Fixes the issues noted in PR48768, where the and/or/xor instruction had been promoted to avoid i8/i16 partial-dependencies, but the test against zero had not.
We can almost certainly relax this fold to work for any truncation, although it breaks a number of existing folds (notable movmsk folds which tend to rely on the truncate to determine the demanded bits/elts in the source vector).
There is a reverse combine in TargetLowering.SimplifySetCC so we must wait until after legalization before attempting this.
Users can reset any external index set by previous fragments by
putting a `None` for the external block, e.g:
```
Index:
External: None
```
Differential Revision: https://reviews.llvm.org/D100106
The previous code calculated the first ldtilecfg by dominating all AMX registers' def. This may result in the ldtilecfg being inserted into a loop.
This patch try to calculate the nearest point where all shapes of AMX registers are reachable.
Reviewed By: LuoYuanke
Differential Revision: https://reviews.llvm.org/D99010
FP16 to FP32 converts can be handled in MVE lane interleaving, much like
the sext/zext lowering we do. This expands the pass with fpext and
fptrunc handling, and basic fp operations allowing more efficient
lowering of fp vectors.
Differential Revision: https://reviews.llvm.org/D97292
The patch makes two updates to the arm-block-placement pass:
- Handle arbitrarily nested loops
- Extends the search (for t2WhileLoopStartLR) to the predecessor of the
preHeader.
Differential Revision: https://reviews.llvm.org/D99649
The `linalg.index` operation provides access to the iteration indexes of immediately enclosing linalg operations. It takes a dimension `dim` attribute and returns the iteration index in the given dimension. Having `linalg.index` allows us to unify `linalg.generic` and `linalg.indexed_generic` and also enables index access in named operations.
Differential Revision: https://reviews.llvm.org/D100292
This reverts commit cca9b5985c.
Buildbot reported an error for CodeGen/AArch64/machine-combiner-fmul-dup.mir:
*** Bad machine code: Virtual register killed in block, but needed live out. ***
- function: indexed_2s
- basic block: %bb.0 entry (0x640fee8)
Virtual register %7 is used after the block.
*** Bad machine code: Virtual register defs don't dominate all uses. ***
- function: indexed_2s
- v. register: %7
LLVM ERROR: Found 2 machine code errors.
This patch adds DUP+FMUL => FMUL_indexed pattern to InstCombiner.
FMUL_indexed is normally selected during instruction selection, but it
does not work in cases when VDUP and VMUL are in different basic
blocks.
Differential Revision: https://reviews.llvm.org/D99662
That code is unused since it's check-in in 2010 (and I believe it would leak
memory when called as it releases the passed unique_ptr), so let's delete it.
Reviewed By: vsk
Differential Revision: https://reviews.llvm.org/D100212
When LLDB's DWARF parser is parsing the member DIEs of a struct/class it
currently fully resolves the types of static member variables in a class before
adding the respective `VarDecl` to the record.
For record types fully resolving the type will also parse the member DIEs of the
respective class. The other way of resolving is just 'forward' resolving the type
which will try to load only the minimum amount of information about the type
(for records that would only be the name/kind of the type). Usually we always
resolve types on-demand so it's rarely useful to speculatively fully resolve
them on the first use.
This patch changes makes that we only 'forward' resolve the types of static
members. This solves the fact that LLDB unnecessarily loads debug information
to parse the type if it's maybe not needed later and it also avoids a crash where
the parsed type might in turn reference the surrounding class that is currently
being parsed.
The new test case demonstrates the crash that might happen. The crash happens
with the following steps:
1. We parse class `ToLayout` and it's members.
2. We parse the static class member and fully resolve its type
(`DependsOnParam2<ToLayout>`).
3. That type has a non-static class member `DependsOnParam1<ToLayout>` for which
LLDB will try to calculate the size.
4. The layout (and size)`DependsOnParam1<ToLayout>` turns depends on the
`ToLayout` size/layout.
