This patch adds the definition of the LASX instructions, providing
support only for assembly and disassembly, similar to D154183.

Depends on D154183

Reviewed By: SixWeining

Differential Revision: https://reviews.llvm.org/D154195

This patch adds the definition for the `LSX` registers and instructions.
It also adds handling for new immediate operands in the AsmParser. This
patch ensures that llvm-mc and llvm-objdump correctly handle the `LSX`
instructions.

We expand those pseudo-instructions `vrepli.{b,h,w,d}` in the
MCCodeEmitter. This increases the readability of the output when
generating assembly files.

Reviewed By: SixWeining

Differential Revision: https://reviews.llvm.org/D154183

Left some includes around from debugging the last patch that I forgot to
take out, so here's the patch taking them out.

This patch switches from using PTRACE_SEIZE within the subprocess
benchmark runner for llvm-exegesis as PTRACE_SEIZE was introduced in
Linux kernel 3.4. Some LLVM users were reporting build failures as they
are using Kernel versions older than 3.4 (such as on CentOS/RHEL 6),
hence the patch.

Having the SubprocessMemory class available on Android was causing build
failures in downstream builds as Android doesn't implement the SystemV
IPC specification that supplies shm_open and other functions that the
class relies on. This patch simply makes it unavailable on Android using
preprocessor directives.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D154662

This reverts commit c3192196.

Reverting my second attempt at https://reviews.llvm.org/D139453
changing which dyld notification method is being used.
The Intel macOS CI bot is still failing with this
rewrite at https://green.lab.llvm.org/green/view/LLDB/job/lldb-cmake/
I'll need to set up an Intel macOS system running a matching
OS version to debug this directly I think.

There are a number of more test combinations here that
can be done together and reduce the number of instructions.

https://reviews.llvm.org/D143191

This reverts commit 09feee55.

Revert because of a testbot failure:
  https://lab.llvm.org/buildbot/#/builders/5/builds/34931

On Darwin systems, the dynamic linker dyld has an empty function
it calls when binaries are added/removed from the process.  lldb puts
a breakpoint on this dyld function to catch the notifications.  The
function arguments are used by lldb to tell what is happening.

The linker has a natural representation when the addresses of
binaries being added/removed are in the pointer size of the process.
There is then a second function where the addresses of the binaries
are in a uint64_t array, which the debugger has been using before -
dyld allocates memory for the array, copies the values in to it,
and calls it for lldb's benefit.

This changes to using the native notifier function, with pointer-sized
addresses.

This is the second time landing this change; this time correct the
size of the image_count argument, and add a fallback if the
notification function "lldb_image_notifier" can't be found.

Differential Revision: https://reviews.llvm.org/D139453

In OrcAArch64::writeIndirectStubsBlock, masks the high bits of the immediate
operand to the stub's ldr instruction so that negative offsets to the stub
pointer do not overflow.

No testcase -- this fixes most of the OrcLazy testcases for AArch64 (at least on
Darwin), but we still need to fix the exception-handling test before we can turn
them on.

With different linker, the offset of strings can change.

Reviewed By: maksfb

Differential Revision: https://reviews.llvm.org/D154670

With different linker, the offset of strings can change.

Reviewed By: maksfb

Differential Revision: https://reviews.llvm.org/D154669

I've temporarily reverted D154272 (which this commit depends on)
due to a build breakage, so also reverting this for now.

This reverts commit c1e28385.

Similar to what we do for orc.b. Another patch will expose this
as a builtin in clang.

This reverts my commit 015dabd7
due to breaking non-glibc builds.

Fixed invalid assembly, where references were not correct.

Reviewed By: rafauler

Differential Revision: https://reviews.llvm.org/D154667

This reverts commit a4a26374.

This was causing some problems with the CPU build and CUDA buildbot.
Revert until I can figure out what those issues are and fix them. I
believe it is just some CMake.

Converting between a complex<f32> to i64 could be useful for handling interop
between the `arith` and `complex` dialects.

Reviewed By: jpienaar

Differential Revision: https://reviews.llvm.org/D154663

Also, make it possible for new Targets which haven't been added to
the TargetList yet to check for interruption, and add a few more
places in building modules where we can check for interruption.

