home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.pc-clone.32bit
- Path: sparky!uunet!van-bc!dogtag!rwj
- From: rwj@dogtag.uucp (Rob Janes)
- Subject: scsi tape problems on isc 3.0
- Message-ID: <1993Jan11.013407.2303@dogtag.uucp>
- Reply-To: rwj@bsoh.wimsey.bc.ca
- Organization: BSOH Enterprises
- Distribution: na
- Date: Mon, 11 Jan 1993 01:34:07 GMT
- Lines: 53
-
- I've been kind of banging my head against the wall here- I can't figure
- out how to get this archive viper tape drive recognized by the kernel.
- A scsi probe reveals scsi 1 and 2 are hard disks, and scsi 4 is the
- archive viper. The light on the viper comes on, so it looks like the
- hardware is working to some extent.
-
- I followed the instructions in the manual for tape devices, however,
- it's for the previous release of interactive. So what I've done is
- 1 added to pack.d/gendev/space.c
- extern int tape_init(); under disk_init()
- and
- tape_init, to the gdev_init_routines table.
- 2 fixed up the pack.d/tape/space.c that wouldn't compile, even after
- following the instructions and removing unused driver entries.
- I ended up copying the one from the web, and using that.
- 3 rebuilt the kernel, rebooted, there was no /dev/tape, so I made
- one, major 41, minor (scsi id 4*16+8) 72. Guess the node.d/tape wasn't
- set up.
- none of the above did the trick, so I looked a bit harder.
- 4 copied the tp0t drivers from the web, and installed them.
- made sure the sdevice and mdevice entries for tape and tp0t matched
- those on the web. Assigned an unused major number for tp0t in mdevice.
- (Where do these drivers come from? They don't seem to be on the base
- system. They look like they're from interactive, they've been in all
- the kernels since the upgrade to ix 3.0.)
- this didn't work.
- 5 compared pack.d/gendev/space.c on my machine and the web-
- 3 defines were different, mainly
- GDEV_MAXDRIVS to 14 on web, 13 on my machine
- GDEV_MINORMAPS to 224 on web, something smaller on my machine
- GDEV_MAJORMAPS to 58 on web, 1 on my machine.
- also, web's had an extra entry had been added to the gdev_iobufs table,
- as indicated by the instructions about GDEV_MAXDRIVS.
- So, changed my defines, and the table, to match the web's.
- Still didn't work.
- 6 The major number for tp0t is 62, whereas the major number for tp0t on
- the web is 57. major 58 is lockd on web, probably some nfs function.
- It seems likely nfs stuff would be routed through the hpdd, so maybe
- GDEV_MAJORMAPS has to be set large enough to encompass all the major
- numbers it has to handle. So I bumped GDEV_MAJORMAPS in
- pack.d/gendev/space.c up to 62.
- Still didn't work.
- So I'm at a loss. I'm going to try one more thing. I'm going to bump
- GDEV_MAJORMAPS up to 63, on the assumption that it needs the extra
- room, and nfs has nothing to do with the hpdd (high performance disk drv).
- I don't think it's going to work though.
- ... later ... - It didn't!
- Any suggestions would be much appreciated.
- Thanks!
- Rob Janes
- --
- Rob Janes
- rwj@bsoh.web.apc.org
-