[sane-devel] 4 years and It's still the same error

[Karl Heinz Kremer]

I also suspect a SCSI error, the backend receives a buffer containing
garbage, two bytes of this buffer are used as the "function level" -
that's EPSON's version of the scanner command set. For the perfection 1200
this command level should be B7. Because most EPSON scanners use the same
subset of commands, it is usually OK to fall back to the B3 level. This
is done mainly for new scanners, so that you can at least use the basic
scan functions.

Because the backend sees nothing that resembles a correct command level
it uses it's default. This alone does not cause the segfault. I don't
know where the segfault happens, but it's certainly possible that it's
in my plug-in. 

What SCSI Adapter are you using? Is the scanner the only device connected
to this adapter? Do you have access to an active terminator? Have you 
used one yet? 

The Perfection 1200 has an internal and automatic termination, but this may
not work correctly. I usually try to terminate my devices with an external
terminator. 

How long is your SCSI cable?

Do you get any errors on the console or in the message log? Does "dmesg"
display any SCSI related error messages? 

I used my Perfection 1200 two days ago with the the pre1 1.0.6 build and
kernel version 2.4.10. So I know that it works.

Karl Heinz

Henning Meier-Geinitz <henning at meier-geinitz.de> said:

> Hi,
> 
> Karl Heinz Kremer will probably comment on this more specifically.
> Just some comments:
> 
> On Mon, Nov 05, 2001 at 08:44:45AM +0100, jebga at tin.it wrote:
> [...]
> > [epson] type    Ð 0xd0
> > [epson] level   Ü 0xdc
> > [epson] no option equipment installed
> > [epson] Unknown type  or level @, using B3
> 
> This seems to be the problem. There are some hits in the sane-devel
> archives about this error and my impression was, that it may be some
> sort of SCSI problem.
> 
> > [epson] open_scanner()xsane at xevilstar:/root$ xsane
> > [epson] Unknown type H or level @, using B3
> 
> This time the type is different (?)
> 
> > Segmentation fault
> 
> This shouldn't happen in any case. Karl Heinz, do you think the
> segfault happens in epson.c or sanei_scsi.c?
> 
> > xsane at xevilstar:/root$xsane at xevilstar:/root$ xscanimage
> > [epson] Unknown type ? or level @, using B3
> 
> Yet another type. Maybe the backend gets garbage from the SCSI-system
> or the variable is uninitialized?
> 
> > pls help I've started four years ago with a phantom mikrotek scanner an
d
> > now I'm here with an epson, four years and the error is always the same
> 
> The error was the same with the Mikrotek? I can't believe that. Or are
> you referring to the "segmentation fault"? This just means, that there
> is a bug in the code that lead to an forbidden access to memeory.
> 
> Bye,
>   Henning
> _______________________________________________
> Sane-devel mailing list
> Sane-devel at www.mostang.com
> http://www.mostang.com/mailman/listinfo/sane-devel
> 



--