summaryrefslogtreecommitdiff
path: root/compiler/powerpc64
Commit message (Expand)AuthorAgeFilesLines
* + common assembler optimizer base class for powerpc and powerpc64florian2021-04-111-2/+2
* * patch by Marģers to unify internal error numbers, resolves #37888florian2020-10-134-6/+6
* Replace outdated cgop2string function by tcgsize2str function from cgbase un...pierre2020-08-252-7/+7
* Fix EXTDEBUG code: tai_comment requires use of strpnew function with constan...pierre2020-08-251-1/+1
* * reworked usage of tcgnotnode.handle_locjumpflorian2020-08-051-2/+2
* * fixed ppc64 triple architecture namesjonas2020-08-051-2/+2
* * mark all external assemblers using an LLVM tool using af_llvmjonas2020-07-191-0/+48
* * fix several double ;florian2020-05-241-1/+1
* * ppc64le: fix storing homogeneous 64 bit float parameters in case they arejonas2020-05-011-3/+4
* * ELFv2: fixed homogeneous aggregate parameter passing when they shouldjonas2020-04-301-4/+32
* * split i/t_darwin from i/t_bsd, as they don't have that much in commonjonas2020-03-291-2/+2
* * safecall support is meanwhile generic, so enabled for all linux targetsflorian2020-01-311-0/+1
* * disable cs_opt_regvar on all platforms when compiled for LLVM (LLVM doesjonas2020-01-291-2/+2
* * added is_normal_fieldvarsym() helper and use itjonas2019-12-241-3/+2
* * renamed getintparaloc to getcgtempparalocjonas2019-12-241-2/+2
* * common naming for fpu_none stringflorian2019-12-241-1/+2
* * patch by J. Gareth Moreton, issue #36271, part 3: support for the other a...florian2019-11-101-2/+14
* somehow committing went wrong, second part of last commit:florian2019-09-031-1/+0
* * fixed size of base stack frame for ELFv2jonas2019-08-261-2/+6
* Fix internal error by using same code to set checkoverflow variable as in po...pierre2019-08-051-1/+1
* * factored out the conditions under which add nodes need to performjonas2019-08-031-5/+1
* Systematically include fpcdefs.inc at sart of all units used by compilerpierre2019-07-031-0/+2
* * synchronised with trunk till r42118jonas2019-05-251-1/+1
|\
| * * disable regular array -> dynamic array type coversion support unlessjonas2019-05-251-1/+1
* | * synchronised with trunk till r42049jonas2019-05-125-5/+31
|\ \ | |/
| * * fix case completeness and unreachable code warnings in compiler that wouldjonas2019-05-125-5/+31
* | * synchronised with trunk till r41725jonas2019-03-171-2/+2
|\ \ | |/
| * Fix tmoddiv4.pp test failure with -O1 option for powerpc64 as suggested by J...pierre2019-03-051-2/+2
* | * synchronised with trunk till r41537jonas2019-03-011-2/+2
|\ \ | |/
| * * track va_uses_float_reg also for varargs functions without varargsjonas2019-02-251-2/+2
* | * synchronised with trunk till r41423jonas2019-02-231-29/+23
|\ \ | |/
| * * keep track of whether a routine has a C-style variadic parameter in thejonas2019-02-231-29/+23
* | * changed create_hlcodegen into a procvar, so that we don't have to insertjonas2018-11-291-3/+2
* | * synchronised with trunk till r40348jonas2018-11-181-4/+0
|\ \ | |/
| * - get rid of MaxOps, it is redundant with max_operandsflorian2018-11-021-4/+0
* | + fpc_eh_return_data_regno() intrinsic to get the return register numbersjonas2018-10-281-4/+13
|/
* Fix for bug report #34380pierre2018-10-181-0/+5
* * keep track of the temp position separately from the offset in references,jonas2018-04-221-12/+12
* * replaced the saved_XXX_registers arrays with virtual methods insidenickysn2018-04-192-17/+14
* Fix compilation with -dEXTDEBUGpierre2018-03-101-3/+6
* Remove overflow check that cannot happen because of typespierre2018-03-101-2/+0
* Generate error if CR field of F var is zero, as it leads to invalid instruct...pierre2018-03-101-0/+2
* * factored out check to determine whether a variable can be subscripted injonas2018-01-011-7/+2
* * only take into account the location of the parameter at the callee side tojonas2017-05-211-2/+4
* * fixed check to determine whether a record parameter can be subscriptedjonas2017-04-261-4/+4
* * use pocalls_cdecl and cstylearrayofconst more consistently instead ofjonas2017-02-251-3/+3
* * reverted r35424, wasn't ready for commit yetjonas2017-02-111-3/+2
* * fixed check to determine whether a record parameter can be subscriptedjonas2017-02-111-2/+3
* * renamed t<cpuname>procinfo to tcpuprocinfo for all targets, so we canjonas2016-12-163-14/+14
* * fix some more occured->occuredmarco2016-12-131-2/+2