Zachary Turner [Mon, 10 Oct 2016 21:36:23 +0000 (21:36 +0000)]
Revert "Disallow ArrayRef assignment from temporaries."
This reverts commit r283798, as it causes static asserts on
MSVC 2015 with the following errors:
ArrayRefTest.cpp(38): error C2338: Assigning from single prvalue element
ArrayRefTest.cpp(41): error C2338: Assigning from single xvalue element
ArrayRefTest.cpp(47): error C2338: Assigning from an initializer list
Zachary Turner [Mon, 10 Oct 2016 21:24:34 +0000 (21:24 +0000)]
Rename llvm::apply -> llvm::apply_tuple.
llvm::cl already has a function called llvm::apply() so this is
causing an ODR violation. The STLExtras version should win the
vote on which one gets to be called apply() since it is named
after the equivalent STL function, but since renaiming the cl
version is more difficult, let's do this for now to get the
bots green.
Adrian Prantl [Mon, 10 Oct 2016 17:53:33 +0000 (17:53 +0000)]
Teach llvm::StripDebugInfo() about global variable !dbg attachments.
This is a regression introduced by the global variable ownership
reversal performed in r281284.
[ARM] Fix invalid VLDM/VSTM access when targeting Big Endian with NEON
The instructions VLDM/VSTM can only access word-aligned memory
locations and produce alignment fault if the condition is not met.
The compiler currently generates VLDM/VSTM for v2f64 load/store
regardless the alignment of the memory access. Instead, if a v2f64
load/store is not word-aligned, the compiler should generate
VLD1/VST1. For each non double-word-aligned VLD1/VST1, a VREV
instruction should be generated when targeting Big Endian.
George Rimar [Mon, 10 Oct 2016 10:51:38 +0000 (10:51 +0000)]
[Object/ELF] - Do not crash on invalid Header->e_shoff value.
sections_begin() may return unalignment pointer when Header->e_shoff isinvalid.
That may result in a crash in clients, for example we have one in LLD:
assert((PtrWord & ~PointerBitMask) == 0 &&
"Pointer is not sufficiently aligned");
fails when trying to push_back Elf_Shdr* (unaligned) into TinyPtrVector.
Patch forces check for alignment of Header->e_shoff.
Chris Dewhurst [Mon, 10 Oct 2016 08:53:06 +0000 (08:53 +0000)]
This pass, fixing an erratum in some LEON 2 processors ensures that the SDIV instruction is not issued, but replaced by SDIVcc instead, which does not exhibit the error. Unit test included.
Commit in the name of:Coby Tayree
1.'v' constraint for (x86) non-avx arch imitates the already implemented 'x' constraint, i.e. allows XMM{0-15} & YMM{0-15} depending on the apparent arch & mode (32/64).
2.for the avx512 arch it allows [X,Y,Z]MM{0-31} (mode dependent)
This patch applies the needed changes to clang
clang patch: https://reviews.llvm.org/D25004
Brian Gesiak [Mon, 10 Oct 2016 01:22:06 +0000 (01:22 +0000)]
[lit] Remove (or allow specific) unused imports
Summary:
Using Python linter flake8 on the utils/lit reveals several linter
warnings designated "F401: Unused import". Fix or silence these
warnings.
Some of these unused imports are legitimate, while some are part of lit's API.
For example, users of lit expect to be able to access `lit.formats.ShTest` in
their `lit.cfg`, despite the module hierarchy for that symbol actually being
`lit.formats.shtest.ShTest`. To silence linter errors for these lines,
include a "noqa" directive.
Brian Gesiak [Mon, 10 Oct 2016 01:19:27 +0000 (01:19 +0000)]
[lit] Fix undefined symbol ArgumentError
Summary:
`ArgumentError` is not defined by the Python standard library.
Executing this line of code would throw a exception, but not the
intended one. It would throw a `NameError` exception, since `ArgumentError`
is undefined.
Use `ValueError` instead, which is defined by the Python standard
library.
Brian Gesiak [Mon, 10 Oct 2016 01:18:14 +0000 (01:18 +0000)]
[lit] Remove semicolons in Python code
Summary:
Semicolons aren't necessary as statement terminators in Python, and
each of these uses are superfluous as they appear at the end of a line.
The convention is to not use semicolons where not needed, so remove them.
DAG: Setting Masked-Expand-Load as a variant of Masked-Load node
Masked-expand-load node represents load operation that loads a variable amount of elements from memory according to amount of "true" bits in the mask and expands the loaded elements according to their position in the mask vector.
Right now, the node is used in intrinsics for VEXPAND* instructions.
The work is done towards implementation of masked.expandload and masked.compressstore intrinsics.
Craig Topper [Sat, 8 Oct 2016 18:49:57 +0000 (18:49 +0000)]
[AVX-512] Fix a bug in getLargestLegalSuperClass where we inflated to VR128X/VR256X even when VLX isn't supported.