5. Clang will calculate the record layout/size for `ToLayout` even though we are
currently parsing it and it's missing it's non-static member.
The created is missing the offset for the yet unparsed non-static member. If we
later try to get the offset we end up hitting different asserts. Most common is
the one in `TypeSystemClang::DumpValue` where it checks that the record layout
has offsets for the current FieldDecl.
```
assert(field_idx < record_layout.getFieldCount());
```
Fixed rdar://67910011
Reviewed By: shafik
Differential Revision: https://reviews.llvm.org/D100180
This patch fixes an issue with the SVE prefetch and qinc/qdec intrinsics
that take an `enum` argument, but where the builtin prototype encodes
these as `int`. Some code in SemaDecl found the mismatch and chose
to forget about the builtin altogether, which meant that any future
code using that builtin would fail. The code that forgets about the
builtin was actually obsolete after D77491 and should have been removed.
This patch now removes that code.
This patch also fixes another issue with the SVE prefetch intrinsic
when built with C++, where the builtin didn't accept the correct
pointer type, which should be `const void *`.
Reviewed By: tambre
Differential Revision: https://reviews.llvm.org/D100046
Spilling the fp or bp to scratch could overwrite VGPRs of inactive
lanes. Fix that by using only the active lanes of the scavenged VGPR.
This builds on the assumptions that
1. a function is never called with exec=0
2. lanes do not die in a function, i.e. exec!=0 in the function epilog
3. no new lanes are active when exiting the function, i.e. exec in the
epilog is a subset of exec in the prolog.
Differential Revision: https://reviews.llvm.org/D96869
Spilling SGPRs to scratch uses a temporary VGPR. LLVM currently cannot
determine if a VGPR is used in other lanes or not, so we need to save
all lanes of the VGPR. We even need to save the VGPR if it is marked as
dead.
The generated code depends on two things:
- Can we scavenge an SGPR to save EXEC?
- And can we scavenge a VGPR?
If we can scavenge an SGPR, we
- save EXEC into the SGPR
- set the needed lane mask
- save the temporary VGPR
- write the spilled SGPR into VGPR lanes
- save the VGPR again to the target stack slot
- restore the VGPR
- restore EXEC
If we were not able to scavenge an SGPR, we do the same operations, but
everytime the temporary VGPR is written to memory, we
- write VGPR to memory
- flip exec (s_not exec, exec)
- write VGPR again (previously inactive lanes)
Surprisingly often, we are able to scavenge an SGPR, even though we are
at the brink of running out of SGPRs.
Scavenging a VGPR does not have a great effect (saves three instructions
if no SGPR was scavenged), but we need to know if the VGPR we use is
live before or not, otherwise the machine verifier complains.
Differential Revision: https://reviews.llvm.org/D96336
The .rgba vector component accessors are supported in OpenCL C 3.0.
Previously, the diagnostic would check `OpenCLVersion` for version 2.2
(value 220) and report those accessors are an OpenCL 2.2 feature.
However, there is no "OpenCL C version 2.2", so change the check and
diagnostic text to 3.0 only.
A spurious `OpenCLVersion` argument was passed into the diagnostic;
remove that.
Differential Revision: https://reviews.llvm.org/D99969
This patch adds the memory operands for indexed loads so
that certain optimizations can take place.
Differential Revision: https://reviews.llvm.org/D100215/
Change-Id: I539fcf046ca4ad1e7df1d893f57d751419d8364d
Summary: The tags DW_LANG_C_plus_plus_14 and DW_LANG_C_plus_plus_11, introduced in Dwarf-5, are unexpected in previous versions. Fixing the mismathing doesn't have any drawbacks for any other debuggers, but helps dbx.
Reviewed By: aprantl, shchenz
Differential Revision: https://reviews.llvm.org/D99250
The function did not handle every case. In some cases this
caused assertion failure.
After the fix the function returns DependentTy if the exact
return type can not be determined.
It seems that clang itself does not call the function in the
affected cases but some checker or other code may call it.
