RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-15 07:59:33 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
1,245 Commits 4 Branches 5 Tags
Commit Graph

710 Commits

Author SHA1 Message Date
Chris Lattner
7d0854c71e Remove debug info :( 
llvm-svn: 1280
 2001-11-12 21:52:57 +00:00
Chris Lattner
a6ada3716c Fix bug in new assertion 
llvm-svn: 1279
 2001-11-12 21:48:38 +00:00
Ruchira Sasanka
3484bbb97c Fixed a bug with pervious ColorCallArg 
llvm-svn: 1278
 2001-11-12 20:54:19 +00:00
Ruchira Sasanka
67478b7484 Canged ColorCallArg so that when a call arg is directly pushed on to stack 
for argument passing, that instruction is directly added to the InstructionsBefore
set of the called machine instruction - i.e., it is not reordered.

llvm-svn: 1277
 2001-11-12 20:31:47 +00:00
Chris Lattner
3dc54c19b3 Add an assertion check 
llvm-svn: 1276
 2001-11-12 20:30:32 +00:00
Chris Lattner
2dce653a78 I don't know what I was thinking 
llvm-svn: 1275
 2001-11-12 20:30:12 +00:00
Chris Lattner
ab0f8a05f6 Print percentages for profiling info better 
llvm-svn: 1274
 2001-11-12 20:13:14 +00:00
Chris Lattner
184137a419 Fix minor bug 
llvm-svn: 1273
 2001-11-12 20:13:04 +00:00
Ruchira Sasanka
d6c5e0f535 For phi elimination, now we are generating only one instruction 
using cpValue2Value instead of
passing vector to createCopyInstructionByType.

llvm-svn: 1272
 2001-11-12 19:42:27 +00:00
Ruchira Sasanka
3e193c7709 Changed insert instructions only for CONSTANTS - back to original 
llvm-svn: 1271
 2001-11-12 19:32:04 +00:00
Vikram S. Adve
a3f0115c71 Disable use of the Phi machine instruction which is no longer needed 
for register allocation.

llvm-svn: 1270
 2001-11-12 18:54:11 +00:00
Vikram S. Adve
624846d3fd Eliminate most uses of the machine instruction vector for each LLVM instr, 
since some m. instr. may be generated by LLVM instrs. in other blocks.
Handle non-SSA (anti and output) edges and true edges uniformly by
working with machine instructions alone.

llvm-svn: 1269
 2001-11-12 18:53:43 +00:00
Chris Lattner
29ad087a83 Print profile info if exit() is called 
llvm-svn: 1268
 2001-11-12 16:28:48 +00:00
Chris Lattner
28afda8ca4 Hack a structure profiling option together 
llvm-svn: 1267
 2001-11-12 16:19:45 +00:00
Ruchira Sasanka
426d6e2b10 Added phi elimination code 
llvm-svn: 1265
 2001-11-12 14:45:33 +00:00
Ruchira Sasanka
9b7638e108 Added phi elimination code - not final 
llvm-svn: 1264
 2001-11-12 14:44:50 +00:00
Vikram S. Adve
1fc185b8e7 Improved printing routines. 
llvm-svn: 1263
 2001-11-12 14:19:47 +00:00
Vikram S. Adve
3a5bf13976 Major improvement to how nodes are built for a BB. 
LLVM instruction is no longer recorded in each node, but BB is.

llvm-svn: 1262
 2001-11-12 14:18:01 +00:00
Vikram S. Adve
31372d2b26 Adjust all stack variable offsets to be aligned by size. 
llvm-svn: 1261
 2001-11-12 05:17:23 +00:00
Vikram S. Adve
bb424a4fe1 Fix dumb bug in alignment adjustment code! 
llvm-svn: 1260
 2001-11-12 05:16:39 +00:00
Vikram S. Adve
2b5e92123c Fix bogus code that was eliminating needed args. 
llvm-svn: 1259
 2001-11-11 23:11:36 +00:00
Ruchira Sasanka
661bf856f5 Fixed FP and SP usage. SP for outgoing args - for everything else FP 
llvm-svn: 1258
 2001-11-11 22:37:51 +00:00
Ruchira Sasanka
9d20df621c cahnged StackPointer accesses to FramePointer 
llvm-svn: 1257
 2001-11-11 21:49:37 +00:00
Vikram S. Adve
51265e7072 Fix errors in computing downgrowing offsets, and in 
computing size of extra outgoing args.

llvm-svn: 1256
 2001-11-11 21:23:33 +00:00
Vikram S. Adve
b98d36462b Down-growing offsets from FP should start at 0, not -1. 
llvm-svn: 1255
 2001-11-11 21:22:11 +00:00
Vikram S. Adve
0b0bd6e749 Only add true dep. edges from an earlier to a later instruction. 
This wasn't a problem until we started putting copies for Phi values
that produced cycles in the SchedGraph!