This seems to have been responsible for the XMM16-31 spills observed in PR29112. With this fixed the test case has been modified to no longer have a spill of XMM16.
Teresa Johnson [Sat, 8 Oct 2016 16:11:42 +0000 (16:11 +0000)]
[ThinLTO] Record calls to aliases
Summary:
When there is a call to an alias in the same module, we were not
adding a call edge. So we could incorrectly think that the alias
was dead if it was inlined in that function, despite having a
reference imported elsewhere. This resulted in unsats at link time.
Mehdi Amini [Sat, 8 Oct 2016 04:44:18 +0000 (04:44 +0000)]
ThinLTO: handles modules with empty summaries
We need to add an entry in the combined-index for modules that have
a hash but otherwise empty summary, this is needed so that we can
get the hash for the module.
Also, if no entry is present in the combined index for a module, we
need to skip it when trying to compute a cache entry.
Zachary Turner [Sat, 8 Oct 2016 01:12:01 +0000 (01:12 +0000)]
[pdb] Dump Module Symbols to Yaml.
This is the first step towards round-tripping symbol information,
and thusly being able to write symbol information to a PDB.
This patch writes the symbol information for each compiland to
the Yaml when running in pdb2yaml mode. There's still some loose
ends, such as what to do about relocations (necessary in order to
print linkage names), how to print enums with friendly names, and
how to give the dumper access to the StringTable, but this is a
good first start.
I mistakenly invered the condition assuming it was a typo. I am now
removing it because it doesn't seem to be a problem anymore (plus it's a
dirty hack).
Dylan McKay [Sat, 8 Oct 2016 01:01:49 +0000 (01:01 +0000)]
[AVR] Expand MULHS for all types
Once MULHS was expanded, this exposed an issue where the condition
register was thought to be 16-bit. This caused an attempt to copy a
16-bit register to an 8-bit register.
Hal Finkel [Sat, 8 Oct 2016 00:26:54 +0000 (00:26 +0000)]
[llvm-opt-report] Don't leave space for opts that never happen
Because screen space is precious, if an optimization (vectorization, for
example) never happens, don't leave empty space for the associated markers on
every line of the output. This makes the output much more compact, and allows
for the later inclusion of markers for more (although perhaps rare)
optimizations.
Gor Nishanov [Sat, 8 Oct 2016 00:22:50 +0000 (00:22 +0000)]
[coroutines] Store an address of destroy OR cleanup part in the coroutine frame.
Summary:
If heap allocation of a coroutine is elided, we need to make sure that we will update an address stored in the coroutine frame from f.destroy to f.cleanup.
Before this change, CoroSplit synthesized these stores after coro.begin:
```
store void (%f.Frame*)* @f.resume, void (%f.Frame*)** %resume.addr
store void (%f.Frame*)* @f.destroy, void (%f.Frame*)** %destroy.addr
```
In those cases where we did heap elision, but were not able to devirtualize all indirect calls, destroy call will attempt to "free" the coroutine frame stored on the stack. Oops.
Now we use select to put an appropriate coroutine subfunction in the destroy slot. As bellow:
Kyle Butt [Fri, 7 Oct 2016 22:33:20 +0000 (22:33 +0000)]
Codegen: Tail-duplicate during placement.
The tail duplication pass uses an assumed layout when making duplication
decisions. This is fine, but passes up duplication opportunities that
may arise when blocks are outlined. Because we want the updated CFG to
affect subsequent placement decisions, this change must occur during
placement.
In order to achieve this goal, TailDuplicationPass is split into a
utility class, TailDuplicator, and the pass itself. The pass delegates
nearly everything to the TailDuplicator object, except for looping over
the blocks in a function. This allows the same code to be used for tail
duplication in both places.
This change, in concert with outlining optional branches, allows
triangle shaped code to perform much better, esepecially when the
taken/untaken branches are correlated, as it creates a second spine when
the tests are small enough.
Issue from previous rollback fixed, and a new test was added for that
case as well. Issue was worklist/scheduling/taildup issue in layout.
Issue from 2nd rollback fixed, with 2 additional tests. Issue was
tail merging/loop info/tail-duplication causing issue with loops that share
a header block.
swifterror: Don't compute swifterror vregs during instruction selection
The code used llvm basic block predecessors to decided where to insert phi
nodes. Instruction selection can and will liberally insert new machine basic
block predecessors. There is not a guaranteed one-to-one mapping from pred.
llvm basic blocks and machine basic blocks.
Therefore the current approach does not work as it assumes we can mark
predecessor machine basic block as needing a copy, and needs to know the set of
all predecessor machine basic blocks to decide when to insert phis.
Instead of computing the swifterror vregs as we select instructions, propagate
them at the end of instruction selection when the MBB CFG is complete.
When an instruction needs a swifterror vreg and we don't know the value yet,
generate a new vreg and remember this "upward exposed" use, and reconcile this
at the end of instruction selection.
This will only happen if the target supports promoting swifterror parameters to
registers and the swifterror attribute is used.
projects / llvm / log