Reviewed By: hokein
Differential Revision: https://reviews.llvm.org/D95244
We saw some big compiling time impact after enabling the debug entry value
feature for X86 platform(D73534). Compiling time goes from 900s->1600s with
our testcase. It is caused by allocating/freeing the memory busily.
'using FwdRegWorklist = MapVector<unsigned, SmallVector<FwdRegParamInfo, 2>>;'
The value for this map is vector, and we miss the reference when access the
element. The same happens for `auto CalleesMap = MF->getCallSitesInfo();` which is a DenseMap.
Reviewed by: djtodoro, flychen50
Differential Revision: https://reviews.llvm.org/D100162
Without the fix we get:
06:31:09 In file included from ../../clang-tools-extra/unittests/clang-tidy/ClangTidyDiagnosticConsumerTest.cpp:3:
06:31:09 ../utils/unittest/googletest/include/gtest/gtest.h:1392:11: error: comparison of integers of different signs: 'const int' and 'const unsigned int' [-Werror,-Wsign-compare]
06:31:09 if (lhs == rhs) {
06:31:09 ~~~ ^ ~~~
06:31:09 ../utils/unittest/googletest/include/gtest/gtest.h:1421:12: note: in instantiation of function template specialization 'testing::internal::CmpHelperEQ<int, unsigned int>' requested here
06:31:09 return CmpHelperEQ(lhs_expression, rhs_expression, lhs, rhs);
06:31:09 ^
06:31:09 ../../clang-tools-extra/unittests/clang-tidy/ClangTidyDiagnosticConsumerTest.cpp:60:3: note: in instantiation of function template specialization 'testing::internal::EqHelper<false>::Compare<int, unsigned int>' requested here
06:31:09 EXPECT_EQ(4, Errors[0].Message.FileOffset);
06:31:09 ^
06:31:09 ../utils/unittest/googletest/include/gtest/gtest.h:1924:63: note: expanded from macro 'EXPECT_EQ'
06:31:09 EqHelper<GTEST_IS_NULL_LITERAL_(val1)>::Compare, \
06:31:09 ^
06:31:09 ../utils/unittest/googletest/include/gtest/gtest.h:1392:11: error: comparison of integers of different signs: 'const int' and 'const unsigned long' [-Werror,-Wsign-compare]
06:31:09 if (lhs == rhs) {
06:31:09 ~~~ ^ ~~~
06:31:09 ../utils/unittest/googletest/include/gtest/gtest.h:1421:12: note: in instantiation of function template specialization 'testing::internal::CmpHelperEQ<int, unsigned long>' requested here
06:31:09 return CmpHelperEQ(lhs_expression, rhs_expression, lhs, rhs);
06:31:09 ^
06:31:09 ../../clang-tools-extra/unittests/clang-tidy/ClangTidyDiagnosticConsumerTest.cpp:64:3: note: in instantiation of function template specialization 'testing::internal::EqHelper<false>::Compare<int, unsigned long>' requested here
06:31:09 EXPECT_EQ(1, Errors[0].Message.Ranges.size());
06:31:09 ^
06:31:09 ../utils/unittest/googletest/include/gtest/gtest.h:1924:63: note: expanded from macro 'EXPECT_EQ'
06:31:09 EqHelper<GTEST_IS_NULL_LITERAL_(val1)>::Compare, \
06:31:09 ^
06:31:09 2 errors generated.
The current variable name isn't used anywhere else, which indicates it's
a typo. Let's fix it before someone copy+pastes it somewhere else.
Reviewed By: Jim
Differential Revision: https://reviews.llvm.org/D39157
Say we have
%1=min(%a,%b)
%2=min(%b,%c)
%3=min(%2,%a)
The optimization will try to reassociate the later one so that we can rewrite it to %3=min(%1, %c) and remove %2.
But if %2 has another uses outside of %3 then we can't remove %2 and end up with:
%1=min(%a,%b)
%2=min(%b,%c)
%3=min(%1, %c)
This doesn't harm by itself except it is not profitable and changes IR for no good reason.
