LGTM
http://codereview.chromium.org/1606019/diff/5001/6001 File src/arm/codegen-arm.cc (right): http://codereview.chromium.org/1606019/diff/5001/6001#newcode1014 src/arm/codegen-arm.cc:1014: // Move the rhs to r0. Does this comment still hold? http://codereview.chromium.org/1606019/diff/5001/6001#newcode6400 src/arm/codegen-arm.cc:6400: How about having scratch registers here as well, and then never use any explicit register below? Register t0 = VirtualFrame::scratch0(); Register t1 = ... ... http://codereview.chromium.org/1606019/diff/5001/6001#newcode6413 src/arm/codegen-arm.cc:6413: // This code can't cope with other register allocations yet. Doesn't this assert apply for the GenericBinaryOpStub as a whole? If this move it to the top. http://codereview.chromium.org/1606019/diff/5001/6001#newcode6433 src/arm/codegen-arm.cc:6433: // This code can't cope with other register allocations yet. Ditto. http://codereview.chromium.org/1606019/diff/5001/6001#newcode6439 src/arm/codegen-arm.cc:6439: if (lhs.is(r1)) { Can't you just drop the if/else and use: __ sub(r0, lhs, Operand(rhs), SetCC); // Return if no overflow. __ Ret(vc); __ sub(r0, lhs, Operand(rhs)); // Revert optimistic subtract. http://codereview.chromium.org/1606019/show -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev To unsubscribe, reply using "remove me" as the subject.
