pfile@sprite.berkeley.edu (Rob Pfile) (03/24/91)
Hi folx, I had sort of a crazy idea the other day. I do a lot of file transfers between the unix/internet world to Macs, and I usually connect a serial cable to a sparcstation and zmodem the stuff over at 19.2Kbaud. A couple of weeks ago, I finally got A/UX for my macintosh at home, and started formatting disks for use with A/UX, etc, and noticing that A/UX can mount BSD 4.2 filesystems, thought maybe that somehow I could get my mac to mount a sunOS 4.1 filesystem in order to transfer large programs (like binaries for X, gnu stuff, etc) fast. I started playing around with this today, and just tried the naive approach of using mount on A/UX. It did not work. Bunch of errors about not being able to determine the type of the filesystem...I am used to the sun world of /dev/sd1[a..g], etc, and am pretty unfamiliar with (what I assume is) the system-Vish /dev/dsk/ type stuff. In particular, I did not know which slice to try (the sunOS disk has a, b and g partitions) and know that these slices probably have nothing to do with the sunOS partitions. I am particularly curious as to why there are the "extra" slices cXd0s30, 32 in addition to cXd0s3, and why trying cXd0s30 gave a different error: invalid file system type. Anyway, I realize that an A/UX disk is fundamentally different from a sunOS disk given the way they are partitioned. Upon reflecting upon this post, I realize that the slice and partition information is probably contained in the partition map, which I assume is fixed on the disk in both cases, and has to be read by the driver to determine where the partitions physically reside. Trying to mount the sunOS disk referring to it as an AUX device, then of course makes no sense (but hey, it was worth a shot) So, it looks like I need to write a special driver. After failing with the mount command, I started trying to read the label off of the sunOS disk. What I need to do is seek to cyl 0, head 0, sector 0 and read the scsi block. How can this be accomplished given the partition problem above? I can't remember if the raw disk devices are of the form cXd0sX or just cXd0 which would make more sense. Has anyone tried this? Any ideas? Rob Pfile pfile@sprite.berkeley.edu