For DSE with MemorySSA it is beneficial to manually traverse the
defining access, instead of using a MemorySSA walker, so we can
better control the number of steps together with other limits and
also weed out invalid/unprofitable paths early on.

This patch requires a follow-up patch to be most effective, which I will
share soon after putting this patch up.

This temporarily XFAIL's the limit tests, because we now explore more
MemoryDefs that may not alias/clobber the killing def. This will be
improved/fixed by the follow-up patch.

This patch also renames some `Dom*` variables to `Earlier*`, because the
dominance relation is not really used/important here and potentially
confusing.

This patch allows us to aggressively cut down compile time, geomean
-O3 -0.64%, ReleaseThinLTO -1.65%, at the expense of fewer stores
removed. Subsequent patches will increase the number of removed stores
again, while keeping compile-time in check.

http://llvm-compile-time-tracker.com/compare.php?from=d8e3294118a8c5f3f97688a704d5a05b67646012&to=0a929b6978a068af8ddb02d0d4714a2843dd8ba9&stat=instructions

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D86486

A build on `sparcv9-sun-solaris2.11` with `-DLLVM_ENABLE_PIC=Off` failed
linking `LLVMPolly.so`:

  [2277/2297] Linking CXX shared module lib/LLVMPolly.so
  FAILED: lib/LLVMPolly.so
  [...]
  ld: fatal: relocation error: R_SPARC_H44: file tools/polly/lib/CMakeFiles/obj.Polly.dir/Analysis/DependenceInfo.cpp.o: symbol .data._ZL16__gthread_active (section): invalid shared object relocation type: ABS44 code model unsupported
  [...]

As on many other targets, one cannot link non-PIC objects into a shared
object on Solaris/sparcv9.

The following patch avoids this by not building the library without PIC.
It allowed the build to finish.

Differential Revision: https://reviews.llvm.org/D85627

This reverts commit 8d5f64c4.

Thanks to Eli Friedma for pointing out that this check is not appropiate here,
this check will be moved to the Lint pass.

There is no justification for changing vcc_lo to vcc
when shrinking V_CNDMASK, and such a change could
later confuse live variable analysis.

Make sure the original register is preserved.

Differential Revision: https://reviews.llvm.org/D86541

Differential Revision: https://reviews.llvm.org/D86606

We had already specified that second argument `n` of this intrinsic is `n > 0`,
but now add to this that the result is a poison value if this is not the case.

Differential Revision: https://reviews.llvm.org/D86637

Currently, an undef value is reduced to 0 when it is added to a set of potential values.
This patch introduces a flag for under values. By this, for example, we can merge two states `{undef}`, `{1}` to `{1}` (because we can reduce the undef to 1).

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D85592

Enable default outlining when the function has the minsize attribute
and we're targeting an m-class core.

Differential Revision: https://reviews.llvm.org/D82951

This reverts commit 99364552.

That commit caused failed assertions e.g. like this:

$ cat alloca.c
a;
b() {
  float c;
  d();
  a = __builtin_alloca(d);
  c = e();
  f(a);
  return c;
}
$ clang -target aarch64-linux-gnu -c alloca.c -O2
clang: ../lib/Target/AArch64/AArch64InstrInfo.cpp:3446: void
llvm::emitFrameOffset(llvm::MachineBasicBlock&,
llvm::MachineBasicBlock::iterator, const llvm::DebugLoc&, unsigned int,
unsigned int, llvm::StackOffset, const llvm::TargetInstrInfo*,
llvm::MachineInstr::MIFlag, bool, bool, bool*):
Assertion `(DestReg != AArch64::SP || Bytes % 16 == 0) &&
"SP increment/decrement not 16-byte aligned"' failed.

Reviewed By: rriddle

Differential Revision: https://reviews.llvm.org/D86648

This patch allows to pass the gpu module name to SPIR-V
module during conversion. This has many benefits as we can lookup
converted to SPIR-V kernel in the symbol table.

In order to avoid symbol conflicts, `"__spv__"` is added to the
gpu module name to form the new one.

Reviewed By: mravishankar

Differential Revision: https://reviews.llvm.org/D86384

Implements the assemble and disassemble support of RISCV Vector
extension zvamo instructions, base on the 0.9 spec version.

Reviewed  by HsiangKai

Differential Revision: https://reviews.llvm.org/D85069

Fix the ARM backend's analyzeBranch so it doesn't ignore predicated
return instructions, and make the MachineVerifier rule more strict.

Differential Revision: https://reviews.llvm.org/D40061

Summary: It was passing on my local machine due to previously-written
files cached in the test output folder.

This patch introduces a hook to encode descriptor set
and binding number into `spv.globalVariable`'s symbolic name. This
allows to preserve this information, and at the same time legalize
the global variable for the conversion to LLVM dialect.

This is required for `mlir-spirv-cpu-runner` to convert kernel
arguments into LLVM.