llvm-svn: 1254
 2001-11-11 01:23:27 +00:00
Ruchira Sasanka
bf9791fe90 Changed code to ignore Phi Nodes in PhyRegAlloc 
llvm-svn: 1253
 2001-11-10 21:21:36 +00:00
Ruchira Sasanka
bf341d2d5d Corrected reodering code for instructions inserted before calls 
llvm-svn: 1252
 2001-11-10 21:20:43 +00:00
Chris Lattner
053906abc0 Initail checking of structure swapper 
llvm-svn: 1248
 2001-11-10 07:28:25 +00:00
Chris Lattner
79fd0836bb Initial checkin of structure mutator 
llvm-svn: 1246
 2001-11-10 07:26:31 +00:00
Vikram S. Adve
ca73b36000 Several bug fixes in printing constants, particularly with unsized 
arrays and multi-dimensional arrays.

llvm-svn: 1239
 2001-11-10 02:03:06 +00:00
Vikram S. Adve
cbc2cec180 Must generate an instruction for GetElementPtr if single user is not 
a memory instruction!

llvm-svn: 1238
 2001-11-10 01:05:26 +00:00
Ruchira Sasanka
e98f033070 Disabled some debuggin messages 
llvm-svn: 1237
 2001-11-10 00:26:55 +00:00
Ruchira Sasanka
699d104906 Coalesing bug fix - now checks for the same regType instead of the same 
regClass since FP class has two reg Types.

llvm-svn: 1236
 2001-11-10 00:20:24 +00:00
Ruchira Sasanka
c32ebd1a4e No major change 
llvm-svn: 1235
 2001-11-09 23:49:42 +00:00
Ruchira Sasanka
a4a64f3755 Added code for correct reordering of call arguments 
llvm-svn: 1234
 2001-11-09 23:49:14 +00:00
Chris Lattner
37955a1d28 New pass to find types in use by a program 
llvm-svn: 1232
 2001-11-09 05:27:21 +00:00
Vikram S. Adve
68306457ae Added class MachineCacheInfo. 
llvm-svn: 1229
 2001-11-09 02:20:18 +00:00
Vikram S. Adve
6c5f14ebbc Add support to print constant arrays and structures. 
Align data larger than an L1 cache line on L1 cache line boundary.

llvm-svn: 1228
 2001-11-09 02:19:29 +00:00
Vikram S. Adve
91f30ac019 Do the same for float->int that we did for int->float earlier. 
Also check IsPointerType() in addition to IsIntegral() in several places.

llvm-svn: 1227
 2001-11-09 02:18:16 +00:00
Vikram S. Adve
61173418a7 Added function UltraSparcInstrInfo::CreateCodeToCopyFloatToInt. 
llvm-svn: 1226
 2001-11-09 02:16:40 +00:00
Vikram S. Adve
8766f0ad37 Added class UltraSparcCacheInfo. 
llvm-svn: 1225
 2001-11-09 02:16:04 +00:00
Vikram S. Adve
1b7faa892c Bug fix: Need to include the stack offset "bias" in assembly code itself. 
Added class UltraSparcCacheInfo.
Added function to convert float to int by copying via memory.

llvm-svn: 1224
 2001-11-09 02:15:52 +00:00
Vikram S. Adve
cc746868d7 Use node->getOpCode() instead of node->getMachineInstr()->getOpCode(). 
Much nicer.

llvm-svn: 1223
 2001-11-09 02:14:20 +00:00
Chris Lattner
de13795a04 Remove false optimization that basically broke everything 
llvm-svn: 1219
 2001-11-09 01:08:10 +00:00
Chris Lattner
7969ed3025 * Add better caching of data to avoid silly recusions 
* Only check to see if uses of instructions can be converted for expressions... so we don't look at all of the uses of a constant.  This was making the code unnecessarily conservative

llvm-svn: 1218
 2001-11-08 22:06:31 +00:00
Chris Lattner
d0f1f10bec Add method decl 
llvm-svn: 1217
 2001-11-08 20:55:05 +00:00
Chris Lattner
fdc70e08d0 Improve raising significantly 
llvm-svn: 1214
 2001-11-08 20:19:56 +00:00
Chris Lattner
41f173e9d0 Print the method that makes an instruction invalid 
llvm-svn: 1213
 2001-11-08 19:29:51 +00:00
Ruchira Sasanka
5ebbfbdd2e corrected insertCode4Spilled ... bug. 
llvm-svn: 1212
 2001-11-08 19:11:30 +00:00