On 02/07/2014 01:49 PM, Peter Maydell wrote: > This is the fourth set of A64 Neon support patches. This set > provides complete implementations of the categories: > * scalar x indexed > * vector x indexed > * scalar three-different > and fills in all the previously missing instructions in: > * vector 3-reg-same > * scalar 3-reg-same > > It includes adding support in softfloat for "fused multiply, > add and then divide by two", because the A64 VRECPS instruction > requires it. > > By my reckoning the parts that are still missing for Neon are: > * parts of scalar shift-immediate and vector shift-immediate > * parts of scalar-2-misc > * most of vector-3-diff > * parts of vector-2-misc
Except for the nit in 7/8, Reviewed-by: Richard Henderson <[email protected]> r~
