[patch][ARM] Don't generate redundant zero_extend before smlabb

[Richard Sandiford]

Richard Sandiford <richard.sandiford@linaro.org> writes:
> Could this go in during stage 3?  Like you say, the current maddhisi4
> expander (which is known to optabs) generates "noncanonical" rtx.
> That seems like a bug.

Sorry to follow up on myself, but to be more specific: as Paolo said
recently, "noncanonical rtx" in this type of situation really means
"invalid rtx", so having optabs generate what is effectively invalid
rtx seems like a correctness bug.

Richard