If a symbol points to the end of a fragment, instead of searching for
fixups in that fragment, search in the next fragment.

Fixes spurious assembler error with subtarget change next to "la"
pseudo-instruction, or expanded equivalent.

Alternate proposal to fix the problem discussed in
https://reviews.llvm.org/D58759.

Testcase by Ana Pazos.

Differential Revision: https://reviews.llvm.org/D58943

llvm-svn: 355946

llvm-svn: 355945

llvm-svn: 355942

Prior to this change, the "Symbol" field of a relocation would always be
assumed to be a symbol name, and if no such symbol existed, the
relocation would reference index 0. This confused me when I tried to use
a literal symbol index in the field: since "0x1" was not a known symbol
name, the symbol index was set as 0. This change falls back to treating
unknown symbol names as integers, and emits an error if the name is not
found and the string is not an integer.

Note that the Symbol field is optional, so if a relocation doesn't
reference a symbol, it shouldn't be specified. The new error required a
number of test updates.

Reviewed by: grimar, ruiu
Differential Revision: https://reviews.llvm.org/D58510

llvm-svn: 355938

Expand MULO with constant power of two operand into a shift. The
overflow is checked with (x << shift) >> shift == x, where the right
shift will be logical for umulo and arithmetic for smulo (with
exception for multiplications by signed_min).

Differential Revision: https://reviews.llvm.org/D59041

llvm-svn: 355937

This makes SandyBridge inherit back to Westmere/Nehalem.

Make bdver1-4 inherit from each other and btver2 inherit from btver1.

llvm-svn: 355935

This will change as part of the fix for the regressions in D58017.

llvm-svn: 355933

llvm-svn: 355932

Add a note about legacy FunctionPassManager to the LLVM tutorial.

It seems to confuse some people, worth adding a warning to the tutorial
to elaborate and suggest using `llvm::legacy::FunctionPassManager` for
now. Not a perfect solution but hopefully will avoid confusion
in the meantime.

Differential Revision: https://reviews.llvm.org/D59258

llvm-svn: 355930

I recently discovered a bug in llvm-cxxfilt introduced in r353743 but
was fixed later incidentally due to r355031. Specifically, llvm-cxxfilt
was attempting to call .back() on an empty string any time there was a
new line in the input. This was causing a crash in my debug builds only.
This patch simply adds a test that explicitly tests that llvm-cxxfilt
handles empty lines correctly. It may pass under release builds under
the broken behaviour, but it fails at least in debug builds.

Reviewed by: mattd

Differential Revision: https://reviews.llvm.org/D58785

llvm-svn: 355929

[FileCheck]Remove assertions that prevent matching an empty string at file start before CHECK-NEXT/SAME

This patch removes two assertions that were preventing writing of a test
that checked an empty line followed by some text. For example:

CHECK: {{^$}}
CHECK-NEXT: foo()

The assertion was because the current location the CHECK-NEXT was
scanning from was the start of the buffer. A similar issue occurred with
CHECK-SAME. These assertions don't protect against anything, as there is
already an error check that checks that CHECK-NEXT/EMPTY/SAME don't
appear first in the checks, and the following code works fine if the
pointer is at the start of the input.

Reviewed by: probinson, thopre, jdenny
Differential Revision: https://reviews.llvm.org/D58784

llvm-svn: 355928

Targets can potentially emit more efficient code if they know address
computations never overflow. For example ILP32 code on AArch64 (which only has
64-bit address computation) can ignore the possibility of overflow with this
extra information.

llvm-svn: 355926

Summary:
1. \t => '  '
2. test commit access

Reviewers: Higuoxing, liangdzou

Reviewed By: Higuoxing, liangdzou

Subscribers: kristina, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59243

llvm-svn: 355924

Summary: This helps resolve https://bugs.llvm.org/show_bug.cgi?id=31564

Reviewers: jhenderson, grimar

Reviewed By: jhenderson

Subscribers: rupprecht, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59175

llvm-svn: 355922

The code might intend to replace puts("") with putchar('\n') even if the
return value is used. It failed because use_empty() was used to guard
the whole block. While returning '\n' (putchar('\n')) is technically
correct (puts is only required to return a nonnegative number on
success), doing this looks weird and there is really little benefit to
optimize puts whose return value is used. So don't do that.

llvm-svn: 355921

This was found when we generated COPY from G8RC to F8RC in
EmitInstrWithCustomInserter without checking proper architecture,
we silently generated mtvsrd, which require P8 and up.

