 florian
|
b178b08ba7
Merged revisions 11665-11738 via svnmerge from
|
17 years ago |
|
florian
|
1afb1aa9cc
+ ror/rol functions
|
17 years ago |
 yury
|
4dac183601
* Bring back tcg.getflagregister on Florian's request, but made it non-abstract. It may be used for IA64 support in future (if IA64 will live so long :).
|
17 years ago |
|
yury
|
0e34e98b82
* Removed abstract modifier for some optional (not needed for every CPU target) methods in tcg and tparamanager classes. These methods throw internal error if not overrode.
|
17 years ago |
|
yury
|
8ce526d760
* Removed unused abstract method tcg.getflagregister
|
17 years ago |
 Jonas Maebe
|
a60a957420
* fixed setting bitpacked record fields straddling their natural boundaries
|
17 years ago |
|
yury
|
ec943198fd
* Properly fill treference.alignment when variable is loaded by tcgloadnode. It allows code generator to insert unaligned handling if needed.
|
17 years ago |
|
Jonas Maebe
|
f36e5411af
* split cpu64bit compiler define into
|
17 years ago |
|
Jonas Maebe
|
8349cde7db
* changed byte/word/longbool to be Delphi-compatible (+ similar changes
|
17 years ago |
|
Jonas Maebe
|
e72fab1e43
* perform floating point type conversions directly in assignment nodes
|
17 years ago |
 peter
|
b5e417bd19
* increase the temp requirements for saving MM registers with the full
|
17 years ago |
|
Jonas Maebe
|
bc545b077f
- removed internalerror in case a subsetref value is loaded on a cpu
|
17 years ago |
|
peter
|
6058b2c247
* renamed t_times to executionweight and moved it to
|
17 years ago |
|
Jonas Maebe
|
b4b069908f
- removed some unnecessary cg. prefixes when calling tcg methods
|
17 years ago |
|
peter
|
f704ee2d70
* remove duplicate temp allocation for register saving
|
17 years ago |
|
Jonas Maebe
|
032352d98b
+ darwin/x86_64 support
|
18 years ago |
|
florian
|
0b83cb5761
* align locations where mm registers are stored to be preserved
|
18 years ago |
|
florian
|
00d6a03b2c
+ default code now preserves mm registers
|
18 years ago |
|
Jonas Maebe
|
b103605bc3
* added comment clarifying why g_indirect_sym_load doesn't have to
|
18 years ago |
|
Jonas Maebe
|
335bc9fd46
+ PIC support for darwin/ppc32 (-Cg works now, no regressions in test
|
18 years ago |
 tom_at_work
|
b0bb3f94a7
* improve powerpc64/linux stub for external procedures in units: also consider GOT change in call to this external procedure
|
18 years ago |
|
peter
|
58e9fc0368
* only freetemp when there were registers saved
|
18 years ago |
 daniel
|
75af9f2a4d
* Fixed another placed where chained paralocs where not being processed.
|
18 years ago |
|
daniel
|
9adb202a92
* Rework the constexprint to allow operations from low(int64) to high(qword).
|
18 years ago |
|
florian
|
8e3ade9f1f
Merged revisions 7567 via svnmerge from
|
18 years ago |
|
florian
|
a3d906f1ca
- removed several ;;
|
18 years ago |
|
Jonas Maebe
|
a244d54d6f
* fixed a_load_ref_reg_unaligned in case reg is used in ref
|
18 years ago |
|
Jonas Maebe
|
1e10424c6b
* fixed a_load_reg_ref_unaligned() so it doesn't destroy the source
|
18 years ago |
|
Jonas Maebe
|
528eb2ec50
* shl/shr count on sparc is "mod <bitsize>" -> fix subsetref routines to
|
18 years ago |
|
Jonas Maebe
|
a0b57eddb5
* new internal set format for big endian systems. Advantages:
|
18 years ago |
