Kernel messages: ``bad page table'' and ``swap_free: weirdness''

Kernel messages: ``bad page table'' and ``swap_free: weirdness''

Post by Jan Havem » Sun, 17 Jul 1994 21:44:57



After replacing my 8x1M RAM by 4x4M I started getting segmentation faults
when using more than some 10M of memory. The kernel started spitting out
error messages, most often ``swap_free: weirdness'' or
``Trying to free/duplicate nonexistent swap-page'' and every now and then
``Bad page table: [00...CD0]=10000000'' where the ... stands for 3 random
hex digits.
I've got the Linux 1.0.9 kernel installed using the Slackware 2.0.0 setup,
the swap-partition was initialised ok using mkswap.
Is this an indication that the memory chips can't handle Unix? They're 70ns
chips, and with DOS or Windows stuff using lots of memory there are no
problems (so somehow I'll have to convince the shop to let me trade them
in for new ones then).

Any help to tell me what I can do about this will be appreciated.

Jan.

     ============================================================


                                     phone: +31-30-534504
                                     fax:   +31-30-537555

          Dept. of Physics and Astronomy, Utrecht University
          PO box 80.000, 3508 TA  Utrecht, The Netherlands

     ============================================================

 
 
 

Kernel messages: ``bad page table'' and ``swap_free: weirdness''

Post by richa » Mon, 18 Jul 1994 01:49:49


: After replacing my 8x1M RAM by 4x4M I started getting segmentation faults
: when using more than some 10M of memory. The kernel started spitting out
: error messages, most often ``swap_free: weirdness'' or
: ``Trying to free/duplicate nonexistent swap-page'' and every now and then
: ``Bad page table: [00...CD0]=10000000'' where the ... stands for 3 random
: hex digits.
: I've got the Linux 1.0.9 kernel installed using the Slackware 2.0.0 setup,
: the swap-partition was initialised ok using mkswap.
: Is this an indication that the memory chips can't handle Unix? They're 70ns
: chips, and with DOS or Windows stuff using lots of memory there are no
: problems (so somehow I'll have to convince the shop to let me trade them
: in for new ones then).

unix puts a lot more stress on memory than dos/windows.  the problem
could well be the chips.

 
 
 

1. Error message 'can't read bad track table' during installation

I'm attempting to install OpenServer 5.0.0. on a 200mhz Pentium 32mb Ram
4.2GB Hard disk.
Both the Cdrom and Hard disk are EIDE and I used (after lots of failed
attempts) the following boostring :

biosgeom restart ahslink=wd disable=dptr

incorporating oss451b

The hard disk already contains 3GB of Windows95 which i wish to preserve
and am therefore using the interactive
fdisk/divvy setup.

I have managed to get to setting a unix partition and activating it,
however when the program leaves fdisk it returns
the message :

Can't read bad track table

I cant' seem to get past this stage as the installation program only
provides me with fdisk again. I can if necessary wipe the whole disk, but I
dont' want to, I would like to preserve what I already have, any ideas
anyone?

2. IP alias on 1 linux BOX

3. 'Bad page table' error and console

4. freebsd support Compaq Integrated NetFlex-3/P network card?

5. 'info' displays hyphens in man pages as '^i' instead of '-'

6. parport0: unable to get hardware to register plip0

7. 'IRQ timeout' message and 'DMA disabled' with 'ls -l'

8. Linux freezes occassiona

9. 'mismatched read page pointers' kernel messages - on my poor thing only?

10. why does 'admintool' hide my 'auto_master' table ?

11. 'ld' and 'bad magic number'

12. A minimal 'man' page for the 'ip' command

13. all man pages show '-' chars as '???'