As noted when #82404 was pushed (canonicalizing `sitofp` -> `uitofp`),
different signedness on fp casts can have dramatic performance
implications on different backends.
So, it makes to create a reliable means for the backend to pick its
cast signedness if either are correct.
Further, this allows us to start canonicalizing `sitofp`- > `uitofp`
which may easy middle end analysis.
Closes#86141
Allow using atomicrmw fadd, fsub, fmin, and fmax with vectors of
floating-point type. AMDGPU supports atomic fadd for <2 x half> and <2 x
bfloat> on some targets and address spaces.
Note this only supports the proper floating-point operations; float
vector typed xchg is still not supported. cmpxchg still only supports
integers, so this inserts bitcasts for the loop expansion.
I have support for fp vector typed xchg, and vector of int/ptr
separately implemented but I don't have an immediate need for those
beyond feature consistency.
Building the Apple way turns off plugin support, meaning we don't need
to export unloadable symbols from all executables. While deadstripping
effects aren't expected to change, enabling this across all tools
prevents the creation of export tries. This saves us ~3.5 MB in just the
universal build of `clang`.
-V prints the version information in both BSD and GNU ar/ranlib.
BSD ranlib rejects -v while -v enables verbose output in GNU ar but is
another alias for --version in GNU ranlib. The GNU ranlib behavior is
inconsistent: `ranlib -v` is different from `ar -sv`. But it's not a
major concern in practice:
* Users typically use ranlib solely for creating archive symbol tables,
and they don't need verbose output.
* Verbose output in ranlib seems a no-op.
* GNU ar creates an archive symbol table by default. Many ranlib uses
have been eliminated.
* Modern linkers like lld/ELF (since version 14) and mold don't rely on
archive symbol tables anymore.
https://reviews.llvm.org/D71554 introduced -v. This patch removes it so
that `llvm-ranlib -v` and `llvm-ranlib -version` lead to errors (GNU
ranlib rejects `-version` as well). -V is added as an alias for
--version.
Close#87654
Pull Request: https://github.com/llvm/llvm-project/pull/87661
This adds support for using the L and H argument modifiers for twinword
operands in inline asm code, such as in:
```
%1 = tail call i64 asm sideeffect "rd %pc, ${0:L} ; srlx ${0:L}, 32, ${0:H}", "={o4}"()
```
This is needed by the Linux kernel.
--compress-sections is similar to --compress-debug-sections but applies
to arbitrary sections.
* `--compress-sections <section>=none`: decompress sections
* `--compress-sections <section>=[zlib|zstd]`: compress sections with zlib/zstd
Like `--remove-section`, the pattern is by default a glob, but a regex
when --regex is specified.
For `--remove-section` like options, `!` prevents matches and is not
dependent on ordering (see `ELF/wildcard-syntax.test`). Since
`--compress-sections a=zlib --compress-sections a=none` naturally allows
overriding, having an order-independent `!` would be confusing.
Therefore, `!` is disallowed.
Sections within a segment are effectively immutable. Report an error for
an attempt to (de)compress them. `SHF_ALLOC` sections in a relocatable
file can be compressed, but linkers usually reject them.
Link: https://discourse.llvm.org/t/rfc-compress-arbitrary-sections-with-ld-lld-compress-sections/71674
Pull Request: https://github.com/llvm/llvm-project/pull/85036
The goal is to have ability to change logic compile time based on PGO
data.
Our primary application is removing UBSAN checks from hot code.
Then we'd like to use this for libc++ hardening and regular debug
asserts.
Previous attempt is #84214.
Benefits from special intrinsic vs #84214:
1. Resulting binary is 3% faster than removing traps (on
"test-suite/MultiSource/Benchmarks" with PGO+ThinLTO)
2. Intrinsic can be used from source code to change behavior from C/C++
program. E.g. enabling asserts in cold code.
3. Easier to match basic blocks.
RFC:
https://discourse.llvm.org/t/rfc-add-llvm-experimental-hot-intrinsic-or-llvm-hot/77641
---------
Co-authored-by: Nikita Popov <npopov@redhat.com>
General cleanup in LangRef (and two outdated comments in LLParser.cpp)
with the aim of making it easier to understand some of the terminology
and subtle idiosyncrasies related to metadata in the IR.
I'm still not happy with the fact that "node" is used both informally
and with a particular category of metadata in mind, depending on the
context. This also bleeds into the type names in the implementation.
There are also several places where names from the implementation appear
in the document with no other context or definition. In some cases I
added a parenthetical to section titles to tie the two together, but I
don't think this is ideal.
I also think it might be useful to define the "abstract" metadata
classes like "DIScope" in the document, so the hierarchy of metadata
node kinds is direct, and so we can avoid repetitive descriptions of all
of the members of on part of the hierarchy. This inheritance doesn't
have to be in terms of C++ classes, but using the same names as the
implementation seems helpful, and we already do it for many other
things.
Finally I added sections for the specialized nodes which are implemented
in the IR but didn't have documentation in LangRef yet. These could use
some work, and I admit I didn't dig too deep into the specifics beyond
enumerating the fields, but I think we would ideally always have a
LangRef section for every kind of node.
