The existing memory manager API can not be shared between objects when linking
concurrently (since there is no way to know which concurrent allocations were
performed on behalf of which object, and hence which allocations would be safe
to finalize when finalizeMemory is called). For now, we can work around this by
requiring a new memory manager for each object.

This change only affects the concurrent version of the ORC APIs.

llvm-svn: 341579

Section address mappings can be applied using the RuntimeDyld instance passed to
the RuntimeDyld::MemoryManager::notifyObjectLoaded method. Proving an alternate
route via RuntimeDyldObjectLinkingLayer2 is redundant.

llvm-svn: 341578

This was from a missed review comment from D51338.

llvm-svn: 341577

llvm-svn: 341575

llvm-svn: 341574

MRI scripts have two comment chars, * and ;, but only the latter was
supported before.

Also allow leading spaces before comment chars (and before any command
string), and allow comments after a command.

Differential Revision: https://reviews.llvm.org/D51338

llvm-svn: 341571

Keeping the compile units in memory is expensive. For the single
threaded case we allocate them in the analyze part and deallocate them
again once we've finished cloning. This poses a problem in the single
threaded case where we did all the analysis first followed by all the
cloning. This meant we had all the link context in memory right after
analyzing finished.

This patch changes the way we order work in the single threaded case.
Instead of doing all the analysis and cloning in serial, we now
interleave the two so we can deallocate the memory as soon as a file is
processed. The result is binary identical and peak memory usage went
down from 13.43GB to 5.73GB for a debug build of trunk clang.

Differential revision: https://reviews.llvm.org/D51618

llvm-svn: 341568

llvm-svn: 341567

llvm-svn: 341565

llvm-svn: 341561

This fold is needed to avoid a regression when we try
to recommit rL300977. 
We can't see the most basic win currently because 
demanded bits changes the patterns:
https://rise4fun.com/Alive/plpp

llvm-svn: 341559

Summary:
I added a few ARM64 memset codegen tests in r341406 and r341493, and annotated
where the generated code was bad. This patch fixes the majority of the issues by
requesting that a 2xi64 vector be used for memset of 32 bytes and above.

The patch leaves the former request for f128 unchanged, despite f128
materialization being suboptimal: doing otherwise runs into other asserts in
isel and makes this patch too broad.

This patch hides the issue that was present in bzero_40_stack and bzero_72_stack
because the code now generates in a better order which doesn't have the store
offset issue. I'm not aware of that issue appearing elsewhere at the moment.

<rdar://problem/44157755>

Reviewers: t.p.northover, MatzeB, javed.absar

Subscribers: eraman, kristof.beyls, chrib, dexonsmith, llvm-commits

Differential Revision: https://reviews.llvm.org/D51706

llvm-svn: 341558

These tests demonstrate a missing fold that would
also be needed to avoid a regression when we try 
to recommit rL300977.

llvm-svn: 341557

llvm-svn: 341555

Add the __msan_va_arg_origin_tls TLS array to keep the origins for variadic function parameters.
Change the instrumentation pass to store parameter origins in this array.

This is a reland of r341528.

test/msan/vararg.cc doesn't work on Mips, PPC and AArch64 (because this
patch doesn't touch them), XFAIL these arches.
Also turned out Clang crashed on i80 vararg arguments because of
incorrect origin type returned by getOriginPtrForVAArgument() - fixed it
and added a test.

llvm-svn: 341554

Standardise on check lines:
* CHECK-ASM
* CHECK-OBJ
* CHECK-ASM-AND-OBJ

This allows for the addition of tests involving symbol operands, which will
not result in identical instructions in both assembly and disassembled object 
output.

This commit doesn't exploit this reworking to increase test coverage of symbol
operands - that will come in a future patch.

llvm-svn: 341546

I'm preparing to add the same functionality both here and to the DAG 
version of this code in D51696 / D51433, so try to make those cases 
as similar as possible to avoid bugs.

llvm-svn: 341545

llvm-svn: 341541

llvm-svn: 341537

The MCID::Flag enumeration now has more than 32 items, this means that
the hasPropertyBundle argument 'Mask' can overflow.

This patch changes the argument to be 64 bits instead.

Patch by Mikael Nilsson.

Differential Revision: https://reviews.llvm.org/D51596

llvm-svn: 341536

There is no need to create preheaders in the analysis stage, we only
need them when adjusting the branches. Also, the only cases we need to
create our own preheaders is when they have more than 1 predecessors or
PHI nodes (even with only 1 predecessor, we could have an LCSSA phi
node). I have simplified the conditions and added some assertions to be
sure. Because we know the inner and outer loop need to be tightly
nested, it is sufficient to check if the inner loop preheader is the
outer loop header to check if we need to create a new preheader.

Reviewers: efriedma, mcrosier, karthikthecool

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D51703

llvm-svn: 341533

Currently it has a set KnownBlocks that marks blocks as having cached
answers and a map FirstSpecialInsts that maps these blocks to first
special instructions in them. The value in the map is always non-null,
and for blocks that are known to have no special instructions the map
does not have an instance.

