Add option to `opt` to run the `ProfileInjectorPass` before the passes opt would run, and then `ProfileVerifierPass` after. This will then be a mode in which we run tests on a specialized buildbot, with the goal of finding passes that drop (and, later, corrupt) profile information.
Currently TargetLibraryInfo assumes ~everything is available, and
specific triples (or the -disable-builtin) flag, opt-out. This is a
backwards system, where features should only be positively be enabled
when known. Add this flag to help migrate tests with no triple.
This implements the result of the discussion at:
https://discourse.llvm.org/t/rfc-report-fatal-error-and-the-default-value-of-gencrashdialog/73587
There are two different use cases for report_fatal_error, so replace it
with two functions reportFatalInternalError() and
reportFatalUsageError(). The former indicates a bug in LLVM and
generates a crash dialog. The latter does not. The names have been
suggested by rnk and people seemed to like them.
This replaces a lot of the usages that passed an explicit value for
GenCrashDiag. I did not bulk replace remaining report_fatal_error usage
-- they probably require case by case review for which function to use.
This is needed in llvm-reduce to avoid perturbing the uselistorder in
intermediate steps. Really llvm-reduce wants pure serialization with
no dependency on the pass manager. There are other optimizations mixed
in to the serialization here depending on metadata in the module, which
is also bad.
Part of #63621
After https://github.com/llvm/llvm-project/pull/131217 was submitted,
time-passes.ll fails because `opt` prints `-time-report` when
`ManagedTimerGlobals` is destroyed. `ManagedTimerGlobals` stores
`TimerGroup`s in an unordered map, so the ordering of the output
`TimerGroup`s depends on the underlying iterator.
To fix this, we do what Clang does and use
`llvm::TimerGroup::printAll(...)`, which *is* deterministic. This is
also what Clang does. This does put move analysis section before the
pass section for `-time-report`, but again, this is also what Clang
currently does.
These date back to when the non-intrinsic format of variable locations
was still being tested and was behind a compile-time flag, so not all
builds / bots would correctly run them. The solution at the time, to get
at least some test coverage, was to have tests opt-in to non-intrinsic
debug-info if it was built into LLVM.
Nowadays, non-intrinsic format is the default and has been on for more
than a year, there's no need for this flag to exist.
(I've downgraded the flag from "try" to explicitly requesting
non-intrinsic format in some places, so that we can deal with tests that
are explicitly about non-intrinsic format in their own commit).
This is meant as a preparation for PR #130988 "[AMDGPU] Implement IR
expansion for frem instruction" which implements the expansion of
another instruction in this pass. The more general name seems more
appropriate given this change and quite reasonable even without it.
The module currently stores the target triple as a string. This means
that any code that wants to actually use the triple first has to
instantiate a Triple, which is somewhat expensive. The change in #121652
caused a moderate compile-time regression due to this. While it would be
easy enough to work around, I think that architecturally, it makes more
sense to store the parsed Triple in the module, so that it can always be
directly queried.
For this change, I've opted not to add any magic conversions between
std::string and Triple for backwards-compatibilty purses, and instead
write out needed Triple()s or str()s explicitly. This is because I think
a decent number of them should be changed to work on Triple as well, to
avoid unnecessary conversions back and forth.
The only interesting part in this patch is that the default triple is
Triple("") instead of Triple() to preserve existing behavior. The former
defaults to using the ELF object format instead of unknown object
format. We should fix that as well.
Added an extension point after vectorizer passes in the PassBuilder.
Additionally, added extension points before and after vectorizer passes
in `buildLTODefaultPipeline`. Credit goes to @mshockwave for guiding me
through my first LLVM contribution (and my first open source
contribution in general!) :)
- Implemented `registerVectorizerEndEPCallback`
- Implemented `invokeVectorizerEndEPCallbacks`
- Added `VectorizerEndEPCallbacks` SmallVector
- Added a command line option `passes-ep-vectorizer-end` to
`NewPMDriver.cpp`
- `buildModuleOptimizationPipeline` now calls
`invokeVectorizerEndEPCallbacks`
- `buildO0DefaultPipeline` now calls `invokeVectorizerEndEPCallbacks`
- `buildLTODefaultPipeline` now calls BOTH
`invokeVectorizerStartEPCallbacks` and `invokeVectorizerEndEPCallbacks`
- Added LIT tests to `new-pm-defaults.ll`, `new-pm-lto-defaults.ll`,
`new-pm-O0-ep-callbacks.ll`, and `pass-pipeline-parsing.ll`
- Renamed `CHECK-EP-Peephole` to `CHECK-EP-PEEPHOLE` in
`new-pm-lto-defaults.ll` for consistency.
This code is intended for developers that wish to implement and run
custom passes after the vectorizer passes in the PassBuilder pipeline.
For example, in #91796, a pass was created that changed the induction
variables of vectorized code. This is right after the vectorization
passes.
