Implement protection against the stack clash attack [0] through inline stack
probing.
Probe stack allocation every PAGE_SIZE during frame lowering or dynamic
allocation to make sure the page guard, if any, is touched when touching the
stack, in a similar manner to GCC[1].
This extends the existing `probe-stack' mechanism with a special value `inline-asm'.
Technically the former uses function call before stack allocation while this
patch provides inlined stack probes and chunk allocation.
Only implemented for x86.
[0] https://www.qualys.com/2017/06/19/stack-clash/stack-clash.txt
[1] https://gcc.gnu.org/ml/gcc-patches/2017-07/msg00556.html
This a recommit of 39f50da2a3 with correct option
flags set.
Differential Revision: https://reviews.llvm.org/D68720
This reverts commit 39f50da2a3.
The -fstack-clash-protection is being passed to the linker too, which
is not intended.
Reverting and fixing that in a later commit.
Implement protection against the stack clash attack [0] through inline stack
probing.
Probe stack allocation every PAGE_SIZE during frame lowering or dynamic
allocation to make sure the page guard, if any, is touched when touching the
stack, in a similar manner to GCC[1].
This extends the existing `probe-stack' mechanism with a special value `inline-asm'.
Technically the former uses function call before stack allocation while this
patch provides inlined stack probes and chunk allocation.
Only implemented for x86.
[0] https://www.qualys.com/2017/06/19/stack-clash/stack-clash.txt
[1] https://gcc.gnu.org/ml/gcc-patches/2017-07/msg00556.html
Differential Revision: https://reviews.llvm.org/D68720
This reverts commits f41ec709d9 and 5fedc2b410. On some buildbots, Clang :: Driver/crash-report.c is broken with:
```
Command Output (stderr):
--
/home/buildslave/ps4-buildslave1/clang-with-thin-lto-ubuntu/llvm-project/clang/test/Driver/crash-report.c:48:11: error: CHECK: expected string not found in input
// CHECK: Preprocessed source(s) and associated run script(s) are located at:
^
<stdin>:1:1: note: scanning from here
/home/buildslave/ps4-buildslave1/clang-with-thin-lto-ubuntu/llvm-project/clang/test/Driver/crash-report.c:50:1: error: unknown type name 'BAZ'
```
Example: http://lab.llvm.org:8011/builders/clang-with-thin-lto-ubuntu/builds/21321/steps/test-stage1-compiler/logs/stdio
Previously, when the above '#pragma clang __debug' were used, Driver::generateCompilationDiagnostics() wouldn't work as expected.
The 'clang -E' process created for diagnostics would crash, because it would reach again the intended crash in Pragma.cpp, PragmaDebugHandler::HandlePragma() while preprocessing.
When generating crash diagnostics, we now disable the intended crashing behavior with a new cc1 flag -disable-pragma-debug-crash.
Notes:
- #pragma clang __debug llvm_report_fatal isn't currently tested by crash-report.c, because it needs exit() to be handled differently in -fintegrated-cc1 mode. See https://reviews.llvm.org/D73742 for an upcoming fix.
- This is also needed to further validate that -MF is removed from the 'clang -E ' crash diagnostic cmd-line (currently not the case). See https://reviews.llvm.org/D74076 for an upcoming fix.
Differential Revision: https://reviews.llvm.org/D74070
Summary:
- Similar to other targets, instead of passing a toolchain, a driver
argument should be passed into `arm::getARMTargetFeatures`. Aslo, that
routine should honor the specified triple. Refactor
`arm::getARMFloatABI` with 2 separate interfaces. One has the original
parameters and the other uses the driver and the specified triple.
- That fixes an issue when target & features are queried during the
offload compilation, where the specified triple should be checked
instead of a effective triple. A previously failed test is re-enabled.
Subscribers: kristof.beyls, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D74020
Summary:
As a first step this implementation enables compilation of the offload
code.
Reviewers: ABataev
Subscribers: ebevhan, Anastasia, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D74048
Summary:
- The device compilation needs to have a consistent source code compared
to the corresponding host compilation. If macros based on the
host-specific target processor is not properly populated, the device
compilation may fail due to the inconsistent source after the
preprocessor. So far, only the host triple is used to build the
macros. If a detailed host CPU target or certain features are
specified, macros derived from them won't be populated properly, e.g.
`__SSE3__` won't be added unless `+sse3` feature is present. On
Windows compilation compatible with MSVC, that missing macros result
in that intrinsics are not included and cause device compilation
failure on the host-side source.
- This patch addresses this issue by introducing two `cc1` options,
i.e., `-aux-target-cpu` and `-aux-target-feature`. If a specific host
CPU target or certain features are specified, the compiler driver will
append them during the construction of the offline compilation
actions. Then, the toolchain in `cc1` phase will populate macros
accordingly.
- An internal option `--gpu-use-aux-triple-only` is added to fall back
the original behavior to help diagnosing potential issues from the new
behavior.
Reviewers: tra, yaxunl
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73942
AMDGPU and x86 at least both have separate controls for whether
denormal results are flushed on output, and for whether denormals are
implicitly treated as 0 as an input. The current DAGCombiner use only
really cares about the input treatment of denormals.