Differential Revision: https://reviews.llvm.org/D154542

This is an alternate approach to the patches proposed in D153897 and
D153794. Rather than exporting a single header that can be included on
the GPU in all circumstances, this patch chooses to instead generate a
separate set of headers that only provides the declarations. This can
then be used by external tooling to set up what's on the GPU. This
leaves room for header hacks for offloading languages without needing to
worry about the `libc` implementation.

Currently this generates a set of headers that only contain the
declarations. These will then be installed to a new clang resource
directory called `llvm_libc_wrappers/` which will house the shim code.
We can then automaticlaly include this from `clang` when offloading to
wrap around the headers while specifying what's on the GPU.

Reviewed By: jdoerfert, JonChesterfield

Differential Revision: https://reviews.llvm.org/D154036

Since the LMUL data that is needed to create an instrument is
avaliable statically from vsetivli and vsetvli instructions,
LMUL instruments can be automatically generated so that clients
of the tool do no need to manually insert instrument comments.

Instrument comments may be placed after a vset{i}vli instruction,
which will override instrument that was automatically inserted.
As a result, clients of llvm-mca instruments do not need to update
their existing instrument comments. However, if the instrument
has the same LMUL as the vset{i}vli, then it is reccomended to
remove the instrument comment as it becomes redundant.

Differential Revision: https://reviews.llvm.org/D154526

This precommit patch shows MaxSafeDepBytesDist is 24 when it should
be 8.

Differential Revision: https://reviews.llvm.org/D154173

This patch attempts to fix a dead lock when loading modules in a Scripted
Process.

This issue was triggered by loading the modules after the process did resume,
but before the process actually stop, causing the language runtime mutex to
be locked by a separate thread, responsible to unwind the stack (using
the runtime unwind plan), while the module loading thread was trying to
notify the runtimes of the newly loaded module.

To address that, this patch moves the module loading logic to be done before
sending the stop event, to prevent the dead lock situation described above.

Differential Revision: https://reviews.llvm.org/D154649



Signed-off-by: Med Ismail Bennani <ismail@bennani.ma>

Summary:
To reduce memory footprint changed so that we process and write out TUs first,
reset DIEBuilder and process CUs. CUs are processed in buckets. First bucket
contains all the CUs with cross CU references. Rest processd one at a time.

clang-17 build in debug mode, by clang-17.
before
8:25.81 real, 834.37 user, 86.03 sys, 0 amem, 79525064 mmem
8:02.20 real, 820.46 user, 81.81 sys, 0 amem, 79501616 mmem
7:52.69 real, 802.01 user, 83.99 sys, 0 amem, 79534392 mmem

after
7:49.35 real, 822.04 user, 66.19 sys, 0 amem, 34934260 mmem
7:42.16 real, 825.46 user, 63.52 sys, 0 amem, 34951660 mmem
7:46.71 real, 821.11 user, 63.14 sys, 0 amem, 34981164 mmem

Differential Revision: https://phabricator.intern.facebook.com/D45883198

Summary:
Changed how we handle writing out .dwo and .dwp files. We now write out DWO
sections sooner and destroy DIEBuilder. This should decrease memory footprint.

Ran on clang-17 build in debug mode with split-dwarf.
before
8:07.49 real,   664.62 user,    69.00 sys,      0 amem, 41601612 mmem
8:07.06 real,   669.60 user,    68.75 sys,      0 amem, 41822588 mmem
8:00.36 real,   664.14 user,    66.36 sys,      0 amem, 41561548 mmem

after
8:21.85 real,   682.23 user,    69.64 sys,      0 amem, 39379880 mmem
8:04.58 real,   671.62 user,    66.50 sys,      0 amem, 39735800 mmem
8:10.02 real,   680.67 user,    67.24 sys,       0 amem, 39662888 mmem

Differential Revision: https://phabricator.intern.facebook.com/D45458889

Summary:

* Some cleanup and minor fixes for the new debug information re-writer before moving on
to productatization.

* The new rewriter wasn't handling binary with DWARF5 and DWARF4 with
-fdebug-types-sections.

* Removed dead cross cu reference code.

* Added support for DW_AT_sibling.

* With the new re-writer abbrev number can change which can lead to offset of Type
Units changing. Before we would just copy raw data. Changed to write out Type
Unit List. This is generated by gdb-add-index.

* Fixed how bolt handles gdb-index generated by gdb-11 with types sections.
Simplified logic that handles variations of gdb-index.

* Clang can generate two type units with the same hash, but different content. LLD
does not de-duplicate when ThinLTO is involved. Changed so that TU hash and
offset are used to make TU's unique.

