Softfloat supports two kinds of FP tininess detection: before and after rounding (corresponding to the two behaviours permitted by IEEE754). The ARM architecture mandates detection before rounding, so set the flag appropriately. I had to add a setter function for this, since there wasn't one already.
Peter Maydell (2): softfloat: Add setter function for tininess detection mode target-arm: Detect tininess before rounding for FP operations fpu/softfloat.h | 4 ++++ target-arm/helper.c | 4 ++++ 2 files changed, 8 insertions(+), 0 deletions(-)
