Date: Sun, 21 Jun 1998 06:52:33 -0400 (EDT) From: Peter Dufault <dufault@hda.com> To: ken@plutotech.com (Kenneth D. Merry) Cc: darrylo@sr.hp.com, ken@plutotech.com, freebsd-scsi@FreeBSD.ORG Subject: Re: Rolling CAM in, what is still needed? Message-ID: <199806211052.GAA27439@hda.hda.com> In-Reply-To: <199806210307.VAA20507@panzer.plutotech.com> from "Kenneth D. Merry" at "Jun 20, 98 09:07:42 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> Darryl Okahata wrote...
> > "Kenneth D. Merry" <ken@plutotech.com> wrote:
> >
> > > - How many more programs use this (scsireq) interface?
> >
> > Does /sbin/scsi work (or has it been ported)? If so, great. I've
> > got some perl scripts that fdisk/disklabel/newfs arbitrary disks, and
> > they rely upon scsi(8).
>
> Well, scsi(8) has not been ported, it has been replaced by
> camcontrol. camcontrol should be able to do everything that scsi(8) did,
> and more.
People can also dig up "scsinew.tgz" which is a slicing-dicing version
of scsi(8) that is a superset but can run as a server, generate
Tcl data structures, and has better abilities for command scripts,
can generate code stubs from the ascii CDB description, etc:
> # ndefects plist glist
> ndefects "37 0 0:3 {plist} v:1 {vlist} v:1 5:3 0 0 0 0 4:i2 0" \
> -i 4 "*i2 {Defect list length} i2"
>
> # defects plist glist size
> defects "37 0 0:3 {plist} v:1 {vlist} v:1 5:3 0 0 0 0 {xfer} v:i2 0" \
> -i v "s4 ( {Cylinder} i3 {Head} i1 {Sector} i4)"
This lets scsi(8) recognize commands such as this when set to Tcl
output mode:
> rt# ./scsinew -f sd0 -S tcl -C caps -I
> scsi> ndefects 1 1
> 0 {80}
> scsi> defects 84 1 1 84
> 0 {{771 0 18} {771 0 19} {1671 4 87} {1703 3 80} {1704 3 80} {1704 3 81}
{1866 1 95} {1922 0 47} {1932 3 33} {2184 3 8} }
> scsi>
Or this for code output:
> scsi> ndefects
> errval dev_ndefects(struct scsi_link *sc_link, const int input_arg0,
> const int input_arg1, int *output_arg0)
> {
> int ret;
> static const u_char cdb_tmplt[] =
> {0x37,0x00,0x05,0x00,0x00,0x00,0x00,0x00,0x04,0x00,};
> u_char cdb[sizeof(cdb_tmplt)];
> u_char data[4];
> memcpy(cdb, cdb_tmplt, sizeof(cdb));
> scsi_bits_set(cdb + 2, input_arg0, 1, 3); /* "plist" get b1 */
> scsi_bits_set(cdb + 2, input_arg1, 1, 4); /* "vlist" get b1 */
> if ( (ret = scsi_scsi_cmd(sc_link,
> (const struct scsi_generic *)cdb,
> sizeof(cdb),
> data,
> 4,
> dev_retries,
> 2000,
> NULL,
> SCSI_DATA_IN)) == 0)
> {
> *output_arg0=scsi_2_byte(data + 2); /* "Defect list length" put i2 */
>
> }
> return ret;
>}
I don't know where it is archived now. There wasn't much interest
in this the last time I showed it around.
As I've said in the past, IMHO programs are better than libraries.
Peter
--
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
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806211052.GAA27439>