Re: segfault with Microtek Scanmaker 330

Bernd Schroeder (bernd@aquila.muc.de)
Wed, 29 Jul 1998 00:51:34 +0200

Hi,

On Mon, Jul 27, 1998 at 11:31:54AM -0100, Norman Uhlmann wrote:
> Hi all,
>
> I recompiled the kernel and the sane package, with
> sg.h #define SG_BIG_BUFF 32768 and it works now.

Good to hear that a mode exists that works for you. However, the
problem that the backend returns an error for bigger buffer sizes still
exists.

The backend itself doesn't call a function in the sane_read() or
scsi_read_image() function (which report the 'out of memory' error), that
might cause this error. Most likely it is a write()-call that is issued
in one of the functions in sanei/sanei_scsi.c. You can verify this, if
you enable debugging with SANE_DEBUG_SANEI_SCSI set to 128, for example.

If you see a debug message similar to 'SG_BIG_BUF inconsistency', then
there is still a problem with the value of SG_BIG_BUF.

There are several options:

- Try it with a value for SG_BIG_BUF that is lower than the recommended value
(for example 64 KB or 120 KB)

- The handling of the buffers has changed since version 0.1 of the
backend. Try it with the backend that is included in sane-0.74.
If the problem doesn't disappear further debugging is required.

> After that I got brave and compiled with
> sg.h #define SG_BIG_BUFF 512000 but
> it doesn't work :-((.

To my knowledge the maximum value for SG_BIG_BUFF is 128KB - 0.5KB.

Bernd

-- 
Bernd Schroeder 
Email: mailto:bernd@aquila.muc.de
PGP public key available: mailto:pgp@aquila.muc.de | Subject: send key 

--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com