dcd@tc.fluke.COM (David Dyck) (12/27/89)
Warning about 1.5.0 (maybe I did something wrong,
maybe I didn't......)
I was able to extract/patch all the 1.5.0 postings
to match the crc's except a few of the
doc and man files, and have built new library and kernel,
mm, fs, and tools.
Unfortunately, when I booted the new kernel, the superblock of my
hard disk partition was no longer readable to
1.5.0 and protected mode 1.3. BE CAREFUL!
I had been running protected mode 1.3, and have had ZERO problems
with the minix partitions due to minix. I find it suspicious
that the first time I boot minix 1.5.0 that the superblock got
destroyed. Only the superblock of my 'root' partition (/dev/hd6)
was corrupted, I was able to access my /usr partition (/dev/hd7 ok)
using 1.3 later, but no luck yet with /dev/hd6.
I just wanted to WARN other's to be careful, but hopefully they
will not have the same problem. I would look into this more,
but I felt that it warented a QUICK posting immediately in case
there was some REAL problem. If not, sorry for the scare.
Here are some other observations about 1.5.0.
I noticed by comparing the source files mentioned in ast's liborder
and the lib files posted, that the library functions
memmove.s, strcol.s and strxfrm.s are not posted.
Also the run script for posix does not include the _exit.c ( it
should include [_a-z]*.c instead of just [a-z]*.z.
Also some makefiles reference /lib/libc.a and others refernece /usr/lib/libc.a
more later.....
David Dyck
Domain: dcd@tc.fluke.COM
Voice: +1 206 356 5807
UUCP: {uw-beaver,decwrl,microsof,sun}!fluke!dcd
Snail: John Fluke Mfg. Co. / P.O. Box 9090 / Everett WA 98206-9090flong@sdsu.edu (Fred J. E. Long) (01/03/90)
In article <13505@fluke.COM> dcd@tc.fluke.COM (David Dyck) writes: > >Warning about 1.5.0 (maybe I did something wrong, > maybe I didn't......) > >I was able to extract/patch all the 1.5.0 postings >to match the crc's except a few of the >doc and man files, and have built new library and kernel, >mm, fs, and tools. > >Unfortunately, when I booted the new kernel, the superblock of my >hard disk partition was no longer readable to >1.5.0 and protected mode 1.3. BE CAREFUL! > I'm having a similar problem. I compiled libc.a and rebuilt the kernel with seemingly no problems, but for some reason minix 1.5.0 says one of my filesystems is bad, but works OK with the other partition. After trying bios_wini among other things, somehow the superblock on my "bad" partition got cleared. I had to use Norton Utilities to rebuild it by hand. I still don't know what's wrong. I'm going to try using the old libc.a from 1.3 and see what happens. FL ----- Fred J. E. Long ARPA: flong@midgard.ucsc.edu, flong@sdsu.edu UUCP: ...!ucsd!sdsu!flong -- Fred J. E. Long ARPA: flong@midgard.ucsc.edu, flong@sdsu.edu UUCP: ...!ucsd!sdsu!flong
Peter_Van_Epp@cc.sfu.ca (01/04/90)
>From: flong@SDSU.EDU >Subject: Re: warning about 1.5.0 (root partition trashed) > >In article <13505@fluke.COM> dcd@tc.fluke.COM (David Dyck) writes: >> >>Warning about 1.5.0 (maybe I did something wrong, > maybe I didn't......) >> >>I was able to extract/patch all the 1.5.0 postings >>to match the crc's except a few of the >>doc and man files, and have built new library and kernel, >>mm, fs, and tools. >> >>Unfortunately, when I booted the new kernel, the superblock of my >>hard disk partition was no longer readable to >>1.5.0 and protected mode 1.3. BE CAREFUL! >> > >I'm having a similar problem. I compiled libc.a and rebuilt the kernel >with seemingly no problems, but for some reason minix 1.5.0 says one >of my filesystems is bad, but works OK with the other partition. After >trying bios_wini among other things, somehow the superblock on my "bad" <partition got cleared. I had to use Norton Utilities to rebuild it by >hand. I still don't know what's wrong. I'm going to try using the >old libc.a from 1.3 and see what happens. > >FL >----- >Fred J. E. Long >ARPA: flong@midgard.ucsc.edu, flong@sdsu.edu >UUCP: ...!ucsd!sdsu!flong >-- I can probably shed some light on this, if the partition that was trashed/ wouldn't mount was the first partition on your hard disk then you need to modify xt_wini.c (and probably at_wini.c too but I don't know!) to not bother checking for an old style partition (the code isn't here so I can't give the exact symbol name but send me mail if you need to know), the problem is that the code is looking for a partition type of 0x80 to indicate an old style partition needing the adjustment (that STUPID_WINI_ADJUST was supposed to set but doesn't!), I suspect that the 1.3 Fdisk sets this but since I didn't use Minix fdisk to set up my disk mine isn't. Forcing this adjustment always allowed me to mount my hard drive, when the unmodfied code reported an invalid file system. I suspect that the superblock corruption is happening when the wini driver for some reason tries to write something to the disk when it isn't correctly aligned on the partition. I intend on backing everything up and using the 1.5.0 fdisk to make a new style non adjusted partition later and then remove the patch in xt_wini.c. Hope this helps!