Summary:
Suppose we visit symbols in this order:
1. weak definition of foo in a lazy object
2. reference of foo
3 (optional). definition of foo
bfd/gold allows 123 but not 12.
Current --warn-backrefs implementation will report both cases as a backward reference. With this change, both 123 (intended) and 12 (unintended) are allowed. The usage of weak definitions usually imply there are also global definitions, so the trade-off is justified.
Reviewers: ruiu, espindola
Subscribers: emaste, arichardson, llvm-commits
Differential Revision: https://reviews.llvm.org/D46624
llvm-svn: 332061
We found current sampleFDO had a performance issue when triaging a regression.
For a callsite with inline instance in the profile, even if hot callsite inliner
cannot inline it, it may still execute enough times and should not be treated as
cold in regular inliner later. However, currently if such callsite is not inlined
by hot callsite inliner, and the BB where the callsite locates doesn't get
samples from other instructions inside of it, the callsite will have no profile
metadata annotated. In regular inliner cost analysis, if the callsite has no
profile annotated and its caller has profile information, it will be treated as
cold.
The fix changes the isCallsiteHot check and chooses to compare
CallsiteTotalSamples with hot cutoff value computed by ProfileSummaryInfo.
Differential Revision: https://reviews.llvm.org/D45377
llvm-svn: 332058
This commit adds a wrapper for std::distance() which works with ranges.
As it would be a common case to write `distance(predecessors(BB))`, this
also introduces `pred_size()` and `succ_size()` helpers to make that
easier to write.
Differential Revision: https://reviews.llvm.org/D46668
llvm-svn: 332057
The bitwidth of the operation should always be wider than the result width of the truncate since we don't recurse through any width changing operations.
llvm-svn: 332055
The previous CL changes the order of output sections, which causes address changes in test cases.
Review: https://reviews.llvm.org/D46730
llvm-svn: 332054
This implements a new table-gen emitter to create tables for
a wasm disassembler, and a dissassembler to use them.
Comes with 2 tests, that tests a few instructions manually. Is also able to
disassemble large .wasm files with objdump reasonably.
Not working so well, to be addressed in followups:
- objdump appears to be passing an incorrect starting point.
- since the disassembler works an instruction at a time, and it is
disassembling stack instruction, it has no idea of pseudo register assignments.
These registers are required for the instruction printing code that follows.
For now, all such registers appear in the output as $0.
Patch by Wouter van Oortmerssen
Differential Revision: https://reviews.llvm.org/D45848
llvm-svn: 332052
Summary:
This feature is required for proper libFuzzer support.
Adding SI_NETBSD to the list of OSes, fixes breakage in several
libFuzzer tests. It has been debugged with aid from kcc@.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, vitalybuka, kcc
Reviewed By: vitalybuka
Subscribers: kubamracek, llvm-commits, #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D46729
llvm-svn: 332051
Clang's codegen now uses 128-bit masked load/store intrinsics in IR. The backend will widen to 512-bits on AVX512F targets.
So this patch adds patterns to detect codegen's widening and patterns for AVX512VL that don't get widened.
We may be able to drop some of the old patterns, but I leave that for a future patch.
llvm-svn: 332049
Replace decltype(memcpy) with decltype(__asan_memcpy) because memcpy
has not been defined in any headers on RTEMS. Similarly for memmove
and memset.
Differential Revision: https://reviews.llvm.org/D46625
llvm-svn: 332047
Summary:
adding function: `Cursor.get_included_file` , so the C API's `clang_getIncludedFile` function is available on the python binding interface
also adding test to unittests
related ticket: https://bugs.llvm.org/show_bug.cgi?id=15223
Reviewers: mgorny, arphaman, jbcoe
Reviewed By: jbcoe
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D46383
Patch by jlaz (József Láz)
llvm-svn: 332045
This reverts commit SVN r331889, which could trigger failed
assertions for cases where the snprintf function is declared
with a vaguely differing signature (e.g. being defined as
static inline), see PR37408.
llvm-svn: 332043
Summary: Move and correct LLVMDIBuilderCreateTypedef. This is the last API in DIBuilderBindings.h, so it is being removed and the C API will now be re-exported from IRBindings.h.
