James Molloy [Thu, 12 Feb 2015 15:54:14 +0000 (15:54 +0000)]
[LoopRerolling] Be more forgiving with instruction order.
We can't solve the full subgraph isomorphism problem. But we can
allow obvious cases, where for example two instructions of different
types are out of order. Due to them having different types/opcodes,
there is no ambiguity.
Tim Northover [Thu, 12 Feb 2015 15:12:13 +0000 (15:12 +0000)]
Triple: refactor redundant code.
Should be no functional change, since most of the logic removed was
completely pointless (after some previous refactoring) and the rest
duplicated elsewhere.
Andrea Di Biagio [Thu, 12 Feb 2015 14:17:24 +0000 (14:17 +0000)]
[TTI] Teach the cost heuristic how to query TLI to check if a zext/trunc is 'free' for the target.
Now that SimplifyCFG uses TTI for the cost heuristic, we can teach BasicTTIImpl
how to query TLI in order to get a more accurate cost for truncates and
zero-extends.
Before this patch, the basic cost heuristic in TargetTransformInfoImplCRTPBase
would have conservatively returned a 'default' TCC_Basic for all zero-extends,
and TCC_Free for truncates on native types.
This patch improves the heuristic so that we query TLI (if available) to get
more accurate answers. If TLI is available, then methods 'isZExtFree' and
'isTruncateFree' can be used to check if a zext/trunc is free for the target.
Added more test cases to SimplifyCFG/X86/speculate-cttz-ctlz.ll.
With this change, SimplifyCFG is now able to speculate a 'cheap' cttz/ctlz
immediately followed by a free zext/trunc.
Benjamin Kramer [Thu, 12 Feb 2015 13:47:29 +0000 (13:47 +0000)]
MathExtras: Parametrize count(Trailing|Leading)Zeros on the type size.
Otherwise we will always select the generic version for e.g. unsigned
long if uint64_t is typedef'd to 'unsigned long long'. Also remove
enable_if hacks in favor of static_assert.
Asiri Rathnayake [Thu, 12 Feb 2015 13:37:28 +0000 (13:37 +0000)]
ARM: Fix another regression introduced in r223113
The changes in r223113 (ARM modified-immediate syntax) have broken
instructions like:
mov r0, #~0xffffff00
The problem is that I've added a spurious range check on the immediate
operand to ensure that it lies between INT32_MIN and UINT32_MAX. While
this range check is correct in theory, it causes problems because the
operand is stored in an int64_t (by MC). So valid 32-bit constants like
\#~0xffffff00 become out of range. The solution is to simply remove this
range check. It is not possible to validate the range of the immediate
operand with the current setup because: 1) The operand is stored in an
int64_t by MC, 2) The immediate can be of the forms #imm, #-imm, #~imm
or even #((~imm)) etc. So we just chop the value to 32 bits and use it.
Also noted that the original range check was note tested by any of the
unit tests. I've added a new test to cover #~imm kind of operands.
Dmitry Vyukov [Thu, 12 Feb 2015 09:55:28 +0000 (09:55 +0000)]
tsan: do not instrument not captured values
I've built some tests in WebRTC with and without this change. With this change number of __tsan_read/write calls is reduced by 20-40%, binary size decreases by 5-10% and execution time drops by ~5%. For example:
Using KORTESTW for comparison i1 value with zero was wrong since the instruction tests 16 bits.
KORTESTW may be used with KSHIFTL+KSHIFTR that clean the 15 upper bits.
I removed (X86cmp i1, 0) pattern and zero-extend i1 to i8 and then use TESTB.
There are some cases where i1 is in the mask register and the upper bits are already zeroed.
Then KORTESTW is the better solution, but it is subject for optimization.
Meanwhile, I'm fixing the correctness issue.
[X86] A heuristic to estimate the size impact for converting stack-relative parameter movs to pushes
This gives a rough estimate of whether using pushes instead of movs is profitable, in terms of size.
We go over all calls in the MachineFunction and compute:
a) For each callsite that can not use pushes, the penalty of not having a reserved call frame.
b) For each callsite that can use pushes, the gain of actually replacing the movs with pushes (and the potential penalty of having to readjust the stack).
