Decide on which syntax variants for the assembler should be supported
Currently, the supported syntax is somehow mixed.
JALR for example is using the the three operand versions like the following: JALR x1, x2, 4
. LLVM and presumably gas also accept two operand versions JALR x1, x2
as well as offset versions like JALR x1, 0(x2)
. This last variant is similar to what we currently use (exclusively) for load and store operations (e.g., LW x2, 1(x4)
, SW x5, 0x30(x10)
).