This patchset adds support to the decodetree generator for "named fields", where one field can refer to some other already extracted field, as well as to portions of the instruction word. The specific case where I want this is for some load/store insns in the A64 decoder:
# Load/store with an unsigned 12 bit immediate, which is scaled by the # element size. The function gets the sz:imm and returns the scaled immediate. # For vectors, opc bit 1 (insn bit 23) is effectively bit 2 of the size. %uimm_scaled 10:12 sz:3 !function=uimm_scaled @ldst_uimm .. ... . .. .. ............ rn:5 rt:5 &ldst_imm unpriv=0 p=0 w=0 imm=%uimm_scaled STR_i sz:2 111 0 01 00 ............ ..... ..... @ldst_uimm sign=0 ext=0 LDR_i 00 111 0 01 01 ............ ..... ..... @ldst_uimm sign=0 ext=1 sz=0 LDR_i 01 111 0 01 01 ............ ..... ..... @ldst_uimm sign=0 ext=1 sz=1 LDR_i 10 111 0 01 01 ............ ..... ..... @ldst_uimm sign=0 ext=1 sz=2 LDR_i 11 111 0 01 01 ............ ..... ..... @ldst_uimm sign=0 ext=1 sz=3 LDR_i 00 111 0 01 10 ............ ..... ..... @ldst_uimm sign=1 ext=0 sz=0 LDR_i 01 111 0 01 10 ............ ..... ..... @ldst_uimm sign=1 ext=0 sz=1 LDR_i 10 111 0 01 10 ............ ..... ..... @ldst_uimm sign=1 ext=0 sz=2 LDR_i 00 111 0 01 11 ............ ..... ..... @ldst_uimm sign=1 ext=1 sz=0 LDR_i 01 111 0 01 11 ............ ..... ..... @ldst_uimm sign=1 ext=1 sz=1 Here we need to manually decode the sz field in bits 31:30 because of the complexity of the sign/ext and the parts of the encode space that are UNDEF (or are prefetch). And we want to use a !function to do the "scale the immediate offset by the size of the datatype" so we can use the same LDR_i and STR_i trans_ functions that we already have for the unscaled-immediate loads and stores. But at the moment you can't re-decode bits in a %field that are fixed in the instruction pattern, and you can't refer to the already-decoded sz value directly. This patchset implements the syntax used above where the %field can refer to another field, e.g. 'sz:2'. Patch 1 fixes a trivial bug in the check.sh script that meant that failures weren't reported up to meson. thanks -- PMM Peter Maydell (6): tests/decodetree/check.sh: Exit failure for all failures docs: Document decodetree named field syntax scripts/decodetree: Pass lvalue-formatter function to str_extract() scripts/decodetree: Implement a topological sort scripts/decodetree: Implement named field support tests/decode: Add tests for various named-field cases docs/devel/decodetree.rst | 33 +++- tests/decode/err_field1.decode | 2 +- tests/decode/err_field10.decode | 7 + tests/decode/err_field7.decode | 7 + tests/decode/err_field8.decode | 8 + tests/decode/err_field9.decode | 14 ++ tests/decode/succ_named_field.decode | 19 +++ scripts/decodetree.py | 239 +++++++++++++++++++++++++-- tests/decode/check.sh | 1 + 9 files changed, 310 insertions(+), 20 deletions(-) create mode 100644 tests/decode/err_field10.decode create mode 100644 tests/decode/err_field7.decode create mode 100644 tests/decode/err_field8.decode create mode 100644 tests/decode/err_field9.decode create mode 100644 tests/decode/succ_named_field.decode -- 2.34.1