This is a NFC patch to add assert when we call copyPhysReg, in case
someone accidentally generate COPY between G8RC to F8RC for P7 and
below.

llvm-svn: 355920

llvm-svn: 355917

Differential revision: https://reviews.llvm.org/D59081

llvm-svn: 355914

Revert rL355906: [SLP] Remove redundancy of performing operand reordering twice: once in buildTree() and later in vectorizeTree().

This is a refactoring patch that removes the redundancy of performing operand reordering twice, once in buildTree() and later in vectorizeTree().
To achieve this we need to keep track of the operands within the TreeEntry struct while building the tree, and later in vectorizeTree() we are just accessing them from the TreeEntry in the right order.

This patch is the first in a series of patches that will allow for better operand reordering across chains of instructions (e.g., a chain of ADDs), as presented here: https://www.youtube.com/watch?v=gIEn34LvyNo

Patch by: @vporpo (Vasileios Porpodas)

Differential Revision: https://reviews.llvm.org/D59059
........

Reverted due to buildbot failures that I don't have time to track down.

llvm-svn: 355913

llvm-svn: 355912

Because I don't understand svn.

llvm-svn: 355908

Created a test/CodeGen/ARM/ParallelDSP folder.

llvm-svn: 355907

[SLP] Remove redundancy of performing operand reordering twice: once in buildTree() and later in vectorizeTree().

This is a refactoring patch that removes the redundancy of performing operand reordering twice, once in buildTree() and later in vectorizeTree().
To achieve this we need to keep track of the operands within the TreeEntry struct while building the tree, and later in vectorizeTree() we are just accessing them from the TreeEntry in the right order.

This patch is the first in a series of patches that will allow for better operand reordering across chains of instructions (e.g., a chain of ADDs), as presented here: https://www.youtube.com/watch?v=gIEn34LvyNo

Patch by: @vporpo (Vasileios Porpodas)

Differential Revision: https://reviews.llvm.org/D59059

llvm-svn: 355906

llvm-svn: 355905

Differential revision: https://reviews.llvm.org/D59140

llvm-svn: 355904

Summary:
Add support for immediate operand in S_ENDPGM

Change-Id: I0c56a076a10980f719fb2a8f16407e9c301013f6

Reviewers: alexshap

Subscribers: qcolombet, arsenm, kzhuravl, jvesely, wdng, nhaehnle, yaxunl, tpr, t-tye, eraman, arphaman, Petar.Avramovic, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59213

llvm-svn: 355902

This is addressing the issue that we're not modeling the cost of clib functions
in TTI::getIntrinsicCosts and thus we're basically addressing this fixme:
    
// FIXME: This is wrong for libc intrinsics.

To enable analysis of clib functions, we not only need an intrinsic ID and
formal arguments, but also the actual user of that function so that we can e.g.
look at alignment and values of arguments. So, this is the initial plumbing to
pass the user of an intrinsinsic on to getCallCosts, which queries
getIntrinsicCosts.

Differential Revision: https://reviews.llvm.org/D59014

llvm-svn: 355901

These two values correspond to the 'Empty' and 'Tombstone' special
keys defined by DenseMapInfo<int64_t>, which means that neither one
can be used as a key in DenseMap<int64_t, anything>. Hence, if you try
to use either of those values as an int literal, IntInit::get() fails
an assertion when it tries to insert them into its static cache of
int-literal objects.

Fixed by replacing the DenseMap with a std::map, which doesn't intrude
on the space of legal values of the key type.

Reviewers: nhaehnle, hfinkel, javedabsar, efriedma

Reviewed By: efriedma

Subscribers: fhahn, efriedma, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59016

llvm-svn: 355900

These are closely modeled on similar tests for the ilp32 ABI. Like those
tests, we group together tests that should be common cross lp64, lp64+lp64f,
and lp64+lp64f+lp64d ABIs.

llvm-svn: 355899

Summary:
I'm trying to push D59198 but it seems that `git-llvm push` can't handle the fact
that I add a new directory in the patch:

```
> git llvm push -n
Pushing 1 commit:
  e7c0a9bd136 Correctly look up declarations in inline namespaces
Traceback (most recent call last):
  File "llvm/utils/git-svn//git-llvm", line 431, in <module>
    args.func(args)
  File "llvm/utils/git-svn//git-llvm", line 385, in cmd_push
    clean_svn(svn_root)
  File "llvm/utils/git-svn//git-llvm", line 201, in clean_svn
    os.remove(os.path.join(svn_repo, filename))
IsADirectoryError: [Errno 21] Is a directory: '.git/llvm-upstream-svn/lldb/trunk/packages/Python/lldbsuite/test/expression_command/inline-namespace'
```

This patch just uses shutil to delete the directory instead of trying to use `os.remove`
which only works for files.

