Commit Graph

88993 Commits

Author SHA1 Message Date
Andrew Trick
4e1fb18940 MIsched: Improve the interface to SchedDFS analysis (subtrees).
Allow the strategy to select SchedDFS. Allow the results of SchedDFS
to affect initialization of the scheduler state.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173425 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 06:33:57 +00:00
Andrew Trick
827de0520e ArrayRef reverse iterators.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173424 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 06:33:52 +00:00
Andrew Trick
2f0a4e1aad whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173423 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 06:33:50 +00:00
NAKAMURA Takumi
52d4de971f lit/ShUtil.py: Deprecate '!' in shell expression. It is not sh-compatible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173421 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 06:30:36 +00:00
Andrew Trick
bfb8223e2b SchedDFS: Initial support for nested subtrees.
This is mostly refactoring, along with adding an instruction count
within the subtrees and ensuring we only look at data edges.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173420 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 06:02:44 +00:00
Chandler Carruth
baf868b9b8 Switch this code away from Value::isUsedInBasicBlock. That code either
loops over instructions in the basic block or the use-def list of the
value, neither of which are really efficient when repeatedly querying
about values in the same basic block.

What's more, we already know that the CondBB is small, and so we can do
a much more efficient test by counting the uses in CondBB, and seeing if
those account for all of the uses.

Finally, we shouldn't blanket fail on any such instruction, instead we
should conservatively assume that those instructions are part of the
cost.

Note that this actually fixes a bug in the pass because
isUsedInBasicBlock has a really terrible bug in it. I'll fix that in my
next commit, but the fix for it would make this code suddenly take the
compile time hit I thought it already was taking, so I wanted to go
ahead and migrate this code to a faster & better pattern.

The bug in isUsedInBasicBlock was also causing other tests to test the
wrong thing entirely: for example we weren't actually disabling
speculation for floating point operations as intended (and tested), but
the test passed because we failed to speculate them due to the
isUsedInBasicBlock failure.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173417 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 05:40:09 +00:00
Andrew Trick
178f7d08a4 MISched: Add SchedDFSResult to ScheduleDAGMI to formalize the
interface and allow other strategies to select it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173413 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 04:01:04 +00:00
Jack Carter
801c583883 This patch implements parsing the .word
directive for the Mips assembler.

Contributer: Vladimir Medic
 


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173407 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 01:31:34 +00:00
Akira Hatanaka
d2047c6001 [mips] Set flag neverHasSideEffects flag on some of the floating point instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173401 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 00:20:39 +00:00
Andrew Trick
b12a771992 SchedDFS: Refactor and tweak the subtree selection criteria.
For sanity, create a root when NumDataSuccs >= 4. Splitting large
subtrees will no longer be detrimental after my next checkin to handle
nested tree. A magic number of 4 is fine because single subtrees
seldom rejoin more than this. It makes subtrees easier to visualize
and heuristics more sane.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173399 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 00:12:57 +00:00
Andrew Trick
6d28299b9d SchedDFS: Constify interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173398 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25 00:12:55 +00:00
Jakob Stoklund Olesen
dd4fc446b5 Avoid creating duplicate CFG edges in the IfConversion pass.
Patch by Stefan Hepp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173395 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 23:59:08 +00:00
Renato Golin
b3755e7fa2 Moving Cost Tables up to share with other targets
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173382 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 23:01:00 +00:00
Michael Gottesman
7c24e61a2b Added comment to ObjCARC elaborating what is meant by the term 'Provenance' in 'Provenance Analysis'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173374 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 21:35:00 +00:00
Hal Finkel
78e1057371 Start cleanup of PPC register definitions using foreach loops.
No functionality change intended.

This captures the first two cases GPR32/64. For the others, we need
an addition operator (if we have one, I've not yet found it).

Based on a suggestion made by Tom Stellard in the AArch64 review!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173366 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 20:43:18 +00:00
Saleem Abdulrasool
ed7fcf4fd9 [bugpoint] make tool selection messages unique
Change messages to help identify which interpreter was actually selected (safe
vs testing).

Signed-off-by: Saleem Abdulrasool <compnerd@compnerd.org>
Reviewed-by: Chandler Carruth <chandlerc@gmail.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173360 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 16:49:14 +00:00
Saleem Abdulrasool
78d3097263 [bugpoint] set Message after tool configuration
Set the message returned after the GCC runner has been constructed as otherwise
the message will be overwritten by the construction of the runner, resulting in
misleading messages.

Signed-off-by: Saleem Abdulrasool <compnerd@compnerd.org>
Reviewed-by: Chandler Carruth <chandlerc@gmail.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173359 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 16:49:12 +00:00
Benjamin Kramer
d5a80c7358 Reapply chandlerc's r173342 now that the miscompile it was triggering is fixed.
Original commit message:
Plug TTI into the speculation logic, giving it a real cost interface
that can be specialized by targets.

