RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-02 02:14:23 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
51,610 Commits 4 Branches 5 Tags
Commit Graph

31990 Commits

Author SHA1 Message Date
Dan Gohman
8f7b263087 Fix an abort on a store of an empty struct member. getValue returns 
null in the case of an empty struct, so don't try to call getNumValues
on it.

llvm-svn: 81180
 2009-09-08 01:44:02 +00:00
Nick Lewycky
5536fa68c5 Verify types. Invalid types can be constructed when assertions are off. 
Make the verifier more robust by avoiding unprotected cast<> calls. Notably,
Assert1(isa<>); cast<> is not safe as Assert1 does not terminate the program.

llvm-svn: 81179
 2009-09-08 01:23:52 +00:00
Chris Lattner
ee1d1d3173 Fix PR4882, by making MemCpyOpt not dereference removed stores to get the 
context for the newly created operations.
 
Patch by Jakub Staszak!

llvm-svn: 81175
 2009-09-08 00:27:14 +00:00
Chris Lattner
37dbbde91b fix ComputeMaskedBits handling of zext/sext/trunc to work with vectors. 
This fixes PR4905

llvm-svn: 81174
 2009-09-08 00:13:52 +00:00
Chris Lattner
1247f36e13 add some comments to describe the invariants. 
llvm-svn: 81173
 2009-09-08 00:06:16 +00:00
Dan Gohman
1147a7ba87 Reappy r80998, now that the GlobalOpt bug that it exposed on MiniSAT is fixed. 
llvm-svn: 81172
 2009-09-07 23:54:19 +00:00
Dan Gohman
572ecc26b6 Fix a thinko: When lowering fneg with xor, bitcast the operands 
from floating-point to integer first, and bitcast the result
back to floating-point. Previously, this test was passing by
falling back to SelectionDAG lowering. The resulting code isn't
as nice, but it's correct and CodeGen now stays on the fast path.

llvm-svn: 81171
 2009-09-07 23:47:14 +00:00
Dan Gohman
34f89141ae Don't commit stores with addresses that have indices that are not 
compile-time constant integers or that are out of bounds for their
corresponding static array types. These can cause aliasing that
GlobalOpt assumes won't happen.

llvm-svn: 81165
 2009-09-07 22:44:55 +00:00
Dan Gohman
b8c9d1974c Don't commit addresses of aggregate values. This avoids problems with 
an aggregate store overlapping a different aggregate store, despite
the stores having distinct addresses.

llvm-svn: 81164
 2009-09-07 22:42:05 +00:00
Dan Gohman
f1dd8d2a5b Fix GlobalOpt to avoid committing a store if the address getelementptr 
is missing the inbounds flag. This is slightly conservative, but it
avoids problems with two constants pointing to the same address but
getting distinct entries in the Memory DenseMap.

llvm-svn: 81163
 2009-09-07 22:40:13 +00:00
Dan Gohman
e44305a42b Preserve the InBounds flag when evaluating a getelementptr instruction 
into a getelementptr ConstantExpr.

llvm-svn: 81162
 2009-09-07 22:34:43 +00:00
Dan Gohman
9268f469f5 Simplify this code by using hasDefinitiveInitializer(). 
llvm-svn: 81161
 2009-09-07 22:31:26 +00:00
Nick Lewycky
1528b90211 Express this in the canonical way. 
llvm-svn: 81157
 2009-09-07 21:50:24 +00:00
Nick Lewycky
1bde81d01d Homogenize whitespace. 
llvm-svn: 81156
 2009-09-07 20:44:51 +00:00
Duncan Sands
5c77da0572 Using a signal handler that does nothing should be 
equivalent to SIG_IGN.

llvm-svn: 81144
 2009-09-07 05:58:25 +00:00
Daniel Dunbar
b258ff7cee Disable MDNode uniquing. 
- Hopefully this unbreaks some llvm-gcc bootstraps.

llvm-svn: 81141
 2009-09-07 04:05:49 +00:00
Eric Christopher
6acdc39b86 Fix comment. 
llvm-svn: 81138
 2009-09-06 22:20:54 +00:00
Anton Korobeynikov
6c6f65f3c1 Do not create calls via PLT in compilation callback - this is higly platform 
dependent. Hopefully, this will fix PR3801.

llvm-svn: 81132
 2009-09-06 20:21:48 +00:00
Duncan Sands
357614a0ca Do not try to override non-virtual methods, especially 
when the new method gives the same result as the original
(as far as I can see).  This will hopefully pacify icc.

llvm-svn: 81131
 2009-09-06 20:02:00 +00:00
Duncan Sands
6efbe6150b Remove a left over bit of code with no effect. 
llvm-svn: 81128
 2009-09-06 19:29:07 +00:00
Duncan Sands
9041cda7c7 Avoid warnings if assertions are off. 
llvm-svn: 81127
 2009-09-06 19:28:38 +00:00
Duncan Sands
265ebcdcfc Remove unreachable code. 
llvm-svn: 81126
 2009-09-06 19:27:53 +00:00
Duncan Sands
942d35c184 Simplify. Testing shows that this is not equivalent to BBI = CR.CaseBB + 1. 
llvm-svn: 81124
 2009-09-06 18:03:32 +00:00
Duncan Sands
39ca207077 Avoid an unused variable warning when assertions are 
disabled.

llvm-svn: 81122
 2009-09-06 16:27:34 +00:00
Duncan Sands
0d5fa5f018 Mark more constants unsigned, as warned about by icc (#68). 
Patch by Erick Tryzelaar.

llvm-svn: 81116
 2009-09-06 12:56:52 +00:00
Duncan Sands
598fe699d0 Remove some not-really-used variables, as warned 
about by icc (#593, partial).  Patch by Erick Tryzelaar.

