> The coolscan backend suffered from a very similar problem on SGI. The
> workaround has been to limit the size of the data chunks read from the scanner
> to 64K instead of 'sanei_scsi_max_request_size'.
>
> Probably more a workaround than a fix but since there was no perceivable
> performance penalty I kept it that way.
Hi Didier,
it looks like the modulo(%) function:
s->row_bufsize - (s->row_bufsize % row_len)
makes problems with such big arguments (8*1024*1024)
and returns 0 in the described case.
I replaced it by a (long)-integer division in the umax-backend.
Bye
Oliver
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com