> This is a correct table: > > Linux GRUB Mach/Hurd (ext2fs) Mach/Hurd (ufs) > > hda (hd0) hd0 hd0 > hda1 (hd0,0) hd0s1 hd0a > hda2 (hd0,1) hd0s2 hd0b > > The questions of device-naming are too many. Is it better to change > how to name devices? Especially, the difference between ext2fs > partition and ufs partition is confusing for Hurd beginners.
So what is the difference? I thought that ufs is a file system, like ext2fs, so how can the way of naming a partition depend on a file system? And where do BSD partitions fit into the scheme of things? And what is a "slice"? "Mr White" wrote: "But for example a /dev/hdc disk on linux on a system where there is no /dev/hdb (i.e. two master disks, no slaves), grub would call it (hd1)." This is very interesting. When I was trying boot GRUB by manipulating the device number I tried 0x82, as my IDE floppy is /dev/hdc. But I don't have a /dev/hdb, so should this, perhaps, have been 0x81? (It probably still wouldn't have worked, because 0xff, meaning "use the same boot device" didn't work, but this knowledge might be useful later ...) What about the other kinds of partition names? Do any of them not count missing devices? (Sounds like a bad way of doing things to me: what if you have a device that the BIOS sometimes detects and sometimes doesn't; this sort of thing happens a lot.) Edmund
