Date: Wed, 10 Jul 1996 17:07:16 -0400 From: "Clement T. Cole" <clemc@echo.ccc.com> To: jkh@cdrom.com, clemc@echo.ccc.com Subject: Re: FreeBSD 2.1 and Adaptec 1542 SCSI Strangeness Message-ID: <9607101707.aa00246@echo.ccc.com.> Resent-Message-ID: <7676.837073643@time.cdrom.com>
next in thread | raw e-mail | index | archive | help
More info...
As I type this message, I'm trying an end-around that gets farther...
but fails basically in the same way -- data corruption.
I think I have the ``goods'' on the fact that there is an SCSI
I/O problem (maybe in overlapped operations) in the 2.1 driver
on the ISA 1542 driver, since this works under FreeBSD 2.0.5
on this hardware (as well as a few other OS's).
Note I have a second copy of the 2.1 CD (that I borrowed)
since the boot floppy trick is hanging in the SCSI driver... I need
two copies so I can get the original copy of the system in memory.
Set Up..
0.) Using took my working 386/FreeBSD 2.0.5 system.
put one of the FreeBSD 2.1 disk1's in its cdrom.
put the other FreeBSD 2.1 disk1's in the ``target'' (486)
system's cdrom
make sure you have a 2.1 ``root'' floppy.
1.) Boot DOS on the 486 and then use fbsdboot to load the 2.1 kernel.
2.) Using the ftp install, get the ``dists'' from the 386 across the
ethernet (and the root from the floppy)
What I'm doing.... after using the local CDROM to load the kernel,
the only SCSI activity is the DISK (and the 1452). So only a single
device is on the bus, but whatever overlap the driver does
with a single drive are occuring (We are also now generating ethernet
traffic).
Results... I get >>much<< further - but I eventually panic
with an inode corruption. Clearly a SCSI operations is not being
handled properly. This is different from an ``hang'' mind you, but
I bet its related.
Clem
PS Maybe you guys should add me to your ``QA'' group. Clearly,
I try things other folks may not ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9607101707.aa00246>