What is bad it triggers next iteration which finds out that optimization is applicable to %2 and %3 and generates:
%1=min(%a,%b)
%2=min(%b,%c)
%3=min(%1,%c)
%4=min(%2,%a)
and so on...
The solution is to prevent optimization in the first place if intermediate result (%2) has side uses and
known to be not removed.
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D100170
The first one is the real parameters of the coroutine function, the
other one just for copying parameters to the coroutine frame.
Considering the following c++ code:
```
struct coro {
...
};
coro foo(struct test & t) {
...
co_await suspend_always();
...
co_await suspend_always();
...
co_await suspend_always();
}
int main(int argc, char *argv[]) {
auto c = foo(...);
c.handle.resume();
...
}
```
Function foo is the standard coroutine function, and it has only
one parameter named t (ignoring this at first),
when we use the llvm code to compile this function, we can get the
following ir:
```
!2921 = distinct !DISubprogram(name: "foo", linkageName:
"_ZN6Object3fooE4test", scope: !2211, file: !45, li\
ne: 48, type: !2329, scopeLine: 48, flags: DIFlagPrototyped |
DIFlagAllCallsDescribed, spFlags: DISPFlagDefi\
nition | DISPFlagOptimized, unit: !44, declaration: !2328,
retainedNodes: !2922)
!2924 = !DILocalVariable(name: "t", arg: 2, scope: !2921, file: !45,
line: 48, type: !838)
...
!2926 = !DILocalVariable(name: "t", scope: !2921, type: !838, flags:
DIFlagArtificial)
```
We can find there are two `the same` DIVariable named t in the same
dwarf scope for foo.resume.
And when we try to use llvm-dwarfdump to dump the dwarf info of this
elf, we get the following output:
```
0x00006684: DW_TAG_subprogram
DW_AT_low_pc (0x00000000004013a0)
DW_AT_high_pc (0x00000000004013a8)
DW_AT_frame_base (DW_OP_reg7 RSP)
DW_AT_object_pointer (0x0000669c)
DW_AT_GNU_all_call_sites (true)
DW_AT_specification (0x00005b5c "_ZN6Object3fooE4test")
0x000066a5: DW_TAG_formal_parameter
DW_AT_name ("t")
DW_AT_decl_file ("/disk1/yifeng.dongyifeng/my_code/llvm/build/bin/coro-debug-1.cpp")
DW_AT_decl_line (48)
DW_AT_type (0x00004146 "test")
0x000066ba: DW_TAG_variable
DW_AT_name ("t")
DW_AT_type (0x00004146 "test")
DW_AT_artificial (true)
```
The elf also has two 't' in the same scope.
But unluckily, it might let the debugger
confused. And failed to print parameters for O0 or above.
This patch will make coroutine parameters and move
parameters use the same DIVar and try to fix the problems
that I mentioned before.
Test Plan: check-clang
Reviewed By: aprantl, jmorse
Differential Revision: https://reviews.llvm.org/D97533
XSCMPUQP is not available for pre-P9 subtargets. This patch will lower
them into libcall for correct behavior on power7/power8.
Reviewed By: steven.zhang
Differential Revision: https://reviews.llvm.org/D92083
1. Redefine vpopc and vfirst IR intrinsic so it could adapt on
clang tablegen generator which always appends a type for vl
in IntrinsicType of clang codegen.
2. Remove `c` type transformer and add `u` and `l` for unsigned long
and long type.
Authored-by: Roger Ferrer Ibanez <rofirrim@gmail.com>
Co-Authored-by: Zakk Chen <zakk.chen@sifive.com>
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D100120
Support the following instructions.
1. Mask load and store
2. Vector Strided Instructions
3. Vector Indexed Store Instructions
Authored-by: Roger Ferrer Ibanez <rofirrim@gmail.com>
Co-Authored-by: Zakk Chen <zakk.chen@sifive.com>
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D99965
Kristof Beyls