I loved Hackers - I thought it was an entertaining film. I don't expect films abut computers to portray speccy, social misfits or reading through lines of source code or endless tweaks and recompiles.
At least hackers actually had something of substance in it. When my brother watched it he enjoyed purely at an aesthetic level, Whilst I also got something out of the fact that Phreak dialed the operator using the old tapping the cradle trick.
It also showed them poring over reams of hard copy and quoted the Hackers' Manifesto.
On the down side "28.8 BPS modem"... "killer refresh rate" on a TFT screen... purlease.
Alledgedly Echelon uses powerful computers called Dictionaries (dig that crazy name) that grep all phone conversations for 'suspect' keywords.
If you're really being paranoid then you need to scramble all telephone calls and faxes and none of this bullshit stuff you get from places like SpyMaster (a business security firm in England playing on the paranoia and ignorance of yer average suit) which merely shift the signal 90 degrees out of phase.
To get the x86 into single stepping mode you OR the psw with a special number (ie set a single bit), this means it will generate an interrupt every single instruction. To turn in off you just AND the psw with an inverted (NOT-ted)version of this special number.
From the Minix src code (only because I've got it to hand, it'd be the same for Linux)
src/kernel/const.h
(line 9) for INTEL chips #define TRACEBIT 0x100 /* OR this with psw in proc[] for tracing */
(line 109) for M68000 #define TRACEBIT 0x8000
and from my src code for a minix debugger
/* Set trace bit */ child_proc->p_reg.psw |= TRACEBIT;
... and later...
/* Clear the tracebit */ child_proc->p_reg.psw &= ~TRACEBIT;
As a project at University we had to implement a debugger in Minix such that executing
debug
would run the program using an execdbg (like execle but slightly modified) call and run the program *htrough* the debugger. At the heart of it was an extra couple of lines in the interupt handler that caught an interrupt either after every instruction was executed *OR* (and this is the bit that's important here) caught an interrupt we'd placed.
What we'd done was implement a feature which given an address woudl copy that instruction out, store it somewhere and then replace it with an interrupt. When that interrupt was trapped the progarm counter would be wound back one and the interrupt replaced by the correct instruction. Execution would be resumed on some input from the user.
It wouldn't be very hard for the VM, on loading the progarm, to scan for 'dodgy' instruction.s These could be replaced by an interrupt and pushed onto the end of a linked-list. Should this interrupt be trapped then the program counter would be wound back one and the interrup replaced by the head of the linked list.
Slashdot Mirror
No it hasn't. If you're talking about Ming this (my) project's been around for longer.
I haven't even touched it since June 'cos I went off travelling round the world.
Actually this project was started before Ming.
Which I link from my page.
I started this project about a year ago and the current version has been sitting around for about 6 months doing nothing whilst I went travelling.
Read
/ fi nal_project.html#_Toc486319426
http://www.2shortplanks.com/Flash/stuff/Writeup
for a history
Simon
The 'whatos' software, that works out what Operating System a remote machine is running on, is also written in Perl.
I loved Hackers - I thought it was an entertaining film. I don't expect films abut computers to portray speccy, social misfits or reading through lines of source code or endless tweaks and recompiles.
... "killer refresh rate" on a TFT screen ... purlease.
At least hackers actually had something of substance in it. When my brother watched it he enjoyed purely at an aesthetic level, Whilst I also got something out of the fact that Phreak dialed the operator using the old tapping the cradle trick.
It also showed them poring over reams of hard copy and quoted the Hackers' Manifesto.
On the down side "28.8 BPS modem"
Kickin' sound track though
Didn't Douglas Coupland mention something like this in Microserfs? I think he referred to it as Dyspraxia.
I seem to remember some quote from an NSA official that went along the lines of
... pretty good"
"PGP is just that
Alledgedly Echelon uses powerful computers called Dictionaries (dig that crazy name) that grep all phone conversations for 'suspect' keywords.
If you're really being paranoid then you need to scramble all telephone calls and faxes and none of this bullshit stuff you get from places like SpyMaster (a business security firm in England playing on the paranoia and ignorance of yer average suit) which merely shift the signal 90 degrees out of phase.
To get the x86 into single stepping mode you OR the psw with a special number (ie set a single bit), this means it will generate an interrupt every single instruction. To turn in off you just AND the psw with an inverted (NOT-ted)version of this special number.
/* Set trace bit */
...
/* Clear the tracebit */
From the Minix src code (only because I've got it to hand, it'd be the same for Linux)
src/kernel/const.h
(line 9) for INTEL chips
#define TRACEBIT 0x100
/* OR this with psw in proc[] for tracing */
(line 109) for M68000
#define TRACEBIT 0x8000
and from my src code for a minix debugger
child_proc->p_reg.psw |= TRACEBIT;
... and later
child_proc->p_reg.psw &= ~TRACEBIT;
As a project at University we had to implement a debugger in Minix such that executing
debug
would run the program using an execdbg (like execle but slightly modified) call and run the program *htrough* the debugger. At the heart of it was an extra couple of lines in the interupt handler that caught an interrupt either after every instruction was executed *OR* (and this is the bit that's important here) caught an interrupt we'd placed.
What we'd done was implement a feature which given an address woudl copy that instruction out, store it somewhere and then replace it with an interrupt. When that interrupt was trapped the progarm counter would be wound back one and the interrupt replaced by the correct instruction. Execution would be resumed on some input from the user.
It wouldn't be very hard for the VM, on loading the progarm, to scan for 'dodgy' instruction.s These could be replaced by an interrupt and pushed onto the end of a linked-list. Should this interrupt be trapped then the program counter would be wound back one and the interrup replaced by the head of the linked list.
... was in Edge magazine over here in Britain this month. Verbatim. Which one came first?