the time the data flow between the computer
> | and the scanner goes smoothly.  However, due to whatever reason,
> | whether it be solar flares or stray neutrons, sometimes the
> | communications don't work as expected.  The driver tries to handle
> | most types of errors but not all.  When this message is seen,
> | something weird happened.  Please contact the maintaner listed at the
> | top of this file.
>
> Well, that would be
>
> | Copyright (C) 1999, 2000 David E. Nelson <[EMAIL PROTECTED]>
>
> Add a link to the linux-usb mailing list instead?

Yes.

> | SUPPORTED SCANNERS
> |
> | NOTE: Just because a product is listed here does not mean that
> | applications exist that support the product.  It's in the hopes that
> | this will allow developers a means to produce applications that will
> | support the listed USB products.
> |
> | At the time of this writing, the following scanners were supported by
> | scanner.c:
> |
> |  Acer
> |      Prisa Acerscan 620U & 640U (!)
>
> [...]
>
> Should we really duplicate the list that is already in scanner.h? The
> manufacturer and product names are mentioned in the comments.

This list is getting too large to be useful.

> | MODULE PARAMETERS
> |
> | If you have a device that you wish to experiment with or try using
> | this driver with, but the Vendor and Product ID's are not coded in,
> | don't despair.  If the driver was compiled as a module, you can pass
> | options to the driver.  Simply add
> |
> |   options scanner vendor=0x#### product=0x****
> |
> | to the /etc/modules.conf file replacing the #'s and the *'s with the
> | correct ID's.  The ID's can be retrieved from the messages file or
> | using `cat /proc/bus/usb/devices`. Note that USB /proc support must be
>
> USB /proc support? Is this superceded by usbfs now?
>
> | enabled during kernel configuration.  If the 'scanner' module is
> | already loaded into memory, it must be reloaded for the module
> | parameters to take effect.  In essence, `rmmod scanner; modprobe
> | scanner` must be performed.

Delete that section. Those who don't know that they need proc mounted
on /proc shouldn't be configuring a machine anyway.

> | **NOTE**: In later kernels (2.3.38+), a new filesystem was introduced,
> | usbdevfs.  To mount the filesystem, issue the command (as root):
>
> Isn't this called "usbfs" now?

In 2.5. We should not change the name in a stable series.

[..]
> Just remove the email address/web page link?

Yes.

        Merry Christmas
                Oliver



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to