Following discussions in #110443, and the following earlier discussions
in https://lists.llvm.org/pipermail/llvm-dev/2017-October/117907.html,
https://reviews.llvm.org/D38482, https://reviews.llvm.org/D38489, this
PR attempts to overhaul the `TargetMachine` and `LLVMTargetMachine`
interface classes. More specifically:
1. Makes `TargetMachine` the only class implemented under
`TargetMachine.h` in the `Target` library.
2. `TargetMachine` contains target-specific interface functions that
relate to IR/CodeGen/MC constructs, whereas before (at least on paper)
it was supposed to have only IR/MC constructs. Any Target that doesn't
want to use the independent code generator simply does not implement
them, and returns either `false` or `nullptr`.
3. Renames `LLVMTargetMachine` to `CodeGenCommonTMImpl`. This renaming
aims to make the purpose of `LLVMTargetMachine` clearer. Its interface
was moved under the CodeGen library, to further emphasis its usage in
Targets that use CodeGen directly.
4. Makes `TargetMachine` the only interface used across LLVM and its
projects. With these changes, `CodeGenCommonTMImpl` is simply a set of
shared function implementations of `TargetMachine`, and CodeGen users
don't need to static cast to `LLVMTargetMachine` every time they need a
CodeGen-specific feature of the `TargetMachine`.
5. More importantly, does not change any requirements regarding library
linking.
cc @arsenm @aeubanks
The early simplication pipeline is used in non-LTO and (Thin/Full)LTO
pre-link
stage. There are some passes that we want them in non-LTO mode, but not
at LTO
pre-link stage. The control is missing currently. This PR adds the
support. To
demonstrate the use, we only enable the internalization pass in non-LTO
mode for
AMDGPU because having it run in pre-link stage causes some issues.
As discussed in this
[proposal](https://github.com/llvm/wg-hlsl/pull/62/files?short_path=ac6e592#diff-ac6e59276afe8016e307eedc5c835f534c0cb353707760b44df0fa9d905a5cf8).
We had to bring back the legacy pass manager interface for the
scalarizer pass. Two reasons for this:
1. The DirectX backend is still using the legacy pass manager
2. The new PM isn't hooked up in clang yet via `BackendUtil.cpp`'s
`AddEmitPasses` That means even if we add a `buildCodeGenPipeline` we
won't be able to benefit from the new pass manager's scalarizer pass
interface.
The remaining changes are hooking up the scalarizer pass to the DirectX
backend, updating the DirectX test cases,
and allowing the `optdriver` to not block the legacy invocation of the
scalarizer pass.
Future work still needs to be done to allow the scalarizer pass to
handle target specific intrinsics.
closes#105178
Building with -DLLVM_ENABLE_EXPORTED_SYMBOLS_IN_EXECUTABLES=Off should not
prevent use of opt plugins.
This fix uses the approach implemented in
https://github.com/llvm/llvm-project/pull/101741.
rdar://135841478
* Move the setup_host_tool calls to the directories of their tool.
Although it works to call it in libclc, it can only appear in a single
location so it fails the "what if everyone did this?" test and causes
problems for downstream code that also wants to use native versions of
these tools from other projects.
* Correct the TARGET "${${tool}_target}" check. "${${tool}_target}" may
be set to the path to the executable, which works in dependencies but
cannot be tested using if(TARGET). For lack of a better alternative,
just check that "${${tool}_target}" is non-empty and trust that if it
is, it is set to a meaningful value. If somehow it turns out to be a
valid target, its value will still show up in error messages anyway.
* Account for llvm-spirv possibly being provided in-tree. Per
https://github.com/KhronosGroup/SPIRV-LLVM-Translator?tab=readme-ov-file#llvm-in-tree-build
it is possible to drop llvm-spirv into LLVM and have it built as part of
LLVM's build. In this configuration, cross builds of LLVM require a
native version of llvm-spirv to be built.
Fix the builds with LLVM_TOOL_LLVM_DRIVER_BUILD enabled.
LLVM_ENABLE_EXPORTED_SYMBOLS_IN_EXECUTABLES is not completely
compatible with export_executable_symbols as the later will be ignored
if the previous is set to NO.
Fix the issue by passing if symbols need to be exported to
llvm_add_exectuable so the link flag can be determined directly
without calling export_executable_symbols_* later.
`LLVM_ENABLE_EXPORTED_SYMBOLS_IN_EXECUTABLES` is not completely
compatible with `export_executable_symbols` as the later will be ignored
if the previous is set to NO.
Fix the issue by passing if symbols need to be exported to
`llvm_add_exectuable` so the link flag can be determined directly
without calling `export_executable_symbols_*` later.
This patch changes all single quotes in the help text to double quotes
for the sake of cross platform correctness.
