Right now, the counters are added in regards of the number of successors
for a given BasicBlock: it's good when we've only 1 or 2 successors (at
least with BranchInstr). But in the case of a switch statement, the
BasicBlock after switch has several predecessors and we need know from
which BB we're coming from.
So the idea is to revert what we're doing: add a PHINode in each block
which will select the counter according to the incoming BB. They're
several pros for doing that:
- we fix the "switch" bug
- we remove the function call to "__llvm_gcov_indirect_counter_increment"
and the lookup table stuff
- we replace by PHINodes, so the optimizer will probably makes a better
job.
Patch by calixte!
Differential Revision: https://reviews.llvm.org/D51619
llvm-svn: 341977
Frontends emit 'unused' coverage mapping records for functions which are
provably unused in a TU. These unused records contain a single counter
with CounterKind::Zero. However, a function may be unused in one TU and
used in another. When this happens, prefer the records with a full set
of counters instead of arbitrarily picking the first loaded record.
There is no impact on the single-TU case. In the multiple-TU case, this
resolves issues causing a function to appear unused when it's not.
Testing: check-{llvm,clang,compiler-rt}
rdar://42981322
llvm-svn: 339194
Summary:
This patch introduces `llvm_profile_set_dir_mode` and `llvm_profile_get_dir_mode` to
the compiler-rt profile API.
Originally, profile data was placed into a directory that was created with a hard-coded
mode value of 0755 (for non-win32 builds). In certain cases, it can be helpful to create
directories with a different mode other than 0755. This patch introduces set/get
routines to allow users to specify a desired mode. The default remains at 0755.
Reviewers: void, probinson
Reviewed By: probinson
Subscribers: probinson, dberris, cfe-commits
Differential Revision: https://reviews.llvm.org/D49953
llvm-svn: 338456
Summary:
Add __llvm_profile_get_filename interface to get the profile filename,
which can be used for identifying which profile file belongs to an app
when multiple binaries are instrumented and dumping profiles into the
same directory. The filename includes the path.
Reviewers: davidxl
Subscribers: delcypher, #sanitizers, llvm-commits
Differential Revision: https://reviews.llvm.org/D49529
llvm-svn: 337482
Summary:
This will make the behavior of __gcov_flush match the GCC behavior.
I would like to rename __gcov_flush to __llvm_gcov_flush (in case of programs linking to libraries built with different compilers), but I guess we can't for compatibility reasons.
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: samsonov, vitalybuka, pcc, kcc, junbuml, glider, fhahn, eugenis, dvyukov, davidxl, srhines, chh, llvm-commits, #sanitizers
Differential Revision: https://reviews.llvm.org/D48538
llvm-svn: 336365
__gcov_flush is hidden.
For applications to dump profiling data of selected .so files,
they can use dlsym to find and call llvm_gcov_flush in each .so file.
Differential Revision: https://reviews.llvm.org/D45454
llvm-svn: 336019
This patch fixes the following issues:
(1) The strong definition of the merge hook function was not working which
breaks the online value profile merging. This patch removes the weak
attribute of VPMergeHook and assigns the value dynamically.
(2) Truncate the proifle file so that we don't have garbage data at the end of
the file.
(3) Add new __llvm_profile_instrument_target_value() interface to do the value
profile update in batch. This is needed as the original incremental by 1
in __llvm_profile_instrument_target() is too slow for online merge.
Differential Revision: https://reviews.llvm.org/D44847
llvm-svn: 328987
Programs without any code in them should export the exact same set of
symbols as programs with code, at least on Darwin. This is done to make
text-based API verification possible for certain Darwin frameworks.
llvm-svn: 326351
The test expects to see the pattern "load add store" twice
in a row.
However the test contains actually several instances of this pattern.
Currently the first pattern appears in the output twice.
But after LLVM change https://reviews.llvm.org/D42691 some addtional
modification is done and one of another pattern has been hoisted in
function entry block.
As a result we have another order of meeting this pattern.
The test is updated to meet this change.
llvm-svn: 324571
This includes a few nice bits of refactoring (e.g splitting out the
exclusive locking code into a common utility).
