"l" 32 bit floating point constant in .lit4
MDMX instruction operands (note that while these use the FP register
- fields, they accept both $fN and $vN names for the registers):
+ fields, they accept both $fN and $vN names for the registers):
"O" MDMX alignment offset (OP_*_ALN)
"Q" MDMX vector/scalar/immediate source (OP_*_VSEL and OP_*_FT)
"X" MDMX destination register (OP_*_FD)