You've tried my python example right? I just finished making the display
changes, but I should be able to finish the statistics and verification
by tomorrow. it's bedtime for me :)

On Thu, 2009-06-25 at 10:57 +0900, tof wrote: 
> hello
> 
> 
> thanks everyone for the help
> 
> 
> 
>  >> 1) what if, a memory address is referenced by multiple dumps
>  > just put the addresses next to each other - order should not matter
> 
> 
> we must do more :
> 
> about 80% of the lines values are corrupt, we need to filter them.
> we will have 2-10 different values for each line
> 
> the way to detect the corrupt values is to classify into 3 groups
> 
> if the line (4words x 4 bytes) has a 16 bit repeating pattern, put in group 3 
> (lowest prio)
> else if the line has a 32 bit repeating pattern, put in group 2
> else if the line has no repeating pattern, put in group1
> 
> 
> after all the lines with the same adress are classified in groups :
> 
> if there are values in group1, take the most recurrent value
> else if there are values in group2, take the most recurrent value
> else if there are values in group3, take the most recurrent value
> 
> 
> 
> Finally we need to mark non present lines in the middle, to have a consistant 
> reading.
> example of output :
> 
> 0x085edf00: 085ee2dc 2200dfd8 085fc3ac 0801eec0
> 0x085edf10: ???????? ???????? ???????? ????????
> 0x085edf20: 09eff6ac 00000041 00000400 00001000
> 
> 
> but it would be a good idea also to stripp this filling lines if they are too 
> much at a time, example of output :
> 
> 
> 0x085edf00: 085ee2dc 2200dfd8 085fc3ac 0801eec0
> 0x085edf10: ???????? ???????? ???????? ????????
> 0x085edf20: 09eff6ac 00000041 00000400 00001000
> ..........
> 0x08700000: 085ee2dc 2200dfd8 085fc3ac 0801eec0
> 0x08700010: ???????? ???????? ???????? ????????
> 0x08700020: 09eff6ac 00000041 00000400 00001000
> 
> 
> in attachement, a log collection example (one of the first i did, not very 
> much correct values)
> 
> 
> 
>  >> 2) what should be the output format, in what file structure
>  > plaintext is fine. the point of this program is to sort the addresses
>  > for a human to read and analyze. no special format is required
> 
> yes, i confirm
> 
> 
> sto
> 
> 
> 
> Cory Walker a écrit :
> > This is what I assume, but feel free to correct me tof:
> >> 1) what if, a memory address is referenced by multiple dumps
> > just put the addresses next to each other - order should not matter
> >> 2) what should be the output format, in what file structure
> > plaintext is fine. the point of this program is to sort the addresses
> > for a human to read and analyze. no special format is required
> > 
> > On Wed, 2009-06-24 at 22:37 +0300, Bahattin TOZYILMAZ wrote:
> >> i can help. i can write a code that will search a directory for plaintext
> >> memory dumps and sort adresses. it will probably be dead slow and a litle
> >> bit buggy, but it will work(i hope). give me (and others) a deadend. and 
> >> can
> >> you answer that questions
> >> 1) what if, a memory address is referenced by multiple dumps
> >> 2) what should be the output format, in what file structure
> >>
> >> regards,
> >> baha
> >> _______________________________________________
> >> Linux4nano-dev mailing list
> >> [email protected]
> >> https://mail.gna.org/listinfo/linux4nano-dev
> >> http://www.linux4nano.org
> > 
> > 
> > _______________________________________________
> > Linux4nano-dev mailing list
> > [email protected]
> > https://mail.gna.org/listinfo/linux4nano-dev
> > http://www.linux4nano.org
> > 
> _______________________________________________
> Linux4nano-dev mailing list
> [email protected]
> https://mail.gna.org/listinfo/linux4nano-dev
> http://www.linux4nano.org


_______________________________________________
Linux4nano-dev mailing list
[email protected]
https://mail.gna.org/listinfo/linux4nano-dev
http://www.linux4nano.org

Reply via email to