Ahmed Bougacha [Thu, 12 Feb 2015 06:15:29 +0000 (06:15 +0000)]
[CodeGen] Don't blindly combine (fp_round (fp_round x)) to (fp_round x).
We used to do this DAG combine, but it's not always correct:
If the first fp_round isn't a value preserving truncation, it might
introduce a tie in the second fp_round, that wouldn't occur in the
single-step fp_round we want to fold to.
In other words, double rounding isn't the same as rounding.
We would crash if we couldn't locate a Function that either Location's
Value belonged to. Now we just print out a debug message and return
conservatively.
Chandler Carruth [Thu, 12 Feb 2015 02:30:56 +0000 (02:30 +0000)]
[slp] Fix a nasty bug in the SLP vectorizer that Joerg pointed out.
Apparently some code finally started to tickle this after my
canonicalization changes to instcombine.
The bug stems from trying to form a vector type out of scalars that
aren't compatible at all. In this example, from x86_mmx values. The code
in the vectorizer that checks for reasonable types whas checking for
aggregates or vectors, but there are lots of other types that should
just never reach the vectorizer.
Debugging this was made more confusing by the lie in an assert in
VectorType::get() -- it isn't that the types are *primitive*. The types
must be integer, pointer, or floating point types. No other types are
allowed.
I've improved the assert and added a helper to the vectorizer to handle
the element type validity checks. It now re-uses the VectorType static
function and then further excludes weird target-specific types that we
probably shouldn't be touching here (x86_fp80 and ppc_fp128). Neither of
these are really reachable anyways (neither 80-bit nor 128-bit things
will get vectorized) but it seems better to just eagerly exclude such
nonesense.
I've added a test case, but while it definitely covers two of the paths
through this code there may be more paths that would benefit from test
coverage. I'm not familiar enough with the SLP vectorizer to synthesize
test cases for all of these, but was able to update the code itself by
inspection.
Hal Finkel [Thu, 12 Feb 2015 01:02:52 +0000 (01:02 +0000)]
[PowerPC] Mark jumps as expensive (using using CR bits)
On PowerPC, which has a full set of logical operations on (its multiple sets
of) condition-register bits, it is not profitable to break of complex
conditions feeding a jump into multiple jumps. We can turn off this feature of
CGP/SDAGBuilder by marking jumps as "expensive".
Zachary Turner [Thu, 12 Feb 2015 00:05:49 +0000 (00:05 +0000)]
Revert "Change Path::filename_pos() to skip the drive letter."
This reverts commit 228874. For some reason users reported
seeing Clang taking up 25+GB of memory and bringing down
machines with this change. Reverting until we figure it out.
I mistakenly thought the liveness of each "RetVal(F, i)" depended only on F. It
actually depends on the index too, which means we need to be careful about how
the results are combined before return. In particular if a single Use returns
Live, that counts for the entire object, at the granularity we're considering.
Zachary Turner [Wed, 11 Feb 2015 21:16:35 +0000 (21:16 +0000)]
Change Path::filename_pos() to skip the drive letter.
For Windows, filename_pos() tries to find the filename by
searching for separators after the last :. Instead, it should
really check for the only location that a : is valid, which is
in the second character, and search for separators after that.
Mehdi Amini [Wed, 11 Feb 2015 19:54:44 +0000 (19:54 +0000)]
Reassociate: cannot negate a INT_MIN value
Summary:
When trying to canonicalize negative constants out of
multiplication expressions, we need to check that the
constant is not INT_MIN which cannot be negated.
Adrian Prantl [Wed, 11 Feb 2015 17:45:05 +0000 (17:45 +0000)]
Generalize DIBuilder's createReplaceableForwardDecl() to a more flexible
createReplaceableCompositeType() that allows to create non-forward-declared
temporary nodes.
Jan Wen Voung [Wed, 11 Feb 2015 16:12:50 +0000 (16:12 +0000)]
Gold-plugin: Broaden scope of get/release_input_file to scope of Module.
Summary:
Move calls to get_input_file and release_input_file out of
getModuleForFile(). Otherwise release_input_file may end up
unmapping a view of the file while the view is still being
used by the Module (on 32-bit hosts).
Jonas Paulsson [Wed, 11 Feb 2015 16:10:31 +0000 (16:10 +0000)]
Fix SelectionDAG compile time issue with alias analysis.