Follow on from #84915 which adds the DbgRecord function variants. The C API
changes were reviewed in #85657.
# C API
Update the LLVMDIBuilderInsert... functions to insert DbgRecords instead
of debug intrinsics.
LLVMDIBuilderInsertDeclareBefore
LLVMDIBuilderInsertDeclareAtEnd
LLVMDIBuilderInsertDbgValueBefore
LLVMDIBuilderInsertDbgValueAtEnd
Calling these functions will now cause an assertion if the module is in the
wrong debug info format. They should only be used when the module is in "new
debug format".
Use LLVMIsNewDbgInfoFormat to query and LLVMSetIsNewDbgInfoFormat to change the
debug info format of a module.
Please see https://llvm.org/docs/RemoveDIsDebugInfo.html#c-api-change
(RemoveDIsDebugInfo.md) for more info.
# OCaml bindings
Add set_is_new_dbg_info_format and is_new_dbg_info_format to the OCaml bindings.
These can be used to set and query the current debug info mode. These will
eventually be removed, but are useful while we're transitioning between old and
new debug info formats.
Add string_of_lldbgrecord, like string_of_llvalue but prints DbgRecords.
In test dbginfo.ml, unconditionally set the module debug info to the new mode
and update CHECK lines to check for DbgRecords. Without this change the test
crashes because it attempts to insert DbgRecords (new default behaviour of
llvm_dibuild_insert_declare_...) into a module that is in the old debug info
mode.
Currently, the builtins used for implementing `va_list` handling
unconditionally take their arguments as unqualified `ptr`s i.e. pointers
to AS 0. This does not work for targets where the default AS is not 0 or
AS 0 is not a viable AS (for example, a target might choose 0 to
represent the constant address space). This patch changes the builtins'
signature to take generic `anyptr` args, which corrects this issue. It
is noisy due to the number of tests affected. A test for an upstream
target which does not use 0 as its default AS (SPIRV for HIP device
compilations) is added as well.
When `+sve` is passed in the command line, if the Architecture being
targeted is V8.6A/V9.1A or later, `+f32mm` is also added. This enables
FEAT_32MM, however at the time of writing no CPU's support this. This
leads to the FEAT_32MM instructions being compiled for CPU's that do not
support them.
This commit removes the automatic enablement, however the option is
still able to be used by passing `+f32mm`.
This change:
- Updates the existing Clang User's Manual section on SPGO so that it
describes how to use llvm-profgen to perform SPGO on Windows. This is
new functionality implemented in #83972.
- Fixes a minor typo in the existing llvm-profgen invocation example.
- Adds an LLVM release note on this new functionality in llvm-profgen.
Here we introduce three new GMIR instructions to cover a set of trap
intrinsics. The idea behind it is that generic intrinsics shouldn't be
used with G_INTRINSIC opcode.
These new instructions can match perfectly with existing trap ISD nodes.
It allows X86, AArch64, RISCV and Mips to reuse SelectionDAG patterns for
selection and avoid manual selection. However AMDGPU is an exception. It
selects traps during legalization regardless SelectionDAG or GlobalISel.
Since there are not many places where traps are used, this change
attempts to clean up all the usages of G_INTRINSIC with trap intrinsics. So,
there is no stage when both G_TRAP and
G_INTRINSIC_W_SIDE_EFFECTS(@llvm.trap) are allowed.
Follow on from #84915 which adds the DbgRecord function variants.
Update the LLVMDIBuilderInsert... functions to insert DbgRecords instead
of debug intrinsics.
LLVMDIBuilderInsertDeclareBefore
LLVMDIBuilderInsertDeclareAtEnd
LLVMDIBuilderInsertDbgValueBefore
LLVMDIBuilderInsertDbgValueAtEnd
Calling these functions will now cause an assertion if the module is in the
wrong debug info format. They should only be used when the module is in "new
debug format".
Use LLVMIsNewDbgInfoFormat to query and LLVMSetIsNewDbgInfoFormat to change the
debug info format of a module.
Please see https://llvm.org/docs/RemoveDIsDebugInfo.html#c-api-change
(RemoveDIsDebugInfo.md) for more info.
Add --skip-symbol and --skip-symbols options that allow to skip symbols
when executing other options that can change the symbol's name, binding
or visibility, similar to an existing option --keep-symbol that keeps a
symbol from being removed by other options.
SIMachineFunctionInfo has a scan of the function body for inline asm
which may use AGPRs, or callees in SIMachineFunctionInfo. Move this
into the attributor, so it actually works interprocedurally.
Could probably avoid most of the test churn if this bothered to avoid
adding this on subtargets without AGPRs. We should also probably
try to delete the MIR scan in usesAGPRs but it seems to be trickier
to eliminate.
Another trivial rename patch, the last big one for now, which renamed
DPMarkers to DbgMarkers. This required the field `DbgMarker` in
`Instruction` to be renamed to `DebugMarker` to avoid a clash, but
otherwise was a simple string substitution of `s/DPMarker/DbgMarker` and
a manual renaming of `DPM` to `DM` in the few places where that acronym
was used for debug markers.
