Differential revision: https://reviews.llvm.org/D43457

llvm-svn: 327381

clang-3.8 complains that constructor for '...' must explicitly
initialize the const member. Newer clangs and gcc seem to be fine with
this, but explicitly initializing the member does not hurt.

llvm-svn: 327380

llvm-svn: 327379

This finishes PR35877.

INSERT BEFORE used similar to INSERT AFTER,
it inserts sections before the given target section.

Differential revision: https://reviews.llvm.org/D44380

llvm-svn: 327378

Its a follow up for r327374 to fix BB.

llvm-svn: 327377

This is part of PR36515.

With some linkerscripts it is possible to get file offset overlaps
and overflows. Currently LLD checks overlaps in checkNoOverlappingSections().
And also we allow broken output with --no-inhibit-exec.
Problem is that sometimes final offset of sections is completely broken
and we calculate output file size wrong and might crash.

Patch implements check to verify that there is no output section
which offset exceeds file size.

Differential revision: https://reviews.llvm.org/D43819

llvm-svn: 327376

splitMergedValStore will split a store into two if target prefers this, or if
-force-split-store is passed.

This patch adds the missing handling for endianness in this function along
with a test case.

Review: Eli Friedman
https://reviews.llvm.org/D44396

llvm-svn: 327375

This fixes PR36598.

LLD currently crashes when we have empty output section
with SHF_LINK_ORDER flag. This might happen if we place an 
empty synthetic section in the linker script, but keep output
section alive with the use of additional symbol, for example.

The patch fixes the issue by dropping all special flags
for empty sections.

Differential revision: https://reviews.llvm.org/D44376

llvm-svn: 327374

isAvailableAtLoopEntry duplicates logic of `properlyDominates` after checking invariance.
This patch replaces this logic with invocation of this method which is more profitable
because it supports caching.

Differential Revision: https://reviews.llvm.org/D43997

llvm-svn: 327373

Summary: Fixes PR36557.

Reviewers: trentxintong, spatel

Subscribers: mstorsjo, llvm-commits

Differential Revision: https://reviews.llvm.org/D44083

llvm-svn: 327372

Add more debug information for peephole optimization passes.

These would only be enabled for debug version binary and could help
analyzing why some optimization opportunities were missed.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327371

This new pass eliminate identical move:

  MOV rA, rA

This is particularly possible to happen when sub-register support
enabled. The special type cast insn MOV_32_64 involves different
register class on src (i32) and dst (i64), RA could generate useless
instruction due to this.

This pass also could serve as the bast for further post-RA optimization.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327370

Currently, there is no ALU32 bswap support in eBPF ISA.

BSWAP on i32 was set to EXPAND which would need about eight instructions
for single BSWAP.

It would be more efficient to promote it to i64, then doing BSWAP on i64.
For eBPF programs, most of the promotion are zero extensions which are
likely be elimiated later by peephole optimizations.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327369

This patch relax the subregister definition check on Phi node.
Previously, we just cancel the optimizatoin when the definition is Phi
node while actually we could further check the definitions of incoming
parameters of PHI node.

This helps catch more elimination opportunities.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327368

The current zero extension elimination was restricted to operands of
comparison. It actually could be extended to more cases.

For example:

  int *inc_p (int *p, unsigned a)
  {
    return p + a;
  }

'a' will be promoted to i64 during addition, and the zero extension could
be eliminated as well.

For the elimination optimization, it should be much better to start
recognizing the candidate sequence from the SRL instruction instead of J*
instructions.

This patch makes it an generic zero extension elimination pass instead of
one restricted with comparison.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327367

There is a mistake in current code that we "break" out the optimization
when the first operand of J*_RR doesn't qualify the elimination. This
caused some elimination opportunities missed, for example the one in the
testcase.

The code should just fall through to handle the second operand.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327366

The current subregister definition check stops after the MOV_32_64
instruction.

This means we are thinking all the following instruction sequences
are safe to be eliminated:

  MOV_32_64 rB, wA
  SLL_ri    rB, rB, 32
  SRL_ri    rB, rB, 32

However, this is *not* true. The source subregister wA of MOV_32_64 could
come from a implicit truncation of 64-bit register in which case the high
bits of the 64-bit register is not zeroed, therefore we can't eliminate
above sequence.

For example, for i32_val, we shouldn't do the elimination:

  long long bar ();

  int foo (int b, int c)
  {
    unsigned int i32_val = (unsigned int) bar();

    if (i32_val < 10)
      return b;
    else
      return c;
  }

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327365

Improve the test accuracy by adding more check directives.

