a45a216ff6
Make the verifier API more complete and useful.
...
Patch contributed by Reid Spencer
llvm-svn: 12609
2004-04-02 15:45:08 +00:00
467cb2bc5d
Add support for constant expr casts
...
llvm-svn: 12564
2004-03-31 02:56:11 +00:00
b7897ce431
Fix a latent bug in select constantexpr handling that was broke 253.perlbmk
...
llvm-svn: 12562
2004-03-30 22:51:03 +00:00
3cdc27c868
Handle -0.0 correctly
...
llvm-svn: 12530
2004-03-29 19:51:24 +00:00
817175f883
Add a bunch of methods that should have been added a long time ago.
...
llvm-svn: 12526
2004-03-29 02:37:53 +00:00
5f126b74d3
Add another check
...
llvm-svn: 12525
2004-03-29 00:29:36 +00:00
587a9f1416
Fix an assertion
...
Contributed by Reid Spencer
llvm-svn: 12524
2004-03-29 00:17:20 +00:00
04d6d9d606
Add two methods which have been needed for a long time: Type::get(Un)signedVersion
...
llvm-svn: 12522
2004-03-26 21:43:22 +00:00
9ece94b02b
Catch some more cases of broken code. The loop extractor seems to be creating
...
situations where there is a branch that goes to a block in another function.
llvm-svn: 12379
2004-03-14 03:23:54 +00:00
b870ca7766
verifyFunction has been broken for a long time now. Fix it.
...
llvm-svn: 12377
2004-03-14 03:16:15 +00:00
071a5e5649
Rename the intrinsic enum values for llvm.va_* from Intrinsic::va_* to
...
Intrinsic::va*. This avoid conflicting with macros in the stdlib.h file.
llvm-svn: 12356
2004-03-13 00:24:00 +00:00
a447feac33
Keep transitively-required passes alive for queries to work after the initial
...
user pass is destroyed.
llvm-svn: 12332
2004-03-12 06:16:28 +00:00
75648e7ae4
Add support for checking the select instruction
...
llvm-svn: 12325
2004-03-12 05:54:31 +00:00
8337d4462f
Know the opcode name of the select instruction
...
llvm-svn: 12324
2004-03-12 05:54:20 +00:00
6e415c06c2
Add support for select constant expressions. Use reserve a bit more to avoid
...
memory wasteage.
llvm-svn: 12323
2004-03-12 05:54:04 +00:00
16375e30b0
Add new function
...
llvm-svn: 12322
2004-03-12 05:53:41 +00:00
6ea4b52903
Teach the constant folder how to do select instructions
...
llvm-svn: 12321
2004-03-12 05:53:32 +00:00
52bd5cb967
Print select instructions correctly
...
llvm-svn: 12320
2004-03-12 05:53:14 +00:00
1df51ed4e5
Fix indentation.
...
llvm-svn: 12298
2004-03-11 23:53:51 +00:00
96cdd27b65
Adjust to new interface
...
llvm-svn: 12231
2004-03-08 18:51:45 +00:00
ea0789ca7b
Implement test/Regression/Assembler/2004-03-07-FunctionAddressAlignment.llx
...
llvm-svn: 12218
2004-03-08 06:17:35 +00:00
c3ecaae16c
Remove Module::mutateConstantPointerRef, which is now thankfully dead!
...
This is one small step towards the complete obliteration of
ConstantPointerRef's entirely!! Woot!
llvm-svn: 12216
2004-03-08 06:16:10 +00:00
9f4918ab93
Eliminate nightmarish API
...
llvm-svn: 12214
2004-03-08 06:11:10 +00:00
c566ca36a1
Doxygenify comments.
...
llvm-svn: 12071
2004-03-02 00:22:19 +00:00
b22d09cc5e
* If a badref has a name, print it out for ease of debugging
...
* Doxygenify (some) comments
llvm-svn: 12057
2004-03-01 19:48:13 +00:00
43640d7138
Fix -debug-pass=Executions, which relied on Function, Module, and BasicBlock
...
being annotable
llvm-svn: 12013
2004-02-29 22:37:04 +00:00
465a5cc291
Fix typo in comment
...
llvm-svn: 11966
2004-02-28 21:55:18 +00:00
3705370df1
Fix Regression/Assembler/2004-02-27-SelfUseAssertError.ll
...
llvm-svn: 11913
2004-02-27 17:28:25 +00:00
120035898b
Make sure that at least one virtual method is defined in a .cpp file to avoid
...
having the compiler emit RTTI and vtables to EVERY translation unit.
llvm-svn: 11871
2004-02-26 07:24:18 +00:00
d996e543eb
Make the verifier a little more explicit about this problem.
...
llvm-svn: 11811
2004-02-24 22:06:07 +00:00
a9eddae213
Get all instruction definitions
...
llvm-svn: 11706
2004-02-22 06:25:38 +00:00
ecc7d6e27a
Adjust to recent changes
...
llvm-svn: 11514
2004-02-17 03:03:47 +00:00
04b60fe5ad
Move the folding of gep null, 0, 0, 0 to a place where it can be shared and
...
enjoyed by all, fixing a fixme. Add an assert
llvm-svn: 11505
2004-02-16 20:46:13 +00:00
303ca1e509
By default, llvm.setjmp/llvm.longjmp intrinsics get lowered to their libc counterparts
...
llvm-svn: 11483
2004-02-15 22:24:51 +00:00
25530a12a1
Refactor code. Now the intrinsic lowering pass tries to recycle preexisting
...
prototypes, even if they don't precisely match what it would prefer to use.