First attempt at implementing -fsemantic-interposition.
Rely on GlobalValue::isInterposable that already captures most of the expected
behavior.
Rely on a ModuleFlag to state whether we should respect SemanticInterposition or
not. The default remains no.
So this should be a no-op if -fsemantic-interposition isn't used, and if it is,
isInterposable being already used in most optimisation, they should honor it
properly.
Note that it only impacts architecture compiled with -fPIC and no pie.
Differential Revision: https://reviews.llvm.org/D72829
Summary: With OpenMP offloading host compilation is done in two phases to capture host IR that is passed to all device compilations as input. But it turns out that we currently run entire LLVM optimization pipeline on host IR on both compilations which may have unpredictable effects on the resulting code. This patch fixes this problem by disabling LLVM passes on the first compilation, so the host IR that is passed to device compilations will be captured right after front end.
Reviewers: ABataev, jdoerfert, hfinkel
Reviewed By: ABataev
Subscribers: guansong, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73721
See
https://docs.google.com/document/d/1xMkTZMKx9llnMPgso0jrx3ankI4cv60xeZ0y4ksf4wc/preview
for background discussion.
This adds a warning, flags and pragmas to limit the number of
pre-processor tokens either at a certain point in a translation unit, or
overall.
The idea is that this would allow projects to limit the size of certain
widely included headers, or for translation units overall, as a way to
insert backstops for header bloat and prevent compile-time regressions.
Differential revision: https://reviews.llvm.org/D72703
Currently there are 4 different mechanisms for controlling denormal
flushing behavior, and about as many equivalent frontend controls.
- AMDGPU uses the fp32-denormals and fp64-f16-denormals subtarget features
- NVPTX uses the nvptx-f32ftz attribute
- ARM directly uses the denormal-fp-math attribute
- Other targets indirectly use denormal-fp-math in one DAGCombine
- cl-denorms-are-zero has a corresponding denorms-are-zero attribute
AMDGPU wants a distinct control for f32 flushing from f16/f64, and as
far as I can tell the same is true for NVPTX (based on the attribute
name).
Work on consolidating these into the denormal-fp-math attribute, and a
new type specific denormal-fp-math-f32 variant. Only ARM seems to
support the two different flush modes, so this is overkill for the
other use cases. Ideally we would error on the unsupported
positive-zero mode on other targets from somewhere.
Move the logic for selecting the flush mode into the compiler driver,
instead of handling it in cc1. denormal-fp-math/denormal-fp-math-f32
are now both cc1 flags, but denormal-fp-math-f32 is not yet exposed as
a user flag.
-cl-denorms-are-zero, -fcuda-flush-denormals-to-zero and
-fno-cuda-flush-denormals-to-zero will be mapped to
-fp-denormal-math-f32=ieee or preserve-sign rather than the old
attributes.
Stop emitting the denorms-are-zero attribute for the OpenCL flag. It
has no in-tree users. The meaning would also be target dependent, such
as the AMDGPU choice to treat this as only meaning allow flushing of
f32 and not f16 or f64. The naming is also potentially confusing,
since DAZ in other contexts refers to instructions implicitly treating
input denormals as zero, not necessarily flushing output denormals to
zero.
This also does not attempt to change the behavior for the current
attribute. The LangRef now states that the default is ieee behavior,
but this is inaccurate for the current implementation. The clang
handling is slightly hacky to avoid touching the existing
denormal-fp-math uses. Fixing this will be left for a future patch.
AMDGPU is still using the subtarget feature to control the denormal
mode, but the new attribute are now emitted. A future change will
switch this and remove the subtarget features.
These driver options perform some checking and delegate to MC options -x86-align-branch* and -x86-branches-within-32B-boundaries.
Reviewed By: skan
Differential Revision: https://reviews.llvm.org/D72463
The option will limit debug info by only emitting complete class
type information when its constructor is emitted.
This patch changes comparisons with LimitedDebugInfo to use the new
level instead.
Differential Revision: https://reviews.llvm.org/D72427
With this patch, the clang tool will now call the -cc1 invocation directly inside the same process. Previously, the -cc1 invocation was creating, and waiting for, a new process.
This patch therefore reduces the number of created processes during a build, thus it reduces build times on platforms where process creation can be costly (Windows) and/or impacted by a antivirus.
It also makes debugging a bit easier, as there's no need to attach to the secondary -cc1 process anymore, breakpoints will be hit inside the same process.
Crashes or signaling inside the -cc1 invocation will have the same side-effect as before, and will be reported through the same means.
This behavior can be controlled at compile-time through the CLANG_SPAWN_CC1 cmake flag, which defaults to OFF. Setting it to ON will revert to the previous behavior, where any -cc1 invocation will create/fork a secondary process.
At run-time, it is also possible to tweak the CLANG_SPAWN_CC1 environment variable. Setting it and will override the compile-time setting. A value of 0 calls -cc1 inside the calling process; a value of 1 will create a secondary process, as before.
