Summary:
We found a nondeterministic behavior when doing online profile merging
for multi-process applications. The application forks a sub-process and
sub-process sets to get SIGKILL when the parent process exits,
The first process gets the lock, and dumps the profile. The second one
will mmap the file, do the merge and write out the file. Note that before
the merged write, we truncate the profile.
Depending on the timing, the child process might be terminated
abnormally when the parent exits first. If this happens:
(1) before the truncation, we will get the profile for the main process
(2) after the truncation, and before write-out the profile, we will get
0 size profile.
(3) after the merged write, we get merged profile.
This patch temporarily suspend the SIGKILL for PR_SET_PDEATHSIG
before profile-write and restore it after the write.
This patch only applies to Linux system.
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: xur, llvm-commits
Differential Revision: https://reviews.llvm.org/D29954
llvm-svn: 295364
(Part-1 merging API is in profile runtime)
This patch implements a portable file opening API
with exclusive access for the process. In-process
profile merge requires profile file update to be
atomic/fully sychronized.
llvm-svn: 271864
licensees actually see in the toolchain we deliver to them. This will
reduce the set of local patches we have to maintain. The triple is
not changing. (The term ORBIS is an internal code name for PS4.)
llvm-svn: 269672
Summary:
Hi David, SCE folks,
What is implemented in this patch is enough for the upstream libprofile to
work for PGO with the PS4 game codebase I tested ("game7" for you SCE
folks; this is with a standalone build of compiler-rt).
The first change, which is simple, is to stub out gethostname. PS4
doesn't have a simple analog for this that doesn't bring in extra
OS libraries, so for now we do not support `%h` expansion.
This is consistent with internal B#136272.
The second change implies future work, but is a simple change at present.
PS4 does not have `getenv`, so for now we will introduce a shim.
This obviously makes it impossible for many of the tests to be run since
they require setting `LLVM_PROFILE_FILE=`.
I see two paths forward:
1. In the tests we are already wrapping execution with `%run` and so by
setting a PS4-specific expansion for `%run` we can pass the information
in another way We can adapt the getenv shim as appropriate.
We will need to experiment with this internally.
Maggie, Phillip, Filipe? Any ideas? Maybe ping me internally since we
may need to get into some PS4 vagaries. I'm thinking a fake getenv
library that uses some side channel for communication.
2. Another possibility which is more verbose is to use a separate clang
invocation with `-profile-generate=<filename>` to set the filename in
each test.
This might require redundant clang invocations though which may be
undesirable for upstream. David, thoughts?
Also, this is a fairly libprofile-specific workaround, so it e.g.
doesn't help Filipe's ASan work.
Overall, this approach sounds like a bit of a hack to me.
Small detail:
InstrProfilingPort.h seems like the natural place for the getenv shim,
but GCDAProfiling.c needs it as well. InstrProfilingUtil.h is currently
the only header common between InstrProfilingFile.c and GCDAProfiling.c.
I can move the shim to InstrProfilingPort.h and add an include to
GCDAProfiling.c as per your preference David.
Reviewers: davidxl, MaggieYi, phillip.power, filcab
Subscribers: simon.f.whittaker, slingn, probinson, llvm-commits
Differential Revision: http://reviews.llvm.org/D17676
llvm-svn: 262527
When the file is initialized, this patch checks whether the path
specifies a directory. If so, it creates the directory tree before
truncating the file.
Use default.profdata instead of pgo-data for default indexed profile name.
llvm-svn: 241824
Renato Golin