This fixes: CBackend/2004-02-15-PreexistingExternals.llx compiling it into:
ltmp_0_30 = memcpy(l14_C, 4u, 17);
ltmp_1_30 = memcpy(((int *)l27_A), ((unsigned )(long)l27_B), ((int )123u));
instead of:
ltmp_0_30 = memcpy(l14_C, 4u, 17);
ltmp_1_27 = l43_memcpy(l27_A, l27_B, 123u);
Which does the wrong thing as you could imagine.
llvm-svn: 11481
2004-02-15 22:16:39 +00:00
76b2ff4ded
Adjustments to support the new ConstantAggregateZero class
...
llvm-svn: 11474
2004-02-15 05:55:15 +00:00
9fba3da879
Add a new ConstantAggregateZero class, to fix PR239. This makes zero
...
initializers for constant structs and arrays take constant space, instead of
space proportinal to the number of elements. This reduces the memory usage of
the LLVM compiler by hundreds of megabytes when compiling some nasty SPEC95
benchmarks.
llvm-svn: 11470
2004-02-15 05:53:04 +00:00
015e821279
ConstantArray::get and ConstantStruct::get now just return pointers to
...
'Constant', instead of specific subclass pointers. In the future, these will
return an instance of ConstantAggregateZero if all of the inputs are zeros.
llvm-svn: 11467
2004-02-15 04:14:47 +00:00
37a716fa80
Remove dependence on return type of ConstantStruct::get
...
llvm-svn: 11466
2004-02-15 04:07:32 +00:00
c75bf528c1
Remove dependence on the return type of ConstantArray::get
...
llvm-svn: 11463
2004-02-15 04:05:58 +00:00
5d56c478ba
Keep a cache of non-abstract null arrays and structs. This speeds up llvm-dis
...
from 16.57 -> 13.46s on 129.compress.
llvm-svn: 11462
2004-02-15 02:46:46 +00:00
0c48a03374
A target that doesn't support these intrinsics will still meet spec (the
...
intrinsic will always produce zero), but it will behave unexpectedly, so
warn like GCC does.
llvm-svn: 11444
2004-02-14 04:52:06 +00:00
dc632111bf
Add llvm.memset/frameaddress/returnaddress intrinsics.
...
llvm-svn: 11431
2004-02-14 02:47:17 +00:00
5ed171e317
Add support for the llvm.memmove intrinsic
...
Patch graciously contributed by Reid Spencer!
llvm-svn: 11355
2004-02-12 18:11:20 +00:00
17d028d52b
Implement the llvm.memcpy intrinsic
...
llvm-svn: 11349
2004-02-12 17:01:09 +00:00
2be4744c5e
Remove obsolete method
...
llvm-svn: 11302
2004-02-11 01:17:33 +00:00
54b582f11d
Remove the statistics
...
llvm-svn: 11243
2004-02-09 21:01:23 +00:00
478fac37e4
Speed up type resolution some more. On the testcase in PR224, for example,
...
this speeds up a release llvm-as from 21.95s to 11.21s, because before it
would do an expensive traversal of the type-graph of every type resolved.
llvm-svn: 11242
2004-02-09 20:23:44 +00:00
aedce9d6b1
Implement the hashing scheme in an attempt to speed up the "slow" case in
...
type resolution. Unfortunately it doesn't help.
Also delete some dead debugging code.
llvm-svn: 11237
2004-02-09 18:32:40 +00:00
eecaba0c33
Code cleanup in preparation for later changes. Now that ContainedTy's are
...
consistent across the various type classes, we can factor out a LOT more
almost-identical code. Also, add a couple of temporary statistics.
llvm-svn: 11232
2004-02-09 16:35:14 +00:00
cbd34b2ae6
Now that all of the derived types have disciplined interfaces, we can eliminate
...
all of the ad-hoc storage of contained types. This allows getContainedType to
not be virtual, and allows us to entirely delete the TypeIterator class.
llvm-svn: 11230
2004-02-09 05:40:24 +00:00
ac6db755c3
Adjust to the changed StructType interface. In particular, getElementTypes() is gone.
...
llvm-svn: 11228
2004-02-09 04:37:31 +00:00
fa829be4d3
Start using the new and improve interface to FunctionType arguments
...
llvm-svn: 11224
2004-02-09 04:14:01 +00:00
2eaf58775e
Two problems with these lines of code:
...
1. The "work" was not in the assert, so it was punishing the optimized release
2. getNamedFunction is _very_ expensive in large programs. It is not designed
to be used like this, and was taking 7% of the execution time of the code
generator on perlbmk.
Since the assert "can never fail", I'm just killing it.
llvm-svn: 11214
2004-02-09 00:59:07 +00:00
5b6892e37c
Fix PR215: [bcwriter] Problem compactifying ConstantPointerRefs
...
Have I ever mentioned how much I _hate_ constantpointerrefs?
llvm-svn: 11212
2004-02-09 00:15:41 +00:00
85195147e1
vi failed me again. :)
...
llvm-svn: 11206
2004-02-08 21:52:30 +00:00
144f0efd4c
Rename the invoke 'except' destination to the 'unwind' destination
...
llvm-svn: 11205
2004-02-08 21:52:04 +00:00
fae8ab3088
rename the "exceptional" destination of an invoke instruction to the 'unwind' dest
...
llvm-svn: 11202
2004-02-08 21:44:31 +00:00
bb5f0db8d6
Adjust to the new BB ctor
...
llvm-svn: 11113
2004-02-04 03:57:50 +00:00
9cda4bf8e5
When splitting a basic block, insert the new half immediately after the first
...
half.
llvm-svn: 11110
2004-02-04 03:21:31 +00:00
d8a232b716
If changing a parent, don't add then remove the object from the leak detector
...
llvm-svn: 11106
2004-02-04 01:06:38 +00:00
c4c7ea5288
In BasicBlock::splitBasicBlock, just use islist::splice to move the instructions,
...
instead of a loop that is really inefficient with large basic blocks.
This speeds up the inliner pass on the testcase in PR209 from 13.8s to 2.24s
which still isn't exactly speedy, but is a lot better. :)
llvm-svn: 11105
2004-02-03 23:11:21 +00:00
bcade4b8e6
Floating point negates are -0.0 - X, not 0.0 - X
...
llvm-svn: 11084
2004-02-02 20:21:29 +00:00
ac80ea4970
Second half of a fix for PR218 & test/Regression/Assembler/2004-02-01-NegativeZero.llx.
...
