Jeppe Johansen
|
15150aea04
Initial experimental support for GCRA.
|
há 10 anos atrás |
sergei
|
9b9ea15d68
+ m68k: initial implementation of do_spill_replace (supports only MOVE instructions for now). Tested with qemu for Coldfire target.
|
há 11 anos atrás |
Jonas Maebe
|
246caf3526
* never alias/coalesce cpu registers that are not usable by the register
|
há 11 anos atrás |
florian
|
23c724f885
* prevent a temp. register allocated during spilling being immediately spilled again, resolves #25164
|
há 11 anos atrás |
florian
|
829764e96b
+ replace registers to spill in mov instructions and convert them into ldr/str
|
há 11 anos atrás |
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...
|
há 11 anos atrás |
Károly Balogh
|
70f9b098e0
no longer define a segment field for m68k
|
há 11 anos atrás |
sergei
|
d2a9308181
+ SPARC: implemented register spill replacement.
|
há 11 anos atrás |
svenbarth
|
32a18199c5
Print aX or aregX for address registers if DEBUG_REGISTERLIFE is defined.
|
há 12 anos atrás |
florian
|
dc567acc9a
* accumulate register weight instead of taking only the maximum
|
há 12 anos atrás |
florian
|
c2f7ccb253
* coalesce more aggressive
|
há 12 anos atrás |
sergei
|
8b8553991a
+ MIPS: prevent coalescing written-to registers with $sp,$fp,$zero and $at.
|
há 12 anos atrás |
florian
|
23c84fd27c
* more fixes for segment register usage
|
há 12 anos atrás |
florian
|
b85addb451
+ color segment register
|
há 12 anos atrás |
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.
|
há 12 anos atrás |
pierre
|
322b793506
Try to do something for m68k integer/address registers; not working yet :!(
|
há 12 anos atrás |
svenbarth
|
2ada9a528b
Fix a critical bug in the register allocator (at least for CPUs with seperate address
|
há 12 anos atrás |
florian
|
95830d5a4a
+ DEBUG_REGISTERLIFE define
|
há 13 anos atrás |
florian
|
f75e39e660
+ aliasing of registers, allows to split live ranges
|
há 13 anos atrás |
florian
|
f9dd5bf5e6
+ live_start and live_end_properties
|
há 13 anos atrás |
florian
|
c2d067940a
* get rid of hash chars, modern tools and IDEs can use normal comments as documentation if needed
|
há 13 anos atrás |
florian
|
b78f9826fb
* call AllocMem instead of GetMem/FillChar
|
há 13 anos atrás |
florian
|
ecdec2e431
* disable reg. allocator optimization introduced in r21812 for now because it caused some trouble, see #22405
|
há 13 anos atrás |
masta
|
f9cdf3d4ca
Be more specific in spilling debug messages
|
há 13 anos atrás |
masta
|
aef81cad68
Add debug information about register spilling
|
há 13 anos atrás |
florian
|
9833f642fe
* coalesce if either u or v are precolored, tested on arm and i386, it gives for both a small advantage
|
há 13 anos atrás |
Jonas Maebe
|
0fc422f244
* moved definition of maxcpuregister and tcpuregisterset from cgbase to
|
há 13 anos atrás |
Jonas Maebe
|
6f336a5d13
* the register allocator always knows the exact number of the first imaginary
|
há 13 anos atrás |
Jonas Maebe
|
aee5380ae0
* merged trunk up to r20882
|
há 13 anos atrás |
florian
|
6211a83865
* write variable locations of 64 bit integer values on 32 bit targets correctly
|
há 14 anos atrás |