[MLton-commit] r5758
Matthew Fluet
fluet at mlton.org
Wed Jul 11 04:50:23 PDT 2007
Float/int conversions require int size suffix to disambiguiate memory operands
----------------------------------------------------------------------
U mlton/trunk/mlton/codegen/amd64-codegen/amd64.fun
----------------------------------------------------------------------
Modified: mlton/trunk/mlton/codegen/amd64-codegen/amd64.fun
===================================================================
--- mlton/trunk/mlton/codegen/amd64-codegen/amd64.fun 2007-07-11 11:13:58 UTC (rev 5757)
+++ mlton/trunk/mlton/codegen/amd64-codegen/amd64.fun 2007-07-11 11:50:21 UTC (rev 5758)
@@ -2007,16 +2007,16 @@
str "2", str "s", Size.layout dstsize],
Operand.layout src,
Operand.layout dst)
- | SSE_CVTSFP2SI {src, srcsize, dst, ...}
+ | SSE_CVTSFP2SI {src, srcsize, dst, dstsize, ...}
=> bin (str "cvt",
seq [str "s", Size.layout srcsize,
- str "2", str "si"],
+ str "2", str "si", Size.layout dstsize],
Operand.layout src,
Operand.layout dst)
- | SSE_CVTSI2SFP {src, dst, dstsize, ...}
+ | SSE_CVTSI2SFP {src, srcsize, dst, dstsize, ...}
=> bin (str "cvt",
seq [str "si",
- str "2", str "s", Size.layout dstsize],
+ str "2", str "s", Size.layout dstsize, Size.layout srcsize],
Operand.layout src,
Operand.layout dst)
| SSE_MOVD {src, dst, ...}
More information about the MLton-commit
mailing list