Basically we store floating point values as their integral components, instead of relying
on the semantics of floating point < to differentiate between values. This is likely to
make the map search be faster anyway.
llvm-svn: 11064
2004-02-01 22:49:04 +00:00
125ed54fdc
Fix a crasher bug in my constant folding rewrite
...
llvm-svn: 11044
2004-02-01 01:23:19 +00:00
241ed4c500
Fix a problem brian ran into with the bytecode reader asserting. It turns
...
out that the problem was actually the writer writing out a 'null' value
because it didn't normalize it. This fixes:
test/Regression/Assembler/2004-01-22-FloatNormalization.ll
llvm-svn: 10967
2004-01-23 00:55:21 +00:00
2cbec5de52
Let subclasses implement ModuleProvider::materializeModule() which is based on
...
their implementation of book-keeping for which functions need to be materialized
and which don't.
llvm-svn: 10943
2004-01-21 22:54:50 +00:00
c70b3f63cc
SlotCalculator.h moved
...
llvm-svn: 10931
2004-01-20 19:50:34 +00:00
083c99e347
Major changes. Now we only compactify individual type planes if it is in
...
fact "profitable" to do so. This makes compactification "free" for small
programs (ie, it is completely disabled) and even helps large programs by
not having to encode pointless compactification planes.
On 176.gcc, this saves 50K from the bytecode file, which is, alas only
a couple percent.
This concludes my head bashing against the bytecode format, at least for
now.
llvm-svn: 10922
2004-01-20 00:57:32 +00:00
ca1f8afe85
Add support for building the compactiontable for bytecode files. This shrinks
...
the bytecode file for 176.gcc by about 200K (10%), and 254.gap by about 167K,
a 25% reduction. There is still a lot of room for improvement in the encoding
of the compaction table.
llvm-svn: 10913
2004-01-18 21:07:07 +00:00
f78819d9ad
Eliminate special case handling for CPR's
...
Fix some problem cases where I was building the slot calculator in bytecode
writer mode instead of asmwriter mode.
llvm-svn: 10911
2004-01-18 21:03:06 +00:00
677af4a1cb
Bytecode format for LLVM 1.2 no longer explicitly encodes zeros in primitive
...
type planes. This saves about 5k on 176.gcc, and is needed for a subsequent
patch of mine I'm working on.
llvm-svn: 10908
2004-01-17 23:25:43 +00:00
2d0e6f6326
Cleanups & efficiency improvements
...
llvm-svn: 10888
2004-01-15 20:24:09 +00:00
21699e8f55
The bcwriter does not want ConstantPointerRef's to be indexed, and the asmwriter never did!
...
llvm-svn: 10885
2004-01-15 18:47:15 +00:00
de36304453
When emitting bytecode, handle strings specially. Do not add the characters
...
that make up the strings to the slotcalculator.
llvm-svn: 10862
2004-01-14 23:34:39 +00:00
2d6902b4b6
Hrm, apparently I missed lowering this intrinsic. :(
...
llvm-svn: 10858
2004-01-14 20:41:29 +00:00
e8701f690f
Fix bug in previous checkin: Demorgan would be ashamed of me, I need to
...
lay off the crack.
llvm-svn: 10855
2004-01-14 17:51:53 +00:00
e8dfccaabd
Implement ConstantArray::isString
...
llvm-svn: 10851
2004-01-14 17:06:38 +00:00
02062823f7
Fix check. PHI nodes must be handled specially, of course.
...
llvm-svn: 10842
2004-01-14 05:42:52 +00:00
ec5089a223
Tighten up verifier checks. The result of an invoke instruction only
...
dominates the normal destination, not the exceptional dest (ie, the result
of a call is undefined on an exception)
llvm-svn: 10841
2004-01-14 04:25:59 +00:00
6089bb1f19
The only clients of the slot calculator are now the asmwriter and bcwriter.
...
Since this really only makes sense for these two, change hte instance variable
to reflect whether we are writing a bytecode file or not. This makes it
reasonable to add bcwriter specific stuff to it as necessary.
llvm-svn: 10837
2004-01-14 02:49:34 +00:00
061da2f059
Implement a bunch of symbolic constant folding opportunities. This implements
...
testcase test/Regression/Assembler/ConstantExprFold.llx
Note that these kinds of things only rarely show up in source code, but are
exceedingly common in the intermediate stages of algorithms like SCCP. By
folding things (especially relational operators) that use symbolic constants,
we are able to speculatively fold more conditional branches, which can
lead to some big simplifications.
It would be easy to add a lot more special cases here, so if you notice
SCCP missing anything "obvious", you know what to make smarter. :)
llvm-svn: 10812
2004-01-13 05:51:55 +00:00
dab8b6bb50
Implement new Instruction::isRelational method
...
llvm-svn: 10810
2004-01-12 23:18:25 +00:00
1dd054c41b
Rearrange and comment code better. No functionality changes
...
llvm-svn: 10808
2004-01-12 22:07:24 +00:00
5a945e3616
Rename ConstantHandling.* -> ConstantFolding.*
...
Move a bunch of (now) private stuff from ConstantFolding.h into
ConstantFolding.cpp.
This _finally_ gets us to a place where we have a sane constant folder. The
rules are:
1. LLVM clients now use ConstantExpr::get* methods to fold constants. If they
cannot be folded, a constantexpr is created, so these methods always return
valid Constant*'s.
2. The implementation of ConstantExpr::get* uses the functions exposed by
ConstantFolding.h to try to fold constants. If they cannot be folded,
they should return a null pointer.
3. The implementation of ConstantFolding can do whatever it wants, and only
has one client (Constants.cpp)
This cuts down on the wierd dependencies, and eliminates the two interfaces.
