RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 04:02:41 +01:00
Code Issues Projects Releases Wiki Activity
7,279 Commits 4 Branches 5 Tags 704 MiB
78b2442ec4
Commit Graph

141 Commits

Author SHA1 Message Date
Brian Gaeke
af3bc2b895 Factory methods for function passes now return type FunctionPass *. 
Get rid of RegisterLLC, which can't handle FunctionPasses anyway.

llvm-svn: 7836
 2003-08-14 06:04:49 +00:00
Vikram S. Adve
1414fba288 All constant-evaluation code now unified into 
TargetInstrInfo::ConvertConstantToIntType().

llvm-svn: 7390
 2003-07-29 19:50:12 +00:00
Chris Lattner
92dfc1ed32 Making this code const-correct would be a pain, so I'll hack it. 
llvm-svn: 7350
 2003-07-26 23:29:51 +00:00
Chris Lattner
9c569f48a0 Remove unnecessary casts 
llvm-svn: 7250
 2003-07-23 15:17:51 +00:00
Chris Lattner
67138b2c59 Remove using decl 
llvm-svn: 7246
 2003-07-23 14:55:59 +00:00
Vikram S. Adve
d7b18ee501 Choose register instead of immediate for ConstantExpr in ChooseRegOrImmed. 
llvm-svn: 7112
 2003-07-06 20:33:21 +00:00
John Criswell
258dfc0319 Merged in autoconf branch. This provides configuration via the autoconf 
system.

llvm-svn: 7014
 2003-06-30 21:59:07 +00:00
Chris Lattner
493103ea3b Actually, change it to use explicit new/delete, which is more likely to be 
optimized INTO an alloca

llvm-svn: 6727
 2003-06-16 22:29:09 +00:00
Chris Lattner
c32989e03a Remove two using decls 
Remove usage of alloca

llvm-svn: 6725
 2003-06-16 22:18:28 +00:00
Brian Gaeke
4681b27a92 Isolate machine-dependent use of <alloca.h> in "Support/Alloca.h", 
so that we can easily change its use to be conditional on the result of
an autoconf test later.

llvm-svn: 6723
 2003-06-16 21:54:01 +00:00
Misha Brukman
ba5c703619 Do not hastily change the Opcode from 'r' to 'i' type if we're not actually 
SETTING the operand to be an immediate or have verified that one of the operands
is really a SignExtended or Unextended immediate value already, which warrants
an 'i' opcode.

llvm-svn: 6662
 2003-06-07 02:34:43 +00:00
Misha Brukman
b126ea43eb I have finally seen the light. The code to change the opcode must live higher in 
the loop, and in both cases. In the first case, it is a VReg that is a constant
so it may be actually converted to a constant. In the second case, it is already
a constant, but then if it doesn't change its type (e.g. to become a register
and have the value loaded from memory if it is too large to live in its
instruction field), we must change the opcode BEFORE the 'continue', otherwise
we miss the opportunity.

llvm-svn: 6602
 2003-06-04 04:54:06 +00:00
Misha Brukman
7a5eac8485 Moved code to modify the opcode from 'reg' to 'imm' form to a more logical place. 
llvm-svn: 6563
 2003-06-03 03:18:20 +00:00
Vikram S. Adve
d43c10b583 Added MachineCodeForInstruction object as an argument to 
TmpInstruction constructors because every TmpInstruction object has
to be registered with a MachineCodeForInstruction to prevent leaks.
This simplifies the user's code.

llvm-svn: 6469
 2003-05-31 07:41:24 +00:00
Misha Brukman
b1b8a6732e When converting virtual registers to immediate constants, change the opcode. 
llvm-svn: 6452
 2003-05-30 20:36:27 +00:00
Chris Lattner
3e687b8b65 Remove unneccesary &* 
llvm-svn: 5871
 2003-04-23 16:36:11 +00:00
Chris Lattner
301e1307d2 Move sparc specific code into the Sparc backend 
llvm-svn: 5317
 2003-01-15 21:36:50 +00:00
Chris Lattner
ec5237b2fe Fix bug in previous checkin 
llvm-svn: 5310
 2003-01-15 20:32:15 +00:00
Chris Lattner
c934e3fa5f Minor changes 
llvm-svn: 5302
 2003-01-15 19:47:53 +00:00
Chris Lattner
4c178a5732 Remove dead code 
llvm-svn: 5298
 2003-01-15 19:22:38 +00:00
Chris Lattner
b27d60ccf1 Rename MachineInstrInfo -> TargetInstrInfo 
llvm-svn: 5272
 2003-01-14 22:00:31 +00:00
Chris Lattner
dfa3c21f1c More renamings of Target/Machine*Info to Target/Target*Info 
llvm-svn: 5204
 2002-12-29 03:13:05 +00:00
Chris Lattner
0141841bc4 Fix warning 
llvm-svn: 4649
 2002-11-09 00:49:43 +00:00
Chris Lattner
c3e30fed7b Use higher level method 
llvm-svn: 4386
 2002-10-29 17:25:41 +00:00
Chris Lattner
7d6709e213 MachineInstrInfo no longer #includes this header, so we must 
llvm-svn: 4366
 2002-10-28 23:54:47 +00:00
Chris Lattner
fdca057f8d Remove usage of MachineBasicBlock::get 
llvm-svn: 4342
 2002-10-28 19:01:16 +00:00
Chris Lattner
7cdc2977f0 Change MachineBasicBlock's to not be Annotations, instead they are kept as 
part of a linked list tracked by MachineFunction.  MachineBasicBlock::get
is now linear time instead of constant time, and thus is deprecated!

