 marco
|
185b9a216d
--- Merging r29639 into '.':
|
před 10 roky |
 sergei
|
ae627a4ba8
* tcgx86.a_op_const_reg_reg: optimize trivial cases before considering more complex ones.
|
před 11 roky |
 nickysn
|
daf71e6d88
* fixed multiplication by 64-bit constants on x86_64, mantis #26230
|
před 11 roky |
|
nickysn
|
6e7dd647fd
* Tcgx86.g_concatcopy: optimization for the case where source.segment is
|
před 11 roky |
|
nickysn
|
1eb24f9d98
+ remove the segment prefix in make_simple_ref on i8086, in case it is the same
|
před 11 roky |
|
sergei
|
07e90aaa24
+ Implemented IEEE 754-compliant checking for unordered results of floating-point compares on x86 targets. Mantis #9362.
|
před 11 roky |
 florian
|
18233f2c0c
* fix compilation of i8086 compiler
|
před 11 roky |
|
florian
|
d88d644925
+ support for FMA intrinsic: if there is no hardware support, the compiler throws an error.
|
před 11 roky |
|
nickysn
|
afcb8c3f10
* initialize DS from CS in the interrupt procedure entry code in tiny model,
|
před 11 roky |
|
nickysn
|
7a873f495d
* use the stack segment for SP-relative addresses in Tcgx86.g_concatcopy in
|
před 11 roky |
|
nickysn
|
ee20ecfcf2
* tcgx86.make_simple_ref: in i8086 memory models with DS<>SS, add an SS: segment
|
před 11 roky |
|
nickysn
|
b36a98ce51
+ Tcgx86.g_concatcopy: added support for memory models with DS<>SS
|
před 11 roky |
|
nickysn
|
e9e4cc8bcb
* Tcgx86.g_concatcopy: use boolean variables saved_ds and saved_es to keep track
|
před 11 roky |
|
nickysn
|
096bba90da
* fixed Tcgx86.g_concatcopy not using the correct push instruction size when
|
před 11 roky |
 pierre
|
044891fd94
Avoid range check error in a_op_const_reg_reg
|
před 11 roky |
|
nickysn
|
229ebf33a8
+ support getting the address of far references on i8086 in
|
před 11 roky |
|
florian
|
37b34edbc6
* do not reuse registers when calculating addresses in pic mode
|
před 11 roky |
|
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.
|
před 11 roky |
|
sergei
|
0d3f36eebf
- Remove references to global variable 'cg' from methods of tcg and some of its descendants.
|
před 11 roky |
|
florian
|
28a8dc42c6
* fix tcgx86.a_op_reg_reg on x86-64 and make use of it
|
před 11 roky |
|
florian
|
7d7bf1d877
+ make use of SHLX/SHRX on CPUs supporting BMI2
|
před 11 roky |
|
florian
|
48ae2d215a
+ concatcopy variants using sse and avx, only activated if optimization for size is done because at least on an i7-4770 it has shown no benefit
|
před 11 roky |
|
nickysn
|
85dd9e5789
+ added a size parameter to optimize_op_const and do a sign extension of the 'a' parameter up from the specified size, so that things like (i and $ffffffff) get optimized away the same way as (i and -1)
|
před 11 roky |
|
sergei
|
6b1f021fcf
- Removed remaining optimizations that duplicate functionality of tcg.optimize_op_const.
|
před 11 roky |
|
nickysn
|
1f7b14f38e
* fixed the conversion of powerof2 mul/imuls to lea in tcgx86.a_op_const_reg_reg
|
před 11 roky |
|
nickysn
|
cc1faad3d2
+ make use of the 3-operand form of IMUL on i386 and x86_64
|
před 11 roky |
|
nickysn
|
5d75bf4f92
* always use IMUL instead of MUL on i386, when doing a 32->32 multiplication
|
před 11 roky |
|
nickysn
|
820b667e1b
- rm the OP_IMUL to SHL optimization from tcgx86.a_op_const_reg and tcgx86.a_op_const_ref, because it's already done in optimize_op_const, which is called in the beginning of both methods.
|
před 11 roky |
|
florian
|
e210d5f30e
+ cpu_capabilites for x86_64 and i386
|
před 11 roky |
|
sergei
|
f2096de53a
* tcg386.g_proc_exit: instead of recalculating stack size, use current_procinfo.final_localsize which was calculated in g_proc_entry.
|
před 11 roky |
