Re: 2.4.1-pre8 losing pages

[Peter Horton]

On Fri, Jan 26, 2001 at 03:20:33AM +0100, Xuan Baldauf wrote:
> 
> Peter Horton wrote:
> 
> > I'm experiencing repeatable corruption whilst writing large volumes of
> > data to disk. Kernel version is 2.4.1-pre8, on an 850MHz AMD Athlon on an
> > ASUS A7V (VIA KT133 chipset) motherboard 128M RAM (tested with 'memtest86'
> > for 10 hours).
> >
> 
> So what output does following bash script produce?
> 

Well this is the script I've been testing with ...

   #!/bin/bash -x
   set -e
   uname -a
   rm -f test test[a-z]
   dd if=/dev/urandom of=test bs=1024k count=128
   for I in a b c d e f g h i j k l m n o p q r s t u v w x y z; do
	   cp test test$I
   done
   md5sum test*

... this is the kinda output I get on most runs :-

   Linux mole-rat 2.4.1-pre10 #1 Fri Jan 26 08:48:55 GMT 2001 i686 unknown
   ...
   aa6a64589748321899bab2b66f71427f  testt
   aa6a64589748321899bab2b66f71427f  testu
   aa6a64589748321899bab2b66f71427f  testv
   9dde1bed276e32a1f9af98c87ab05978  testw
   aa6a64589748321899bab2b66f71427f  testx
   aa6a64589748321899bab2b66f71427f  testy
   aa6a64589748321899bab2b66f71427f  testz
   mole-rat:~# cmp testw testx
   testw testx differ: char 110862337, line 433772
   mole-rat:~# cmp -i $(( 110862336 + 4096 )) testw testx
   mole-rat:~# echo $(( 110862336 % 4096 ))
   0

> 
> I cannot reproduce your behaviour in 2.4.1-pre9.
> 

No, I can't find anybody else who can either. Maybe I've got a dodgy CPU
:-(

P.