The old constanthandling interface was especially bad for clients to use
because almost none of them took the failure condition into consideration,
thus leading to obscure problems.
llvm-svn: 10807
2004-01-12 21:13:12 +00:00
6ff6ceaf75
Remove a whole bunch more ugliness. This is actually getting to the point of
...
this whole refactoring: allow constant folding methods to return something
other than predefined classes, allow them to return generic Constant*'s.
llvm-svn: 10806
2004-01-12 21:02:29 +00:00
2361bd02a1
Move ConstantHandling.h into lib/VMCore and out of include/llvm
...
llvm-svn: 10805
2004-01-12 20:48:11 +00:00
e4867d75dd
Move this file to lib/VMCore
...
llvm-svn: 10804
2004-01-12 20:47:29 +00:00
f8348c3549
Eliminate usage of the wierd overloaded operator constant folders
...
llvm-svn: 10802
2004-01-12 20:41:05 +00:00
0bba771f49
Eliminate ConstantFoldShiftInstruction reference
...
llvm-svn: 10801
2004-01-12 20:40:42 +00:00
9eb2b527ad
Fix out of date comment, remove use of ConstantExpr::getShift
...
llvm-svn: 10794
2004-01-12 19:12:58 +00:00
5645e8a82d
Make ConstantExpr::get work for shifts as well
...
llvm-svn: 10790
2004-01-12 19:04:55 +00:00
0df14b3f2b
Move llvm::ConstantFoldInstruction from here to Transforms/Utils
...
llvm-svn: 10784
2004-01-12 18:25:03 +00:00
aacf2a5833
Implement: Assembler/2004-01-11-getelementptrfolding.llx
...
llvm-svn: 10759
2004-01-11 23:56:33 +00:00
d248950a8f
Urg, remove testing code.
...
llvm-svn: 10757
2004-01-11 23:30:03 +00:00
0180590bea
Fix a regression that I introduced yesterday. :(
...
llvm-svn: 10756
2004-01-11 23:29:26 +00:00
65f901e23b
Check in two changes:
...
The first change (which is disabled) compactifies all of the function constant
pools into the global constant pool, in an attempt to reduce the amount of
duplication and overhead. Unfortunately, as the comment indicates, this is
not yet a win, so it is disabled.
The second change sorts the typeid's so that those types that can be used
by instructions in the program appear earlier in the table than those that
cannot (such as structures and arrays). This causes the instructions to
be able to use the dense encoding more often, saving about 5K on 254.gap.
This is only a .65% savings though, unfortunately. :(
llvm-svn: 10754
2004-01-10 23:46:13 +00:00
570c69ce01
Hrm, another minor cleanup, which I missed before
...
llvm-svn: 10753
2004-01-10 21:42:24 +00:00
03c71fa8c5
Minor cleanup
...
llvm-svn: 10752
2004-01-10 21:40:29 +00:00
59f1ef40b0
Add support for new intrinsic
...
llvm-svn: 10701
2004-01-06 05:33:02 +00:00
3d903f0c0b
Add VMCore and code generator support for debugging intrinsics. By default
...
code generators completely ignore them.
llvm-svn: 10691
2004-01-05 05:36:30 +00:00
09bd1a06c1
Add some comments, add new getGlobalVariable method
...
llvm-svn: 10671
2003-12-31 08:43:01 +00:00
be3596c8ca
* Make Module::getTypeName const
...
* Add new Module::getTypeByName method
* Group methods in Module.cpp better
llvm-svn: 10668
2003-12-31 07:09:33 +00:00
efeb39e3c0
Make the lookup method const.
...
llvm-svn: 10667
2003-12-31 07:08:19 +00:00
eb537b2401
Fix PR198
...
llvm-svn: 10659
2003-12-31 03:19:37 +00:00
af3f8f3b42
Add comments
...
llvm-svn: 10658
2003-12-31 02:50:02 +00:00
5d236005b0
Clean up a lot of the code I added yesterday by exposing the IntrinsicLowering
...
implementation from the TargetMachine directly.
llvm-svn: 10636
2003-12-28 21:23:38 +00:00
24650d2b5e
Move into the VMCore library
...
llvm-svn: 10623
2003-12-28 08:30:20 +00:00
d1ea9cd345
The recalclulate method was a nasty hack that was once used by the -cee pass,
...
which never worked itself. The cee pass still doesn't work, but it doesn't use
this method anymore anyway, so eliminate the method.
llvm-svn: 10302
2003-12-07 00:55:32 +00:00
00f5167693
Completely rewrite domset, idom, and domtree implementation. Now it is based
...
on the algorithm for directly computing immediate dominators presented in this
paper:
A Fast Algorithm for Finding Dominators in a Flowgraph
T. Lengauer & R. Tarjan, ACM TOPLAS July 1979, pgs 121-141.
This _substantially_ speeds up construction of all dominator related information.
Post-dominators to follow.
llvm-svn: 10301
2003-12-07 00:38:08 +00:00
f5569bb9a5
Fix const correctness
...
llvm-svn: 10229
2003-11-25 21:28:00 +00:00
ad70d4a15b
Relax constrains on GEP type indexes
...
llvm-svn: 10228
2003-11-25 21:21:46 +00:00
3ae303cc8c
Check return types of functions
...
llvm-svn: 10146
2003-11-21 22:32:23 +00:00
dcdc371522
Fix problem Reid was having with external globals and nonexternal linkage
...
llvm-svn: 10133
2003-11-21 20:33:27 +00:00
189d19fb04
Finegrainify namespacification
...
llvm-svn: 10131
2003-11-21 20:23:48 +00:00
7e5e456d96
Fix some problems with assertions printing
...
llvm-svn: 10129
2003-11-21 17:35:51 +00:00
9d72c2f229
improve error message
...
llvm-svn: 10128
2003-11-21 17:06:29 +00:00
a29600046d
Minor cleanups and simplifications
...
llvm-svn: 10127
2003-11-21 16:52:05 +00:00
2af517281d
Start using the nicer terminator auto-insertion API
...
llvm-svn: 10111
2003-11-20 18:25:24 +00:00
9ab47d1625
Somehow I forgot poor little UnwindInst
...
*sniff*
llvm-svn: 10108
2003-11-20 18:11:56 +00:00
0e03ab67c3
* Finegrainify namespacification
...
* Add new constructors to allow insertion of terminator instructions at the
end of basic blocks.
* Move a ReturnInst method out-of-line, so that the vtable and type info don't
need to be emitted to every translation unit that uses the class.
llvm-svn: 10107
2003-11-20 17:45:12 +00:00
2447ef59c4
Don't crash if we are printing an orphaned basic block!
