aboutsummaryrefslogtreecommitdiff
path: root/gcc/expmed.c
AgeCommit message (Expand)Author
1998-11-30(emit_store_flag): Don't try to negate STORE_FLAG_VALUE.kenner
1998-11-29Use gen_rtx_FOO.kenner
1998-11-29Include system.h and remove now-redundant includes or definitionskenner
1998-06-07(init_expmed): Allocate all RTX in memory we'll free.kenner
1998-05-20(expand_divmod): Save last divison constant and if rem is same as div,kenner
1998-05-13(expand_mult): Use 0 as add_target if should preserve subexpressions.kenner
1998-04-19(do_cmp_and_jump): New function.kenner
1998-03-22(extract_bit_field): Don't confuse SUBREG_WORD with endian adjustment inkenner
1997-11-08(store_split_bit_field): Force ADDRESSOF into register.kenner
1997-10-18(extract_bit_field): Don't make flag_force_mem disable extzv forkenner
1997-09-23(expand_divmod): If compute_mode is not same as mode, handle casekenner
1997-08-08Include <stdio.h> before include files that formerly used STDIO_PROTO.eggert
1997-08-07(expand_divmod, emit_store_flag): Add default case to switch.kenner
1997-05-06reformat a little to match GNU coding standards.mrs
1997-02-15(expand_divmod): prefer divmod in same mode over div in wider mode.kenner
1997-01-27(emit_store_flag): Properly test for STORE_FLAG_VALUE of just sign bit.kenner
1997-01-04(store_fixed_bit_field): If not SLOW_UNALIGNED_ACCESS, treatkenner
1996-12-12 * expmed.c (emit_store_flag_force): Fix typos/thinkos.law
1996-10-21(emit_store_flag_force): New function.tege
1996-09-25(store_bit_field): don't make flag_force_mem disable insvwilson
1996-07-17(expand_mult_highpart): Revert last change.tege
1996-07-03formatting tweaksmrs
1996-06-28(expand_mult_highpart): Use op1 not wide_op1 in expansion ofkenner
1996-06-27(extract_bit_field): Check TRULY_NOOP_TRUNCATION beforewilson
1996-05-17(store_split_bit_field): Don't assume the alignment of VALUE is thekenner
1996-05-05Fix typo in comment.kenner
1996-03-30(expand_divmod, case TRUNC_DIV_EXPR): Move some codetege
1996-03-27(negate_rtx): Fix typo in previous change.kenner
1996-03-24(negate_rtx): Don't try to negate a constant ourself; instead callkenner
1996-03-19(emit_store_flag): If expanding (GE X 0) will need twokenner
1996-01-17(extract_bit_field): For multi-word bitfield, clobber target beforewilson
1996-01-16(extract_bit_field): Don't abort if not MEM_IN_STRUCT_P.kenner
1995-12-17(expand_mult_highpart): When doing widening multiply,tege
1995-12-15(store_bit_field): Don't use insv for BLKmode value.kenner
1995-12-14(expand_divmod, case TRUNC_DIV_EXPR): Only rejecttege
1995-12-13(expand_divmod): Don't use TARGET if it's the wrong mode.kenner
1995-07-13(SHIFT_COUNT_TRUNCATED): Use #ifdef not #if.kenner
1995-06-15Update FSF address.kenner
1995-06-14(emit_store_flag): Always set LAST.kenner
1995-06-08(store_split_bit_field): When adjust arg in BYTES_BIT_ENDIAN case, usekenner
1995-06-08(expand_divmod): Always check result of emit_store_flag.kenner
1995-06-01(extract_fixed_bit_field): For REG case, compute total_bits from modekenner
1995-06-01(expand_mult_highpart): Use wide_op1 for all multiplies.kenner
1995-05-16Fix typos in comments.kenner
1995-05-02If expensive optimizations, use different pseudos when making store flag insns.meissner
1995-04-03Get correct bit offset for big endian (SUBREG (REG))meissner
1995-02-13(store_split_bit_field): If value is VOIDmode, pass word_mode to force_reg.dje
1995-02-11(expand_divmod): Always call expand_twoval_binop with psuedos askenner
1995-02-06(mul_cost, div_cost): New variables.tege
1995-01-14Add 1995 to copyright.dje