Encoding for the 'b' or 'bl' instruction.

r2 -> rj r3 -> rd.

r1 -> rk r2 -> rj r3 -> rd.

ACMPGED -> fcmp.sle.d.

ACMPGEF -> fcmp.sle.s.

ACMPGTD -> fcmp.slt.d.

ACMPGTF -> fcmp.slt.s.

aliases.

64-bit.

hardware nop.

2.2.10.

64-bit FP.

-0x800 <= v < 0.

0 < v <= 0xFFF.

$n(REG) where int32 < n.

other 64 (could subdivide further).

$n(REG) where int12 < n <= int32.

other 32.

must be the last.

$n(REG) where n <= int12.

12 bit signed.

$tlsvar.

32 bit signed, low 12 bits 0.

not use.

mark flags.

number of floating point registers.

p expands to multiple instructions, but does NOT use REGTMP.

number of general registers.

must be a multiple of 32.

must be a multiple of 32.

only eight registers are needed.

must be a multiple of 32.

only four registers are needed.

the last defined register.

must be a multiple of 32.

-1 disables passing the first argument in register.

context for closures.

G in loong64.

not use.

reserved for runtime, duffzero and duffcopy.

reserved for runtime, duffcopy.

used by the assembler.

set to zero.