Together with Abel Deuring I have been doing some testing
using "direct IO" to speed data transfers with SANE when
using lk 2.4 kernels. Abel has had the code in place for
a while (both in 1.0.3 and 1.0.4) but it hasn't yet been
used in "anger". Anyway I found a small problem and have
been unable to contact Abel. Since I don't have CVS access
perhaps someone could apply the patch below.
--- sane-backends-1.0.4/sanei/sanei_scsi.c Thu Nov 30 15:02:30 2000
+++ sane-backends-1.0.4/sanei/sanei_scsi.c_dio Sun Apr 1 10:38:55 2001
@@ -2022,7 +2022,7 @@
else
{
/* check for errors, but let the sense_handler decide.... */
- if ( (req->sgdata.sg3.hdr.info && SG_INFO_CHECK != 0)
+ if ( (req->sgdata.sg3.hdr.info & SG_INFO_CHECK != 0)
|| (req->sgdata.sg3.hdr.sb_len_wr > 0 &&
((req->sgdata.sg3.sense_buffer[0] & 0x7f) != 0)))
For any adventurous people who have a fast scsi scanner and
are running lk 2.4 then to test this you will need to get
sg version 3.1.18 (at http://www.torque.net/sg ), apply
the above patch and rebuild SANE. First try a scan to make
sure everything is still working. Then do
"echo 1 > /proc/scsi/sg/allow_dio" and try to scan again.
It works for me, but given the data transfer speed of my
UMAX 1220S, it is hardly likely to make a measurable
performance impact. I would be interested in reports.
Doug Gilbert
-- 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 : Sun Apr 01 2001 - 08:50:45 PDT