sergei
|
c0eafe38aa
* Make trgobj.uses_registers method return True when registers of appropriate type are specified in list of used registers (after assembler blocks), but not used otherwise. This is a missing piece of r30011, enabling code generator to actually save/restore such registers. Resolves #28421.
|
9 rokov pred |
florian
|
199b1861d6
* fix wrong comparisons in trgobj.conservative to enable better coalescing,
|
10 rokov pred |
florian
|
31709b1b43
* check registers of ait_regalloc for a valid color when translating and compiling with -dextdebug
|
10 rokov pred |
Jonas Maebe
|
e12dff9a5e
* support top_para in trgllvm.get_spill_temp()
|
10 rokov pred |
Jonas Maebe
|
e750678f37
* factored out spilling code that deals with operand types into virtual
|
10 rokov pred |
Jonas Maebe
|
197e5194b9
* don't crash if we have to spill a register of the last instruction in an
|
10 rokov pred |
Jonas Maebe
|
67b8aceaee
* synchronized with privatetrunk till r30095
|
10 rokov pred |
sergei
|
31fddaafe8
+ New subtype of tai_regalloc, "ra_markused". It marks a physical register as used in procedure, triggering saving/restoring that register if it is non-volatile.
|
10 rokov pred |
Jonas Maebe
|
5c75b6dd6b
* synchronised with trunk up till r28402
|
11 rokov pred |
sergei
|
9b9ea15d68
+ m68k: initial implementation of do_spill_replace (supports only MOVE instructions for now). Tested with qemu for Coldfire target.
|
11 rokov pred |
Jonas Maebe
|
bacd303208
* synchronized with trunk up to r27758
|
11 rokov pred |
Jonas Maebe
|
246caf3526
* never alias/coalesce cpu registers that are not usable by the register
|
11 rokov pred |
Jonas Maebe
|
e9268a0a14
* synchronised with trunk up till r26975
|
11 rokov pred |
florian
|
23c724f885
* prevent a temp. register allocated during spilling being immediately spilled again, resolves #25164
|
11 rokov pred |
florian
|
829764e96b
+ replace registers to spill in mov instructions and convert them into ldr/str
|
11 rokov pred |
Károly Balogh
|
587acf6452
fix infinite spilling on m68k, patch by Florian, additional IE workaround by me, based on a similar hack in the ARM cg...
|
11 rokov pred |
Károly Balogh
|
70f9b098e0
no longer define a segment field for m68k
|
11 rokov pred |
sergei
|
d2a9308181
+ SPARC: implemented register spill replacement.
|
11 rokov pred |
Jonas Maebe
|
b7803ab974
+ llvm support for the register allocator. While llvm works with virtual
|
11 rokov pred |
Jonas Maebe
|
5ef93e85b8
+ added extra "orgsupreg" parameter to do_spill_read/do_spill_written/
|
11 rokov pred |
Jonas Maebe
|
5599870a4e
* moved the register colouring and spill temp allocation to a virtual method,
|
11 rokov pred |
svenbarth
|
32a18199c5
Print aX or aregX for address registers if DEBUG_REGISTERLIFE is defined.
|
12 rokov pred |
florian
|
dc567acc9a
* accumulate register weight instead of taking only the maximum
|
12 rokov pred |
florian
|
c2f7ccb253
* coalesce more aggressive
|
12 rokov pred |
sergei
|
8b8553991a
+ MIPS: prevent coalescing written-to registers with $sp,$fp,$zero and $at.
|
12 rokov pred |
florian
|
23c84fd27c
* more fixes for segment register usage
|
12 rokov pred |
florian
|
b85addb451
+ color segment register
|
12 rokov pred |
svenbarth
|
d15304c25e
Also check for the type of the register when replacing it. In certain circumstances this can still lead to access to invalid memory resulting either in an access violation or invalid registers.
|
12 rokov pred |
pierre
|
322b793506
Try to do something for m68k integer/address registers; not working yet :!(
|
12 rokov pred |
svenbarth
|
2ada9a528b
Fix a critical bug in the register allocator (at least for CPUs with seperate address
|
12 rokov pred |