Differential Revision: https://reviews.llvm.org/D69825
which is the default TLS model for non-PIC objects. This allows large/
many thread local variables or a compact/fast code in an executable.
Specification is same as that of GCC. For example, the code model
option precedes the TLS size option.
TLS access models other than local-exec are not changed. It means
supoort of the large code model is only in the local exec TLS model.
Patch By KAWASHIMA Takahiro (kawashima-fj <t-kawashima@fujitsu.com>)
Reviewers: dmgreen, mstorsjo, t.p.northover, peter.smith, ostannard
Reviewd By: peter.smith
Committed by: peter.smith
Differential Revision: https://reviews.llvm.org/D71688
All 130+ f_Group flags that take an argument allow it after a '=',
except for fdebug-complation-dir. Add a Joined<> alias so that
it behaves consistently with all the other f_Group flags.
(Keep the old Separate flag for backwards compat.)
In the backend, this feature is implemented with the function attribute
"patchable-function-entry". Both the attribute and XRay use
TargetOpcode::PATCHABLE_FUNCTION_ENTER, so the two features are
incompatible.
Reviewed By: ostannard, MaskRay
Differential Revision: https://reviews.llvm.org/D72222
Summary:
Every powerpc64le platform uses elfv2.
For powerpc64, the environments "elfv1" and "elfv2" were added for
FreeBSD ELFv1->ELFv2 migration in D61950. FreeBSD developers have
decided to use OS versions to select ABI, and no one is relying on the
environments.
Also use elfv2 on powerpc64-linux-musl.
Users can always use -mabi=elfv1 and -mabi=elfv2 to override the default
ABI.
Reviewed By: adalava
Differential Revision: https://reviews.llvm.org/D72352
-mpacked-stack is currently not supported with -mbackchain, so this should
result in a compilation error message instead of being silently ignored.
Review: Ulrich Weigand
gcc/config/{i386,s390} support -mnop-mcount. We currently only support
-mnop-mcount for SystemZ. The function attribute "mnop-mcount" is
ignored on other targets.
gcc/config/{i386,s390} support -mfentry. We currently only support
-mfentry for X86 and SystemZ. TargetOpcode::FENTRY_CALL is not handled
on other targets.
% clang -target aarch64 -pg -mfentry a.c -c
fatal error: error in backend: Not supported instr: <MCInst 21>
-mfentry, -mrecord-mcount, and -mnop-mcount were invented for Linux
ftrace. Linux uses $(call cc-option-yn,-mrecord-mcount) to detect if the
specific feature is available. Reject unsupported features so that Linux
build system will not wrongly consider them available and cause
build/runtime failures.
Note, GCC has stricter checks that we do not implement, e.g. -fpic/-fpie
-fnop-mcount is not allowed on x86, -fpic/-fpie -mfentry is not allowed
on x86-32.
GCC's x86 and s390 ports support -mrecord-mcount. Other ports reject the
option.
aarch64-linux-gnu-gcc: error: unrecognized command line option ‘-mrecord-mcount’
Allowing this option can cause failures when building Linux kernel for
aarch64, powerpc64, etc, which will think the feature is available if
the clang command returns 0.
Method '-[NSCoder decodeValueOfObjCType:at:]' is not only deprecated
but also a security hazard, hence a loud check.
Differential Revision: https://reviews.llvm.org/D71728
Recognize -mrecord-mcount from the command line and add a function attribute
"mrecord-mcount" when passed.
Only valid on SystemZ (when used with -mfentry).
Review: Ulrich Weigand
https://reviews.llvm.org/D71627
Summary: Per D62731, the behavior of clang with `-frounding-math` is no worse than when the rounding flag was completely ignored, so remove this unnecessary warning.
Reviewers: mibintc, chandlerc, echristo, rjmccall, kpn, erichkeane, rsmith, andrew.w.kaylor
Reviewed By: mibintc
Subscribers: merge_guards_bot, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71671
On Darwin, when used for generating a linked binary from a source file
(through an intermediate object file), the driver will invoke `cc1` to
generate a temporary object file. The temporary remark file will now be
emitted next to the object file, which will then be picked up by
`dsymutil` and emitted in the .dSYM bundle.
This is available for all formats except YAML since by default, YAML
doesn't need a section and the remark file will be lost.
Our build system does not handle randomly named files created during
the build well. We'd prefer to write compilation output directly
without creating a temporary file. Function parameters already
existed to control this behavior but were not exposed all the way out
to the command line.
Patch by Zachary Henkel!
Differential revision: https://reviews.llvm.org/D70615
Recognize -mpacked-stack from the command line and add a function attribute
"mpacked-stack" when passed. This is needed for building the Linux kernel.
If this option is passed for any other target than SystemZ, an error is
generated.
Review: Ulrich Weigand
https://reviews.llvm.org/D71441
This matches https://gcc.gnu.org/onlinedocs/gcc/AArch64-Options.html
> -momit-leaf-frame-pointer
> -mno-omit-leaf-frame-pointer
>
> Omit or keep the frame pointer in leaf functions. The former behavior is the default.
