Mercurial > hg > openjdk > jdk7 > hotspot
changeset 2542:ccf072cdba91
7046893: LP64 problem with double_quadword in c1_LIRAssembler_x86.cpp
Summary: Fixed invalid casts in address computation
Reviewed-by: kvn, never
Contributed-by: thomas.salter@unisys.com
| author | iveresov |
|---|---|
| date | Tue, 24 May 2011 15:30:05 -0700 |
| parents | 7523488edce5 |
| children | 28a9fe9534ea |
| files | src/cpu/x86/vm/c1_LIRAssembler_x86.cpp |
| diffstat | 1 files changed, 1 insertions(+), 1 deletions(-) [+] |
line wrap: on
line diff
--- a/src/cpu/x86/vm/c1_LIRAssembler_x86.cpp Tue May 24 12:54:51 2011 -0700 +++ b/src/cpu/x86/vm/c1_LIRAssembler_x86.cpp Tue May 24 15:30:05 2011 -0700 @@ -47,7 +47,7 @@ static jlong* double_quadword(jlong *adr, jlong lo, jlong hi) { // Use the expression (adr)&(~0xF) to provide 128-bits aligned address // of 128-bits operands for SSE instructions. - jlong *operand = (jlong*)(((long)adr)&((long)(~0xF))); + jlong *operand = (jlong*)(((intptr_t)adr) & ((intptr_t)(~0xF))); // Store the value to a 128-bits operand. operand[0] = lo; operand[1] = hi;