This is a non-intrusive fix for
https://bugs.llvm.org/show_bug.cgi?id=51476 intended for backport
to the 13.x release branch. It expands on the current hack by
distinguishing between CmpValue of 0, 1 and 2, where 0 and 1 have
the obvious meaning and 2 means "anything else". The new optimization
from D98564 should only be performed for CmpValue of 0 or 1.
For main, I think we should switch the analyzeCompare() and
optimizeCompare() APIs to use int64_t instead of int, which is in
line with MachineOperand's notion of an immediate, and avoids this
problem altogether.
Differential Revision: https://reviews.llvm.org/D108076
`CompilerInstance` is a more appropriate place for a key component of
the frontend like `Semantics`.
This change opens a path for us to introduce new frontend actions that
will also run semantics, but for which inheriting from
`PrescanAndSemaAction` wouldn't make much sense. For example, for
code-gen actions we plan to introduce a dedicate hierarchy of action
classes.
I've also added a doxyment for `CompilerInstance` to add a bit of
context for this change (and also make future refactoring more informed).
As `CompilerInstance` in Flang has been inspired by its counterpart in
Clang, this comment is roughly a verbatim copy of the comment in Clang
(with some adjustments from me). Credits to Daniel Dunbar for the great
design and the original comment.
Differential Revision: https://reviews.llvm.org/D108035
Currently we only support %z and %ll width modifiers,
but surprisingly not %l. This makes it impossible to print longs
(sizeof(long) not necessary equal to sizeof(size_t)).
We had some printf's that printed longs with %zu,
but that's wrong and now with __attribute__((format)) in place
they are flagged by compiler. So we either have a choice of
doing static_cast<uptr>(long) everywhere or add %l.
Adding %l looks better, that's a standard modifier.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D108066
Previously we just used {}, but that doesn't work in situations
like this.
if (1)
_MM_EXTRACT_FLOAT(d, x, n);
else
...
The semicolon would terminate the if.
`LLVM :: Other/lit-quoting.txt` currently `FAIL`s on Solaris:
llvm/test/Other/lit-quoting.txt:8:9: error: CHECK2: expected string not found in input
CHECK2: {{^a\[b\\c$}}
^
<stdin>:1:1: note: scanning from here
a[b
^
This happens because echo with backslashes or special characters is
unportable, as extensively documented in the Autoconf manual. In the case
at hand, `echo 'a[b\c'` yields `a[b\c` on Linux, but `a[b` (no newline) on
Solaris.
This patch fixes this by using the portable alternative suggested in the
Autoconf manual.
Tested on `amd64-pc-solaris2.11`, `sparcv9-sun-solaris2.11`, and
`x86_64-pc-linux-gnu`.
Differential Revision: https://reviews.llvm.org/D108031
Only one person committed with these email addresses, so there's no need to use
the map-different-names-for-one-email-address syntax.
No behavior change.
Differential Revision: https://reviews.llvm.org/D108007
This tests code starting from smin/smax, as opposed to the icmp/select
form. Also adds a ARM MVE phase ordering test for vectorizing to
sadd.sat from the original IR.
While the changes are extensive, they basically fall into a few
categories:
1) Moving the TestDialect itself.
2) Updating C++ code in tablegen to explicitly use ::mlir, since it
will be put in a headers that shouldn't expect a 'using'.
3) Updating some generic MLIR Interface definitions to do the same thing.
4) Updating the Tablegen generator in a few places to be explicit about
namespaces
5) Doing the same thing for llvm references, since we no longer pick
up the definitions from mlir/Support/LLVM.h
Differential Revision: https://reviews.llvm.org/D88251
This is enabled by default. Drop explicit uses in preparation for
removing the option.
Also drop RUN lines that are now the same (typically modulo a
-verify-memoryssa option).
This patch unify optimizeELF_x86_64_GOTAndStubs and optimizeMachO_x86_64_GOTAndStubs into a pure optimize_x86_64_GOTAndStubs
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D108025
The current LIR does not deal with runtime-determined memset-size. This patch
utilizes SCEV and check if the PointerStrideSCEV and the MemsetSizeSCEV are equal.
Before comparison the pass would try to fold the expression that is already
protected by the loop guard.
Testcase file `memset-runtime.ll`, `memset-runtime-debug.ll` added.
This patch deals with proper loop-idiom. Proceeding patch wants to deal with SCEV-s
that are inequal after folding with the loop guards.