-mno-omit-leaf-frame-pointer is currently a no-op because
TargetOptions::DisableFramePointerElim is only considered for non-leaf
functions.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D71167
Very few ELF platforms still use .ctors/.dtors now. Linux (glibc: 1999-07),
DragonFlyBSD, FreeBSD (2012-03) and Solaris have supported .init_array
for many years. Some architectures like AArch64/RISC-V default to
.init_array . GNU ld and gold can even convert .ctors to .init_array .
It makes more sense to flip the CC1 default, and only uses
-fno-use-init-array on platforms that don't support .init_array .
For example, OpenBSD did not support DT_INIT_ARRAY before Aug 2016
(86fa57a279)
I may miss some ELF platforms that still use .ctors, but their
maintainers can easily diagnose such problems.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D71393
Serialized remarks contain debug locations for each remark, by storing a
file path, a line, and a column.
Also, remarks support being embedded in a .dSYM bundle using a separate
section in object files, that is found by `dsymutil` through the debug
map.
In order for tools to map addresses to source and display remarks in the
source, we need line tables, and in order for `dsymutil` to find the
object files containing the remark section, we need to keep the debug
map around.
Differential Revision: https://reviews.llvm.org/D71325
This is a follow up patch to use the OpenMP-IR-Builder, as discussed on
the mailing list ([1] and later) and at the US Dev Meeting'19.
[1] http://lists.flang-compiler.org/pipermail/flang-dev_lists.flang-compiler.org/2019-May/000197.html
Reviewers: kiranchandramohan, ABataev, RaviNarayanaswamy, gtbercea, grokos, sdmitriev, JonChesterfield, hfinkel, fghanim
Subscribers: ppenzin, penzn, llvm-commits, cfe-commits, jfb, guansong, bollu, hiraditya, mgorny
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69922
The -fsplit-dwarf-inlining option does not conform to DWARF5 standard.
It creates children for Skeleton compilation unit. We need default behavior
to be DWARF5 compatible. Thus set default state for -fsplit-dwarf-inlining
into "false".
Differential Revision: https://reviews.llvm.org/D71304
This adds a check for the usage of -foptimization-record-file with
multiple -arch options. This is not permitted since it would require us
to rename the file requested by the user to avoid overwriting it for the
second cc1 invocation.
This patch allows for -o to be used with -c when compiling with clang
interface stubs enabled. This is because the second file will be an
intermediate ifs stubs file that is the text stub analog of the .o file.
Both get produces in this case, so two files.
Why are we doing this? Because we want to support the case where
interface stubs are used bu first invoking clang like so:
clang -c <other flags> -emit-interface-stubs foo.c -o foo.o
...
clang -emit-interface-stubs <.o files> -o libfoo.so
This should generate N .ifs files, and one .ifso file. Prior to this
patch, using -o with the -c invocation was not possible. Currently the
clang driver supports generating a a.out/.so file at the same time as a
merged ifs file / ifso file, but this is done by checking that the final
job is the IfsMerge job. When -c is used, the final job is a Compile job
so what this patch does is check to figure out of the job type is
TY_IFS_CPP.
Differential Revision: https://reviews.llvm.org/D70763
Patch was reverted because https://bugs.llvm.org/show_bug.cgi?id=44048
The original patch is modified to set the strictfp IR attribute
explicitly in CodeGen instead of as a side effect of IRBuilder.
In the 2nd attempt to reapply there was a windows lit test fail, the
tests were fixed to use wildcard matching.
Differential Revision: https://reviews.llvm.org/D62731
Skip distro detection when we're not running on Linux, or when the target triple is not Linux. This saves a few OS calls for each invocation of clang.exe.
Differential Revision: https://reviews.llvm.org/D70467
Summary:
Removed the ```-fforce-experimental-new-constant-interpreter flag```, leaving
only the ```-fexperimental-new-constant-interpreter``` one. The interpreter
now always emits an error on an unsupported feature.
Allowing the interpreter to bail out would require a mapping from APValue to
interpreter memory, which will not be necessary in the final version. It is
more sensible to always emit an error if the interpreter fails.
Reviewers: jfb, Bigcheese, rsmith, dexonsmith
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D70071
GCC 8 implements -fmacro-prefix-map. Like -fdebug-prefix-map, it replaces a string prefix for the __FILE__ macro.
-ffile-prefix-map is the union of -fdebug-prefix-map and -fmacro-prefix-map
Reviewed By: rnk, Lekensteyn, maskray
Differential Revision: https://reviews.llvm.org/D49466
When the driver is targeting multiple architectures at once, for things
like Universal Mach-Os, we need to emit different remark files for each
cc1 invocation to avoid overwriting the files from a different
invocation.
For example:
$ clang -c -o foo.o -fsave-optimization-record -arch x86_64 -arch x86_64h
will create two remark files:
* foo-x86_64.opt.yaml
* foo-x86_64h.opt.yaml
and a follow-up NFC rearrangement as it's causing a crash on valid. Testcase is on the original review thread.
This reverts commits af57dbf12e and e6584b2b7b
When the driver is targeting multiple architectures at once, for things
like Universal Mach-Os, we need to emit different remark files for each
cc1 invocation to avoid overwriting the files from a different
invocation.
