Jonas Maebe
|
f78e2072a1
* fix for r41416: 64 bit values on 32 bit platforms can of course also be
|
6 years ago |
Jonas Maebe
|
af36913469
* don't use cg64.a_load64_loc_cgpara() for 64 bit records passed in
|
6 years ago |
Jonas Maebe
|
04d295f186
* first step towards supporting 32 bit targets with the LLVM code generator:
|
6 years ago |
Jonas Maebe
|
2b3fb18327
* fixed thlcg2ll.location_force_mmregscalar() after r40633
|
6 years ago |
Jonas Maebe
|
377d4e1b58
* make the size parameter of thlcgobj.location_force_mmregscalar() a
|
6 years ago |
florian
|
31f78ea2b6
+ implementation of the vectorcall calling convention by J. Gareth Moreton
|
7 years ago |
florian
|
b1dff29cbf
* removed unused units
|
8 years ago |
Jonas Maebe
|
a25ebbba3e
+ added volatility information to all memory references
|
8 years ago |
Jonas Maebe
|
0afbe85aab
* various memory reference alignment fixes
|
8 years ago |
Károly Balogh
|
56fea8d047
* have regvars in address registers for defs which are address types. improves code with reference bases in regvars on m68k. and also reduces data register pressure.
|
9 years ago |
Károly Balogh
|
5cfdd8fbf5
removed a redundant zero extension from location_force_reg, which occured if the source location was a smaller than 32bit subreg, and the destination was 64bit. hregister is always overwritten in the codepath below anyway.
|
9 years ago |
Jonas Maebe
|
0fc1fd6ac1
* replaced current_procinfo.currtrue/falselabel with storing the true/false
|
10 years ago |
Jonas Maebe
|
bd203a5b57
* synchronised with trunk till r30240
|
10 years ago |
Jonas Maebe
|
a02d52f533
* support thlcg2ll.gen_loadfpu_loc_cgpara() from LOC_REGISTER to
|
10 years ago |
Jonas Maebe
|
67b8aceaee
* synchronized with privatetrunk till r30095
|
10 years ago |
Jonas Maebe
|
2ab7f5c35d
* moved x86-specific requirements from the generic bsr/bsf code to the
|
10 years ago |
pierre
|
0a3f00ef2f
Avoid Internal Error 200310121 with -O3 for test/units/system/tabs.pp with ppcarm
|
10 years ago |
Jonas Maebe
|
b745dcc64c
* moved g_external_wrapper() to the hlcg, and also g_intf_wrapper() because
|
11 years ago |
Jonas Maebe
|
a637fbe596
* moved all g_exception_*() methods to hlcgobj and cleaned them up (no more
|
11 years ago |
Jonas Maebe
|
9a683a0288
* converted new_exception to hlcgobj
|
11 years ago |
Jonas Maebe
|
7949bebb8d
* synchronised with r28168 of trunk
|
11 years ago |
sergei
|
847adff94d
* Load double-precision locations from integer registers directly into cgpara if sizes match and destination has a single location.
|
11 years ago |
Jonas Maebe
|
bacd303208
* synchronized with trunk up to r27758
|
11 years ago |
nickysn
|
3e308e0ad7
* 16/8-bit alu fix in thlcg2ll.location_force_reg for the 32->64 conversion of LOC_CONSTANT sources
|
11 years ago |
florian
|
6f14dbdebe
* avoid internal error if an empty record is assigned to a function result, resolves #25895
|
11 years ago |
Jonas Maebe
|
4065483a50
* completed thlcgobj.location_force_fpureg(), use it everywhere and removed
|
11 years ago |
Jonas Maebe
|
81427523ab
* pass a list of (pointers to) the paralocs to hlcgobj.a_call/g_call*, as
|
11 years ago |
sergei
|
9c1f917e3a
* a_call_ref functionality cannot be implemented efficiently at code generator level, because references need specific preparations at earlier points. Moved this support to tcgcallnode and its x86 descendants, and got rid of all ifdef's around.
|
11 years ago |
nickysn
|
f69a76b5b1
* generate shorter code on 8-bit and 16-bit CPUs in thlcg2ll.location_force_reg,
|
11 years ago |
nickysn
|
9ec84d1259
* generate shorter code on 8 and 16-bit CPUs in thlcg2ll.location_force_reg,
|
11 years ago |