On Sun, 31 Mar 2019 20:11:47 PDT (-0700), richard.henderson@linaro.org wrote:
> There's new support in decodetree for pattern groups.
> I believe that patch 5, or something close to it, was
> posted as an RFC as part of the original pattern group
> patch set, but this cleans that up further.
>
>
> r~
>
>
> Richard Henderson (8):
> target/riscv: Name the argument sets for all of insn32 formats
> target/riscv: Use --static-decode for decodetree
> target/riscv: Merge argument sets for insn32 and insn16
> target/riscv: Merge argument decode for RVC shifti
> target/riscv: Use pattern groups in insn16.decode
> target/riscv: Split RVC32 and RVC64 insns into separate files
> target/riscv: Split gen_arith_imm into functional and temp
> target/riscv: Remove spaces from register names
>
> target/riscv/cpu.c | 16 +-
> target/riscv/insn_trans/trans_rvc.inc.c | 347 ------------------------
> target/riscv/insn_trans/trans_rvi.inc.c | 20 +-
> target/riscv/translate.c | 40 ++-
> target/riscv/Makefile.objs | 15 +-
> target/riscv/insn16-32.decode | 28 ++
> target/riscv/insn16-64.decode | 30 ++
> target/riscv/insn16.decode | 154 ++++++-----
> target/riscv/insn32.decode | 10 +-
> 9 files changed, 205 insertions(+), 455 deletions(-)
> delete mode 100644 target/riscv/insn_trans/trans_rvc.inc.c
> create mode 100644 target/riscv/insn16-32.decode
> create mode 100644 target/riscv/insn16-64.decode
Thanks! I'm happy taking this as it stands, since all those decode issues I
pointed out aren't regressions in this patch set. Let me know if you want to
fix those or if you want me to -- as you can probably tell my email queue is a
bit long... :)
I've dropped this on my integration branch, but I probably won't submit a pull
for a week or so.