For example:
$ clang -c -o foo.o -fsave-optimization-record -arch x86_64 -arch x86_64h
will create two remark files:
* foo-x86_64.opt.yaml
* foo-x86_64h.opt.yaml
This started passing target-features on the linker line, not just for RISCV but
for all targets, leading to error messages in Chromium Android build:
'+soft-float-abi' is not a recognized feature for this target (ignoring feature)
'+soft-float-abi' is not a recognized feature for this target (ignoring feature)
See Phabricator review for details.
Reverting until this can be fixed properly.
> Summary:
> 1. enable LTO need to pass target feature and abi to LTO code generation
> RISCV backend need the target feature to decide which extension used in
> code generation.
> 2. move getTargetFeatures to CommonArgs.h and add ForLTOPlugin flag
> 3. add general tools::getTargetABI in CommonArgs.h because different target uses different
> way to get the target ABI.
>
> Patch by Kuan Hsu Chen (khchen)
>
> Reviewers: lenary, lewis-revill, asb, MaskRay
>
> Reviewed By: lenary
>
> Subscribers: hiraditya, dschuff, aheejin, fedor.sergeev, mehdi_amini, inglorion, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, kito-cheng, shiva0217, jrtc27, MaskRay, zzheng, edward-jones, steven_wu, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, rkruppe, PkmX, jocewei, psnobl, benna, Jim, lenary, s.egerton, pzheng, cfe-commits
>
> Tags: #clang
>
> Differential Revision: https://reviews.llvm.org/D67409
This flag decouples specifying the DWARF version from enabling/disabling
DWARF in general (or the gN level - gmlt/limited/standalone, etc) while
still allowing existing -gdwarf-N flags to override this default.
Patch by Caroline Tice!
Differential Revision: https://reviews.llvm.org/D69822
Add options to control floating point behavior: trapping and
exception behavior, rounding, and control of optimizations that affect
floating point calculations. More details in UsersManual.rst.
Reviewers: rjmccall
Differential Revision: https://reviews.llvm.org/D62731
Recognize -mnop-mcount from the command line and add a function attribute
"mnop-mcount"="true" when passed.
When this option is used, a nop is added instead of a call to fentry. This
is used when building the Linux Kernel.
If this option is passed for any other target than SystemZ, an error is
generated.
Review: Ulrich Weigand
https://reviews.llvm.org/D67763
The linker options (e.g. pragma detect_mismatch) are intended for host
compilation only, therefore disable it for device compilation.
Differential Revision: https://reviews.llvm.org/D57829
This reverts commit 004ed2b0d1.
Original commit hash 6d03890384
Summary:
This adds a clang option to disable inline line tables. When it is used,
the inliner uses the call site as the location of the inlined function instead of
marking it as an inline location with the function location.
https://reviews.llvm.org/D67723
This adds a flag to LLVM and clang to always generate a .debug_frame
section, even if other debug information is not being generated. In
situations where .eh_frame would normally be emitted, both .debug_frame
and .eh_frame will be used.
Differential Revision: https://reviews.llvm.org/D67216
Summary:
This adds a clang option to disable inline line tables. When it is used,
the inliner uses the call site as the location of the inlined function instead of
marking it as an inline location with the function location.
See https://bugs.llvm.org/show_bug.cgi?id=42344
Reviewers: rnk
Subscribers: hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D67723
Summary:
A new function pass (Transforms/CFGuard/CFGuard.cpp) inserts CFGuard checks on
indirect function calls, using either the check mechanism (X86, ARM, AArch64) or
or the dispatch mechanism (X86-64). The check mechanism requires a new calling
convention for the supported targets. The dispatch mechanism adds the target as
an operand bundle, which is processed by SelectionDAG. Another pass
(CodeGen/CFGuardLongjmp.cpp) identifies and emits valid longjmp targets, as
required by /guard:cf. This feature is enabled using the `cfguard` CC1 option.
Reviewers: thakis, rnk, theraven, pcc
Subscribers: ychen, hans, metalcanine, dmajor, tomrittervg, alex, mehdi_amini, mgorny, javed.absar, kristof.beyls, hiraditya, steven_wu, dexonsmith, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D65761
Summary:
- As variadic parameters have the lowest rank in overload resolution,
without real usage of `va_arg`, they are commonly used as the
catch-all fallbacks in SFINAE. As the front-end still reports errors
on calls to `va_arg`, the declaration of functions with variadic
arguments should be allowed in general.
Reviewers: jlebar, tra, yaxunl
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69389
Summary:
A necessary step to let build system caching work for its output.
Reviewers: tejohnson, steven_wu
Reviewed by: tejohnson
Subscribers: mehdi_amini, inglorion, dexonsmith, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69406
This is clang part of the patch. It adds -flto-unit flag for thin LTO
builds on Mac and PS4
Differential revision: https://reviews.llvm.org/D68950
llvm-svn: 375224
Remove dead virtual functions from vtables with
replaceNonMetadataUsesWith, so that CGProfile metadata gets cleaned up
correctly.
