On 10/30/2014 08:59 AM, Bastian Koppelmann wrote: > Add instructions of RC opcode format. > Add helper for mul, sha, absdif with signed saturation on overflow. > Add helper for add, sub, mul with unsigned saturation on overflow. > Add microcode generator functions: > * gen_add_CC, which calculates the carry bit. > * gen_addc_CC, which adds the carry bit to the add and calculates the > carry bit. > * gen_absdif, which calculates the absolute difference. > * gen_mul_i64s/u, which mul two 32 bits val into one 64bit reg. > * gen_sh_hi, which shifts two 16bit words in one reg. > * gen_sha_hi, which does a arithmetic shift on two 16bit words. > * gen_sh_cond, which shifts left a reg by one and writes the result of > cond into the lsb. > * gen_accumulating_cond, which ands/ors/xors the result of cond of the > lsbs > with the lsb of the result. > * gen_eqany_bi/hi, which checks ever byte/hword on equality. > > Signed-off-by: Bastian Koppelmann <kbast...@mail.uni-paderborn.de> > --- > v1 -> v2: > - gen_sha_hi: Remove mask for low in case shift_count > 0, since deposit > handles that. > - gen_sha_hi: Remove mask for high in case shift <= 0, since deposit > handles that.
Reviewed-by: Richard Henderson <r...@twiddle.net> r~
