[llvm-commits] [llvm] r49869 - /llvm/trunk/lib/Target/X86/X86InstrInfo.td

[Dan Gohman]

On Apr 17, 2008, at 2:26 PM, Evan Cheng wrote:
> Author: evancheng
> Date: Thu Apr 17 16:26:35 2008
> New Revision: 49869
>
> URL: http://llvm.org/viewvc/llvm-project?rev=49869&view=rev
> Log:
> Fix assembly code for atomic operations.
>
> Modified:
>    llvm/trunk/lib/Target/X86/X86InstrInfo.td
>
> Modified: llvm/trunk/lib/Target/X86/X86InstrInfo.td
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.td?rev=49869&r1=49868&r2=49869&view=diff
>
> =
> =
> =
> =
> =
> =
> =
> =
> ======================================================================
> --- llvm/trunk/lib/Target/X86/X86InstrInfo.td (original)
> +++ llvm/trunk/lib/Target/X86/X86InstrInfo.td Thu Apr 17 16:26:35 2008
> @@ -2548,9 +2548,9 @@
>
> let Defs = [EAX, EFLAGS], Uses = [EAX] in {
> def CMPXCHG32 : I<0xB1, Pseudo, (outs), (ins i32mem:$ptr, GR32:$swap),
> -               "cmpxchgl $swap,$ptr", []>, TB;
> +               "cmpxchg{l} $swap,$ptr", []>, TB;
> def LCMPXCHG32 : I<0xB1, Pseudo, (outs), (ins i32mem:$ptr,  
> GR32:$swap),
> -               "lock cmpxchgl $swap,$ptr",
> +               "lock cmpxchg{l} $swap,$ptr",

Isn't it also necessary to switch the order of the operands
between Intel and AT&T syntax for these?

Dan