...
llvm-svn: 10100
2003-11-20 00:09:43 +00:00
45cfe54683
Really, there was a reason for me factoring that out
...
llvm-svn: 10089
2003-11-19 19:20:20 +00:00
1a02ca932c
*** Significantly speed up type resultion
...
This change speeds up type resolution by checking to see if a type is
recursive, and if it's not, using a more efficient algorithm.
This dramatically reduces bytecode loading time of kc++, reducing time-to-jit
kc++ --version to 17s from 33s
llvm-svn: 10088
2003-11-19 19:10:23 +00:00
a11f78167b
Finegrainify namespacification
...
llvm-svn: 10080
2003-11-19 06:14:38 +00:00
74c55dd4f9
Fix PR130, and testcase test/Regression/Linker/2003-11-18-TypeResolution.ll
...
llvm-svn: 10075
2003-11-19 01:02:52 +00:00
dc2e39191a
Implement == and != correctly. Before they would incorrectly return !=
...
for some constant exprs when they could really be the same value
llvm-svn: 10058
2003-11-17 20:19:35 +00:00
e120a73166
The ConstantPointer class is now gone.
...
llvm-svn: 10057
2003-11-17 19:47:21 +00:00
77f20dc13f
No REALLY, get rid of ConstantPointer references
...
llvm-svn: 10054
2003-11-17 19:21:04 +00:00
4b6addf98a
Eliminate use of the ConstantPointer class
...
llvm-svn: 10053
2003-11-17 19:19:32 +00:00
9d9cbcf246
Constant folding shalt not be built on annotations
...
llvm-svn: 10052
2003-11-17 19:05:17 +00:00
c96e96b26d
Be a bit more tolerant of broken code
...
llvm-svn: 10050
2003-11-17 01:17:04 +00:00
fda1f1835b
When you hand WriteAsOperand a type, it now prints out its symbolic name.
...
llvm-svn: 10042
2003-11-16 23:08:27 +00:00
9f47927b28
When failing a Verifier assertion, use WriteAsOperand to print out the
...
offending types and/or values.
llvm-svn: 10041
2003-11-16 23:07:42 +00:00
00211f1d80
Don't print 'No predecessors!' on the entry block
...
llvm-svn: 10038
2003-11-16 22:59:57 +00:00
ef901291c6
Expose the verifier to opt
...
llvm-svn: 9984
2003-11-13 19:47:29 +00:00
3b93c91836
Detect problems with PHI node operands!
...
llvm-svn: 9916
2003-11-12 07:13:37 +00:00
960707c335
Put all LLVM code into the llvm namespace, as per bug 109.
...
llvm-svn: 9903
2003-11-11 22:41:34 +00:00
a570f7b681
Fix PR95. I'm checking this patch in for Reid Spencer, who figured it out
...
and wrote it up. Thanks!!
llvm-svn: 9832
2003-11-09 19:39:48 +00:00
da8571b464
Add an assertion
...
llvm-svn: 9831
2003-11-09 15:51:07 +00:00
b36e8a1a2e
Add support for constant expr shifts
...
llvm-svn: 9735
2003-11-05 19:53:03 +00:00
e225507193
If we're not checking, don't check!
...
llvm-svn: 9732
2003-11-05 19:09:40 +00:00
41baa98478
Make code gcc 3.4 clean
...
llvm-svn: 9719
2003-11-05 05:15:42 +00:00
7c18b18c46
Fix InstCombine/2003-11-03-VarargsCallBug.ll
...
llvm-svn: 9681
2003-11-03 18:44:58 +00:00
9157ec025b
Did I mention that I _HATE_ CPRs?
...
llvm-svn: 9639
2003-10-31 18:38:06 +00:00
f6925da295
Constant pointer refs are causing these to fail unnecessarily, which is causing
...
a lot of code to be pessimized. I hate CPRs. :(
llvm-svn: 9635
2003-10-31 17:51:16 +00:00
6029d4032c
Use already existing classes for post order iterator.
...
llvm-svn: 9629
2003-10-31 03:02:34 +00:00
8339f7d379
Make use of the new AssemblyAnnotationWriter interface
...
llvm-svn: 9619
2003-10-30 23:41:03 +00:00
5a320efd3b
Output types in reverse postorder. This will allow the ByteCode/Reader
...
to create the minimum number of opaque types for each type with a
cycle in its type graph.
llvm-svn: 9615
2003-10-30 21:04:44 +00:00
f14ead9210
Print the names of more opaque types
...
llvm-svn: 9598
2003-10-30 00:22:33 +00:00
92d60539c8
Make sure to print opaque types names if they are available.
...
llvm-svn: 9597
2003-10-30 00:12:51 +00:00
d478d3d19c
Add more debugging info to help tracing the SlotCalculator.
...
llvm-svn: 9575
2003-10-29 03:12:12 +00:00
8a55262c76
Fix bug: instcombine/2003-10-23-InstcombineNullFail.ll
...
llvm-svn: 9403
2003-10-23 05:21:48 +00:00
2c35b1c94c
Make sure to add a type for va_next operands, which may not other wise
...
occur in the module
llvm-svn: 9336
2003-10-21 17:40:54 +00:00
7267b352fb
Fix problem which effected linking apache
...
llvm-svn: 9335
2003-10-21 17:39:59 +00:00
29265fe981
Added LLVM copyright header.
...
llvm-svn: 9321
2003-10-21 15:17:13 +00:00
4436c49787
Added LLVM copyright notice to Makefiles.
...
llvm-svn: 9312
2003-10-20 22:26:57 +00:00
482202a601
Added LLVM project notice to the top of every C++ source file.
...
Header files will be on the way.
llvm-svn: 9298
2003-10-20 19:43:21 +00:00
d77dd78c52
"Function"[0] == 'F', not 'M'
...
llvm-svn: 9297
2003-10-20 19:10:06 +00:00
b94550e537
Change the Opcode enum for PHI nodes from "Instruction::PHINode" to "Instruction::PHI" to be more consistent with the other instructions.
