Hi Richard,
On 7/1/25 08:59, Richard Henderson wrote:
Currently we have individual opcodes for each integer type.
This makes things difficult in tcg/optimize.c when we want
to transform the opcode stream.
Add TCGOP_TYPE() and use that instead. Merge the opcodes that are
common to both TCG_TYPE_I32 and TCG_TYPE_I64. We could eventually
merge some of the vector opcodes too, but that is a larger job.
Richard Henderson (81):
tcg: Move call abi parameters from tcg-target.h to tcg-target.c.inc
tcg: Replace TCGOP_VECL with TCGOP_TYPE
tcg: Move tcg_op_insert_{after,before} decls to tcg-internal.h
tcg: Copy TCGOP_TYPE in tcg_op_insert_{after,before}
tcg: Add TCGOP_FLAGS
tcg: Add type and flags arguments to tcg_op_supported
target/arm: Do not test TCG_TARGET_HAS_bitsel_vec
target/arm: Use tcg_op_supported
target/tricore: Use tcg_op_supported
tcg: Add tcg_op_deposit_valid
target/i386: Remove TCG_TARGET_extract_tl_valid
target/i386: Use tcg_op_deposit_valid
target/i386: Use tcg_op_supported
tcg: Remove TCG_TARGET_NEED_LDST_LABELS and
TCG_TARGET_NEED_POOL_LABELS
tcg: Rename tcg-target.opc.h to tcg-target-opc.h.inc
tcg/tci: Move TCI specific opcodes to tcg-target-opc.h.inc
tcg: Move fallback tcg_can_emit_vec_op out of line
tcg: Split out tcg-target-has.h and tcg-has.h
tcg: Split out tcg-target-mo.h
tcg: Use C_NotImplemented in tcg_target_op_def
tcg: Change have_vec to has_type in tcg_op_supported
tcg: Reorg process_op_defs
tcg: Remove args_ct from TCGOpDef
tcg: Constify tcg_op_defs
tcg: Validate op supported in opcode_args_ct
tcg: Add TCG_OPF_NOT_PRESENT to opcodes without inputs or outputs
tcg: Pass type and flags to tcg_target_op_def
tcg: Add TCGType argument to tcg_out_op
tcg: Remove TCG_OPF_64BIT
tcg: Drop implementation checks from tcg-opc.h
tcg: Replace IMPLVEC with TCG_OPF_VECTOR
I still have 3 patches to review, but looking at the overall
maybe you can merge up to here, as the first part.
tcg/mips: Expand bswap unconditionally
(or up to here if I get this one reviewed in time).
tcg/i386: Handle all 8-bit extensions for i686
tcg/i386: Fold the ext{8,16,32}[us] cases into {s}extract
tcg/aarch64: Provide TCG_TARGET_{s}extract_valid
tcg/aarch64: Expand extract with offset 0 with andi
tcg/arm: Add full [US]XT[BH] into {s}extract
tcg/loongarch64: Fold the ext{8,16,32}[us] cases into {s}extract
tcg/mips: Fold the ext{8,16,32}[us] cases into {s}extract
tcg/ppc: Fold the ext{8,16,32}[us] cases into {s}extract
tcg/riscv64: Fold the ext{8,16,32}[us] cases into {s}extract
tcg/riscv: Use SRAIW, SRLIW for {s}extract_i64
tcg/s390x: Fold the ext{8,16,32}[us] cases into {s}extract
tcg/sparc64: Use SRA, SRL for {s}extract_i64
tcg/tci: Provide TCG_TARGET_{s}extract_valid
tcg/tci: Remove assertions for deposit and extract
tcg: Remove TCG_TARGET_HAS_{s}extract_{i32,i64}
tcg: Remove TCG_TARGET_HAS_deposit_{i32,i64}
tcg: Remove INDEX_op_ext{8,16,32}*
tcg: Add all_outop[]
tcg: Merge INDEX_op_mov_{i32,i64}
tcg: Convert add to TCGOutOpBinary
tcg: Merge INDEX_op_add_{i32,i64}
tcg: Convert and to TCGOutOpBinary
tcg: Merge INDEX_op_and_{i32,i64}
tcg/optimize: Fold andc with immediate to and
tcg/optimize: Emit add r,r,-1 in fold_setcond_tst_pow2
tcg: Convert andc to TCGOutOpBinary
tcg: Merge INDEX_op_andc_{i32,i64}
tcg: Convert or to TCGOutOpBinary
tcg: Merge INDEX_op_or_{i32,i64}
tcg/optimize: Fold orc with immediate to or
tcg: Convert orc to TCGOutOpBinary
tcg: Merge INDEX_op_orc_{i32,i64}
tcg: Convert xor to TCGOutOpBinary
tcg: Merge INDEX_op_xor_{i32,i64}
tcg/optimize: Fold eqv with immediate to xor
tcg: Convert eqv to TCGOutOpBinary
tcg: Merge INDEX_op_eqv_{i32,i64}
tcg: Convert nand to TCGOutOpBinary
tcg: Merge INDEX_op_nand_{i32,i64}
tcg/loongarch64: Do not accept constant argument to nor
tcg: Convert nor to TCGOutOpBinary
tcg: Merge INDEX_op_nor_{i32,i64}
tcg/arm: Fix constraints for sub
tcg: Convert sub to TCGOutOpSubtract
tcg: Merge INDEX_op_sub_{i32,i64}
tcg: Convert neg to TCGOutOpUnary
tcg: Merge INDEX_op_neg_{i32,i64}
tcg: Convert not to TCGOutOpUnary
tcg: Merge INDEX_op_not_{i32,i64}
Here I'm missing the "Convert X to TCGOutOpY" patches. I haven't seen
any blocking review difficulty, so will proceed soon.
Regards,
Phil.