Add new token factor node and its users to worklist if alias analysis is
turned on, in DAGCombiner::visitTokenFactor(). Alias analysis may cause
a lot of new token factors to be inserted into the DAG, and they need to
be optimized to avoid significant slow-downs.
Marek Olsak [Wed, 11 Feb 2015 14:26:46 +0000 (14:26 +0000)]
R600/SI: Enable a lot of existing tests for VI (squashed commits)
This is a union of these commits:
* R600/SI: Enable more tests for VI which need no changes
* R600/SI: Enable V_BCNT tests for VI
Differences:
- v_bcnt_..._e32 -> _e64
- s_load_dword* inline offset is in bytes instead of dwords
* R600/SI: Enable all tests for VI which use S_LOAD_DWORD
The inline offset is changed from dwords to bytes.
* R600/SI: Enable LDS tests for VI
Differences:
- the s_load_dword inline offset changed from dwords to bytes
- the tests checked very little on CI, so they have been fixed to check all
instructions that "SI" checked
* R600/SI: Enable lshr tests for VI
* R600/SI: Fix divrem64 tests
- "v_lshl_64" was missing "b" before "64"
- added VI-NOT checks
* R600/SI: Enable the SI.tid test for VI
* R600/SI: Enable the frem test for VI
Also, the frem_f64 checking is added for CI-VI.
James Molloy [Wed, 11 Feb 2015 12:15:41 +0000 (12:15 +0000)]
[SimplifyCFG] Swap to using TargetTransformInfo for cost
analysis.
We're already using TTI in SimplifyCFG, so remove the hard-baked "cheapness"
heuristic and use TTI directly. Generally NFC intended, but we're using a slightly
different heuristic now so there is a slight test churn.
Test changes:
* combine-comparisons-by-cse.ll: Removed unneeded branch check.
* 2014-08-04-muls-it.ll: Test now doesn't branch but emits muleq.
* coalesce-subregs.ll: Superfluous block check.
* 2008-01-02-hoist-fp-add.ll: fadd is safe to speculate. Change to udiv.
* PhiBlockMerge.ll: Superfluous CFG checking code. Main checks still present.
* select-gep.ll: A variable GEP is not expensive, just TCC_Basic, according to the TTI.
Daniel Sanders [Wed, 11 Feb 2015 11:28:56 +0000 (11:28 +0000)]
[mips] Merge disassemblers into a single implementation.
Summary:
Currently we have Mips32 and Mips64 disassemblers and this causes the target
triple to affect the disassembly despite all the relevant information being in
the ELF header. These implementations do not need to be separate.
This patch merges them together such that the appropriate tables are checked
for the subtarget (e.g. Mips64 is checked when GP64 is enabled).
The NodeMetadata are maintained in an incremental way. When an edge between
2 nodes has its cost updated, in the course of graph reduction for example,
the NodeMetadata need first to have the old edge cost removed, then the new
edge cost added. Only once the NodeMetadata have been fully updated, it
becomes safe to consider promoting the nodes to the
ConservativelyAllocatable or OptimallyReducible sets. Previously, this
promotion was occuring right after the removing the old cost, and this was
breaking the assumption that a ConservativelyAllocatable should not be
spilled.
This patch also adds asserts to:
- enforces the invariant that a node's reduction can not be downgraded,
- only not provably allocatable or optimally reducible nodes can be spilled.
Justin Bogner [Wed, 11 Feb 2015 02:52:44 +0000 (02:52 +0000)]
InstrProf: Lower coverage mappings by setting their sections appropriately
Add handling for __llvm_coverage_mapping to the InstrProfiling
pass. We need to make sure the constant and any profile names it
refers to are in the correct sections, which is easier and cleaner to
do here where we have to know about profiling sections anyway.
This is really tricky to test without a frontend, so I'm committing
the test for the fix in clang. If anyone knows a good way to test this
within LLVM, please let me know.
Petar Jovanovic [Tue, 10 Feb 2015 23:30:14 +0000 (23:30 +0000)]
Fix makeLibCall argument (signed) in SoftenFloatRes_XINT_TO_FP function
The isSigned argument of makeLibCall function was hard-coded to false
(unsigned). This caused zero extension on MIPS64 soft float.
