[cfe-commits] [PATCH] MIPS built-in function overloading

Simon Atanasyan satanasyan at mips.com
Wed Jul 11 12:22:56 PDT 2012

On Sun, Jul 8, 2012 at 11:10 PM, Simon Atanasyan <satanasyan at mips.com> wrote:
> Some MIPS built-in functions have two versions. The first one is used
> for a register operand. The second one is used for a immediate number.
> For example, the second argument of __builtin_mips_shll_qb can be a
> register operand or a constant expression.
> v4i8 __builtin_mips_shll_qb(v4i8, imm0_7)
> v4i8 __builtin_mips_shll_qb(v4i8, i32)
> Attached patches implement MIPS built-in function overloading (now for
> __builtin_mips_shll_qb only). EmitMipsBuiltinExpr routine checks
> whether the second argument is a constant expression and use
> corresponding DSP instruction to emit call expression.
> Could you please review the patches? Is this approach for built-in
> function overloading correct?



More information about the cfe-commits mailing list