llvm-svn: 81115
 2009-09-06 12:41:19 +00:00
Benjamin Kramer
f41c1c09fa More MSVC warning fixes: 
1. DUPMAX is defined in regcomp.c, no need to redefine it in regutils.
2. MSVC doesn't like snprintf, use _snprintf instead.

llvm-svn: 81114
 2009-09-06 12:26:28 +00:00
Duncan Sands
ab1e3dc9e2 Remove strange 'const' qualifiers, as warned about by icc 
(#411).  Patch by Erick Tryzelaar.

llvm-svn: 81113
 2009-09-06 12:16:26 +00:00
Benjamin Kramer
7155f79983 It's a bool, so treat it like one. Fixes a MSVC warning. 
llvm-svn: 81112
 2009-09-06 12:10:17 +00:00
Duncan Sands
4b92376ba8 Tweak code into an equivalent form for which icc 
doesn't warn about unreachable instructions.  Patch
by Erick Tryzelaar (#111).

llvm-svn: 81110
 2009-09-06 10:53:22 +00:00
Benjamin Kramer
ff8022c220 Remove splint hints to silence warnings from ICC and MSVC. 
llvm-svn: 81108
 2009-09-06 09:29:39 +00:00
Duncan Sands
750b1f5de2 Public and private corrections, warned about by icc (#304). 
Patch by Erick Tryzelaar.

llvm-svn: 81107
 2009-09-06 08:55:57 +00:00
Duncan Sands
10c1356bad Remove some unused variables and methods warned about by 
icc (#177, partial).  Patch by Erick Tryzelaar.

llvm-svn: 81106
 2009-09-06 08:33:48 +00:00
Daniel Dunbar
4d822b9da3 Fix a possible crash call setIsInBounds. 
- I think there are more instances of this, but I think they are fixed in Dan's
   incoming patch. This one was preventing me from doing a bugpoint reduction
   though.

llvm-svn: 81103
 2009-09-06 02:31:36 +00:00
Evan Cheng
a191202da5 Revert r80926. It causes loop unswitch assertion and slow down some JIT tests significantly. 
llvm-svn: 81101
 2009-09-06 02:26:10 +00:00
Daniel Dunbar
452643b58c Revert "Include optional subclass flags, such as inbounds, nsw, etc., ...", this 
breaks MiniSAT on x86_64.

llvm-svn: 81098
 2009-09-06 00:11:24 +00:00
Devang Patel
7b422c1956 Detect VLAs. 
Do not use DenseMap operator[] because it inserts new entry if lookup fails. Use find() to check an entry in a DenseMap first.

llvm-svn: 81058
 2009-09-05 00:34:14 +00:00
Devang Patel
c174e064b2 Ignore malformed global variable debug info. 
llvm-svn: 81055
 2009-09-04 23:59:07 +00:00
Devang Patel
40a0a72be8 While replacing an MDNode elment, properly update MDNode's operand list. 
MDNode's operand list does not include all elements.

llvm-svn: 81045
 2009-09-04 21:32:05 +00:00
Andreas Neustifter
bba38af145 Prevent warnings on compilers for which its not clear that assert won't return. 
llvm-svn: 81044
 2009-09-04 21:22:04 +00:00
Lang Hames
e504e61ab5 Replaces uses of unsigned for indexes in LiveInterval and VNInfo with 
a new class, MachineInstrIndex, which hides arithmetic details from
most clients. This is a step towards allowing the register allocator
to update/insert code during allocation.

llvm-svn: 81040
 2009-09-04 20:41:11 +00:00
Andreas Neustifter
a956e23520 Cleaned up ProfileVerifierPass. 
(See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20090831/086219.html)

llvm-svn: 81007
 2009-09-04 17:15:10 +00:00
Andreas Neustifter
548db47ff4 Converted MaximumSpanningTree algorithm to a generic template, this could go 
into llvm/ADT.

llvm-svn: 81001
 2009-09-04 12:34:44 +00:00
Dan Gohman
8b386a0b63 Include optional subclass flags, such as inbounds, nsw, etc., in the 
Constant uniquing tables. This allows distinct ConstantExpr objects
with the same operation and different flags.

Even though a ConstantExpr "a + b" is either always overflowing or
never overflowing (due to being a ConstantExpr), it's still necessary
to be able to represent it both with and without overflow flags at
the same time within the IR, because the safety of the flag may
depend on the context of the use. If the constant really does overflow,
it wouldn't ever be safe to use with the flag set, however the use
may be in code that is never actually executed.

This also makes it possible to merge all the flags tests into a single test.

llvm-svn: 80998
 2009-09-04 12:08:11 +00:00
Duncan Sands
2037ffaa73 Use delete[] to match new[] (found by valgrind). 
llvm-svn: 80997
 2009-09-04 11:59:43 +00:00
Evan Cheng
46e40befe7 Run branch folding if if-converter make some transformations. 
llvm-svn: 80994
 2009-09-04 07:47:40 +00:00
Evan Cheng
284b38aaa7 Fix comment for consistency sake. 
llvm-svn: 80993
 2009-09-04 07:46:30 +00:00
Jim Grosbach
5a21028ece Whitespace cleanup 
llvm-svn: 80978
 2009-09-04 01:38:51 +00:00
Eric Christopher
a1d56b83f9 If there's a calling convention attach it to the rewind function call. 
llvm-svn: 80976
 2009-09-04 01:14:14 +00:00
Evan Cheng
86e6ec85a3 Funky indentation. 
llvm-svn: 80971
 2009-09-03 23:54:22 +00:00