Re: [PATCH v3 08/16] target/mips: Introduce decode tree bindings for microMIPS ISA

[Richard Henderson]

On 11/26/24 07:59, Philippe Mathieu-Daudé wrote:
> From: Philippe Mathieu-Daudé <f4bug@amsat.org>
>
> Introduce the microMIPS decodetree configs for the 16-bit
> and 32-bit instructions.
>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> ---
>   target/mips/tcg/translate.h               |  2 ++
>   target/mips/tcg/micromips16.decode        | 11 +++++++++++
>   target/mips/tcg/micromips32.decode        | 11 +++++++++++
>   target/mips/tcg/micromips_translate.c     | 14 ++++++++++++++
>   target/mips/tcg/micromips_translate.c.inc |  9 +++++++++
>   target/mips/tcg/meson.build               |  3 +++
>   6 files changed, 50 insertions(+)
>   create mode 100644 target/mips/tcg/micromips16.decode
>   create mode 100644 target/mips/tcg/micromips32.decode
>   create mode 100644 target/mips/tcg/micromips_translate.c
>
> diff --git a/target/mips/tcg/translate.h b/target/mips/tcg/translate.h
> index d1aa811cfa1..2a079cb28d9 100644
> --- a/target/mips/tcg/translate.h
> +++ b/target/mips/tcg/translate.h
> @@ -222,6 +222,8 @@ bool decode_ase_mxu(DisasContext *ctx, uint32_t insn);
>   bool decode_64bit_enabled(DisasContext *ctx);
>   
>   /* decodetree generated */
> +bool decode_isa_micromips16(DisasContext *ctx, uint16_t insn);
> +bool decode_isa_micromips32(DisasContext *ctx, uint32_t insn);
>   bool decode_isa_rel6(DisasContext *ctx, uint32_t insn);
>   bool decode_ase_mips16e_16(DisasContext *ctx, uint16_t insn);
>   bool decode_ase_mips16e_32(DisasContext *ctx, uint32_t insn);
> diff --git a/target/mips/tcg/micromips16.decode 
> b/target/mips/tcg/micromips16.decode
> new file mode 100644
> index 00000000000..d341da16b04
> --- /dev/null
> +++ b/target/mips/tcg/micromips16.decode
> @@ -0,0 +1,11 @@
> +# microMIPS32 16-bit instruction set extensions
> +#
> +# Copyright (C) 2021  Philippe Mathieu-Daudé
> +#
> +# SPDX-License-Identifier: LGPL-2.1-or-later
> +#
> +# Reference: MIPS Architecture for Programmers, Volume II-B
> +#            microMIPS32 Instruction Set
> +#            (Document Number: MD00582)
> +#            microMIPS64 Instruction Set
> +#            (Document Number: MD00594)
> diff --git a/target/mips/tcg/micromips32.decode 
> b/target/mips/tcg/micromips32.decode
> new file mode 100644
> index 00000000000..333ab0969ca
> --- /dev/null
> +++ b/target/mips/tcg/micromips32.decode
> @@ -0,0 +1,11 @@
> +# microMIPS32 32-bit instruction set extensions
> +#
> +# Copyright (C) 2021  Philippe Mathieu-Daudé
> +#
> +# SPDX-License-Identifier: LGPL-2.1-or-later
> +#
> +# Reference: MIPS Architecture for Programmers, Volume II-B
> +#            microMIPS32 Instruction Set
> +#            (Document Number: MD00582)
> +#            microMIPS64 Instruction Set
> +#            (Document Number: MD00594)
> diff --git a/target/mips/tcg/micromips_translate.c 
> b/target/mips/tcg/micromips_translate.c
> new file mode 100644
> index 00000000000..49e90e7eca2
> --- /dev/null
> +++ b/target/mips/tcg/micromips_translate.c
> @@ -0,0 +1,14 @@
> +/*
> + * MIPS emulation for QEMU - microMIPS translation routines
> + *
> + * Copyright (c) 2021 Philippe Mathieu-Daudé <f4bug@amsat.org>
> + *
> + * SPDX-License-Identifier: LGPL-2.1-or-later
> + */
> +
> +#include "qemu/osdep.h"
> +#include "translate.h"
> +
> +/* Include the auto-generated decoders.  */
> +#include "decode-micromips16.c.inc"
> +#include "decode-micromips32.c.inc"
> diff --git a/target/mips/tcg/micromips_translate.c.inc 
> b/target/mips/tcg/micromips_translate.c.inc
> index 26006f84df7..7a884222eed 100644
> --- a/target/mips/tcg/micromips_translate.c.inc
> +++ b/target/mips/tcg/micromips_translate.c.inc
> @@ -3018,6 +3018,15 @@ static int decode_isa_micromips(CPUMIPSState *env, 
> DisasContext *ctx)
>           }
>       }
>   
> +    if (decode_isa_micromips16(ctx, opcode)) {
> +        return 2;
> +    }
> +    opcode <<= 16;
> +    opcode |= translator_lduw(env, &ctx->base, ctx->base.pc_next + 2);
> +    if (decode_isa_micromips32(ctx, opcode)) {
> +        return 4;
> +    }


In the switch just above, you have determined 16-bit vs 32-bit, have you not? I think you 
should make use of that and keep these two cases separate.


r~