This remove some EM_386 specific code from InputSection.cpp and opens
the way for more relaxations.

llvm-svn: 271814

It will also be used for GT_TO_IE relaxations.

llvm-svn: 271813

It will be used for more than just gots.

llvm-svn: 271812

The name map might not be densely packed on disk.  Using a sparse map
will save memory in such situations.

llvm-svn: 271811

This is step 1 of unknown towards fixing PR28001:
https://llvm.org/bugs/show_bug.cgi?id=28001

llvm-svn: 271810

llvm-svn: 271809

llvm-svn: 271808

llvm-svn: 271807

llvm-svn: 271806

llvm-svn: 271805

llvm-svn: 271804

Allows XOP to vectorize BITREVERSE - other targets will follow as their costmodels improve.

llvm-svn: 271803

[X86] Add the VR128L/H and VR256L/H to the list of vector register classes for inline asm constraints. Also fix the comment on the function.

llvm-svn: 271802

non-shared, PIE and shared output mode.

llvm-svn: 271801

CALL_ONCE_... macro in the legacy pass manager with the new
llvm::call_once facility.

Nothing changed sicne the last attempt in r271781 which I reverted in
r271788. At least one of the failures I saw was spurious, and I want to
make sure the other failures are real before I work around them -- they
appeared to only effect ppc64le and ppc64be.

Original commit message of r271781:
----
[LPM] Reinstate r271652 to replace the CALL_ONCE_... macro in the legacy
pass manager with the new llvm::call_once facility.

This reverts commit r271657 and re-applies r271652 with a fix to
actually work with arguments. In the original version, we just ended up
directly calling std::call_once via ADL because of the std::once_flag
argument. The llvm::call_once never worked with arguments. Now,
llvm::call_once is a variadic template that perfectly forwards
everything. As a part of this it had to move to the header and we use
a generic functor rather than an explict function pointer. It would be
nice to use std::invoke here but we don't have it yet. That means
pointer to members won't work here, but that seems a tolerable
compromise.

I've also tested this by forcing the fallback path, so hopefully it
sticks this time.
----

Original commit message of r271652:
----
[LPM] Replace the CALL_ONCE_... macro in the legacy pass manager with
the new llvm::call_once facility.

This facility matches the standard APIs and when the platform supports
it actually directly uses the standard provided functionality. This is
both more efficient on some platforms and much more TSan friendly.

The only remaining user of the cas_flag and home-rolled atomics is the
fallback implementation of call_once. I have a patch that removes them
entirely, but it needs a Windows patch to land first.

This alone substantially cleans up the macros for the legacy pass
manager, and should subsume some of the work Mehdi was doing to clear
the path for TSan testing of ThinLTO, a really important step to have
reliable upstream testing of ThinLTO in all forms.
----

llvm-svn: 271800

llvm-svn: 271799

We were not handling page relative relocations.

llvm-svn: 271798

Windows itanium is nearly identical to windows-msvc (MS ABI for C, itanium for
C++).  Enable the TLS support for the target similar to the MSVC model.

llvm-svn: 271797

The AVX2 v16i16 shift lowering works by unpacking to 2 x v8i32, performing the shift and then truncating the result.

The unpacking is used to place the values in the upper 16-bits so that we can correctly sign-extend for SRA shifts. Unfortunately we weren't ensuring that the lower 16-bits were zero to ensure that SHL correctly shifts in zero bits.

llvm-svn: 271796

llvm-svn: 271795

llvm-svn: 271794

C++ has a builtin type called wchar_t.  Clang also provides a type
called __wchar_t in C mode.

In C mode, wchar_t can be a typedef to unsigned short.

llvm-svn: 271793

It was checking for Union when it should have checked for Interface.

llvm-svn: 271792

llvm-svn: 271791

llvm-svn: 271790

Add the MMX implementation to the SimplifyDemandedUseBits SSE/AVX MOVMSK support added in D19614

Requires a minor tweak as llvm.x86.mmx.pmovmskb takes a x86_mmx argument - so we have to be explicit about the implied v8i8 vector type.

llvm-svn: 271789

There appears to be a strange exception thrown and crash using call_once
on a PPC build bot, and a *really* weird windows link error for
GCMetadata.obj. Still need to investigate the cause of both problems.

Original change summary:
[LPM] Reinstate r271652 to replace the CALL_ONCE_... macro in the legacy
pass manager with the new llvm::call_once facility.

llvm-svn: 271788

llvm-svn: 271787

fixing tis test. There are different configurations running, and they
have subtly different backtraces. I didn't notice that the configs
changed and so I kept occilating between the two.

Now I've looked at two different configs at the same time, and so this
should be much more likely to work.

llvm-svn: 271786

available along side the leak checking, so use the REQUIRES for that.

Also, use %run as other tests do when launching the built binary.

This fixes check-asan for me on Linux and looks like it should fix the
linux sanitizer bots as well.

llvm-svn: 271785

of lines provided with the filecheck output from the previous run. I'll
probably give up after this and get someone with a Windows build to help
me out.

llvm-svn: 271784

This should fail on frame #1 and show what that should actually be.

llvm-svn: 271783

with this, the Windows sanitizer bot will go green!

llvm-svn: 271782

pass manager with the new llvm::call_once facility.

This reverts commit r271657 and re-applies r271652 with a fix to
actually work with arguments. In the original version, we just ended up
directly calling std::call_once via ADL because of the std::once_flag
argument. The llvm::call_once never worked with arguments. Now,
llvm::call_once is a variadic template that perfectly forwards
everything. As a part of this it had to move to the header and we use
a generic functor rather than an explict function pointer. It would be
nice to use std::invoke here but we don't have it yet. That means
pointer to members won't work here, but that seems a tolerable
compromise.

I've also tested this by forcing the fallback path, so hopefully it
sticks this time.

Original commit message:
----
[LPM] Replace the CALL_ONCE_... macro in the legacy pass manager with
the new llvm::call_once facility.

This facility matches the standard APIs and when the platform supports
it actually directly uses the standard provided functionality. This is
both more efficient on some platforms and much more TSan friendly.

The only remaining user of the cas_flag and home-rolled atomics is the
fallback implementation of call_once. I have a patch that removes them
entirely, but it needs a Windows patch to land first.

This alone substantially cleans up the macros for the legacy pass
manager, and should subsume some of the work Mehdi was doing to clear
the path for TSan testing of ThinLTO, a really important step to have
reliable upstream testing of ThinLTO in all forms.

llvm-svn: 271781

[NFC] Check that a parameter of ScheduleTreeOptimizer::isMatrMultPattern contains a correct partial schedule

llvm-svn: 271780

show enough information for me to fix the patterns used here.

llvm-svn: 271779

[X86] Use unsigned types for vector arithmetic in intrinsics to avoid undefined behavior for signed integer overflow.

This is really only needed for addition, subtraction, and multiplication, but I did the bitwise ops too for overall consistency. Clang currently doesn't set NSW for signed vector operations so the undefined behavior shouldn't happen today.

llvm-svn: 271778

llvm-svn: 271777

llvm-svn: 271776

[X86] Use X86ISD::ABS for lowering pabs SSSE3/AVX intrinsics to match AVX512. Should allow those intrinsics to use the EVEX encoded instructions and get the extra registers when available.

llvm-svn: 271775