On Fri, Feb 6, 2026 at 4:50 PM Pierrick Bouvier <[email protected]> wrote: > > On 2/6/26 10:38 AM, Brian Cain wrote: > > Fix the issues reported by Alex Rønne Petersen causing zig test suite > > failures. The 0x0 case is a conflicting "multiple register write" - the > > handler here will cover duplexes but not the more general case of multiple > > register writes. I'll include that in a subsequent patchset. > > > > Brian Cain (3): > > target/hexagon: Fix invalid duplex decoding > > target/hexagon: Reject duplex encodings with duplicate dest registers > > target/hexagon: Return decode failure for invalid non-duplex encodings > > > > linux-user/hexagon/cpu_loop.c | 4 ++++ > > target/hexagon/decode.c | 28 +++++++++++++++++++++++++--- > > target/hexagon/translate.c | 18 ++++++++++++++++-- > > tests/tcg/hexagon/invalid-duplex.c | 28 ++++++++++++++++++++++++++++ > > tests/tcg/hexagon/invalid-dups.c | 23 +++++++++++++++++++++++ > > tests/tcg/hexagon/invalid-encoding.c | 20 ++++++++++++++++++++ > > tests/tcg/hexagon/Makefile.target | 23 ++++++++++++++++++++++- > > 7 files changed, 138 insertions(+), 6 deletions(-) > > create mode 100644 tests/tcg/hexagon/invalid-duplex.c > > create mode 100644 tests/tcg/hexagon/invalid-dups.c > > create mode 100644 tests/tcg/hexagon/invalid-encoding.c > > > > Sounds good. > > I have one suggestion to improve it. For tests, it would be nice to > catch SIGILL, so test returns 0 if it worked as expected, or exit with 1 > if SIGILL was not sent. It's just a suggestion though :)
Tsk - good call. And that way I could combine the test cases. I like it, will do. > Regards, > Pierrick