Original commit message:
Currently, it is hard for the compiler to remove unused C++ virtual
functions, because they are all referenced from vtables, which are referenced
by constructors. This means that if the constructor is called from any live
code, then we keep every virtual function in the final link, even if there
are no call sites which can use it.
This patch allows unused virtual functions to be removed during LTO (and
regular compilation in limited circumstances) by using type metadata to match
virtual function call sites to the vtable slots they might load from. This
information can then be used in the global dead code elimination pass instead
of the references from vtables to virtual functions, to more accurately
determine which functions are reachable.
To make this transformation safe, I have changed clang's code-generation to
always load virtual function pointers using the llvm.type.checked.load
intrinsic, instead of regular load instructions. I originally tried writing
this using clang's existing code-generation, which uses the llvm.type.test
and llvm.assume intrinsics after doing a normal load. However, it is possible
for optimisations to obscure the relationship between the GEP, load and
llvm.type.test, causing GlobalDCE to fail to find virtual function call
sites.
The existing linkage and visibility types don't accurately describe the scope
in which a virtual call could be made which uses a given vtable. This is
wider than the visibility of the type itself, because a virtual function call
could be made using a more-visible base class. I've added a new
!vcall_visibility metadata type to represent this, described in
TypeMetadata.rst. The internalization pass and libLTO have been updated to
change this metadata when linking is performed.
This doesn't currently work with ThinLTO, because it needs to see every call
to llvm.type.checked.load in the linkage unit. It might be possible to
extend this optimisation to be able to use the ThinLTO summary, as was done
for devirtualization, but until then that combination is rejected in the
clang driver.
To test this, I've written a fuzzer which generates random C++ programs with
complex class inheritance graphs, and virtual functions called through object
and function pointers of different types. The programs are spread across
multiple translation units and DSOs to test the different visibility
restrictions.
I've also tried doing bootstrap builds of LLVM to test this. This isn't
ideal, because only classes in anonymous namespaces can be optimised with
-fvisibility=default, and some parts of LLVM (plugins and bugpoint) do not
work correctly with -fvisibility=hidden. However, there are only 12 test
failures when building with -fvisibility=hidden (and an unmodified compiler),
and this change does not cause any new failures for either value of
-fvisibility.
On the 7 C++ sub-benchmarks of SPEC2006, this gives a geomean code-size
reduction of ~6%, over a baseline compiled with "-O2 -flto
-fvisibility=hidden -fwhole-program-vtables". The best cases are reductions
of ~14% in 450.soplex and 483.xalancbmk, and there are no code size
increases.
I've also run this on a set of 8 mbed-os examples compiled for Armv7M, which
show a geomean size reduction of ~3%, again with no size increases.
I had hoped that this would have no effect on performance, which would allow
it to awlays be enabled (when using -fwhole-program-vtables). However, the
changes in clang to use the llvm.type.checked.load intrinsic are causing ~1%
performance regression in the C++ parts of SPEC2006. It should be possible to
recover some of this perf loss by teaching optimisations about the
llvm.type.checked.load intrinsic, which would make it worth turning this on
by default (though it's still dependent on -fwhole-program-vtables).
Differential revision: https://reviews.llvm.org/D63932
llvm-svn: 375094
The final list of OpenMP offload targets becomes known only at the link time and since offload registration code depends on the targets list it makes sense to delay offload registration code generation to the link time instead of adding it to the host part of every fat object. This patch moves offload registration code generation from clang to the offload wrapper tool.
This is the last part of the OpenMP linker script elimination patch https://reviews.llvm.org/D64943
Differential Revision: https://reviews.llvm.org/D68746
llvm-svn: 374937
Summary:
When files often get touched during builds, the mtime based validation
leads to different problems in implicit modules builds, even when the
content doesn't actually change:
- Modules only: module invalidation due to out of date files. Usually causing rebuild traffic.
- Modules + PCH: build failures because clang cannot rebuild a module if it comes from building a PCH.
- PCH: build failures because clang cannot rebuild a PCH in case one of the input headers has different mtime.
This patch proposes hashing the content of input files (headers and
module maps), which is performed during serialization time. When looking
at input files for validation, clang only computes the hash in case
there's a mtime mismatch.
I've tested a couple of different hash algorithms availble in LLVM in
face of building modules+pch for `#import <Cocoa/Cocoa.h>`:
- `hash_code`: performace diff within the noise, total module cache increased by 0.07%.
- `SHA1`: 5% slowdown. Haven't done real size measurements, but it'd be BLOCK_ID+20 bytes per input file, instead of BLOCK_ID+8 bytes from `hash_code`.
- `MD5`: 3% slowdown. Like above, but BLOCK_ID+16 bytes per input file.
Given the numbers above, the patch uses `hash_code`. The patch also
improves invalidation error msgs to point out which type of problem the
user is facing: "mtime", "size" or "content".
rdar://problem/29320105
Reviewers: dexonsmith, arphaman, rsmith, aprantl
Subscribers: jkorous, cfe-commits, ributzka
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67249
> llvm-svn: 374841
llvm-svn: 374895
Summary:
When files often get touched during builds, the mtime based validation
leads to different problems in implicit modules builds, even when the
content doesn't actually change:
- Modules only: module invalidation due to out of date files. Usually causing rebuild traffic.