Reviewed By: lebedev.ri, Whitney
Differential Revision: https://reviews.llvm.org/D107353
https://bugs.llvm.org/show_bug.cgi?id=50069
When clang-tidy sees:
```
if (true) [[unlikely]] {
...
}
```
It thinks the braces are missing and add them again.
```
if (true) { [[unlikely]] {
...
}
}
```
This revision aims to prevent that incorrect code generation
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D105479
clang/docs/tool/dump_format_style.py was not run as part of {D99840}
Bring ClangFormatStyleOptions.rst back in line.
Reviewed By: HazardyKnusperkeks
Differential Revision: https://reviews.llvm.org/D107958
Another simple cleanups set in DSE. CheckCache is removed since 1f1145006b and in consequence KnownNoReads is useless.
Also update description of MemorySSAScanLimit which default value is 150 instead 100.
Differential Revision: https://reviews.llvm.org/D107812
The overload of the constructor will repeatedly fix the member variables that need to be initialized.
Removed the duplicate '{}'.
```
struct A {
A() {}
A(int) {}
int _var; // int _var{}{}; <-- wrong fix
};
```
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D107641
Improves maintainability (edit/modify the tests without recompiling) and
error messages (previously the failure would be a gtest failure
mentioning nothing of the input or desired text) and the option to
improve tests with more checks.
(maybe these tests shouldn't all be in separate files - we could
probably have DWARF yaml that contains multiple errors while still being
fairly maintainable - the various invalid offsets (ref_addr, rnglists,
ranges, etc) could probably be all in one test, but for the simple sake
of the migration I just did the mechanical thing here)
This is a fairly common pattern:
```
%mask = G_CONSTANT iN <mask val>
%add = G_ADD %lhs, %rhs
%and = G_AND %add, %mask
```
We have combines to eliminate G_AND with a mask that does nothing.
If we combined the above to this:
```
%mask = G_CONSTANT iN <mask val>
%narrow_lhs = G_TRUNC %lhs
%narrow_rhs = G_TRUNC %rhs
%narrow_add = G_ADD %narrow_lhs, %narrow_rhs
%ext = G_ZEXT %narrow_add
%and = G_AND %ext, %mask
```
We'd be able to take advantage of those combines using the trunc + zext.
For this to work (or be beneficial in the best case)
- The operation we want to narrow then widen must only be used by the G_AND
- The G_TRUNC + G_ZEXT must be free
- Performing the operation at a narrower width must not produce a different
value than performing it at the original width *after masking.*
Example comparison between SDAG + GISel: https://godbolt.org/z/63jzb1Yvj
At -Os for AArch64, this is a 0.2% code size improvement on CTMark/pairlocalign.
Differential Revision: https://reviews.llvm.org/D107929
Unfortunately Mesa is still using amdgcn-- as the triple for OpenGL,
so we still have the awkward unknown OS case to deal with. Previously
if the HSA ABI intrinsics appeared, we we would not add the ABI
registers to the function. We would emit an error later, but we still
need to produce some compile result. Start adding the registers to any
compute function, regardless of the OS. This keeps the internal state
more consistent, and will help avoid numerous test crashes in a future
patch which starts assuming the ABI inputs are present on functions by
default.
Previously we would allow promotion even if the byval/inalloca
attributes on the call and the callee didn't match.
It's ok if the byval/inalloca types aren't the same. For example, LTO
importing may rename types.
Fixes PR51397.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D107998
AttributeList::hasAttribute() is confusing. In an attempt to change the
name to something that suggests using other methods, fix up some
existing uses.
When cross compiling lldb-server, do not create a host build
for building lldb-tblgeb when LLDB_TABLEGEN_EXE is already
provided. This avoids an expensive and time-consuming build step
if lldb-tblgen was already built previously for host.
Reviewed By: JDevlieghere
Differential Revision: https://reviews.llvm.org/D108053
The approach for handling reductions in the outer most
dimension follows that for inner most dimensions, outlined
below
First, transpose to move reduction dims, if needed
Convert reduction from n-d to 2-d canonical form
Then, for outer reductions, we emit the appropriate op
(add/mul/min/max/or/and/xor) and combine the results.
Differential Revision: https://reviews.llvm.org/D107675
AttributeList::hasAttribute() is confusing, use clearer methods like
hasParamAttr()/hasRetAttr().
Add hasRetAttr() since it was missing from AttributeList.
Nikita Popov