> Hmm, I further revised comment in the patch since in try_combine, i2
> is always after i0/i1.  The original comment is inaccurate about that.

OK.

-- 
Eric Botcazou

Reply via email to