summaryrefslogtreecommitdiffstats
path: root/gcc/combine.c
Commit message (Expand)AuthorAgeFilesLines
* * emit-rtl.c (gen_lowpart_common): Suppress last change if __complex__.law1998-02-191-32/+0
* * combine.c (simplify_rtx): Obey CLASS_CANNOT_CHANGE_SIZE whenlaw1998-02-171-0/+32
* * varasm.c (output_constant_pool): Bring back 'done' label insidelaw1998-02-141-1/+5
* * combine.c (can_combine_p): Handle USEs in PARALLELs.amylaar1998-02-131-0/+37
* * combine.c (simplify_shift_const): Fix typo in last change.law1998-02-011-1/+2
* * combine.c (simplify_shift_const): (lshiftrt (truncate (lshiftrt)))law1998-02-011-0/+22
* * rtl.c (read_rtx): Use atol/atoq based upon disposition oflaw1998-01-281-2/+2
* law1998-01-271-5/+11
* * combine.c (force_to_mode, nonzero_bits): Correctly optimizelaw1998-01-171-10/+51
* * alias.c: Change all uses of gen_rtx(FOO...) to gen_rtx_FOO;rth1998-01-141-128/+136
* Bring in final gcc-2.8.0 changes.law1998-01-141-1/+1
* * bitmap.c (bitmap_clear): Ensure `inline' is at the beginninglaw1997-12-201-0/+3
* * combine.c (force_to_mode): return immediately if operand is a CLOBBER.law1997-12-161-2/+5
* Another handful of gcc-2.8 changes (will it ever go out the door?).law1997-12-111-3/+5
* * combine.c (simplify_rtx, case ABS): Do nothing for ASM_OPERANDS.law1997-12-111-0/+3
* Merge from gcc-2.8law1997-12-071-6/+52
* Fix alpha-linux GNU libc build failures.wilson1997-11-051-5/+10
* Remove last combine patch. rth & jim are working on a new one.law1997-11-031-8/+4
* * combine.c (try_combine): Don't elim_i1 when it is used in i2src.law1997-11-031-4/+8
* Update mainline egcs to gcc2 snapshot 971021.law1997-11-021-32/+108
* * combine.c (can_combine_p): Don't combine with an asm whoserth1997-10-161-0/+7
* * combine.c (try_combine): Use NULL_RTX instead of '0' wherelaw1997-09-221-2/+2
* * Integrate alias analysis changes from jfc@mit.edulaw1997-08-111-0/+1
* (force_to_mode): Don't do anything for a ASM_OPERANDS insn.kenner1997-06-221-3/+4
* Use REG_SET macrosmeissner1997-06-161-8/+4
* Use accessor macros to access arrays based on regno; move many of the arrays ...meissner1997-05-301-29/+29
* reformat a little to match GNU coding standards.mrs1997-05-061-3/+3
* (get_last_value_validate): New argument insn.wilson1997-04-141-7/+21
* (move_deaths): Handle partial REG_DEAD note forwilson1997-04-031-5/+15
* * combine.c (distribute_notes): Handle REG_BR_PROB and REG_EXEC_COUNTdje1997-03-281-0/+8
* Fix typo in comment.kenner1997-03-251-1/+1
* (if_then_else_cond): Call copy_rtx to prevent sharing.wilson1997-03-181-0/+8
* (simplify_rtx, case SUBREG): Fix direction of test when callingkenner1997-03-101-3/+11
* * combine.c (simplify_rtx): Do nothing with (truncate:mode) iflaw1997-03-031-0/+6
* * combine.c (find_split_point): Don't turn a SIGN_EXTEND intolaw1997-02-191-0/+8
* Update comments.kenner1997-02-151-1/+6
* (find_split_point, case SET): Fix error in last change.kenner1997-02-021-2/+2
* Add a case where reg_n_refs is clearedian1997-01-291-1/+3
* (simplify_rtx): Don't do anything with if_then_else_condkenner1997-01-271-52/+56
* (find_split_point, case SET): If SET_SRC is NE andkenner1997-01-041-3/+26
* (nonzero_bits): Ifdef out calls to num_sign_bit_copies.wilson1996-12-171-0/+18
* (simplify_comparison): Use mode_width as shift countwilson1996-12-161-4/+8
* Check SMALL_REGISTER_CLASSES at runtimeian1996-12-131-5/+8
* (try_combine): When splitting an insn, check for the new I2 setting akenner1996-12-111-1/+6
* (combinable_i3pat): Bring back to sync with can_combine_p.kenner1996-12-111-2/+6
* (try_combine): Properly copy an rtvec.kenner1996-11-151-1/+1
* Add some optimizations for TRUNCATE and ZERO_EXTENDian1996-10-291-0/+121
* (can_combine_p): When SMALL_REGISTER_CLASSES is defined, avoidkenner1996-10-231-1/+6
* (can_combine_p): Even if SMALL_REGISTER_CLASSES, know lifetime notkenner1996-09-231-1/+1
* * combine.c (rtx_equal_for_field_assignment_p): Check forlaw1996-08-041-2/+8
OpenPOWER on IntegriCloud