| Age | Commit message (Expand) | Author | Files | Lines |
|---|
| 1995-08-21 | (decrement_and_branch_until_zero): Use arithmetic_comparison_operator to deci... | Richard Kenner | 1 | -2/+5 |
| 1995-06-15 | Change FSF address. | Richard Kenner | 1 | -1/+2 |
| 1995-05-12 | Add better support for passing args in registers; Add loop/jump/function alig... | Michael Meissner | 1 | -1/+69 |
| 1995-04-25 | (untyped_call): Make it return a complex double. | Richard Kenner | 1 | -1/+8 |
| 1995-04-09 | (adddi3, subdi3): Need scratch reg whenever operand 0 in mem and operands 1 n... | Richard Kenner | 1 | -3/+3 |
| 1995-03-28 | (adddi3): Don't treat two non-equal MEMs as non-aliasing. | Richard Kenner | 1 | -2/+2 |
| 1995-02-24 | Define long double versions of sin/cos | Michael Meissner | 1 | -0/+14 |
| 1994-11-25 | Add comment on ffs to say why optimizing it further is a bad idea. | Michael Meissner | 1 | -1/+4 |
| 1994-10-31 | (movsf... | Richard Kenner | 1 | -0/+36 |
| 1994-10-28 | Fix adddi3. | Michael Meissner | 1 | -7/+13 |
| 1994-10-09 | Make sure movsf_mem does not clobber a destination register | Michael Meissner | 1 | -1/+1 |
| 1994-10-05 | Supprt LABEL_REFs when using -fpic. | Michael Meissner | 1 | -4/+2 |
| 1994-10-01 | Prefer float registers for movdf. | Michael Meissner | 1 | -3/+3 |
| 1994-09-26 | Allow memory/memory DImode add/subtract. | Michael Meissner | 1 | -10/+97 |
| 1994-09-20 | Respin DI support to be combine friendly; Allow push of SF without temp reg; ... | Michael Meissner | 1 | -342/+45 |
| 1994-09-15 | Do not use separate 386 patterns for SF/DF/XF/DI pushes; allow SF/DF/XF/DI pu... | Michael Meissner | 1 | -75/+68 |
| 1994-09-13 | More DF/XF/DI fixes. | Michael Meissner | 1 | -48/+72 |
| 1994-09-09 | Merge push patterns for memory/nonmemory | Michael Meissner | 1 | -128/+120 |
| 1994-09-08 | Rewrite movsf/movdf/movdi patterns | Michael Meissner | 1 | -105/+457 |
| 1994-09-03 | Fix DImode move problems. | Michael Meissner | 1 | -28/+33 |
| 1994-08-28 | Fix indexing off arg ptr & push of byte values. | Michael Meissner | 1 | -12/+6 |
| 1994-08-26 | Add -mno-wide-multiply, -mno-move, make addresses more compatible with other ... | Michael Meissner | 1 | -74/+268 |
| 1994-06-15 | Support 32x32->64 bit multiply | Michael Meissner | 1 | -5/+24 |
| 1994-04-25 | (TARGET_NO_FANCY_MATH_387): Implement new option. | Richard Kenner | 1 | -12/+24 |
| 1994-04-09 | (mulhi3): Use SImode for MULT RTL. | Richard Kenner | 1 | -3/+3 |
| 1994-03-25 | i386.md (movhi, movqi): Properly recognized unsigned forms of -1 for | Richard Kenner | 1 | -8/+11 |
| 1993-10-05 | (pushxf,movxf,extenddfxf2,extendsfxf2): Handle XFmode write to MEM by eading... | James Van Artsdalen | 1 | -43/+17 |
| 1993-10-03 | (tstxf_cc, tstxf, cmpxf, cmpxf_cc, cmpxf_ccfpeq, swapxf, movxf, | James Van Artsdalen | 1 | -0/+617 |
| 1993-10-02 | (addhi3): Use byte opcodes when the low byte of a word is known to be zero. | James Van Artsdalen | 1 | -0/+18 |
| 1993-07-30 | (ffssi2 and ffshi2 recognizers): Generate unique labels by hand. | Richard Stallman | 1 | -4/+40 |
| 1993-07-25 | (movstrictqi): Change q<-g alternative to q<-m. | Richard Stallman | 1 | -3/+3 |
| 1993-07-13 | (ffssi2, ffshi2): Assume bsf produces garbage if input is 0. | Richard Stallman | 1 | -8/+17 |
| 1993-06-19 | (call_pop, call, call_value_pop, call_value) | Richard Stallman | 1 | -5/+5 |
| 1993-05-09 | (call matchers): Use call_insn_operand. | Richard Stallman | 1 | -5/+29 |
| 1993-05-06 | (extendsidi2): Use cdq or cltd, not cwtl. | Richard Stallman | 1 | -2/+8 |
| 1993-05-06 | (extendsidi2): Use cdq, not cwtl. | Richard Stallman | 1 | -1/+2 |
| 1993-03-23 | (untyped_call pattern): Fix typo. | Tom Wood | 1 | -1/+1 |
| 1993-03-23 | (untyped_call, untyped_return): New patterns. | Tom Wood | 1 | -0/+117 |
| 1993-03-22 | (movsf,movdf): Use an input reload for MEM-MEM moves. | James Van Artsdalen | 1 | -59/+98 |
| 1993-03-12 | (truncdfsf2): Rewrite to allow FLOAT_REGS in constraints for better register ... | James Van Artsdalen | 1 | -105/+160 |
| 1993-02-08 | (extendsidi2): Use correct mnemonic for cwtl opcode. | James Van Artsdalen | 1 | -1/+1 |
| 1993-01-25 | Delete emacs local-variables block at end. | Richard Kenner | 1 | -10/+0 |
| 1993-01-25 | (extendsidi2): Fix typo in regno. | James Van Artsdalen | 1 | -1/+1 |
| 1993-01-20 | (cmpM_cc): Allow first operand to be a MEM. | James Van Artsdalen | 1 | -60/+98 |
| 1993-01-14 | (cmpstr patterns): Fix mode mismatch in SET. | James Van Artsdalen | 1 | -3/+3 |
| 1993-01-11 | (extendsidi2,zero_extendsidi2): New patterns. | James Van Artsdalen | 1 | -9/+39 |
| 1992-12-31 | (movsf,movdf): Change constraints to use integer regs for mem-mem copies. | James Van Artsdalen | 1 | -4/+12 |
| 1992-12-30 | (bit complement): Constrain inputs and outputs to match, | James Van Artsdalen | 1 | -28/+53 |
| 1992-11-12 | (floatMN2 recognizers): Don't omit mode on a match_operand that might be relo... | James Van Artsdalen | 1 | -13/+42 |
| 1992-11-09 | (movstrsi): mask move count after signed shift. | James Van Artsdalen | 1 | -1/+1 |