llvm-svn: 4337
 2002-10-28 05:30:46 +00:00
Chris Lattner
e309cdbc95 Rename the redundant MachineOperand::getOperandType() to MachineOperand::getType() 
llvm-svn: 4331
 2002-10-28 04:45:29 +00:00
Chris Lattner
dd22b9e8b6 Add #includes now that MachineInstr.h doesn't include llvm/Target/MachineInstrInfo.h 
llvm-svn: 4327
 2002-10-28 02:28:39 +00:00
Chris Lattner
df1d87f5f3 *** empty log message *** 
llvm-svn: 4323
 2002-10-28 02:01:37 +00:00
Chris Lattner
b8c6acfbc0 Rename MachineCodeForBasicBlock to MachineBasicBlock 
llvm-svn: 4318
 2002-10-28 01:41:47 +00:00
Misha Brukman
142795fd17 Changed MachineCodeForMethod' to MachineFunction'. 
llvm-svn: 4301
 2002-10-28 00:28:31 +00:00
Chris Lattner
47508d5c0f Tell PassManager that this pass does not invalidate the CFG so that dominator 
information and Loop info will not have to be recomputed after this runs.

llvm-svn: 4269
 2002-10-23 03:30:47 +00:00
Vikram S. Adve
3a54b695ce Significant improvement: GEP used by a load or store no longer generates 
a separate ADD; instead just use the indexed load/store instruction!
Also, a bug fix: folding a GEP with a leading non-zero index with
its predecessor was incorrect: now it only happens if the predecessor
is pointing to an indexable type (aka SequentialType).

llvm-svn: 4168
 2002-10-14 16:30:55 +00:00
Vikram S. Adve
01e64b5eef Bug fix in folding getElementPtr instructions: don't fold one into 
a predecessor if it has a non-zero first index and the predecessor
ends with a struct index.

llvm-svn: 3982
 2002-09-29 22:55:05 +00:00
Vikram S. Adve
c795b90eac Sign-extend integer constants from original type size to 64 bits! 
llvm-svn: 3958
 2002-09-27 14:26:20 +00:00
Chris Lattner
7c1b7e5889 Add support for GCC 2.96 
llvm-svn: 3777
 2002-09-17 17:23:09 +00:00
Vikram S. Adve
533c888d4a Add a version of ChooseRegOrImmed to handle numerical constants 
introduced by InstrSelection.

llvm-svn: 3735
 2002-09-16 15:15:57 +00:00
Chris Lattner
23bd92aeeb Fix a bug I introduced into the code generator. :( 
llvm-svn: 3694
 2002-09-12 20:27:10 +00:00
Chris Lattner
18e3c4ff50 - Change getelementptr instruction to use long indexes instead of uint 
indexes for sequential types.

llvm-svn: 3681
 2002-09-11 01:21:29 +00:00
Chris Lattner
22bcab6bd5 Make sure the leakdetector is told about temporary Phi nodes also, so it doesn't 
erroneously report them as leaks

llvm-svn: 3624
 2002-09-08 21:19:29 +00:00
Chris Lattner
1c2e1ebb5c Inform the memory leak detector that TmpInstruction objects should not be 
subject to memory leak checking.

llvm-svn: 3623
 2002-09-08 21:08:43 +00:00
Chris Lattner
3ce5b343c5 - Renamed Type::isIntegral() to Type::isInteger() 
- Added new method Type::isIntegral() that is the same as isInteger, but
    also accepts bool.

llvm-svn: 3574
 2002-09-03 01:08:28 +00:00
Vikram S. Adve
b8a0f00662 Use newly-added type inspection support in InstTreeNode subclasses. 
llvm-svn: 3501
 2002-08-24 21:02:09 +00:00
Vikram S. Adve
3190395e10 Moved code here to extract memory instruction indices and to check 
whether FoldGetElemNodes should be called.  This is machine-independent
but was in the Sparc directory.  This is in a new function: GetMemInstArgs.

llvm-svn: 3500
 2002-08-24 21:00:08 +00:00
Anand Shukla
656de9732b Added #include <alloca.h> 
llvm-svn: 3491
 2002-08-23 10:54:26 +00:00
Chris Lattner
20b85d96e4 Eliminated the MemAccessInst class, folding contents into GEP class. 
llvm-svn: 3487
 2002-08-22 23:37:20 +00:00
Vikram S. Adve
0672c43566 dumpNode() does not need to print MachineInstrs. 
llvm-svn: 3444
 2002-08-22 02:59:46 +00:00
Vikram S. Adve
21b0deb4ab Unary Not (boolean and bitwise) is no longer a separate LLVM instruction 
but is instead implemented with XOR.  Note that the InstrForest opLabels
for Not and BNot remain the same, i.e., the XOR is recognized and
represented as a (boolean or bitwise) Not when building the instruction
trees.  But these tree nodes are now binary, not unary.

llvm-svn: 3343
 2002-08-15 14:19:22 +00:00
Chris Lattner
8a4787495a * Removed extraneous #includes 
* Fixed file headers to be consistent with the rest of LLVM
* Other minor fixes

llvm-svn: 3278
 2002-08-09 20:08:06 +00:00