Reviewers: whitequark, harlanhaskins, deadalnix
Reviewed By: whitequark
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46725
llvm-svn: 332041
Checking for complete types is really rather tricky when you consider
the amount of specializations required to check a function type. This
specifically caused PR37407 where we incorrectly diagnosed
noexcept function types as incomplete (but there were plenty of other
cases that would cause this).
This patch removes the complete type checking for now. I'm going
to look into adding a clang builtin to correctly do this for us.
llvm-svn: 332040
This CL is to mitigate R_X86_64_PC32 relocation overflow problems for huge binaries that has near 4G allocated sections.
By examining those binaries, there're 2 issues contributes to the problem:
1). huge ".dynsym" and ".dynstr" stands in the way between .rodata and .text
2). _init_array_start/end are placed at 0 if no ".init_array" presents, this causes .text relocation against them become more prone to overflow.
This CL addresses 1st problem (the 2nd will be addressed in another CL.) by assigning a smaller sortrank to .dynsym and .dynstr thus they no longer stand in between.
llvm-svn: 332038
Summary:
NetBSD can use the approach that exists in FreeBSD, Linux and SunOS.
Pick the FreeBSD one as marking programs with "-z origin" is useful.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, vitalybuka, kcc
Reviewed By: vitalybuka
Subscribers: emaste, fedor.sergeev, llvm-commits, #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D46718
llvm-svn: 332035
If detect-stack-use-after-return is on, initialize fake stack during
AsanThread::Init(), rather than lazily. This is required on Myriad.
From kcc: "There used to be a reason why this was done lazily, but I
don't remember if we still have that reason." Tested on x86.
Differential Revision: https://reviews.llvm.org/D46626
llvm-svn: 332033
length excluding the table header. Instead it must encode the contribution length minus the length
field itself.
Reviewer: JDevliegehere
Differential Revision: https://reviews.llvm.org/D45922
llvm-svn: 332030
Summary:
Experimental data flow tracer for fuzz targets.
Allows to tell which bytes of the input affect which functions of the fuzz target.
We previously attempted to use DFSan directly in the libFuzzer process,
and that didn't work nicely.
Now we will try to collect the data flow information for the seed corpus
in a separate process (using this tracer), and then use it in the regular libFuzzer runs.
Reviewers: morehouse, pcc, Dor1s
Reviewed By: morehouse, Dor1s
Subscribers: delcypher, #sanitizers, llvm-commits
Differential Revision: https://reviews.llvm.org/D46666
llvm-svn: 332029
Summary:
The Itanium ABI requires that the type info for pointer-to-incomplete types to have internal linkage, so that it doesn't interfere with the type info once completed. Currently it also marks the type info name as internal as well. However, this causes a bug with the STL implementations, which use the type info name pointer to perform ordering and hashing of type infos.
For example:
```
// header.h
struct T;
extern std::type_info const& Info;
// tu_one.cpp
#include "header.h"
std::type_info const& Info = typeid(T*);
// tu_two.cpp
#include "header.h"
struct T {};
int main() {
auto &TI1 = Info;
auto &TI2 = typeid(T*);
assert(TI1 == TI2); // Fails
assert(TI1.hash_code() == TI2.hash_code()); // Fails
}
```
This patch fixes the STL bug by emitting the type info name as linkonce_odr when the type-info is for a pointer-to-incomplete type.
Note that libc++ could fix this without a compiler change, but the quality of fix would be poor. The library would either have to:
(A) Always perform strcmp/string hashes.
(B) Determine if we have a pointer-to-incomplete type, and only do strcmp then. This would require an ABI break for libc++.
Reviewers: rsmith, rjmccall, majnemer, vsapsai
Reviewed By: rjmccall
Subscribers: smeenai, cfe-commits
Differential Revision: https://reviews.llvm.org/D46665
llvm-svn: 332028
This commit relands r331905.
r331904 added SrcMgr::CharacteristicKind to the InclusionDirective
callback, this revision updates instances of it in clang-tools-extra.
llvm-svn: 332023
This commit relands r331904.
Adding a SrcMgr::CharacteristicKind parameter to the InclusionDirective
in PPCallbacks, and updating calls to that function. This will be useful
in https://reviews.llvm.org/D43778 to determine which includes are
system
headers.
Differential Revision: https://reviews.llvm.org/D46614
llvm-svn: 332021
Summary:
The SanitizerCommon-lsan-x86_64-Linux test failed due to the address of
the very first allocation ending up in the stack through "delete[]".
