We already do this for 16/32/64 as well as 8-bit add with register/immediate. Might as well do it for 8-bit INC/DEC too.
Differential Revision: https://reviews.llvm.org/D58869
llvm-svn: 355424
We already support 8-bits adds in convertToThreeAddress. But we can also support 8-bit OR if the bits are disjoint. We already do this for 16/32/64.
Differential Revision: https://reviews.llvm.org/D58863
llvm-svn: 355423
If a kernel template has a function as its template parameter, a device function should be
allowed as template argument since a kernel can call a device function. However,
currently if the kernel template is instantiated in a host function, clang will emit an error
message saying the device function is an invalid candidate for the template parameter.
This happens because clang checks the reference to the device function during parsing
the template arguments. At this point, the template is not instantiated yet. Clang incorrectly
assumes the device function is called by the host function and emits the error message.
This patch fixes the issue by disabling checking of device function during parsing template
arguments and deferring the check to the instantion of the template. At that point, the
template decl is already available, therefore the check can be done against the instantiated
function template decl.
Differential Revision: https://reviews.llvm.org/D56411
llvm-svn: 355421
x86-64 is an invalid architecture in triples. Changing it to the correct
triple (x86_64) changes some tests, because SLP is not deemed profitable
any more.
Reviewers: ABataev, RKSimon, spatel
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D58931
llvm-svn: 355420
HIP uses clang-offload-bundler to create fat binary. The bundle for host is empty.
Currently clang-offload-bundler checks if the bundle size is 0 when unbundling.
If so it will exit without unbundling the remaining bundles. This causes
clang-offload-bundler not being able to unbundle fat binaries generated for HIP.
This patch allows bundles size to be 0 when clang-offload-bundler unbundles
input files.
Differential Revision: https://reviews.llvm.org/D58057
llvm-svn: 355419
memory.
If the variable with the constant non-scalar type is firstprivatized in
the target region, the local copy is created with the data copying.
Instead, we allocate the copy in the constant memory and avoid extra
copying in the outlined target regions. This global copy is used in the
target regions without loss of the performance.
llvm-svn: 355418
Revert "[libc++] Fix <atomic> failures on GCC"
Revert "[libc++] Change memory_order to an enum class"
Revert "[libc++] decoupling Freestanding atomic<T> from libatomic.a"
The lldb formatter nededs to be updated. Shafik and Louis will
coordinate to do so.
llvm-svn: 355417
Summary:
The bytemap classes will be used by the primary32 allocator to associate
classes with memory regions. It's similar to the sanitizer_common one
except for the fact that the base (level1) maps are mapped instead of
being static to reduce the memory footprint of an uninitialized allocator.
Reviewers: vitalybuka, eugenis, morehouse, flowerhack, dmmoore415, mcgrathr
Reviewed By: vitalybuka, morehouse
Subscribers: mgorny, delcypher, jfb, #sanitizers, llvm-commits
Tags: #llvm, #sanitizers
Differential Revision: https://reviews.llvm.org/D58723
llvm-svn: 355416
Currently one can concatenate strings using hash(#),
but not lists, although that would be a natural thing to do.
This patch allows one to write something like:
def : A<!listconcat([1,2], [3,4])>;
simply as :
def : A<[1,2] # [3,4]>;
This was missing feature was highlighted by Nicolai
at FOSDEM talk.
Reviewed by: nhaehnle, hfinkel
Differential Revision: https://reviews.llvm.org/D58895
llvm-svn: 355414
When -fno-gpu-rdc is set, device code is compiled, linked, and assembled into fat binary
and embedded as string in object files. The object files are normal object files which
can be linked by host linker. In the linking stage, the object files should not be unbundled
when -fno-gpu-rdc is set since they are normal object files, not bundles. The object files
only need to be unbundled when -fgpu-rdc is set.
Currently clang always unbundles object files, disregarding -fgpu-rdc option.
This patch fixes that.
Differential Revision: https://reviews.llvm.org/D58917
llvm-svn: 355410
Summary:
In https://reviews.llvm.org/D58201, we turned memory_order into an enum
class in C++20 mode. However, we were not casting memory_order to its
underlying type correctly for the GCC implementation, which broke the
build bots. I also fixed a test that was failing in C++17 mode on GCC 5.
