This out parameter called 'result' is of type tcgpara, which is an object, but not a class,
it has no virtual method, which means that the VMT is never used, and thus the
fact that the object is not 'constructed' can be easily missed.
Nevertheless, it could generate problems as the different fields have unassigned
and thus random value, while reset method can dereference location field
if not nil.
Issue found by using -gttt compiler option.
git-svn-id: trunk@47711 -
o added AArch64 regset parsing support in assembler reader, means that "{"
no longer starts comments there (like in the ARM assembler reader)
o added AArch64 indexed SIMD register support and removed old cg hacks
that worked around its absence
git-svn-id: trunk@47116 -
for LLVM, as that results in LLVM assuming exactly the opposite of
what's intended (namely that the extended locations contain valid
rather than invalid bits)
git-svn-id: trunk@43886 -
{$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 -
(like on onther platforms)
o it was passed by reference as a side effect of the fact arrays in C are
pointers, but a dynamic array already is a pointer
git-svn-id: trunk@40573 -
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 -
modulo7, mantis #30381)
o check for array elements of size 0 does not make sense after we've
established it's an array of floats
o since the recursive call to is_hfa_internal() already checks for
floats, we don't have to do it before calling it again for the
elements of an array
o set the result to "true" if the array is small enough to be a HFA
git-svn-id: trunk@34293 -
"don't free even if not registered"; use for defs that may not be written
to a ppu file, but that must nevertheless survive the compilation of the
current module
* mark all defs created for para locations as "don't free even if not
registered", because we don't discard and recalculate all para locations
after a module has been compiled (since that's not needed)
o solves issues if the paralocations for a routine in the interface of
unit A are calculated while the implementation of unit B gets
compiled, and a new reusable type is allocated at that point which
is not used anywhere else (after r32160)
git-svn-id: trunk@32235 -
of in X8, as the ABI allows this deviation (see added comments) and it's
required for some internal RTL code to call interface methods via RTTI
git-svn-id: trunk@29953 -
instead of the frame pointer register:
1) we exactly know the offsets of the temps from the stack pointer
after pass 1 (based on the require parameter stack size for called
routines), while we don't know it for the frame pointer (it depends
on the number of saved registers)
2) temp offsets from the stack pointer are positive while those from
the frame pointer are negative, and we can directly encode much
bigger positive offsets in the instructions
o move the stack pointer register to a virtual register in
loadparentfpn, because many instructions cannot directly operate
on/with the stack pointer
o add the necessary register interference edges for the stack pointer
register
git-svn-id: trunk@29938 -
convention code is triggered before the object structure has been
finalised, so the result can be different in the interface and
implementation. To solve this, something like r20161 has to be
implemented for TP-style objects
git-svn-id: trunk@29930 -
paramgr.pas, tparamanager:
+ add new method "handle_common_ret_in_param"
* extract common code (safecall + record constructor handling) from "ret_in_param" to "handle_common_ret_in_param" and call the latter in "ret_in_param"
[aarch64,arm,avr,i386,x86_64]/cpupara.pas, tCPUparamanager.ret_in_param:
* call "handle_common_ret_in_param" instead of implementing the same check again and again
ncgcal.pas, tcgcallnode.handle_return_value:
* move the check for (record) constructors to "tparamanager.handle_common_ret_in_param"
git-svn-id: trunk@23520 -
that the type of the parameters can be determined automatically
o added compilerproc declarations for all helpers called in the compiler
via their assembler name, so we can look up the corresponding procdef
git-svn-id: trunk@23325 -