Workaround this by performing another allocation. The issue was only
present with optimization enabled, the test would pass with -O0.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D46650
llvm-svn: 332020
In general, it's difficult to poke the ConstantExpr code in CFLAA, since
LLVM is so great at eagerly reducing ConstantExprs. :)
Sadly, this only shows a functional difference from before the patch
because CFLAA has some special logic around taking loads of non-pointers
into account. Namely, with the broken select behavior, CFLAA will
completely fail to take note of @g3. Since CFLAA doesn't have any record
about @g3 when we do an alias query for @g3 and %a, it conservatively
answers MayAlias. When we properly take @g3 into account with the new
select logic, we get NoAlias for this query.
I suspect that the aforementioned "special logic" isn't completely
correct, but this test-case should prevent future wonky aliasing results
from appearing for these flavors of ConstantExprs, so I think it's still
worth having.
llvm-svn: 332017
Added initial support for L2 parallelism in SPMD mode. Note, though,
that the orphaned parallel directives are not currently supported in
SPMD mode.
llvm-svn: 332016
Accessing the members of a large data structures needs a lot of GEPs which
usually have large offsets due to the size of the underlying data structure. If
the offsets are too large to fit into the r+i addressing mode, these GEPs cannot
be sunk to their users' blocks and many extra registers are needed then to carry
the values of these GEPs.
This patch tries to split a large data struct starting from %base like the
following.
Before:
BB0:
%base =
BB1:
%gep0 = gep %base, off0
%gep1 = gep %base, off1
%gep2 = gep %base, off2
BB2:
%load1 = load %gep0
%load2 = load %gep1
%load3 = load %gep2
After:
BB0:
%base =
%new_base = gep %base, off0
BB1:
%new_gep0 = %new_base
%new_gep1 = gep %new_base, off1 - off0
%new_gep2 = gep %new_base, off2 - off0
BB2:
%load1 = load i32, i32* %new_gep0
%load2 = load i32, i32* %new_gep1
%load3 = load i32, i32* %new_gep2
In the above example, the struct is split into two parts. The first part still
starts from %base and the second part starts from %new_base. After the
splitting, %new_gep1 and %new_gep2 have smaller offsets and then can be sunk to
BB2 and folded into their users.
The algorithm to split data structure is simple and very similar to the work of
merging SExts. First, it collects GEPs that have large offsets when iterating
the blocks. Second, it splits the underlying data structures and updates the
collected GEPs to use smaller offsets.
Differential Revision: https://reviews.llvm.org/D42759
llvm-svn: 332015
Summary:
- Adds getters for the line, column, and scope of a DILocation
- Adds getters for the name, size in bits, offset in bits, alignment in bits, line, and flags of a DIType
Reviewers: whitequark, harlanhaskins, deadalnix
Reviewed By: whitequark
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46627
llvm-svn: 332014
Merging data segments produces smaller code sizes because each segment
has some boilerplate. Therefore, merging data segments is generally the
right approach, especially with wasm where binaries are typically
delivered over the network.
However, when analyzing wasm binaries, it can be helpful to get a
conservative picture of which functions are using which data
segments[0]. Perhaps there is a large data segment that you didn't
expect to be included in the wasm, introduced by some library you're
using, and you'd like to know which library it was. In this scenario,
merging data segments only makes the analysis worse.
Alternatively, perhaps you will remove some dead functions by-hand[1]
that can't be statically proven dead by the compiler or lld, and
removing these functions might make some data garbage collect-able, and
you'd like to run `--gc-sections` again so that this now-unused data can
be collected. If the segments were originally merged, then a single use
of the merged data segment will entrench all of the data.
[0] https://github.com/rustwasm/twiggy
[1] https://github.com/fitzgen/wasm-snip
Patch by Nick Fitzgerald!
Differential Revision: https://reviews.llvm.org/D46417
llvm-svn: 332013
This is most useful when using lld-link on a non-Win host (but it might become
useful on Windows too if lld also grows a fansi-escape-codes flag).
Also make the help for --color-diagnostic mention the valid values in ELF and
wasm, and print the flag name with two dashes in diags, since the one-dash form
is seen as a list of many one-letter flags in some contexts.
https://reviews.llvm.org/D46693
llvm-svn: 332012
Fangrui Song