The single quotes previously used in the help message are correct for
bash and most other shells but aren't correct for the Windows command
prompt as it only interprets double quotes in the manner intended in the
help message.
This patch was prompted by the fact that if you aren't aware of the
aformentioned quirk the single quotes result in very confusing error
messages from the command prompt.
SmallPtrSet.h and TimeProfiler.h are unused. CommandLine.h is only
needed for the UseNewDbgInfoFormat declare, which can be moved to the
places that need it.
A std::vector causes a heap allocation and a memcpy of static
initialization data from the rodata section.
Use a static array instead so we can use the static data directly.
The performance of cold functions shouldn't matter too much, so if we
care about binary sizes, add an option to mark cold functions as
optsize/minsize for binary size, or optnone for compile times [1]. Clang
patch will be in a future patch.
This is intended to replace `shouldOptimizeForSize(Function&, ...)`.
We've seen multiple cases where calls to this expensive function, if not
careful, can blow up compile times. I will clean up users of that
function in a followup patch.
Initial version: https://reviews.llvm.org/D149800
[1]
https://discourse.llvm.org/t/rfc-new-feature-proposal-de-optimizing-cold-functions-using-pgo-info/56388
A colleague observes that switching the default value of
LLVM_EXPERIMENTAL_DEBUGINFO_ITERATORS to "On" hasn't flipped the value
in their CMakeCache.txt. This probably means that everyone with an
existing build tree is going to not have support built in, meaning
everyone in LLVM would need to clean+rebuild their worktree when we flip
the switch on... which doesn't sound good.
So instead, just delete the flag and everything it does, making everyone
build and run ~400 lit tests in RemoveDIs mode. None of the buildbots
have had trouble with this, so it Should Be Fine (TM).
(Sending for review as this is changing various comments, and touches
several different areas -- I don't want to get too punchy).
In Bazel, Clang current separates the clang executable into a
clang-driver library, and the actual clang executable. This allows
downstream users to make their own variations of clang, without having
to redo/maintain separate build pipelines.
This adds the same for opt for both CMake and Bazel.
This flag (--try-experimental-debuginfo-iterators) only exists for testing
purposes, to get some RUNlines running in new-debug-info mode before it's
properly supported. The flag isn't something that's going to be useful to
people using llvm 18, so hide it from the options list.
Port CodeGenPrepare to new pass manager and dependency
BasicBlockSectionsProfileReader
Fixes: #75380
Co-authored-by: Krishna-13-cyber <84722531+Krishna-13-cyber@users.noreply.github.com>
Revert e0c554ad87d18dcbfcb9b6485d0da800ae1338d1 "Port CodeGenPrepare to new pass manager (and BasicBlockSectionsProfil… (#75380)"
Revert #75380 and #77054 as they were breaking EXPENSIVE_CHECKS buildbots: https://lab.llvm.org/buildbot/#/builders/104
Port CodeGenPrepare to new pass manager and dependency
BasicBlockSectionsProfileReader
Fixes: #64560
Co-authored-by: Krishna-13-cyber <84722531+Krishna-13-cyber@users.noreply.github.com>
This patch replaces uses of StringRef::{starts,ends}with with
StringRef::{starts,ends}_with for consistency with
std::{string,string_view}::{starts,ends}_with in C++20.
I'm planning to deprecate and eventually remove
StringRef::{starts,ends}with.
Our option to turn on the non-intrinsic form of debug-info
(`--experimental-debuginfo-iterators`) currently requires that LLVM is
built with the `LLVM_EXPERIMENTAL_DEBUGINFO_ITERATORS` cmake flag
enabled, so that some (slight) performance regressions aren't
on-by-default during the prototype/testing period. However, we still
want to be able to _optionally_ run tests, if support is built into
LLVM.
To allow optionally exercising the non-intrinsic debug-info code, this
patch adds `--try-experimental-debuginfo-iterators` to opt, which turns
the `--experimental-debuginfo-iterators` flag on if support is built in,
or leaves it off. This means we can run tests that:
* Use normal dbg.value intrinsics if there's no support, or
* Uses non-instruction DPValues if there is support.
Which means we can start getting test coverage of DPValues/RemoveDIs
behaviour, from in-tree tests, on our RemoveDIs buildbot. All the code
to do with automagically converting from one form to the other landed in
10a9e7442c4c4e.
There are many tests that specify a target triple/CPU flags but no
DataLayout which can lead to IR being generated that has unusual
behaviour. This commit attempts to use the default DataLayout based
on the relevant flags if there is no explicit override on the command
line or in the IR file.
One thing that is not currently possible to differentiate from a missing
datalayout `target datalayout = ""` in the IR file since the current
APIs don't allow detecting this case. If it is considered useful to
support this case (instead of passing "-data-layout=" on the command
line), I can change IR parsers to track whether they have seen such a
directive and change the callback type.
Differential Revision: https://reviews.llvm.org/D141060