The goal here is not to be more aggressive, but to just be more accurate
with very obvious cases. There are instructions which are known to be
truly free and which were not being modeled as such in this code -- see
the regression test which is distilled from an inner loop of zlib.

Everywhere the TTI cost model is insufficiently conservative I've added
explicit checks with FIXME comments to go add proper modelling of these
cost factors.

If this causes regressions, the likely solution is to make TTI even more
conservative in its cost estimates, but test cases will help here.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173357 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 16:44:25 +00:00
Benjamin Kramer
e574246489 ConstantFolding: Add a missing folding that leads to a miscompile.
We use constant folding to see if an intrinsic evaluates to the same value as a
constant that we know. If we don't take the undefinedness into account we get a
value that doesn't match the actual implementation, and miscompiled code.

This was uncovered by Chandler's simplifycfg changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173356 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 16:28:28 +00:00
NAKAMURA Takumi
41d35a335e unittests/SupportTests/Initialize.MultipleThreads: Enable pthread_attr_setstack(3) only on Linux.
I got blamed on darwin11;
unittests/Support/ManagedStatic.cpp:35: error: 'pthread_attr_setstack' was not declared in this scope

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173355 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 15:29:27 +00:00
NAKAMURA Takumi
91e22b0dd3 unittests/SupportTests/Initialize.MultipleThreads: Appease --vg-leak to allocate stack explicitly for glibc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173350 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 14:44:02 +00:00
NAKAMURA Takumi
5bac013d61 lli/RecordingMemoryManager: Free allocated sections in the destructor to satisfy --vg-leak!
FIXME: It could be generalized in MemoryManager.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173349 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 14:12:12 +00:00
Chandler Carruth
eacef325c6 Revert r173342 temporarily. It appears to cause a very late miscompile
of stage2 in a bootstrap. Still investigating....

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173343 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 13:24:24 +00:00
Chandler Carruth
1f255419d4 Plug TTI into the speculation logic, giving it a real cost interface
that can be specialized by targets.

The goal here is not to be more aggressive, but to just be more accurate
with very obvious cases. There are instructions which are known to be
truly free and which were not being modeled as such in this code -- see
the regression test which is distilled from an inner loop of zlib.

Everywhere the TTI cost model is insufficiently conservative I've added
explicit checks with FIXME comments to go add proper modelling of these
cost factors.

If this causes regressions, the likely solution is to make TTI even more
conservative in its cost estimates, but test cases will help here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173342 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 12:39:29 +00:00
Chandler Carruth
47d8f6dca5 Address a large chunk of this FIXME by accumulating the cost for
unfolded constant expressions rather than checking each one
independently.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173341 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 12:05:17 +00:00
Chandler Carruth
681add7a63 Switch the constant expression speculation cost evaluation away from
a cost fuction that seems both a bit ad-hoc and also poorly suited to
evaluating constant expressions.

Notably, it is missing any support for trivial expressions such as
'inttoptr'. I could fix this routine, but it isn't clear to me all of
the constraints its other users are operating under.

The core protection that seems relevant here is avoiding the formation
of a select instruction wich a further chain of select operations in
a constant expression operand. Just explicitly encode that constraint.

Also, update the comments and organization here to make it clear where
this needs to go -- this should be driven off of real cost measurements
which take into account the number of constants expressions and the
depth of the constant expression tree.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173340 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 11:53:01 +00:00
Chandler Carruth
2c107a8020 Rephrase the speculating scan of the conditional BB to be phrased in
terms of cost rather than hoisting a single instruction.

This does *not* change the cost model! We still set the cost threshold
at 1 here, it's just that we track it by accumulating cost rather than
by storing an instruction.

The primary advantage is that we no longer leave no-op intrinsics in the
basic block. For example, this will now move both debug info intrinsics
and a single instruction, instead of only moving the instruction and
leaving a basic block with nothing bug debug info intrinsics in it, and
those intrinsics now no longer ordered correctly with the hoisted value.

Instead, we now splice the entire conditional basic block's instruction
sequence.

This also places the code for checking the safety of hoisting next to
the code computing the cost.

Currently, the only observable side-effect of this change is that debug
info intrinsics are no longer abandoned. I'm not sure how to craft
a test case for this, and my real goal was the refactoring, but I'll
talk to Dave or Eric about how to add a test case for this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173339 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 11:52:58 +00:00
Kostya Serebryany
29f975f8ff [asan] fix 32-bit builds
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173338 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 10:43:50 +00:00
Chandler Carruth
0afa33115c Simplify the PHI node operand rewriting.
Previously, the code would scan the PHI nodes and build up a small
setvector of candidate value pairs in phi nodes to go and rewrite. Once
certain the rewrite could be performed, the code walks the set, and for
each one re-scans the entire PHI node list looking for nodes to rewrite
operands.