As the result SingleSource/Benchmarks/Stanford/FloatMM test and
SingleSource/UnitTests/2005-07-17-INT-To-FP test failed.
The solution was to use the proper argument.
Adrian Prantl [Tue, 10 Feb 2015 23:18:28 +0000 (23:18 +0000)]
Debug Info: Support variables that are described by more than one MMI
table entry. This happens when SROA splits up an alloca and the resulting
allocas cannot be lowered to SSA values because their address is passed
to a function.
Zachary Turner [Tue, 10 Feb 2015 22:43:25 +0000 (22:43 +0000)]
Rewrite llvm-pdbdump in terms of LLVMDebugInfoPDB.
This makes llvm-pdbdump available on all platforms, although it
will currently fail to create a dumper if there is no PDB reader
implementation for the current platform.
It implements dumping of compilands and children, which is less
information than was previously available, but it has to be
rewritten from scratch using the new set of interfaces, so the
rest of the functionality will be added back in subsequent commits.
David Majnemer [Tue, 10 Feb 2015 21:22:05 +0000 (21:22 +0000)]
X86: Make @llvm.frameaddress work correctly with Windows unwind codes
Simply loading or storing the frame pointer is not sufficient for
Windows targets. Instead, create a synthetic frame object that we will
lower later. References to this synthetic object will be replaced with
the correct reference to the frame address.
Zachary Turner [Tue, 10 Feb 2015 21:17:52 +0000 (21:17 +0000)]
Provide DIA implementation of DebugInfoPDB.
This implements DebugInfoPDB when the DIA SDK is present on the system.
Specifically, this means that the following conditions are met:
1) You are building on Windows.
2) You are building with MSVC.
3) Visual Studio did not corrupt the installation of DIA due to a
known issue with side-by-side installations of VS2012 and VS2013.
If all of these conditions are true, you will be able to pass a value
of PDB_Reader::DIA to PDB::createPdbReader().
There are no tests for this yet, as any test will be in the form of a
lit test which tests the llvm-pdbdump.exe, which still needs to be
rewritten in terms of this library.
Aaron Ballman [Tue, 10 Feb 2015 20:13:52 +0000 (20:13 +0000)]
Changing the status code generated by LLVM_BUILTIN_TRAP on Windows to be something categorized as a valid error code. Fixes crashing uses (such as not --crash) with existing sys::Wait behavior.
Tim Northover [Tue, 10 Feb 2015 19:49:18 +0000 (19:49 +0000)]
DeadArgElim: arguments affect all returned sub-values by default.
Unless we meet an insertvalue on a path from some value to a return, that value
will be live if *any* of the return's components are live, so all of those
components must be added to the MaybeLiveUses.
Previously we were deleting arguments if sub-value 0 turned out to be dead.
Add new API for converting temporaries that may self-reference.
Self-referencing nodes are not allowed to be uniqued, so sending them
into `replaceWithUniqued()` is dangerous (and this commit adds
assertions that prevent it).
`replaceWithPermanent()` has similar semantics to `get()` followed by
calls to `replaceOperandWith()`. In particular, if there's a
self-reference, it returns a distinct node; otherwise, it returns a
uniqued one. Like `replaceWithUniqued()` and `replaceWithDistinct()`
(well, it calls out to them) it mutates the temporary node in place if
possible, only calling `replaceAllUsesWith()` on a uniquing collision.
Bill Schmidt [Tue, 10 Feb 2015 19:09:05 +0000 (19:09 +0000)]
[PowerPC] Fix reverted patch r227976 to avoid register assignment issues
See full discussion in http://reviews.llvm.org/D7491.
We now hide the add-immediate and call instructions together in a
separate pseudo-op, which is tagged to define GPR3 and clobber the
call-killed registers. The PPCTLSDynamicCall pass prior to RA now
expands this op into the two separate addi and call ops, with explicit
definitions of GPR3 on both instructions, and explicit clobbers on the
call instruction. The pass is now marked as requiring and preserving
the LiveIntervals and SlotIndexes analyses, and fixes these up after
the replacement sequences are introduced.
Self-hosting has been verified on LE P8 and BE P7 with various
optimization levels, etc. It has also been verified with the
--no-tls-optimize flag workaround removed.