Shifts are expected to be eliminated for zero extension but not for signed
extension.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327364

It is a revert of rL327362 which causes build bot failures with assert like

Assertion `isAvailableAtLoopEntry(RHS, L) && "RHS is not available at Loop Entry"' failed.

llvm-svn: 327363

IsKnownPredicate is updated to implement the following algorithm
proposed by @sanjoy and @mkazantsev :
isKnownPredicate(Pred, LHS, RHS) {
  Collect set S all loops on which either LHS or RHS depend.
  If S is non-empty
    a. Let PD be the element of S which is dominated by all other elements of S
    b. Let E(LHS) be value of LHS on entry of PD.
       To get E(LHS), we should just take LHS and replace all AddRecs that
       are attached to PD on with their entry values.
       Define E(RHS) in the same way.
    c. Let B(LHS) be value of L on backedge of PD.
       To get B(LHS), we should just take LHS and replace all AddRecs that
       are attached to PD on with their backedge values.
       Define B(RHS) in the same way.
    d. Note that E(LHS) and E(RHS) are automatically available on entry of PD,
       so we can assert on that.
    e. Return true if isLoopEntryGuardedByCond(Pred, E(LHS), E(RHS)) &&
                      isLoopBackedgeGuardedByCond(Pred, B(LHS), B(RHS))
Return true if Pred, L, R is known from ranges, splitting etc.
}
This is follow-up for https://reviews.llvm.org/D42417.

Reviewers: sanjoy, mkazantsev, reames
Reviewed By: sanjoy, mkazantsev
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43507

llvm-svn: 327362

Summary:
r327219 added wrappers to std::sort which randomly shuffle the container before sorting.
This will help in uncovering non-determinism caused due to undefined sorting
order of objects having the same key.

To make use of that infrastructure we need to invoke llvm::sort instead of std::sort.

Reviewers: grosser, efriedma, jdoerfert, bollu, sebpop

Reviewed By: sebpop

Subscribers: sebpop, mehdi_amini, llvm-commits, pollydev

Tags: #polly

Differential Revision: https://reviews.llvm.org/D44361

llvm-svn: 327361

Summary:
This change fixes PR36483. The bug was originally introduced by a change
that marked non-prevailing symbols dead. This broke LowerTypeTests
handling of available_externally functions, which are non-prevailing.
LowerTypeTests uses liveness information to avoid emitting thunks for
unused functions.

Marking available_externally functions dead is incorrect, the functions
are used though the function definitions are not. This change keeps them
live, and lets the EliminateAvailableExternally/GlobalDCE passes remove
them later instead.

(Reland with a suspected fix for a unit test failure I haven't been able
to reproduce locally)

Reviewers: pcc, tejohnson

Reviewed By: tejohnson

Subscribers: grimar, mehdi_amini, inglorion, eraman, llvm-commits

Differential Revision: https://reviews.llvm.org/D43690

llvm-svn: 327360

This caused a crash in LTOModule::createInLocalContext.

rdar://37926841

llvm-svn: 327359

Summary:
If there's a callees metadata attached to the indirect call instruction, add CallGraphEdges to the callees mentioned in the metadata when computing FunctionSummary.

* Why this is necessary:
Consider following code example:
```
(foo.c)
static int f1(int x) {...}
static int f2(int x);
static int (*fptr)(int) = f2;
static int f2(int x) {
  if (x) fptr=f1; return f1(x);
}
int foo(int x) {
  (*fptr)(x); // !callees metadata of !{i32 (i32)* @f1, i32 (i32)* @f2} would be attached to this call.
}

(bar.c)
int bar(int x) {
  return foo(x);
}
```

At LTO time when `foo.o` is imported into `bar.o`, function `foo` might be inlined into `bar` and PGO-guided indirect call promotion will run after that. If the profile data tells that the promotion of `@f1` or `@f2` is beneficial, the optimizer will check if the "promoted" `@f1` or `@f2` (such as `@f1.llvm.0` or `@f2.llvm.0`) is available. Without this patch, importing `!callees` metadata would only add promoted declarations of `@f1` and `@f2` to the `bar.o`, but still the optimizer will assume that the function is available and perform the promotion. The result of that is link failure with `undefined reference to @f1.llvm.0`.

This patch fixes this problem by adding callees in the `!callees` metadata to CallGraphEdges so that their definition would be properly imported into.

One may ask that there already is a logic to add indirect call promotion targets to be added to CallGraphEdges. However, if profile data says "indirect call promotion is only beneficial under a certain inline context", the logic wouldn't work. In the code example above, if profile data is like
```
bar:1000000:100000
  1:100000
    1: foo:100000
        1: 100000 f1:100000
```
, Computing FunctionSummary for `foo.o` wouldn't add `foo->f1` to CallGraphEdges. (Also, it is at least "possible" that one can provide profile data to only link step but not to compilation step).

