On Tue, Jan 09, 2001 at 04:03:36PM -0500, Douglas Gilbert wrote:
> by timeouts ** resulting in scsi bus resets. Anyway the problem
> seems to disappear with the recently released SANE 1.0.4 .
> [The original report was based on SANE 1.0.3 and earlier.]
In fact I was wrong when I said this. :-(
> ** SANE's newer sg interface shortens the per command timeout
> from 10 minutes to 10 seconds. Most other OSes have a timeout
> value of 1 minute or more. I suspect 10 seconds may be too short.
Using this patch the problem really did go away.
Tim.
*/
--- sane-backends-1.0.4/sanei/sanei_scsi.c.orig Wed Jan 10 01:41:09 2001
+++ sane-backends-1.0.4/sanei/sanei_scsi.c Wed Jan 10 01:41:13 2001
@@ -1890,7 +1890,7 @@
(const void*) req->sgdata.sg3.hdr.cmdp = cmd;
req->sgdata.sg3.hdr.sbp = &(req->sgdata.sg3.sense_buffer[0]);
/* 10 seconds should be ok even for slow scanners */
- req->sgdata.sg3.hdr.timeout = 10000;
+ req->sgdata.sg3.hdr.timeout = 10 * 60 * 1000;
req->sgdata.sg3.hdr.flags = SG_FLAG_DIRECT_IO;
req->sgdata.sg3.hdr.pack_id = pack_id++;
req->sgdata.sg3.hdr.usr_ptr = 0;
-- 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 : Tue Jan 09 2001 - 17:27:27 PST