I started this last year, but only recently have found the time to ferret out the bugs.
I've tested this vs the gcc execute.exp testsuite, against which there are no regressions. It also continues to boot my old debian sid installation. I had not previously been aware, but I now know this fixes a bug with respect to the purge data cache (PDC) instruction. Helge or Sven, if you have a chance to do a bit more testing in the next week, I would appreciate it. Otherwise we can fix any other minor problems on mainline. This is based on the target/hppa pull request from this week; the whole tree can be found at https://github.com/rth7680/qemu.git tgt-hppa r~ Richard Henderson (19): target/hppa: Use DisasContextBase.is_jmp target/hppa: Begin using scripts/decodetree.py target/hppa: Convert move to/from system registers target/hppa: Convert remainder of system insns target/hppa: Unify specializations of OR target/hppa: Convert memory management insns target/hppa: Convert arithmetic/logical insns target/hppa: Convert indexed memory insns target/hppa: Convert fp multiply-add target/hppa: Convert conditional branches target/hppa: Convert shift, extract, deposit insns target/hppa: Convert direct and indirect branches target/hppa: Convert arithmetic immediate insns target/hppa: Convert offset memory insns target/hppa: Convert fp indexed memory insns target/hppa: Convert halt/reset insns target/hppa: Convert fp fused multiply-add insns target/hppa: Convert fp operate insns target/hppa: Merge translate_one into hppa_tr_translate_insn target/hppa/translate.c | 3218 +++++++++++++++---------------------- target/hppa/Makefile.objs | 8 + target/hppa/insns.decode | 527 ++++++ 3 files changed, 1802 insertions(+), 1951 deletions(-) create mode 100644 target/hppa/insns.decode -- 2.17.2