...
llvm-svn: 9269
2003-10-19 21:34:28 +00:00
5b33748497
Add support for the new varargs intrinsics and instructions
...
llvm-svn: 9226
2003-10-18 05:57:43 +00:00
e5a2565b91
Chris says it's better to assert that TheModule is valid than silently return 0.
...
llvm-svn: 9212
2003-10-17 21:59:57 +00:00
de360f88c8
Return the Module that we just materialized.
...
llvm-svn: 9201
2003-10-17 18:27:13 +00:00
8faf8d93d5
Eliminate some extraneous code in SlotCalculator::insertVal().
...
Rename SlotCalculator::getValSlot() to SlotCalculator::getSlot(),
SlotCalculator::insertValue() to SlotCalculator::getOrCreateSlot(),
SlotCalculator::insertVal() to SlotCalculator::insertValue(), and
SlotCalculator::doInsertVal() to SlotCalculator::doInsertValue().
llvm-svn: 9190
2003-10-17 02:02:40 +00:00
068ad84038
Add support for 'weak' linkage.
...
llvm-svn: 9171
2003-10-16 18:29:00 +00:00
079edeb919
Completely rewrite support for the Value::use_* list. Now, all operations on
...
this list (except use_size()) are constant time. Before the killUse method
(used whenever something stopped using a value) was linear time, and thus
very very slow for large programs.
This speeds GCCAS up _substantially_ on large programs: almost 2x for 176.gcc:
176.gcc: 77.07s -> 37.38s
177.mesa: 7.59s -> 5.57s
252.eon: 21.02s -> 19.52s (*)
253.perlbmk: 11.40s -> 13.05s
254.gap: 7.25s -> 7.42s
252.eon would speed up a whole lot more, but optimization time is being
dominated by the inlining pass, which needs to be fixed.
llvm-svn: 9160
2003-10-16 16:53:07 +00:00
56a8642a42
Enabling incremental bytecode loading in the JIT:
...
* FunctionPassManager ctor now takes in a ModuleProvider
* run() materializes function before running passes on it
llvm-svn: 9126
2003-10-14 21:38:42 +00:00
0c068c097a
Stop using "reverse depth first" order
...
llvm-svn: 9081
2003-10-13 15:21:58 +00:00
e859403772
Speed up TypesEqual by specializing it for all of the derived types, avoiding
...
a lot of virtual method dispatch overhead.
llvm-svn: 9078
2003-10-13 14:55:56 +00:00
1dc3f3fdc0
Avoid creating lots of pointless opaque types, with short lifetimes
...
llvm-svn: 9076
2003-10-13 14:03:36 +00:00
c7bc3bb503
This is a header file, hear me roar
...
llvm-svn: 9075
2003-10-13 05:34:24 +00:00
ab16a659d7
* Remove extraneous #includes
...
* Fix a nasty initializer ordering bug. Any only-CFG passes which registered
themselves before the CFGOnlyAnalysis vector initialized got forgotten and
thus got invalidated and recomputed.
In particular, in my compiled version of gccas, the Loop information pass was
being recomputed unnecessarily.
llvm-svn: 9074
2003-10-13 05:33:01 +00:00
44d2c3514a
Regularize header file comments
...
llvm-svn: 9071
2003-10-13 03:32:08 +00:00
dd99f5bf8f
Rename method
...
llvm-svn: 9055
2003-10-12 18:52:12 +00:00
cc31fddf13
Don't include Config/stdio.h or <stdio.h>.
...
llvm-svn: 9031
2003-10-10 18:46:29 +00:00
fa100531e1
Fix spelling/grammar.
...
llvm-svn: 9023
2003-10-10 17:54:14 +00:00
815ae2bfe3
Add a sanity check for constant expression casts
...
llvm-svn: 8951
2003-10-07 22:19:19 +00:00
794a7490f8
Users can never be null
...
llvm-svn: 8895
2003-10-06 17:33:39 +00:00
df9779c5c2
Be more careful handling PHI nodes, which might be of potentially high degree.
...
This reduces the time to verify a function from eon with a large number of
large PHI nodes from 22996s (6.38 hours) to 10.5499s
llvm-svn: 8866
2003-10-05 17:44:18 +00:00
b50d135710
Type tables are now AbstractTypeUsers. This allows them to merge together
...
constants as necessary due to type resolution. With this change, the
following spec benchmarks now link: 176.gcc, 177.mesa, 252.eon,
253.perlbmk, & 300.twolf. IOW, all SPEC INT and FP benchmarks now link.
llvm-svn: 8853
2003-10-05 00:17:43 +00:00
cc593fb6c6
Rename AbstractModuleProvider -> ModuleProvider, to match the header file name,
...
and because, while the class used by the interface is abstract, the actual
concept is not.
llvm-svn: 8850
2003-10-04 20:14:59 +00:00
3e88ef9f93
Minor random speedup: make Constant::getNullValue only call ConstantFOO::get
...
once!
llvm-svn: 8845
2003-10-03 19:34:51 +00:00
2c6cecf321
Explicit copy ctors are no longer needed
...
llvm-svn: 8844
2003-10-03 19:02:41 +00:00
7f4cfda3ba
Minor cleanups
...
llvm-svn: 8843
2003-10-03 18:57:54 +00:00
0f21ab75fa
This checkin basically amounts to a complete rewrite of the type-resolution
...
machinery. This dramatically simplifies how things works, removes irritating
little corner cases, and overall improves speed and reliability.
Highlights of this change are:
1. The exponential algorithm built into the code is now gone. For example
the time to disassemble one bytecode file from the mesa benchmark went
from taking 12.5s to taking 0.16s.
2. The linker bugs should be dramatically reduced. The one remaining bug
has to do with constant handling, which I actually introduced in
"union-find" checkins.
3. The code is much easier to follow, as a result of fewer special cases.
It's probably also smaller. yaay.
llvm-svn: 8842
2003-10-03 18:46:24 +00:00
29dc65a324
These methods are dead, remove them. Because the bodies will soon be
...
ressurected, just ifdef them out for now.
llvm-svn: 8840
2003-10-03 18:39:57 +00:00
a72b78967f
Dramatically simplify DerivedType::refineAbstractTypeToInternal
...