Reviewers: tejohnson, mehdi_amini, pcc

Reviewed By: tejohnson

Subscribers: inglorion, eraman, llvm-commits

Differential Revision: https://reviews.llvm.org/D44399

llvm-svn: 327358

Currently lld creates plain plt entries when a R_386_PC32 resolves to
a symbol in a shared library. That is a bug (PR36678). Don't depend on
that behavior on this test.

llvm-svn: 327357

Typical example, illformed comparisons (operator== where LHS and
RHS are not compatible). If a symbol matched `operator==` in any
of the object files lldb inserted a generic function declaration
in the ASTContext on which Sema operates. Maintaining the AST
context invariants is fairly tricky and sometimes resulted in
crashes inside clang (or assertions hit).

The real reason why this feature exists in the first place is
that of allowing users to do something like:
(lldb) call printf("patatino")

even if the debug informations for printf() is not available.
Eventually, we might reconsider this feature in its
entirety, but for now we can't remove it as it would break
a bunch of users. Instead, try to limit it to non-C++ symbols,
where getting the invariants right is hopefully easier.

Now you can't do in lldb anymore
(lldb) call _Zsomethingsomething(1,2,3)

but that doesn't seem to be such a big loss.

<rdar://problem/35645893>

llvm-svn: 327356

[LegalizeTypes] In SplitVecOp_TruncateHelper, use GetSplitVector on the input instead of creating new extract_subvectors.

llvm-svn: 327355

llvm-svn: 327354

This diff adjusts the comdat tests after changing the format
of llvm-readobj output for .group sections.

llvm-svn: 327353

Call expressions that return objects by an lvalue reference or an rvalue
reference have a value type in the AST but wear an auxiliary flag of being an
lvalue or an xvalue respectively.

Use the helper method for obtaining the actual return type of the function.

Fixes a crash.

Differential Revision: https://reviews.llvm.org/D44273

llvm-svn: 327352

I forgot to incorporate these comments into the original revision.  This
is just code cleanup addressing the feedback, NFC.

llvm-svn: 327351

llvm-svn: 327350

This diff adjusts the mc tests after changing the format
of llvm-readobj output for .group sections.

llvm-svn: 327349

Properly perform destruction and lifetime extension of such temporaries.

C++ object-type return values of conservatively evaluated functions are now
represented as compound values of well-defined temporary object regions. The
function creates a region that represents the temporary object and will later
be used for destruction or materialization, invalidates it, and returns the
invalidated compound value of the object.

Differential Revision: https://reviews.llvm.org/D44131

llvm-svn: 327348

Differential Revision: https://reviews.llvm.org/D44129

llvm-svn: 327347

This reverts commit r327295 since it was causing the Windows bots to
fail.

llvm-svn: 327346

This patch uses the newly added CFGCXXRecordTypedCall element at the call site
of the caller to construct the return value within the callee directly into the
caller's stack frame. This way it is also capable of populating the temporary
destructor and lifetime extension maps for the temporary, which allows
temporary destructors and lifetime extension to work correctly.

This patch does not affect temporaries that were returned from conservatively
evaluated functions.

Differential Revision: https://reviews.llvm.org/D44124

llvm-svn: 327345

Context can do what Tagged was intended to support (snapshot filesystems),
and less intrusively.
getTaggedFileSystem() no longer needs a filename.

Cleanups while here:
 - code-complete now returns errors as Expected, like other functions
 - added an alias Callback<T> for the usual callback function type

llvm-svn: 327344

This patch adds a new CFGStmt sub-class, CFGCXXRecordTypedCall, which replaces
the regular CFGStmt for the respective CallExpr whenever the CFG has additional
information to provide regarding the lifetime of the returned value.

This additional call site information is represented by a ConstructionContext
(which was previously used for CFGConstructor elements) that provides references
to CXXBindTemporaryExpr and MaterializeTemporaryExpr that surround the call.

This corresponds to the common C++ calling convention solution of providing
the target address for constructing the return value as an auxiliary implicit
argument during function call.

One of the use cases for such extra context at the call site would be to perform
any sort of inter-procedural analysis over the CFG that involves functions
returning objects by value. In this case the elidable constructor at the return
site would construct the object explained by the context at the call site, and
its lifetime would also be managed by the caller, not the callee.

The extra context would also be useful for properly handling the return-value
temporary at the call site, even if the callee is not being analyzed
inter-procedurally.

Differential Revision: https://reviews.llvm.org/D44120

llvm-svn: 327343

Also, move fmul reassociation tests to the same file as other fmul transforms.

llvm-svn: 327342