- Modules + PCH: build failures because clang cannot rebuild a module if it comes from building a PCH.
- PCH: build failures because clang cannot rebuild a PCH in case one of the input headers has different mtime.
This patch proposes hashing the content of input files (headers and
module maps), which is performed during serialization time. When looking
at input files for validation, clang only computes the hash in case
there's a mtime mismatch.
I've tested a couple of different hash algorithms availble in LLVM in
face of building modules+pch for `#import <Cocoa/Cocoa.h>`:
- `hash_code`: performace diff within the noise, total module cache increased by 0.07%.
- `SHA1`: 5% slowdown. Haven't done real size measurements, but it'd be BLOCK_ID+20 bytes per input file, instead of BLOCK_ID+8 bytes from `hash_code`.
- `MD5`: 3% slowdown. Like above, but BLOCK_ID+16 bytes per input file.
Given the numbers above, the patch uses `hash_code`. The patch also
improves invalidation error msgs to point out which type of problem the
user is facing: "mtime", "size" or "content".
rdar://problem/29320105
Reviewers: dexonsmith, arphaman, rsmith, aprantl
Subscribers: jkorous, cfe-commits, ributzka
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67249
llvm-svn: 374841
Summary:
Currently clang does not support -Wa,-W, which suppresses warning
messages in GNU assembler. Add this option for gcc compatibility.
https://bugs.llvm.org/show_bug.cgi?id=43651. Reland with differential
information.
Reviewers: bcain
Reviewed By: bcain
Subscribers: george.burgess.iv, gbiv, llozano, manojgupta, nickdesaulniers, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D68884
llvm-svn: 374834
Currently clang does not support -Wa,-W, which suppresses warning
messages in GNU assembler. Add this option for gcc compatibility.
https://bugs.llvm.org/show_bug.cgi?id=43651
llvm-svn: 374822
The goal is to have 100% fidelity in clang-scan-deps behavior when
--analyze is present in compilation command.
At the same time I don't want to break clang-tidy which expects
__static_analyzer__ macro defined as built-in.
I introduce new cc1 options (-setup-static-analyzer) that controls
the macro definition and is conditionally set in driver.
Differential Revision: https://reviews.llvm.org/D68093
llvm-svn: 374815
Summary:
1. enable LTO need to pass target feature and abi to LTO code generation
RISCV backend need the target feature to decide which extension used in
code generation.
2. move getTargetFeatures to CommonArgs.h and add ForLTOPlugin flag
3. add general tools::getTargetABI in CommonArgs.h because different target uses different
way to get the target ABI.
Patch by Kuan Hsu Chen (khchen)
Reviewers: lenary, lewis-revill, asb, MaskRay
Reviewed By: lenary
Subscribers: hiraditya, dschuff, aheejin, fedor.sergeev, mehdi_amini, inglorion, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, kito-cheng, shiva0217, jrtc27, MaskRay, zzheng, edward-jones, steven_wu, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, rkruppe, PkmX, jocewei, psnobl, benna, Jim, lenary, s.egerton, pzheng, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67409
llvm-svn: 374774
Currently, it is hard for the compiler to remove unused C++ virtual
functions, because they are all referenced from vtables, which are referenced
by constructors. This means that if the constructor is called from any live
code, then we keep every virtual function in the final link, even if there
are no call sites which can use it.
This patch allows unused virtual functions to be removed during LTO (and
regular compilation in limited circumstances) by using type metadata to match
virtual function call sites to the vtable slots they might load from. This
information can then be used in the global dead code elimination pass instead
of the references from vtables to virtual functions, to more accurately
determine which functions are reachable.
To make this transformation safe, I have changed clang's code-generation to
always load virtual function pointers using the llvm.type.checked.load
intrinsic, instead of regular load instructions. I originally tried writing
this using clang's existing code-generation, which uses the llvm.type.test
and llvm.assume intrinsics after doing a normal load. However, it is possible
for optimisations to obscure the relationship between the GEP, load and
llvm.type.test, causing GlobalDCE to fail to find virtual function call
sites.
The existing linkage and visibility types don't accurately describe the scope
in which a virtual call could be made which uses a given vtable. This is
wider than the visibility of the type itself, because a virtual function call
could be made using a more-visible base class. I've added a new
!vcall_visibility metadata type to represent this, described in
TypeMetadata.rst. The internalization pass and libLTO have been updated to
change this metadata when linking is performed.
This doesn't currently work with ThinLTO, because it needs to see every call
to llvm.type.checked.load in the linkage unit. It might be possible to
extend this optimisation to be able to use the ThinLTO summary, as was done
for devirtualization, but until then that combination is rejected in the
clang driver.
To test this, I've written a fuzzer which generates random C++ programs with
complex class inheritance graphs, and virtual functions called through object
and function pointers of different types. The programs are spread across
multiple translation units and DSOs to test the different visibility
restrictions.