This makes use of the new PATypeHolder's to keep types from being deleted
prematurely, instead of the wierd "self reference" garbage. This is easier
to understand and more efficient as well.
llvm-svn: 8834
2003-10-03 04:48:21 +00:00
016418b8c2
Make the PATypeHolder use a simple union-find implementation to handle
...
merging of types. This makes it MUCH more efficient than before, also
making things simpler.
llvm-svn: 8833
2003-10-02 23:35:57 +00:00
22e4ca8863
There is no reason for Value to be an AbstractTypeUser. This just makes things
...
significantly more complete. Instead, just make DerivedType's AbstractTypeUser's,
and make Value contain a PATypeHolder.
llvm-svn: 8828
2003-10-02 19:44:40 +00:00
64d88bc788
Do not use BasicBlock::succ_iterator
...
llvm-svn: 8700
2003-09-24 22:03:22 +00:00
cd1d0e31f5
* No longer initialize the module, a subclass is more suitable to do this
...
because it can add a module ID which we do not have at this time.
* Check to see if the module has been initialized when materializing it.
llvm-svn: 8674
2003-09-22 23:44:13 +00:00
5dac64f629
Rename Function::getEntryNode -> getEntryBlock
...
llvm-svn: 8625
2003-09-20 14:39:18 +00:00
3284ed725a
Reject "llvm." as a function name
...
llvm-svn: 8608
2003-09-19 19:31:41 +00:00
3d72faddeb
Minimal implementation of the abstract ModuleProvider interface.
...
llvm-svn: 8594
2003-09-18 16:17:06 +00:00
c1b1651210
Change the semancics of the dropallrefs method
...
llvm-svn: 8572
2003-09-17 04:58:59 +00:00
bb9d03b9c0
Renamed DominatorTree::Node::getNode() -> getBlock()
...
llvm-svn: 8469
2003-09-11 16:26:13 +00:00
630d2954be
Rework dominator interfaces to handle changes in the post-dominance
...
construction. Now there may be multiple root blocks, and null is a
special node used to mark the "virtual" exit node of a CFG.
llvm-svn: 8461
2003-09-10 20:37:51 +00:00
ff713ae819
Make createVerifierPass return a FunctionPass *.
...
llvm-svn: 8449
2003-09-10 19:37:04 +00:00
216c7b8aed
Fix spell-o's
...
llvm-svn: 8431
2003-09-10 05:29:43 +00:00
f711d34e98
Reverting back to the previous revision.
...
We want to check for length 5 because we might get the "llvm." string as the
name. That string is in the LLVM namespace and should be checked as such.
We also don't have to worry about garbage data because (I believe) the string
class will return a valid value. So, the switch statement will work and we
don't have to worry about the code wandering into segfault land.
llvm-svn: 8419
2003-09-09 16:58:41 +00:00
8d575cb2e1
Modified the check for intrinsic functions. All intrinsic functions have to
...
be at least 6 characters, since something must follow the "llvm." string in the
function name.
This seems to fix an assertion failure with the SingleSource tests, too.
llvm-svn: 8418
2003-09-09 16:36:09 +00:00
04ecefe232
Eliminate support for the llvm.unwind intrinisic, using the Unwind instruction instead
...
llvm-svn: 8411
2003-09-08 19:44:26 +00:00
66d5f575ba
Add support for unwind
...
llvm-svn: 8407
2003-09-08 18:54:36 +00:00
504f9242c3
Add support for volatile loads/stores
...
llvm-svn: 8393
2003-09-08 17:45:59 +00:00
730e047678
Spell `occurrence' correctly.
...
llvm-svn: 8388
2003-09-08 03:08:43 +00:00
20b98d2c93
Add asserts, move code around. This gets the dropAllTypeUses partially implemented
...
llvm-svn: 8361
2003-09-05 05:10:04 +00:00
d15598bfe0
Move functionality preserving changes.
...
Inline containsEquivalent
Remove some //'s
llvm-svn: 8358
2003-09-05 02:39:52 +00:00
5ac07d2221
Refactor what has effectively become copy-and-pasted code
...
llvm-svn: 8357
2003-09-05 02:30:47 +00:00
3f695ff02e
Check in a bunch of minor fixes, plus a whole lot of #if 0'd out code, which will hopefully be enabled in the near future
...
This does not make any functionality changes
llvm-svn: 8355
2003-09-05 02:21:39 +00:00
e4fdce4235
containsEquivalent never returns its argument
...
llvm-svn: 8353
2003-09-04 23:49:53 +00:00
cd27235b23
Remove needless usage of getDescription()
...
llvm-svn: 8351
2003-09-04 23:46:03 +00:00
59be7695ef
Inline method
...
llvm-svn: 8350
2003-09-04 23:43:40 +00:00
b8b10b5c54
Don't try to be sneaky, breaking recursive types in the process
...
llvm-svn: 8348
2003-09-04 23:41:03 +00:00
50745187ef
This is now unnecessary
...
llvm-svn: 8347
2003-09-04 23:38:22 +00:00
13651f0c46
Fix bug where we couldn't print a function without a name
...
llvm-svn: 8341
2003-09-03 17:56:43 +00:00
b89d2afab1
No need to rescan types when they are created.
...
llvm-svn: 8339
2003-09-03 14:44:53 +00:00
cc61a136f7
Another small speedup, this one to: 2.42s
...
llvm-svn: 8335
2003-09-02 22:52:49 +00:00
188afa5feb
Don't return bogus references, and don't add an entry to the Concrete map
...
unless it's not empty!
llvm-svn: 8334
2003-09-02 22:50:02 +00:00
39e5095c8e
Another optimization, speed up the testcase to 2.7s
...
llvm-svn: 8333
2003-09-02 22:15:15 +00:00
835499277b
Removal of explicit stack, which requires the method to be a member (so it can
...
call setAbstract). Now that we just compute abstractness we can also return
the computed value by value instead of as an argument.
llvm-svn: 8332
2003-09-02 21:56:34 +00:00
9a61bbf1d3
Remove the "recursive bit", not only is it unused by anyone, it was also
...
not correctly calculated, and calculating it wrong for fun seems rather
pointless. This also speeds up my favorite testcase by .25 seconds.
llvm-svn: 8330
2003-09-02 21:41:05 +00:00
ce35f14050
Don't bother doing an exhaustive recursive walk if we are at the limit of what
...
we need to know anyway. This reduces the 2002-07-08-HugePerformanceProblem.llx
down to 3.210u:0.010s, which is back in the acceptable range again
llvm-svn: 8323
2003-09-02 20:06:29 +00:00
99b9ddead7
Ugh, fix bugs. Ok, so the last fix wasn't as great as I thought it was.