Reviewers: mehdi_amini, jlebar

Reviewed By: jlebar

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59236

llvm-svn: 355896

Typo `we we're` => `we were` in the pass EarlyCSE

Patch by liangdzou (Liang ZOU)

Differential Revision: https://reviews.llvm.org/D59241

llvm-svn: 355895

Differential Revision: https://reviews.llvm.org/D59126

llvm-svn: 355892

Change from original commit: move test (that uses an X86 triple) into the X86
subdirectory.

Original description:
Gating vectorizing reductions on *all* fastmath flags seems unnecessary;
`reassoc` should be sufficient.

Reviewers: tvvikram, mkuper, kristof.beyls, sdesmalen, Ayal

Reviewed By: sdesmalen

Subscribers: dcaballe, huntergr, jmolloy, mcrosier, jlebar, bixia, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D57728

llvm-svn: 355889

Summary:
Swift now generates PDBs for debugging on Windows. llvm and lldb
need a language enumerator value too properly handle the output
emitted by swiftc.

Subscribers: jdoerfert, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59231

llvm-svn: 355882

For the design in question, overloads seem to be a much simpler and less subtle solution.

This removes ODR issues, and errors of the kind where code that uses the
specialization in question will accidentally and erroneously specialize
the primary template. This only "works" by accident; the program is
ill-formed NDR.

(Found with -Wundefined-func-template.)

Patch by Thomas Köppe!

Differential Revision: https://reviews.llvm.org/D58998

llvm-svn: 355880

This reverts commit r355868.  Breaks hexagon.

llvm-svn: 355873

[X86] Remove ProcModel and ProcFeatures tablegen classes. Move all feature lists into a ProcessorFeatures class.

ProcFeatures was a class that just concatenated two feature lists together and gave it a name. We used it to inherit features between CPUs.

ProcModel took a two CPU feature lists and concatenated them before deferring to ProcessorModel. This was to allow inherited features and specific features to be passed to each CPU.

Both of these allowed for only very rigid CPU inheritance rules.

With this patch we now store all of the lists we were using for inheritance in one object and do any list oncatenation we want there. Then we just pass whatever list we want from this class into the ProcessorModel class for each CPU.

Hopefully this gives us more flexibility to build up feature lists in whatever ways we think make sense. Perhaps untangling ISA flags and tuning flags.

I've only touched the CPUs that were directly affected by the removal of the ProcModel and ProcFeatures classes. We should move more of the feature lists into ProcessorFeatures.

llvm-svn: 355872

After r355865, we should be able to safely select G_EXTRACT_VECTOR_ELT without
running into any problematic intrinsics.

Also add a fix for lane copies, which don't support index 0.

llvm-svn: 355871

Summary: It is incomplete and has no users AFAIK.

Reviewers: pcc, vitalybuka

Subscribers: srhines, kubamracek, mgorny, krytarowski, eraman, hiraditya, jdoerfert, #sanitizers, llvm-commits, thakis

Tags: #sanitizers, #llvm

Differential Revision: https://reviews.llvm.org/D59154

llvm-svn: 355870

AtomicCmpSwapWithSuccess is legalised into an AtomicCmpSwap plus a comparison.
This requires an extension of the value which, by default, is a
zero-extension. When we later lower AtomicCmpSwap into a PseudoCmpXchg32 and then expanded in
RISCVExpandPseudoInsts.cpp, the lr.w instruction does a sign-extension.

This mismatch of extensions causes the comparison to fail when the compared
value is negative. This change overrides TargetLowering::getExtendForAtomicOps
for RISC-V so it does a sign-extension instead.

Differential Revision: https://reviews.llvm.org/D58829
Patch by Ferran Pallarès Roca.

llvm-svn: 355869