TM03%SWTEXAS.BITNET@ricevm1.rice.edu (02/23/90)
In upgrading from 1.3 to 1.5.0 to 1.5.3 in the last two weeks I have reached an impasse. Maybe some kind souls out there can help? 1. I can't get kernel/system.c (for 1.5.3) to have the official crc. Could somebody send me via mail an accurate copy. 2. What should the flags be when compiling the various libs? Is there a difference between compile flags for posix , other, and ansi routines? If so, what are they? I assume the flags should be the same whether its 1.5.0 or 1.5.3. The questions arise because of problems I'm having and because 1.5.0 lib/posix makefile appears to be muddled (?) 3. When system.c and floppy.c are compiled, I get warnings about incompatible pointers. Are these to be expected here (and ignored) or have I got some lib or include problem? 4. Now to my real problem. In reading recent postings others seem to have related stories but I've seen no real answer. The file system on my hard disk was set up two years ago using Minix 1.2 and has been good and stable through 1.3 up until now. My 1.5.x Minix will not mount the system (error is: mount: dev/hd2 is not valid file system ) but my 1.3 Minix still does. What has changed between 1.3 and 1.5 that could make this happen? The hard disk check at boot time reports a bad magic number (for both 1.3 and 1.5.) Have I had a problem all this time and 1.3 has missed it? Any solutions? Thanks in advance for any help. By the way, 1.3 to 1.5.x has been sooooo smooooth compared to other upgrades. My complements to one and all. tom mccabe tm03@swtexas.bitnet
ast@cs.vu.nl (Andy Tanenbaum) (02/23/90)
In article <11934@nigel.udel.EDU> TM03%SWTEXAS.BITNET@ricevm1.rice.edu writes: > The file system on my hard disk was set up two years ago using >Minix 1.2 and has been good and stable through 1.3 up until now. >My 1.5.x Minix will not mount the system (error is: > mount: dev/hd2 is not valid file system ) Chances are it is due to the STUPID_WINI_ADJUST problem. MINIX expects blocks to begin at even sectors. MS-DOS does not. That was always solved in an ad hoc way before. Now Bruce has solved it in a very nice way by providing commands in fdisk to manipulate the origin and size of partitions. Take a good look at fdisk. Andy Tanenbaum (ast@cs.vu.nl)