cmpxchg16b

[arkon at ragestorm.net]

Oh, one more thing,
If you don't use the default address size (that's rax/rbx etc)
Then the assembler will add its own db 0x48,
And if you use eax, then it will add its own db 0x67. and REX must precede
first opcode byte...

So just look at the binary how it ends up and add the REX.W bit in the REX
prefix. That's the fourth bit (0x48).

If you need further help, just drop me a line.

Gil

-----Original Message-----
From: yasm-devel-bounces at tortall.net [mailto:yasm-devel-bounces at tortall.net]
On Behalf Of arkon at ragestorm.net
Sent: Wednesday, June 07, 2006 10:49 PM
To: 'Vikas Kumar'; yasm-devel at tortall.net
Subject: RE: cmpxchg16b

Hey

Luckily this one is easy:
db 0x48
cmpxchg8b [eax] ; use it the same, this will become cmpxchg16b, of course

There you go.
It reminds me the old DOS days where I used to write db 0x66, but now it's
db 0x48 :)

Good luck
Gil
http://ragestorm.net/distorm/


-----Original Message-----
From: yasm-devel-bounces at tortall.net [mailto:yasm-devel-bounces at tortall.net]
On Behalf Of Vikas Kumar
Sent: Wednesday, June 07, 2006 9:07 PM
To: yasm-devel at tortall.net
Subject: Re: cmpxchg16b

Can you not use the direct opcode for it in your code ?Like
db 0x90
(that is for nop, but i dont remember the opcode for cmpxchg)

Vikas
On 6/7/06, Cory Nelson <phrosty at gmail.com> wrote:
> Is support for the amd64 instruction cmpxchg16b planned for future
> releases?  It's a bit frustrating having to use masm for it :)
>
> --
> Cory Nelson
> http://www.int64.org
> _______________________________________________
> yasm-devel mailing list
> yasm-devel at tortall.net
> http://cvs.tortall.net/mailman/listinfo/yasm-devel
>
_______________________________________________
yasm-devel mailing list
yasm-devel at tortall.net
http://cvs.tortall.net/mailman/listinfo/yasm-devel

_______________________________________________
yasm-devel mailing list
yasm-devel at tortall.net
http://cvs.tortall.net/mailman/listinfo/yasm-devel