I've also tried doing bootstrap builds of LLVM to test this. This isn't
ideal, because only classes in anonymous namespaces can be optimised with
-fvisibility=default, and some parts of LLVM (plugins and bugpoint) do not
work correctly with -fvisibility=hidden. However, there are only 12 test
failures when building with -fvisibility=hidden (and an unmodified compiler),
and this change does not cause any new failures for either value of
-fvisibility.
On the 7 C++ sub-benchmarks of SPEC2006, this gives a geomean code-size
reduction of ~6%, over a baseline compiled with "-O2 -flto
-fvisibility=hidden -fwhole-program-vtables". The best cases are reductions
of ~14% in 450.soplex and 483.xalancbmk, and there are no code size
increases.
I've also run this on a set of 8 mbed-os examples compiled for Armv7M, which
show a geomean size reduction of ~3%, again with no size increases.
I had hoped that this would have no effect on performance, which would allow
it to awlays be enabled (when using -fwhole-program-vtables). However, the
changes in clang to use the llvm.type.checked.load intrinsic are causing ~1%
performance regression in the C++ parts of SPEC2006. It should be possible to
recover some of this perf loss by teaching optimisations about the
llvm.type.checked.load intrinsic, which would make it worth turning this on
by default (though it's still dependent on -fwhole-program-vtables).
Differential revision: https://reviews.llvm.org/D63932
llvm-svn: 374539
I noticed that compiling on Windows with -fno-ms-compatibility had the
side effect of defining __GNUC__, along with __GNUG__, __GXX_RTTI__, and
a number of other macros for GCC compatibility. This is undesirable and
causes Chromium to do things like mix __attribute__ and __declspec,
which doesn't work. We should have a positive language option to enable
GCC compatibility features so that we can experiment with
-fno-ms-compatibility on Windows. This change adds -fgnuc-version= to be
that option.
My issue aside, users have, for a long time, reported that __GNUC__
doesn't match their expectations in one way or another. We have
encouraged users to migrate code away from this macro, but new code
continues to be written assuming a GCC-only environment. There's really
nothing we can do to stop that. By adding this flag, we can allow them
to choose their own adventure with __GNUC__.
This overlaps a bit with the "GNUMode" language option from -std=gnu*.
The gnu language mode tends to enable non-conforming behaviors that we'd
rather not enable by default, but the we want to set things like
__GXX_RTTI__ by default, so I've kept these separate.
Helps address PR42817
Reviewed By: hans, nickdesaulniers, MaskRay
Differential Revision: https://reviews.llvm.org/D68055
llvm-svn: 374449
This patch removes the remaining part of the OpenMP offload linker scripts which was used for inserting device binaries into the output linked binary. Device binaries are now inserted into the host binary with a help of the wrapper bit-code file which contains device binaries as data. Wrapper bit-code file is dynamically created by the clang driver with a help of new tool clang-offload-wrapper which takes device binaries as input and produces bit-code file with required contents. Wrapper bit-code is then compiled to an object and resulting object is appended to the host linking by the clang driver.
This is the second part of the patch for eliminating OpenMP linker script (please see https://reviews.llvm.org/D64943).
Differential Revision: https://reviews.llvm.org/D68166
llvm-svn: 374219
Second Landing Attempt:
This patch enables end to end support for generating ELF interface stubs
directly from clang. Now the following:
clang -emit-interface-stubs -o libfoo.so a.cpp b.cpp c.cpp
will product an ELF binary with visible symbols populated. Visibility attributes
and -fvisibility can be used to control what gets populated.
* Adding ToolChain support for clang Driver IFS Merge Phase
* Implementing a default InterfaceStubs Merge clang Tool, used by ToolChain
* Adds support for the clang Driver to involve llvm-ifs on ifs files.
* Adds -emit-merged-ifs flag, to tell llvm-ifs to emit a merged ifs text file
instead of the final object format (normally ELF)
Differential Revision: https://reviews.llvm.org/D63978
llvm-svn: 374061
This patch enables end to end support for generating ELF interface stubs
directly from clang. Now the following:
clang -emit-interface-stubs -o libfoo.so a.cpp b.cpp c.cpp
will product an ELF binary with visible symbols populated. Visibility attributes
and -fvisibility can be used to control what gets populated.
* Adding ToolChain support for clang Driver IFS Merge Phase
* Implementing a default InterfaceStubs Merge clang Tool, used by ToolChain
* Adds support for the clang Driver to involve llvm-ifs on ifs files.
* Adds -emit-merged-ifs flag, to tell llvm-ifs to emit a merged ifs text file
instead of the final object format (normally ELF)
Differential Revision: https://reviews.llvm.org/D63978
llvm-svn: 373538
Summary:
To trigger the index-only Whole Program Devirt support added to LLVM, we
need to be able to specify -fno-split-lto-unit in conjunction with
-fwhole-program-vtables. Keep the default for -fwhole-program-vtables as
-fsplit-lto-unit, but don't error on that option combination.
Reviewers: pcc
Subscribers: mehdi_amini, inglorion, steven_wu, dexonsmith, arphaman, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D68029
llvm-svn: 373370
Nico Weber