mangler@cit-vax.Caltech.Edu (System Mangler) (08/20/86)
Subject: 4.3bsd driver for SI 9900? We recently got our 4.3bsd tape, and I notice that hp.c uses the byte count register to figure out where errors occurred, which doesn't work on SI controllers. There's no lost interrupt timeout either. sc_doseeks is never set. Probably lots of other SI kludges missing too. Unfortunately I'm stuck with a couple of SI 9900's. Does anyone have a patched hp.c for 4.3 bsd? Failing that, how about a good 4.2 driver that I can copy from? The latest 4.2 driver that I have is BRL version 1.9 of two years ago. Don Speck speck@vlsi.caltech.edu seismo!cit-vax!speck
irwin@uiucdcs.CS.UIUC.EDU (08/21/86)
I just talked with SI yesterday, re a driver upgrade for the 9900. We are running 4.2 and are using the SI driver. It has a bug in that it can not handle two bad sectors in a row. We have a disk that has two bad sectors in a row, (in the swap area) they are listed in the bad block list, but the system is reporting hard errors on the second of the two bad blocks. We discovered this many months ago, and was told that they were rewriting the driver to fix this. My call yesterday (to Bill McKee at 408-942-1212) was to obtain the status of the rewrite. I was informed that it is complete, the new driver is to work with 4.2 or 4.3 and has the "two in a row" bug fixed. They are sending us a tape with the new driver. I would assume that if you called them, that they would send the new driver which should work with your 4.3 system. I noted that you said hp.c so I know you are on the mass bus/cmi bus and not on the unibus. He questioned this point when I talked to him.
jim@cs.strath.ac.uk (Jim Reid) (08/22/86)
Try your local SI sales office. SI offer "support" to UNIX sites. They will provide a device driver and utilities for bad block handling, formatting and so on. They may not be up to speed for 4.3 yet, but we have the 4.2 versions and have no complaints. [We've only just been told that the US Govt. will allow us to get 4.3 - we don't hve it yet.] Jim ARPA: jim%cs.strath.ac.uk@ucl-cs.arpa, jim@cs.strath.ac.uk UUCP: jim@strath-cs.uucp, ...!seismo!mcvax!ukc!strath-cs!jim JANET: jim@uk.ac.strath.cs "JANET domain ordering is swapped around so's there'd be some use for rev(1)!"
rick@seismo.CSS.GOV (Rick Adams) (08/26/86)
I have a version of the 4.3 BSD hp.c driver that works with our SI 9400 and 9900 controllers (and CDC 9766, 9775, and 9771 disks). I've also fixed the off by one error in the table for the CDC9775 disk drive. (disktab says 842 cylinders, hp.c says 843. I went with 842, since thats what our old driver used) The 4.3 driver works fine if your controller isn't braindamaged. Unfortunately, our SI controllers are, so we had to fix it. It also works with Emulex SC7000 controllers of Rev C (Emulex fixed the problem in Rev D of the firmware. SI sends you driver hacks. Consider this a plug for Emulex if you wish.) It works fine with disks and controllers that aren't broken (Real DEC RM05s). I also have 4.3BSD versions of the multi-user badblock,diskformat and markbad that I got from BRL for 4.2. I think they originated with SI a long time ago. The problem is with incorrectly calculating the forwarded block when doing bad block forwarding in case those of you with rational hardware were wondering. Send me mail and I'll tell you how to obtain the files. ---rick