This patch renames DPLabel to DbgLabelRecord, in accordance with the
ongoing DbgRecord rename. This rename was fairly trivial, since DPLabel
isn't as widely used as DPValue and has no real conflicts in either its
full or abbreviated name. As usual, the entire replacement was done
automatically, with `s/DPLabel/DbgLabelRecord/` and `s/DPL/DLR/`.
As part of the migration to ptradd
(https://discourse.llvm.org/t/rfc-replacing-getelementptr-with-ptradd/68699),
we need to change the representation of the `inrange` attribute, which
is used for vtable splitting.
Currently, inrange is specified as follows:
```
getelementptr inbounds ({ [4 x ptr], [4 x ptr] }, ptr @vt, i64 0, inrange i32 1, i64 2)
```
The `inrange` is placed on a GEP index, and all accesses must be "in
range" of that index. The new representation is as follows:
```
getelementptr inbounds inrange(-16, 16) ({ [4 x ptr], [4 x ptr] }, ptr @vt, i64 0, i32 1, i64 2)
```
This specifies which offsets are "in range" of the GEP result. The new
representation will continue working when canonicalizing to ptradd
representation:
```
getelementptr inbounds inrange(-16, 16) (i8, ptr @vt, i64 48)
```
The inrange offsets are relative to the return value of the GEP. An
alternative design could make them relative to the source pointer
instead. The result-relative format was chosen on the off-chance that we
want to extend support to non-constant GEPs in the future, in which case
this variant is more expressive.
This implementation "upgrades" the old inrange representation in bitcode
by simply dropping it. This is a very niche feature, and I don't think
trying to upgrade it is worthwhile. Let me know if you disagree.
This is the major rename patch that prior patches have built towards.
The DPValue class is being renamed to DbgVariableRecord, which reflects
the updated terminology for the "final" implementation of the RemoveDI
feature. This is a pure string substitution + clang-format patch. The
only manual component of this patch was determining where to perform
these string substitutions: `DPValue` and `DPV` are almost exclusively
used for DbgRecords, *except* for:
- llvm/lib/target, where 'DP' is used to mean double-precision, and so
appears as part of .td files and in variable names. NB: There is a
single existing use of `DPValue` here that refers to debug info, which
I've manually updated.
- llvm/tools/gold, where 'LDPV' is used as a prefix for symbol
visibility enums.
Outside of these places, I've applied several basic string
substitutions, with the intent that they only affect DbgRecord-related
identifiers; I've checked them as I went through to verify this, with
reasonable confidence that there are no unintended changes that slipped
through the cracks. The substitutions applied are all case-sensitive,
and are applied in the order shown:
```
DPValue -> DbgVariableRecord
DPVal -> DbgVarRec
DPV -> DVR
```
Following the previous rename patches, it should be the case that there
are no instances of any of these strings that are meant to refer to the
general case of DbgRecords, or anything other than the DPValue class.
The idea behind this patch is therefore that pure string substitution is
correct in all cases as long as these assumptions hold.
Remove support for obfuscated bitcode in dsymutil and the DWARF linker.
We no longer support bitcode submissions and the obfuscation support has
been removed from the rest of the compiler.
rdar://123863918
Remove support for obfuscated bitcode in dsymutil and the DWARF linker.
We no longer support bitcode submissions and the obfuscation support has
been removed from the rest of the compiler.
rdar://123863918
Follow on from #84739, which updates the DIBuilder class.
All the functions that have been added are temporary and will be
deprecated in the future. The intention is that they'll help downstream
projects adapt during the transition period.
```
New functions (all to be deprecated)
------------------------------------
LLVMIsNewDbgInfoFormat # Returns true if the module is in the new non-instruction mode.
LLVMSetIsNewDbgInfoFormat # Convert to the requested debug info format.
LLVMDIBuilderInsertDeclareIntrinsicBefore # Insert a debug intrinsic (old debug info format).
LLVMDIBuilderInsertDeclareIntrinsicAtEnd # Same as above.
LLVMDIBuilderInsertDbgValueIntrinsicBefore # Same as above.
LLVMDIBuilderInsertDbgValueIntrinsicAtEnd # Same as above.
LLVMDIBuilderInsertDeclareRecordBefore # Insert a debug record (new debug info format).
LLVMDIBuilderInsertDeclareRecordAtEnd # Same as above.
LLVMDIBuilderInsertDbgValueRecordBefore # Same as above.
LLVMDIBuilderInsertDbgValueRecordAtEnd # Same as above.
```
The existing `LLVMDIBuilderInsert...` functions call through to the
intrinsic versions (old debug info format) currently.
In the next patch, I'll swap them to call the debug records versions
(new debug info format). Downstream users of this API can query and
change the current format using the first two functions above, or can
instead opt to temporarily use intrinsics or records explicitly.
As part of the WebAssembly support work
https://github.com/llvm/llvm-project/pull/82588
As the object files used in the test cases are a single object (just
produced by clang without being processed by wasm-ld), it was determined
to use .o intead of .wasm.
Update the README.txt to reflect that the tool now supports WebAssembly.
