Disable LLVM assertions on Nightly, enable them in "alt" builds.

[SimonSapin]

Per IRC discussion https://mozilla.logbot.info/rust-infra/20171106#c13812170-c13812204

Background: https://internals.rust-lang.org/t/disabling-llvm-assertions-in-nightly-builds/5388/14

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @nikomatsakis (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[aturon]

@bors r+

[bors]

📌 Commit eac34c9 has been approved by aturon

[aturon]

cc @rust-lang/infra, just to keep everyone in the loop on this change.

[kennytm]

I think we also need to update .travis.yml to point the try branch to the alt build so crater and perf can still test with assertion enabled.

[SimonSapin]

@kennytm I’ve pushed another commit. How does it look?

Though I don’t know if perf should have LLVM assertions, since that produces timings different from what everyone will use.

[kennytm]

@SimonSapin Unfortunately, unless we have two try branches we are stuck with either sacrificing perf or crater. At least enabling assertions allows the metrics be comparable before and after this PR.

@bors r=aturon

[bors]

📌 Commit dfb08b6 has been approved by aturon

[michaelwoerister]

@Mark-Simulacrum @alexcrichton Will we back-collect perf.rlo data after this (e.g. by compiling with the former "alt builds" for older nightlies)?

[Mark-Simulacrum]

Since this will require changes to perf.rlo anyway, I should be able to make it so that we properly collect non-alt before and alt now. Ideally, we'd add something like @bors perf which would output dedicated artifacts for perf (and then we wouldn't have to share between crater, which does want assertions, and perf, which doesn't). It's also worth thinking about tracking how "bad" assertions are.

[michaelwoerister]

Ideally, we'd add something like @bors perf which would output dedicated artifacts for perf

It would be nice if perf.rlo tested with debug assertions disabled too since that is closer to how stable releases perform. Also, I'm quite liberal with debug assertions :)

[nikomatsakis]

@Mark-Simulacrum

It's also worth thinking about tracking how "bad" assertions are.

by this do you mean running with assertions both enable and disabled and comparing the results?

[Mark-Simulacrum]

Yes, exactly.

[shepmaster]

@SimonSapin if you happen to push another commit, can you fix the typo "Nigthly"? It exists existed in the PR title and in your first commit.

[oyvindln]

Is it a good idea to merge this as long as #45220 isn't solved, given that it may result in that ending up accessing out-of-bounds memory rather than panicking?

[bors]

⌛️ Testing commit dfb08b6 with merge 9a9ee1b...

[bors]

💔 Test failed - status-travis

[SimonSapin]

[01:12:34] failures:
[01:12:34] thread 'main' panicked at 'Some tests failed', /checkout/src/tools/compiletest/src/main.rs:331:21
[01:12:34] 
[01:12:34] ---- [run-make] run-make/sanitizer-memory stdout ----
[01:12:34]      
[01:12:34] error: make failed
[01:12:34] status: exit code: 2
[01:12:34] command: "make"
[01:12:34] stdout:
[01:12:34] ------------------------------------------
[01:12:34] LD_LIBRARY_PATH="/checkout/obj/build/x86_64-unknown-linux-gnu/test/run-make/sanitizer-memory.stage2-x86_64-unknown-linux-gnu:/checkout/obj/build/x86_64-unknown-linux-gnu/stage2/lib:/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-tools/x86_64-unknown-linux-gnu/release/deps:/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-sysroot/lib/rustlib/x86_64-unknown-linux-gnu/lib:" '/checkout/obj/build/x86_64-unknown-linux-gnu/stage2/bin/rustc' --out-dir /checkout/obj/build/x86_64-unknown-linux-gnu/test/run-make/sanitizer-memory.stage2-x86_64-unknown-linux-gnu -L /checkout/obj/build/x86_64-unknown-linux-gnu/test/run-make/sanitizer-memory.stage2-x86_64-unknown-linux-gnu  -g -Z sanitizer=memory -Z print-link-args uninit.rs | grep -q librustc_msan
[01:12:34] /checkout/obj/build/x86_64-unknown-linux-gnu/test/run-make/sanitizer-memory.stage2-x86_64-unknown-linux-gnu/uninit 2>&1 | grep -q use-of-uninitialized-value
[01:12:34] Makefile:6: recipe for target 'all' failed
[01:12:34] 
[01:12:34] ------------------------------------------
[01:12:34] stderr:
[01:12:34] ------------------------------------------
[01:12:34] warning: unused variable: `y`
[01:12:34]   --> uninit.rs:15:9
[01:12:34]    |
[01:12:34] 15 |     let y = xs[0] + xs[1];
[01:12:34]    |         ^
[01:12:34]    |
[01:12:34]    = note: #[warn(unused_variables)] on by default
[01:12:34]    = note: to avoid this warning, consider using `_y` instead
[01:12:34] 
[01:12:34] make: *** [all] Error 1
[01:12:34] 
[01:12:34] ------------------------------------------
[01:12:34] 
[01:12:34] thread '[run-make] run-make/sanitizer-memory' panicked at 'explicit panic', /checkout/src/tools/compiletest/src/runtest.rs:2505:8
[01:12:34] note: Run with `RUST_BACKTRACE=1` for a backtrace.
[01:12:34] 
[01:12:34] 
[01:12:34] failures:
[01:12:34]     [run-make] run-make/sanitizer-memory
[01:12:34] 
[01:12:34] test result: FAILED. 163 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out

I’m trying to reproduce locally.

