Date: Tue, 10 Mar 1998 17:04:47 -0600 From: "Eric L. Hernes" <erich@lodgenet.com> To: Peter Dufault <dufault@hda.com> Cc: robsch@robkaos.ruhr.de (Robert Schien), freebsd-scsi@FreeBSD.ORG Subject: Re: Scanner support Message-ID: <199803102304.RAA03015@jake.lodgenet.com> In-Reply-To: Your message of "Tue, 10 Mar 1998 15:18:55 EST." <199803102018.PAA22926@hda.hda.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Peter Dufault writes:
>> I set the symlink
>> ln -s uk0 scanner
>>
>> But when I invoke scanimage I get the following message:
>> Mar 10 18:20:11 robkaos /kernel: uk0: physio split the request.. cannot proc
>eed
>> Mar 10 18:20:11 robkaos /kernel: uk0: physio split the request.. cannot proc
>eed
>>
I saw almost the same thing with user level scsi commands to pt0, I traced
it back to scsi_ioctl.c. around line 290, bp gets bzero()ed, so b_kvasize
is forced to zero, regardless of what was passed, in. I patched as follows:
Index: scsi_ioctl.c
===================================================================
RCS file: /cvs/FreeBSD/src/sys/scsi/scsi_ioctl.c,v
retrieving revision 1.29
diff -u -r1.29 scsi_ioctl.c
--- scsi_ioctl.c 1998/02/01 18:09:46 1.29
+++ scsi_ioctl.c 1998/03/09 23:18:17
@@ -292,6 +292,7 @@
bp->b_bcount = len = screq->datalen;
bp->b_screq = screq;
bp->b_sc_link = sc_link;
+ bp->b_kvasize = len;
if (len) {
struct uio auio;
struct iovec aiov;
It's worked on an ncr 53860, and an adaptec 2940 with blocks of
64 bytes and 64k, maybe sizes too. I'd kind of liked to put
some upper bound on len, rather than just using it, but this has
worked.
>Options (if available) are to configure sane to use read/write
>transfers to some device instead for the data transfer, to set it
>to use smaller block sizes (4K will always work), or a questionable
>idea involving hackery to map a buffer to a contiguous address
>using mmap.
Is the contiguous memory for DMA? Have I just been lucky?
>--
>Peter Dufault (dufault@hda.com) Realtime development, Machine control,
>HD Associates, Inc. Safety critical systems, Agency approval
>
>To Unsubscribe: send mail to majordomo@FreeBSD.org
>with "unsubscribe freebsd-scsi" in the body of the message
>
Eric
--
Eric L. Hernes
erich@lodgenet.com
erich@rrnet.com http://rrnet.com/~erich
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199803102304.RAA03015>