...
Now we're back to: 131.730u, 0.330s, which is still quite an improvement,
but still quite unacceptable
llvm-svn: 8322
2003-09-02 19:14:12 +00:00
2a36482e74
Further simplifications
...
llvm-svn: 8321
2003-09-02 16:46:41 +00:00
e0703ff161
Change the code to no longer compute the "type description" immediately when
...
the type is analyzed. Instead, only compute it when requested (with
getDescription), and cached for reuse later.
This dramatically speeds up LLVM in general because these descriptions almost
_never_ need to be constructed. The only time they are used is when a type is
<<'d. Printing of modules by themselves uses other code to print symbolic
types when possible, so these descriptions are really only used for debugging.
Also, this fixes the particularly bad case when lots of types get resolved to
each other, such as during linking of large programs. In these cases, the type
descriptions would be repeatedly recomputed and discarded even though: A. noone
reads the description before it gets resolved, and B. many many resolutions
happen at intermediate steps, causing a HUGE waste of time.
Overall, this makes the getTypeDesc function much more light-weight, and fixes
bug: Assembler/2002-07-08-HugePerformanceProblem.llx, which went from taking
1048.770u/19.150s (which is 17.5 MINUTES, on apoc), to taking 0.020u/0.000s,
which is a nice little speedup. :)
llvm-svn: 8320
2003-09-02 16:35:17 +00:00
e248195b22
Remove assertion which is never reached.
...
llvm-svn: 8318
2003-09-02 06:45:34 +00:00
bd717d897f
Implement new method
...
llvm-svn: 8238
2003-08-31 00:19:28 +00:00
a38adaa88f
Fix bug: Linker/2003-08-28-TypeResolvesGlobal3.ll
...
Use new replaceAllUsesWith stuff in a way that works even though types have not yet been propagated
llvm-svn: 8210
2003-08-29 05:37:22 +00:00
c27038d7b5
Implement "unsafe" replaceAllUsesWWith stuff for use during type resolution.
...
llvm-svn: 8209
2003-08-29 05:36:46 +00:00
6097d94295
Fix bug: Linker/2003-08-28-TypeResolvesGlobal2.ll
...
llvm-svn: 8206
2003-08-29 05:11:34 +00:00
9f15812e88
Refactor code to make it useful outside of Constants.cpp
...
llvm-svn: 8205
2003-08-29 05:09:37 +00:00
ae77242011
Fix bug: Linker/2003-08-28-TypeResolvesGlobal.ll
...
llvm-svn: 8202
2003-08-29 04:49:54 +00:00
ea9891311a
As it turns out, things will be simpler than I first expected. We no longer
...
need any exception handling intrinsics beyond llvm.unwind. (yaay)
llvm-svn: 8145
2003-08-25 22:35:01 +00:00
8068e0cd17
Support new 'any' support for pointer size and endianness
...
llvm-svn: 8119
2003-08-24 13:48:48 +00:00
7da3571af9
rethrow is really the language independent primitive here. "throw" can be written
...
in terms of it and llvm.exc.setcurrent.
Rework the intrinsics.
llvm-svn: 8110
2003-08-24 12:24:08 +00:00
ca3e569269
Fix copy-and-paste-o
...
llvm-svn: 8104
2003-08-24 06:57:53 +00:00
ade9410c99
Initial support for recognizing LLVM exception handling intrinsics
...
llvm-svn: 8102
2003-08-24 05:30:29 +00:00
6d5693566f
Allow specifying the name for the newly split basic block
...
llvm-svn: 8097
2003-08-24 03:41:39 +00:00
180e14cb87
Rename SwitchInst::dest_push_back -> addCase
...
Add new removeCase method
llvm-svn: 8088
2003-08-23 23:14:37 +00:00
1c3430491e
If an "LLVM name" has wierd characters in it, print it out in double quotes instead of prefixing it with %
...
llvm-svn: 8049
2003-08-22 05:40:38 +00:00
be372b91e1
The word `dependent' has no `a'.
...
llvm-svn: 8030
2003-08-21 22:14:26 +00:00
941b06bf3c
Implement Assembler/2003-08-21-ConstantExprCast-Fold.llx
...
llvm-svn: 8015
2003-08-21 19:45:55 +00:00
ddab239262
Implement testcase: Assembler/2003-08-20-ConstantExprGEP-Fold.llx
...
llvm-svn: 7983
2003-08-20 16:11:27 +00:00
f8a77578a2
Do not even attempt to compute dominator information for unreachable blocks
...
llvm-svn: 7963
2003-08-18 22:11:16 +00:00
ab69ecc6d5
Fix linking bug in xboard
...
llvm-svn: 7952
2003-08-18 16:54:48 +00:00
748e9e143b
Add intrinsics for the llvm.sig(set|long)jmp functions
...
llvm-svn: 7949
2003-08-18 15:41:24 +00:00
7eb05a170a
Spell `necessary' correctly.
...
llvm-svn: 7944
2003-08-18 14:43:39 +00:00
2cc4b9dd2c
Add new method to FunctionPassManager to add ImmutablePasses.
...
llvm-svn: 7838
2003-08-14 06:07:57 +00:00
e7e094467d
Dramatically simplify how -time-passes works.
...
This also enables -time-passes for FunctionPassManagers, which allows it to
be used for the JIT
llvm-svn: 7834
2003-08-14 05:20:28 +00:00
cea4d8c2f5
Implement constant folding of casts from boolean constants to other values
...
llvm-svn: 7812
2003-08-13 15:52:25 +00:00
Chris Lattner