Instead, scan the PHI nodes once to check for hazards, and then scan it
a second time to rewrite the operands to selects. No set vector, and
a max of two scans.

The only downside is that we might form identical selects, but
instcombine or anything else should fold those easily, and it seems
unlikely to happen often.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173337 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 10:40:51 +00:00
Kostya Serebryany
63f0846f1e [asan] adaptive redzones for globals (the larger the global the larger is the redzone)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173335 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 10:35:40 +00:00
Chandler Carruth
6cf7a91d12 Give the basic block variables here names based on the if-then-end
structure being analyzed. No functionality changed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173334 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 09:59:39 +00:00
Chandler Carruth
9e62095261 Lift a cheap early exit test above loops and other complex early exit
tests. No need to pay the high cost when we're never going to do
anything.

No functionality changed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173331 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 08:22:40 +00:00
Chandler Carruth
c61bc7a90c Spiff up the comment on this method, making the example a bit more
pretty in doxygen, adding some of the details actually present in
a classic example where this matters (a loop from gzip and many other
compression algorithms), and a cautionary note about the risks inherent
in the transform. This has come up on the mailing lists recently, and
I suspect folks reading this code could benefit from going and looking
at the MI pass that can really deal with these issues.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173329 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 08:05:06 +00:00
NAKAMURA Takumi
b3105b9a9b MipsISelLowering.cpp: Fill unreachable paths to fix warnings. [-Wsometimes-uninitialized]
FIXME: Could they, unreachable(s), be removed?
FIXME: I could prefer the coding standards...

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173325 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 06:08:06 +00:00
NAKAMURA Takumi
00cdf602ae MipsISelLowering.cpp: Fix a warning, take two. [-Wunused-variable]
...and fix a typo, s/#ifdef/#ifndef/

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173324 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 05:54:23 +00:00
NAKAMURA Takumi
d5a336cdb5 MipsISelLowering.cpp: Fix a warning. [-Wunused-variable]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173323 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 05:47:29 +00:00
Craig Topper
b57c292d29 Remove trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173322 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 05:22:40 +00:00
Richard Trieu
0ac7e6f293 Add asserts to SmallVector so that calls to front() and back() only succeed
if the vector is not empty.  This will ensure that calls to these functions
will reference elements in the vector.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173321 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 04:29:24 +00:00
Reed Kotler
8453b3f66a The next phase of Mips16 hard float implementation.
Allow Mips16 routines to call Mips32 routines that have abi requirements
that either arguments or return values are passed in floating point 
registers. This handles only the pic case. We have not done non pic
for Mips16 yet in any form.

The libm functions are Mips32, so with this addition we have a complete
Mips16 hard float implementation.

We still are not able to complete mix Mip16 and Mips32 with hard float.
That will be the next phase which will have several steps. For Mips32
to freely call Mips16 some stub functions must be created.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173320 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 04:24:02 +00:00
Andrew Trick
db4170697f MachineScheduler: enable biasCriticalPath for all DAGs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173318 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 02:09:57 +00:00
Andrew Trick
66658dd9a1 MIsched: Added biasCriticalPath.
Allow schedulers to order DAG edges by critical path. This makes
DFS-based heuristics more stable and effective.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173317 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 02:09:55 +00:00
Michael J. Spencer
e35badad22 [ELF] Add R_X86_64_IRELATIVE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173316 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 02:08:25 +00:00
Bill Wendling
bb08593980 Add a profile for uniquifying the AttributeSet with the AttributeSetNodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173313 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 01:01:34 +00:00
Bill Wendling
8b1f2f3b70 Cleanup the AttributeSetNodes that we create.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173311 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 00:14:46 +00:00
Bill Wendling
3467e30edf Create a new class: AttributeSetNode.
This is a helper class for the AttributeSetImpl class. It holds a set of
attributes that apply to a single element: function, return type, or
parameter.

These are uniqued.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173310 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-24 00:06:56 +00:00
Bill Wendling
a8ab5fc772 Push down the conversion of the alignment from the bit mask to a real number into the attribute implementation class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173304 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-23 23:00:05 +00:00
Bill Wendling
2915a691b9 Remove dead methods.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173302 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-23 22:38:33 +00:00
Eli Bendersky
98202c0974 Fix small typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173298 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-23 22:05:19 +00:00
Tom Stellard
254a83e46c R600: Add a llvm.R600.store.swizzle intrinsics
This intrinsic is translated to ALLOC_EXPORT_WORD1_SWIZ, hence its
name. It is used to store vs/fs outputs

Patch by: Vincent Lejeune

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173297 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-23 21:39:49 +00:00
Tom Stellard
2a3e0d7e76 R600: Simplify stream outputs intrinsic
Patch by: Vincent Lejeune

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173296 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-23 21:39:47 +00:00