This patch removes KnownBlocks as obsolete. Instead, for blocks that
are known to have no special instructions, we just put a nullptr value.
This makes the code much easier to read.

llvm-svn: 341531

Add the __msan_va_arg_origin_tls TLS array to keep the origins for
variadic function parameters.
Change the instrumentation pass to store parameter origins in this array.

llvm-svn: 341528

Previously the alignment on the newly created global strings was not set,
meaning that DataLayout::getPreferredAlignment was free to overalign it
to 16 bytes. This caused unnecessary code bloat with the padding between
variables.

The main example of this happening was the printf->puts optimisation in
SimplifyLibCalls, but as the change here is made in
IRBuilderBase::CreateGlobalString, other globals using this will now be
aligned too.

Differential Revision: https://reviews.llvm.org/D51410

llvm-svn: 341527

This validation patch has been reverted as rL341147 because of conserns raised by
@reames. This revision returns it as is to raise a discussion and address the concerns.

Differential Revision: https://reviews.llvm.org/D51523
Reviewed By: reames

llvm-svn: 341526

Turns out that calling a variadic function with too many (e.g. >100 i64's)
arguments overflows __msan_va_arg_tls, which leads to smashing other TLS
data with function argument shadow values.

getShadow() already checks for kParamTLSSize and returns clean shadow if
the argument does not fit, so just skip storing argument shadow for such
arguments.

llvm-svn: 341525

Summary:
This change adds a `BlockIndexer` type which maintains pointers to
records that belong to the same process+thread pairs. The indexing
happens with order of appearance of records as they are visited.

This version of the indexer currently only supports FDR version 3 logs,
which contain `BufferExtent` records. We will add support for v2 and v1
logs in follow-up patches.

This is another part of D50441.

Reviewers: eizan, kpw, mboerger

Reviewed By: mboerger

Subscribers: mboerger, mgorny, hiraditya, llvm-commits

Differential Revision: https://reviews.llvm.org/D51673

llvm-svn: 341518

llvm-svn: 341517

Function rewriteLoopExitValues contains a check on isValidRewrite which
is needed to make sure that SCEV does not convert the pattern
`gep Base, (&p[n] - &p[0])` into `gep &p[n], Base - &p[0]`. This problem
has been fixed in SCEV long ago, so this check is just obsolete.

This patch converts it into an assertion to make sure that the SCEV will
not mess up this case in the future.

Differential Revision: https://reviews.llvm.org/D51582
Reviewed By: atrick

llvm-svn: 341516

llvm-svn: 341515

In DwarfDebug::collectEntityInfo(), if the label entity is processed in
DbgLabels list, it means the label is not optimized out. There is no
need to generate debug info for it with null position.

llvm-svn: 341513

This basically reverts a change made in r336217, but improves the text of the error message for not allowing IP-relative addressing in 32-bit mode.

Fixes PR38826.

Patch by Iain Sandoe.

llvm-svn: 341512

[llvm-dwp] Fix `UN:` lines (supposed to be `RUN:`) in X86/simple.test and adjust check lines for TYPES:

Reviewers: dblaikie, aprantl

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D51704

llvm-svn: 341510

Reviewers: dblaikie, pcc

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D51707

llvm-svn: 341509

llvm-svn: 341508

llvm-svn: 341505

Forking this change from D51706. This just made it easier to understand llc
output with -debug.

llvm-svn: 341504

The way DIA SDK works is that when you request a symbol, it
gets assigned an internal identifier that is unique for the
life of the session.  You can then use this identifier to
get back the same symbol, with all of the same internal state
that it had before, even if you "destroyed" the original
copy of the object you had.

This didn't work properly in our native implementation, and
if you destroyed an object for a particular symbol, then
requested the same symbol again, it would get assigned a new
ID and you'd get a fresh copy of the object.  In order to fix
this some refactoring had to happen to properly reuse cached
objects.  Some unittests are added to verify that symbol
reuse is taking place, making use of the new unittest input
feature.

llvm-svn: 341503

Occasionally it is useful to have unittest which take inputs.
While we normally try to have this test be more of a lit test
we occasionally don't have tools that can exercise the code
in the right way to test certain things.  LLDB has been using
this style of unit test for a while, particularly with regards
to how it tests core dump and minidump file parsing.  Recently
i needed this as well for the case where we want to test that
some of the PDB reading code works correctly.  It needs to
exercise the code in a way that is not covered by any dumper
and would be impractical to implement in one of the dumpers,
but requires a valid PDB file.  Since this is now needed by
more than one project, it makes sense to have this be a
generally supported thing that unit tests can do, and we just
encourage people to use this sparingly.

Differential Revision: https://reviews.llvm.org/D51561

llvm-svn: 341502

ExecutionSession defaults to creating a new shared pool if none is provided,
so explicitly passing one in is unnecessary.

llvm-svn: 341494