* It is possible to have references within location expression to another DIE.
Fixed it so that relative offset is updated correctly.

* Removed all the code related to patching.

* Removed dead code. Changed how we handling writting out TUs and TU Index. It now
  should fully work for DWARF4 and DWARF5.

* Removed unused arguments from some APIs, changed return type to void, and other
small cleanups.

Test Plan:

Reviewers:

Subscribers:

Tasks:

Tags:

Differential Revision: https://phabricator.intern.facebook.com/D46168257

Summary:
This revision implement new mechanism for DWARFRewriter.
In the new mechanism, we adopt the same way with DWARFLinker did.
By parsing Debug information into IR, we are allowed to handle debug information more flexible.
Now the debug information updating process relies on IR and IR will be written out to binary once the updating finished.

A new class was added: DIEBuilder. This class is responsible for parsing debug information and raising it to the IR level.
This class is also used to write out the .debug_info and .debug_abbrev sections.
Since we output brand new Abbrev section we won't need to always convert low_pc/high_pc into ranges.
When conversion does happen we can also remove low_pc entry.

Differential Revision: https://phabricator.intern.facebook.com/D39484421

Tasks: T117448832

* Relax the AsmParser to accept `.amdhsa_wavefront_size32 0` when the
  `.amdhsa_shared_vgpr_count` directive is present.
* Teach the KD disassembler to respect the setting of
  KERNEL_CODE_PROPERTY_ENABLE_WAVEFRONT_SIZE32 when calculating the
  value of `.amdhsa_next_free_vgpr`.
* Teach the KD disassembler to disassemble COMPUTE_PGM_RSRC3 for gfx90a
  and gfx10+.
* Include "pseudo directive" comments for gfx10 fields which are not
  controlled by any assembler directive.
* Fix disassembleObject failure diagnostic in llvm-objdump to not
  hard-code a comment string, and to follow the convention of not
  capitalizing the first sentence.

Reviewed By: rochauha

Differential Revision: https://reviews.llvm.org/D128014

When LLDB queries the debug names index with a regex, we should use the
`Mangled` class wrapper, which attempts to match regex first against the mangled
name and then against the demangled name. It is important to do so, otherwise
queries like `frame var --regex A::` would never work. This is what is done for
the Apple index as well.

This fixes test/API/lang/cpp/class_static/main.cpp when compiled with DWARF 5.

Differential Revision: https://reviews.llvm.org/D154617

Differential Revision: https://reviews.llvm.org/D154610

Compile time improvements:
https://llvm-compile-time-tracker.com/compare.php?from=a4a2b62495a63516a4f782acff1b19361906546b&to=a408521f71702a5c5fb65077adc23413d8631cfc&stat=instructions:u

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D152744

Fixes invariant memory regressions in future DSE patches.

Also add a flag to print<memoryssa> to not ensure optimized uses to test this.

Noticeable compile time regression [1], but a future DSE change that depends on this more than makes up for it.

[1] https://llvm-compile-time-tracker.com/compare.php?from=9d5466849a770eeab222d5a5890376d3596e8ad6&to=95682dbe11d76a3342870437377216e96b167504&stat=instructions:u

Reviewed By: asbirlea

Differential Revision: https://reviews.llvm.org/D152859

The return type of the recipe must match the array slice provided by
the user. This patch enhance the recipe creation to take into account
the constant slices.

Depends on D154259

Reviewed By: razvanlupusoru

Differential Revision: https://reviews.llvm.org/D154648

Generate code to allocate privates for trivial scalars and arrays.

Reviewed By: razvanlupusoru

Differential Revision: https://reviews.llvm.org/D154259

This is NFC because the function object is stateless anyway. This is
done solely for consistency with surrounding code and this was probably
an oversight in https://reviews.llvm.org/D132505.

Differential Revision: https://reviews.llvm.org/D154612

GCC requires "class" keyword when variable name matches class name.

Reviewed By: ayermolo

Differential Revision: https://reviews.llvm.org/D154654

C89 and C99 list initializers are treated differently than Cpp11 braced
initializers. This patch identifies the C array/struct initializer lists by
finding the preceding equal sign before a left brace, and applies formatting
rules for BracketAlignmentStyle.BlockIndent to those list initializers.

Fixes #57878.

Differential Revision: https://reviews.llvm.org/D153205