Reviewers: EricWF, jfb, mclow.lists
Subscribers: zoecarver
Differential Revision: https://reviews.llvm.org/D58966
llvm-svn: 355409
When --compress-debug-sections is given, llvm-objcopy do not compress
sections that have "ZLIB" header in data. Normally this signature is used
in zlib-gnu compression format. But if zlib-gnu used then the name of the compressed
section should start from .z* (e.g .zdebug_info). If it does not, then it is not
a zlib-gnu format and section should be treated as a normal uncompressed section.
Differential revision: https://reviews.llvm.org/D58908
llvm-svn: 355399
The intention in r355323 has been to implement a no-op resolver in the
HostInfoBase class, which will then be shadowed a an implementation in
the HostInfoPosix class. However, I add the shadowing declaration in
HostInfoPosix.h, and instead had implemented the HostInfoBase function
in HostInfoPosix.cpp. This has lead to undefined symbols on windows, and
a subsequent implementation of a no-op resolver in HostInfoWindows
(r355329).
Since now there is no point on having a no-op resolver in the base
class, I just remove the base declaration altogether, and have
HostInfoPosix implement the (newly-declared) HostInfoPosix version of
that function.
llvm-svn: 355398
The TraceLoggingProvider.h header does work with clang-cl in general
these days with Win SDK 10.0.17763.0, but when compiled in 32 bit x86
mode, with the -Z7 flag, compilation fails with the following error:
fatal error: error in backend: assembler label '' can not be undefined
With older Win SDKs, there are other build failures (regardless of
architecture or the -Z7 flag).
Differential Revision: https://reviews.llvm.org/D58958
llvm-svn: 355397
The gets function has no SrcArgs. Because the default value for isTainted was
false, it didn't mark its DstArgs as tainted.
Patch by Gábor Borsik!
Differential Revision: https://reviews.llvm.org/D58828
llvm-svn: 355396
A SCEV is not low-cost just because you can divide it by a power of 2. We need to also
check what we are dividing to make sure it too is not a high-code expansion. This helps
to not expand the exit value of certain loops, helping not to bloat the code.
The change in no-iv-rewrite.ll is reverting back to what it was testing before rL194116,
and looks a lot like the other tests in replace-loop-exit-folds.ll.
Differential Revision: https://reviews.llvm.org/D58435
llvm-svn: 355393
If zlib is not available, and --compress-debug-sections is passed,
we want to report an error. Currently, it is only reported for
--compress_debug_sections= form of the option.
Fixes the https://bugs.llvm.org/show_bug.cgi?id=40886.
I do not think there is a way to write a test for this.
Differential revision: https://reviews.llvm.org/D58909
llvm-svn: 355391
Summary:
Redecl chains of classes and class templates are not handled well
currently. We want to handle them similarly to functions, i.e. try to
keep the structure of the original AST as much as possible. The aim is
to not squash a prototype with a definition, rather we create both and
put them in a redecl chain.
Reviewers: a_sidorin, shafik, a.sidorin
Subscribers: rnkovacs, dkrupp, Szelethus, gamesh411, jdoerfert, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D58502
llvm-svn: 355390
We can already update multiple files in each update call, this extends it to work with wildcards as well in the same way as update_mca_test_checks.py (to support shells that won't do this for us - windows command prompt etc.)
Differential Revision: https://reviews.llvm.org/D58817
llvm-svn: 355386
When lowering a select_cc node where the true and false values are of type f16,
we can't use a general conditional move because the FP16 instructions do not
support conditional execution. Instead, we must ensure that the condition code
is one of the four supported by the VSEL instruction.
Differential revision: https://reviews.llvm.org/D58813
llvm-svn: 355385
Summary:
In some cases the KILL was causing a hazard to be introduced as these were
scheduled into hazard slots, but don't result in an instruction.
KILL shouldn't be considered for hazard recognition.
Change-Id: Ib6d2a2160f8c94cd0ce611ab198c7e4f46aeffcf
Subscribers: arsenm, kzhuravl, jvesely, wdng, nhaehnle, yaxunl, tpr, t-tye, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D58898
llvm-svn: 355384
If the permission does not permit writing and the standard input is a
terminal, rm gives an annoying prompt. chmod u+w to make the output
directory easier to delete.
llvm-svn: 355382
Craig Topper