resultdef of typeconvnode.typecheck_proc_to_procvar() is reused every time you get an address of a proc. So it is forbidden to change the far/near attribute of resultdef - it leads to mess during code generation.
Instead it is needed to use separate reusable resultdefs with new flags pc_far_address and pc_offset.
git-svn-id: trunk@46470 -
+ added support for constructing target triplets
* pass "-target triplet" when using an LLVM assembler
o removed no longer needed $DARWINVERSION and $ARCH parameters
* consistently use as_clang_gas when clang is used to assembler GAS-style
assembly, and rename as_llcm_clang to as_clang_llvm (for consistency)
* support pipe assembling when using clang on *nix in all cases
git-svn-id: trunk@45807 -
* changed the way OP_NEG and OP_NOT are handled in op_reg_ref, in order to be
consistent with op_reg_reg
* introduced op_reg,op_ref,op_subsetreg,op_subsetref and op_loc for the unary
operations only (OP_NEG,OP_NOT)
git-svn-id: trunk@45302 -
that itself, our LLVM code generator can't handle it, and if it did then
afterwards we would have to spill 90% of those register variables again
to make them SSA)
git-svn-id: trunk@44062 -
D tests/webtbs/tw32219.pp
U compiler/i386/i386att.inc
U compiler/i386/i386atts.inc
U compiler/i386/i386int.inc
U compiler/i386/i386nop.inc
U compiler/i386/i386op.inc
U compiler/i386/i386prop.inc
U compiler/i386/i386tab.inc
U compiler/i8086/i8086att.inc
U compiler/i8086/i8086atts.inc
U compiler/i8086/i8086int.inc
U compiler/i8086/i8086nop.inc
U compiler/i8086/i8086op.inc
U compiler/i8086/i8086prop.inc
U compiler/i8086/i8086tab.inc
U compiler/x86/x86ins.dat
U compiler/x86_64/x8664ats.inc
U compiler/x86_64/x8664att.inc
U compiler/x86_64/x8664int.inc
U compiler/x86_64/x8664nop.inc
U compiler/x86_64/x8664op.inc
U compiler/x86_64/x8664pro.inc
U compiler/x86_64/x8664tab.inc
U tests/utils/avx/avxopcodes.pas
-- Aufzeichnung der Informationen für Zusammenführung von r42706 bis r43443 in ».«:
U .
git-svn-id: trunk@43444 -
A tests/webtbs/tw32219.pp
U tests/utils/avx/avxopcodes.pas
U tests/utils/avx/asmtestgenerator.pas
U compiler/i386/i386tab.inc
U compiler/i386/i386nop.inc
U compiler/i8086/i8086tab.inc
U compiler/i8086/i8086nop.inc
U compiler/x86_64/x8664tab.inc
U compiler/x86_64/x8664nop.inc
U compiler/x86/aasmcpu.pas
U compiler/x86/x86ins.dat
U compiler/x86/rax86.pas
-- Aufzeichnung der Informationen für Zusammenführung von r42706 bis r43306 in ».«:
U .
git-svn-id: trunk@43307 -
globals.pas (size_2_align): ...here.
Change parameter type to asizeuint to avoid overflows.
(var_align_size, const_align_size): Also change parameter type to asizeuint.
i386/cpupara.pas, i8086/cpupara.pas, x86_64/cpupara.pas: Add globals
to implementation uses clause because of size_2_align move.
git-svn-id: trunk@43022 -
private_extern (or plain global in case of PECOFF, as the effect is
the same there): visible across object files, but they become local
when linked into a binary/library. This enables cross-unit inlining
of functions accessig implementation-only symbols.
git-svn-id: trunk@42340 -
Allows compilation of compiler using -dDEBUG_NODE_XML
which will generate a NAME-node-dump.xml file for each
unit, program or library compiled,
containing a XML description of the nodes handled during
compilation of the unit, program or library.
git-svn-id: trunk@42271 -
{$modeswitch arraytodynarray} is active (mantis #35576)
o changed compiler to compile without this modeswitch
o added the modeswitch to a test that depends on it
git-svn-id: trunk@42118 -
procoptions even when it's through an array-of-const parameter
* always call create_varargs_paraloc_info() instead of create_paraloc_info()
in the former case, even when no varargs parameters are specified (because
on some platforms even some non-variadic parameters need to be passed
differently, such as on ARM with gnueabihf)
git-svn-id: trunk@41420 -
hiding the method name of the wrapped routine
o also add a few more '&' prefixes to the generated wrapper code to
prevent issues when keywords are used as identifiers
git-svn-id: trunk@40634 -
hlcgllvm in the uses clause of every unit that calls create_hlcodegen
o prevents dependency cycles that can cause llvm codegen units to init
before the cpu variants, which is bad since the llvm versions have to
override the cpu variants in their init code (+ added checks in the
init code that they are in fact initialised later)
git-svn-id: branches/debug_eh@40410 -
in the i8086-msdos 'ports' unit, but will be enabled on other targets (e.g.
go32v2) in the future as well. 32-bit 'in' and 'out' not inlined on i8086, but
will be on i386 and x86_64.
git-svn-id: trunk@39362 -
and filled it with the dwarf register mapping, used by Open Watcom (Watcom
also uses this mapping on i386, but we don't need to support their debugger on
i386 for now)
git-svn-id: trunk@39097 -
information. Only enabled on i8086 (for now). Does not do much yet, but
hopefully, soon it'll start emitting line number information for i8086-msdos
that is usable by DOS debuggers.
git-svn-id: trunk@39006 -
modify the SI and DI registers. For 'register', this makes it compatible with
Borland C++'s __fastcall calling convention, while 'cdecl' becomes compatible
with Borland C, Microsoft C and Watcom C.
git-svn-id: trunk@38824 -
so that they can still be freed after the reference has been changed
(e.g. in case of array indexing or record field accesses) (mantis #33628)
git-svn-id: trunk@38814 -
tcpuparamanager, very similar to the existing get_volatile_registers_XXX. The
new methods are called get_saved_registers_XXX, where XXX is the register
type ("int", "address", "fpu" or "mm")
git-svn-id: trunk@38794 -
tcg8086.a_load_const_cgpara. This fixes passing const longint parameters in
dx:ax when using the register calling convention on i8086.
git-svn-id: trunk@38785 -
on i8086. This is compatible with Borland C++ 3.1's _fastcall. Note that
passing parameters in multiple registers is not yet implemented, so using this
will cause internal errors, but these will be fixed later (and the 'register'
calling convention was never officially supported on i8086).
git-svn-id: trunk@38784 -
models. Even though, they generate less efficient code, they are useful when
interacting with external APIs (e.g. declaring callbacks for the Win16 API).
git-svn-id: trunk@38726 -
o The @ and Addr() operators in TP or Delphi mode can now be applied to both
near and far procedures and they always produce a CodePointer, regardless of
the call model of the procedure.
o Ofs() and Seg() can now also be applied to both near and far procedures.
o The @ and Addr() operators in non-TP/Delphi modes, as well as the procedure
name itself in TP/Delphi modes now can be applied to both near and far
procedures and produce a near or a far procvar.
git-svn-id: trunk@38691 -
