http://codereview.chromium.org/6530002/diff/1/src/arm/assembler-arm.cc
File src/arm/assembler-arm.cc (right):

http://codereview.chromium.org/6530002/diff/1/src/arm/assembler-arm.cc#newcode1860
src/arm/assembler-arm.cc:1860: mov(ip, Operand(offset));
On 2011/02/16 16:48:54, Rodolph Perfetta wrote:
The offset may be encodable in an add/sub immediate, so it would be
better not
to move offset into ip and let the assembler decide: add(ip, base,
Operand(offset)); ...

Done.

http://codereview.chromium.org/6530002/diff/1/src/arm/assembler-arm.h
File src/arm/assembler-arm.h (right):

http://codereview.chromium.org/6530002/diff/1/src/arm/assembler-arm.h#newcode450
src/arm/assembler-arm.h:450: bool OffsetIsEncodable() const {
Changed name to OffsetIsUint12Encodable On 2011/02/16 16:48:54, Rodolph
Perfetta wrote:
load/store word and byte support a 12bits offset but load/store
halfword and
signed byte support a smaller offset. Furthermore this patch add
support for
vldr/vstr and there the offset is 8bits (shifted left by 2 to get the
value).

http://codereview.chromium.org/6530002/

--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev

Reply via email to