This patch series fixes a number of minor Neon bugs, mostly related to edge cases involving NaN handling.
The patch series includes adding support for float*_min() and float*_max() operations to softfloat, because this makes it easy to get NaN propagation and handling of zeroes correct. I've only added support for float32 and float64 versions because I think the implementation is complex enough that there's no point providing an untested and unused version for float128. (floatx80 doesn't even implement the comparison ops so min/max is definitely a stretch there.) I note in passing that patch 6 (the softfloat one) seems to hang scripts/checkpatch.pl, or at least make it take an unreasonably long time to finish... Peter Maydell (7): target-arm: Make Neon helper routines use correct FP status target-arm/neon_helper.c: Use make_float32/float32_val macros target-arm: Return right result for Neon comparison with NaNs target-arm: Fix VCLE.F32 #0, VCLT.F32 #0 NaN handling target-arm: Correct ABD's handling of negative zeroes softfloat: Add float*_min() and float*_max() functions target-arm: Use new softfloat min/max functions for VMAX,VMIN fpu/softfloat.c | 49 ++++++++++++++++++++++ fpu/softfloat.h | 4 ++ target-arm/helpers.h | 22 +++++----- target-arm/neon_helper.c | 103 ++++++++++++++++++--------------------------- target-arm/translate.c | 60 +++++++++++++++------------ 5 files changed, 139 insertions(+), 99 deletions(-)