Hi, I'll explain a little more background.
A major case which this patch set will be required is when a developer use inline assembly for instructions not yet natively supported by GCC or will not be supported as intrinsics. On such cases, GCC should at least accept corresponding extension via -march (although there's a workaround using ".option" directives in the inline assembly but that's tedious). Thanks, Tsukasa On 2023/08/14 15:09, Tsukasa OI wrote: > PATCH v1: > <https://gcc.gnu.org/pipermail/gcc-patches/2023-August/627220.html> > > > Changes: v1 -> v2 (only in PATCH 3/3) > ====================================== > > Removed: 'Zvkn' -> 'Zvknha' implication (not to cause test failure) > Added: 'Zfa' -> 'F' implication (just I forgot to add in PATCH v1) > > > Thanks, > Tsukasa > > > > > Tsukasa OI (3): > RISC-V: Add stub support for existing extensions (privileged) > RISC-V: Add stub support for existing extensions (vendor) > RISC-V: Add stub support for existing extensions (unprivileged) > > gcc/common/config/riscv/riscv-common.cc | 44 ++++++++++++++++++++++ > gcc/testsuite/gcc.target/riscv/predef-29.c | 35 +++++++++++++++++ > gcc/testsuite/gcc.target/riscv/predef-30.c | 27 +++++++++++++ > gcc/testsuite/gcc.target/riscv/predef-31.c | 31 +++++++++++++++ > 4 files changed, 137 insertions(+) > create mode 100644 gcc/testsuite/gcc.target/riscv/predef-29.c > create mode 100644 gcc/testsuite/gcc.target/riscv/predef-30.c > create mode 100644 gcc/testsuite/gcc.target/riscv/predef-31.c > > > base-commit: 614052dd4ea083e086712809c754ffebd9361316
