Floating-point opcode difference between MASM and JWasm when using operands st,st
Brought to you by:
japheth
Menu
▾
▴
#287 Floating-point opcode difference between MASM and JWasm when using operands st,st
When using st,st on fadd, fmul, fsub, fsubr, fdiv and fdivr, JWasm uses the DCh opcode while MASM uses the D8h opcode.
For example fadd st,st generates D8h C0h in MASM (all versions from 6.00 up) but generates DCh C0h in JWasm. The same occurs if you use any variants like fadd st,st(0) or fadd st(0),st or fadd st(0),st(0)
This appears to be due to the ordering in the tables in instruct.h:
ins (FADD, fadd, 4, OpCls( STI, ST, NONE ), 0, 0, 0, 0xDC, 0xC0, P_87, 0)
insn(FADD, 1, OpCls( ST, STI, NONE ), 0, 0, 0, 0xD8, 0xC0, P_87, 0)
probably should be:
ins (FADD, fadd, 4 OpCls( ST, STI, NONE ), 0, 0, 0, 0xD8, 0xC0, P_87, 0)
insn(FADD, 1, OpCls( STI, ST, NONE ), 0, 0, 0, 0xDC, 0xC0, P_87, 0)
There don't appear to be any issues in changing the order. The same would need to be done for FMUL, FDIV, FDIVR, FSUB and FSUBR.
Ok, will be changed.
Btw, such "issues" ( which are no bugs ) should be opened in the discussion forum.
Verified fixed in v2.11 rc17