Also, a couple of some nits added:
- removed unused comment
- changed to a capital letter in the comment

Reviewed By: mravishankar

Differential Revision: https://reviews.llvm.org/D86515

Previously a NodeRole would generally be prefixed with the `NodeKind`,
we remove this prefix, as it we redundant and made tests more noisy.

Differential Revision: https://reviews.llvm.org/D86636

This makes OpPassManager more of a "container" of passes and not responsible to drive the execution.
As such we also make it constructible publicly, which will allow to build arbitrary pipeline decoupled from the execution. We'll make use of this facility to expose "dynamic pipeline" in the future.

Reviewed By: rriddle

Differential Revision: https://reviews.llvm.org/D86391

This patch updates the type conversion section of the documentation.
It includes the modelling of array strides and the mapping of the
naturally padded structs.

Reviewed By: mravishankar

Differential Revision: https://reviews.llvm.org/D86674

This happens when generating a huge file by LTO, for example, with -gmlt.
When BitNo is > 2^35, ByteNo is overflowed, and an incorrect output offset is overwritten.
This generates ill-formed bitcodes.

Reviewed-by: tejohnson, vitalybuka

Differential Revision: https://reviews.llvm.org/D86645

This patch adds an optional name to SPIR-V module.
This will help with lowering from GPU dialect (so that we
can pass the kernel module name) and will be more naturally
aligned with `GPUModuleOp`/`ModuleOp`.

Reviewed By: mravishankar

Differential Revision: https://reviews.llvm.org/D86386

This patch implements the function prototypes vec_mulh and vec_dive in order to
utilize the vector multiply high (vmulh[s|u][w|d]) and vector divide extended
(vdive[s|u][w|d]) instructions introduced in Power10.

Differential Revision: https://reviews.llvm.org/D82609

error message

/usr/bin/ld: CMakeFiles/mlir-opt.dir/mlir-opt.cpp.o: in function `main':
mlir-opt.cpp:(.text.startup.main+0xb9): undefined reference to `mlir::registerTestDialect(mlir::DialectRegistry&)'

Reviewed By: mehdi_amini

Differential Revision: https://reviews.llvm.org/D86592

Differential Revision: https://reviews.llvm.org/D86640

Previously, we were only emitting regular bindings to weak
dynamic symbols; this diff adds support for the weak bindings too, which
can overwrite the regular bindings at runtime. We also treat weak
defined global symbols similarly -- since they can also be interposed at
runtime, they need to be treated as potentially dynamic symbols.

Note that weak bindings differ from regular bindings in that they do not
specify the dylib to do the lookup in (i.e. weak symbol lookup happens
in a flat namespace.)

Differential Revision: https://reviews.llvm.org/D86572

Previously, the BindingEntry struct could only store bindings to offsets
within InputSections. Since the GOTSection and TLVPointerSections are
OutputSections, I handled those in a separate code path. However, this
makes it awkward to support weak bindings properly without code
duplication. This diff allows BindingEntries to point directly to
OutputSections, simplifying the upcoming weak binding implementation.

Along the way, I also converted a bunch of functions taking references
to symbols to take pointers instead. Given how much casting we do for
Symbol (especially in the upcoming weak binding diffs), it's cleaner
this way.

Differential Revision: https://reviews.llvm.org/D86571

It's roughly like -force_load with some filtering.

Differential Revision: https://reviews.llvm.org/D86181

The re-exports list in a TAPI document can either refer to other inlined
TAPI documents, or to on-disk files (which may themselves be TBD or
regular files.) Similarly, the re-exports of a regular dylib can refer
to a TBD file.

Differential Revision: https://reviews.llvm.org/D85404

Two things needed fixing for that to work:

1. getName() no longer returns null for DylibFiles constructed from TAPIs
2. markSubLibrary() now accepts .tbd as a possible extension

Differential Revision: https://reviews.llvm.org/D86180

This matches ld64's behavior

Differential Revision: https://reviews.llvm.org/D85992

Differential Revision: https://reviews.llvm.org/D86255

 * When _LIBUNWIND_SUPPORT_COMPACT_UNWIND is defined in config.h,
   define it to "1" like the other macros. These macros are still
   checked using "#if defined(...)", however.

 * Include libunwind.h in AddressSpace.hpp before using
   _LIBUNWIND_ARM_EHABI.

 * Rename ProcessFrameHeaderCache to TheFrameHeaderCache, because some
   configurations (e.g. Android / hermetic static libraries) can have
   one cache per shared object in the process. (When there are more
   copies, it's more important not to waste memory in the cache.)

 * Add 3 missing header files to LIBUNWIND_HEADERS.

Differential Revision: https://reviews.llvm.org/D86254

I forgot to squash this with 0b7f6cc7

AArch64, X86 and Mips currently directly consumes these and custom
lowering to produce a libcall, but really these should follow the
normal legalization process through the libcall/lower action.

This will make stateful registrars (e.g. a future TargetProcessControl based
registrar) easier to deal with.