[SimonSapin]

:(

$ ./x.py test src/test/run-make --test-args sanitizer-memory
[…]
running 1 test
test [run-make] run-make/sanitizer-memory ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 163 filtered out

[SimonSapin]

Ah, this test is entirely inside ifdef SANITIZER_SUPPORT. When building with --enable-sanitizers I get errors below. The relevant part seems to be error: aggregate ‘sigaltstack handler_stack’ has incomplete type and cannot be defined.

[ 90%] Built target RTAsan_dynamic.x86_64
CMakeFiles/Makefile2:1636: recipe for target 'lib/asan/CMakeFiles/asan.dir/rule' failed
Makefile:632: recipe for target 'asan' failed

--- stderr
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/ubsan/ubsan_handlers_cxx.cc: In function ‘void __ubsan::HandleCFIBadType(__ubsan::CFICheckFailData*, __ubsan::ValueHandle, bool, __ubsan::ReportOptions)’:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/ubsan/ubsan_handlers_cxx.cc:111:15: warning: ‘CheckKindStr’ may be used uninitialized in this function [-Wmaybe-uninitialized]
   const char *CheckKindStr;
               ^~~~~~~~~~~~
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc: In function ‘int __sanitizer::TracerThread(void*)’:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc:278:22: error: aggregate ‘sigaltstack handler_stack’ has incomplete type and cannot be defined
   struct sigaltstack handler_stack;
                      ^~~~~~~~~~~~~
make[3]: *** [lib/sanitizer_common/CMakeFiles/RTSanitizerCommonLibc.x86_64.dir/sanitizer_stoptheworld_linux_libcdep.cc.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [lib/sanitizer_common/CMakeFiles/RTSanitizerCommonLibc.x86_64.dir/all] Error 2
make[2]: *** Waiting for unfinished jobs....
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/sanitizer_common/sanitizer_posix.cc: In function ‘__sanitizer::fd_t __sanitizer::OpenFile(const char*, __sanitizer::FileAccessMode, __sanitizer::error_t*)’:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/sanitizer_common/sanitizer_posix.cc:215:27: warning: ‘flags’ may be used uninitialized in this function [-Wmaybe-uninitialized]
   fd_t res = internal_open(filename, flags, 0660);
              ~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~
In file included from /home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/asan_interceptors.cc:265:0:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/../sanitizer_common/sanitizer_common_interceptors.inc: In function ‘__sanitizer::uptr __interceptor_ptrace(int, int, void*, void*)’:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/../sanitizer_common/sanitizer_common_interceptors.inc:2810:21: warning: ‘local_iovec.__sanitizer::__sanitizer_iovec::iov_len’ may be used uninitialized in this function [-Wmaybe-uninitialized]
   __sanitizer_iovec local_iovec;
                     ^~~~~~~~~~~
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/asan_interceptors.cc:59:10: warning: ‘local_iovec.__sanitizer::__sanitizer_iovec::iov_base’ may be used uninitialized in this function [-Wmaybe-uninitialized]
     uptr __offset = (uptr)(offset);                                     \
          ^~~~~~~~
In file included from /home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/asan_interceptors.cc:265:0:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/../sanitizer_common/sanitizer_common_interceptors.inc:2810:21: note: ‘local_iovec.__sanitizer::__sanitizer_iovec::iov_base’ was declared here
   __sanitizer_iovec local_iovec;
                     ^~~~~~~~~~~
In file included from /home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/asan_interceptors.cc:265:0:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/../sanitizer_common/sanitizer_common_interceptors.inc: In function ‘__sanitizer::uptr __interceptor_ptrace(int, int, void*, void*)’:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/../sanitizer_common/sanitizer_common_interceptors.inc:2810:21: warning: ‘local_iovec.__sanitizer::__sanitizer_iovec::iov_len’ may be used uninitialized in this function [-Wmaybe-uninitialized]
   __sanitizer_iovec local_iovec;
                     ^~~~~~~~~~~
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/asan_interceptors.cc:59:10: warning: ‘local_iovec.__sanitizer::__sanitizer_iovec::iov_base’ may be used uninitialized in this function [-Wmaybe-uninitialized]
     uptr __offset = (uptr)(offset);                                     \
          ^~~~~~~~
In file included from /home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/asan_interceptors.cc:265:0:
/home/simon/rust/src/libcompiler_builtins/compiler-rt/lib/asan/../sanitizer_common/sanitizer_common_interceptors.inc:2810:21: note: ‘local_iovec.__sanitizer::__sanitizer_iovec::iov_base’ was declared here
   __sanitizer_iovec local_iovec;
                     ^~~~~~~~~~~
make[1]: *** [lib/asan/CMakeFiles/asan.dir/rule] Error 2
make: *** [asan] Error 2
thread 'main' panicked at '
command did not execute successfully, got: exit code: 2

build script failed, must exit now', /home/simon/.cargo/registry/src/github.com-1ecc6299db9ec823/cmake-0.1.26/src/lib.rs:599:4
note: Run with `RUST_BACKTRACE=1` for a backtrace.

thread 'main' panicked at 'command did not execute successfully: "/home/simon/rust/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "build" "--target" "x86_64-unknown-linux-gnu" "-j" "8" "--release" "--features" "panic-unwind jemalloc backtrace profiler" "--manifest-path" "/home/simon/rust/src/libstd/Cargo.toml" "--message-format" "json"
expected success, got: exit code: 101', src/bootstrap/compile.rs:882:8

[gnzlbg]

The comment in .travis.yml still says that alt-builds don't have assertions enabled, but it is actually the other way around 😆

[kennytm]

@gnzlbg Well PRs welcomed 😅

[SimonSapin]

#46330

[michaelwoerister]

@SimonSapin that's a comment you have to dereference twice in order to get to the actual content :D

[gnzlbg]

@michaelwoerister dammit michael! I tried dereferencing it twice and now I am trapped in a loop!