Hopefully the Windows support is fixed now.
Patch by Rainer Orth!
Differential Revision: https://reviews.llvm.org/D40944
llvm-svn: 320731
This includes a few nice bits of refactoring (e.g splitting out the
exclusive locking code into a common utility).
Patch by Rainer Orth!
Differential Revision: https://reviews.llvm.org/D40944
llvm-svn: 320726
Summary:
As discussed in https://github.com/google/oss-fuzz/issues/933,
it would be really awesome to be able to use ThinLTO for fuzzing.
However, as @kcc has pointed out, it is currently undefined (untested)
whether the sanitizers actually function properly with LLD and/or LTO.
This patch is inspired by the cfi test, which already do test with LTO
(and/or LLD), since LTO is required for CFI to function.
I started with UBSan, because it's cmakelists / lit.* files appeared
to be the cleanest. This patch adds the infrastructure to easily add
LLD and/or LTO sub-variants of the existing lit test configurations.
Also, this patch adds the LLD flavor, that explicitly does use LLD to link.
The check-ubsan does pass on my machine. And to minimize the [initial]
potential buildbot breakage i have put some restrictions on this flavour.
Please review carefully, i have not worked with lit/sanitizer tests before.
The original attempt, r319525 was reverted in r319526 due
to the failures in compiler-rt standalone builds.
Reviewers: eugenis, vitalybuka
Reviewed By: eugenis
Subscribers: #sanitizers, pcc, kubamracek, mgorny, llvm-commits, mehdi_amini, inglorion, kcc
Differential Revision: https://reviews.llvm.org/D39508
llvm-svn: 319575
This is a resubmission of r313270. It broke standalone builds of
compiler-rt because we were not correctly generating the llvm-lit
script in the standalone build directory.
The fixes incorporated here attempt to find llvm/utils/llvm-lit
from the source tree returned by llvm-config. If present, it
will generate llvm-lit into the output directory. Regardless,
the user can specify -DLLVM_EXTERNAL_LIT to point to a specific
lit.py on their file system. This supports the use case of
someone installing lit via a package manager. If it cannot find
a source tree, and -DLLVM_EXTERNAL_LIT is either unspecified or
invalid, then we print a warning that tests will not be able
to run.
Differential Revision: https://reviews.llvm.org/D37756
llvm-svn: 313407
This patch is still breaking several multi-stage compiler-rt bots.
I already know what the fix is, but I want to get the bots green
for now and then try re-applying in the morning.
llvm-svn: 313335
This patch simplifies LLVM's lit infrastructure by enforcing an ordering
that a site config is always run before a source-tree config.
A significant amount of the complexity from lit config files arises from
the fact that inside of a source-tree config file, we don't yet know if
the site config has been run. However it is *always* required to run
a site config first, because it passes various variables down through
CMake that the main config depends on. As a result, every config
file has to do a bunch of magic to try to reverse-engineer the location
of the site config file if they detect (heuristically) that the site
config file has not yet been run.
This patch solves the problem by emitting a mapping from source tree
config file to binary tree site config file in llvm-lit.py. Then, during
discovery when we find a config file, we check to see if we have a
target mapping for it, and if so we use that instead.
This mechanism is generic enough that it does not affect external users
of lit. They will just not have a config mapping defined, and everything
will work as normal.
On the other hand, for us it allows us to make many simplifications:
* We are guaranteed that a site config will be executed first
* Inside of a main config, we no longer have to assume that attributes
might not be present and use getattr everywhere.
* We no longer have to pass parameters such as --param llvm_site_config=<path>
on the command line.
* It is future-proof, meaning you don't have to edit llvm-lit.in to add
support for new projects.
* All of the duplicated logic of trying various fallback mechanisms of
finding a site config from the main config are now gone.
One potentially noteworthy thing that was required to implement this
change is that whereas the ninja check targets previously used the first
method to spawn lit, they now use the second. In particular, you can no
longer run lit.py against the source tree while specifying the various
`foo_site_config=<path>` parameters. Instead, you need to run
llvm-lit.py.
Differential Revision: https://reviews.llvm.org/D37756
llvm-svn: 313270
