Asan doesn't report invalid write

[jeremysalwen]

I have a test that perform an invalid write inside unsafe code. However, this program just crashes with a segfault when asan is enabled. No backtrace, or error message besides "Segmentation Fault" is output, and GDB crashes instead of breaking on the error. The only way I was able to get a valid backtrace was with valgrind.

In my experience with asan for C++, this error would definitely have been caught. Is this a bug in the Rust asan integration? I am happy to help gather more debug information about what is happening here.

Valgrind trace:

jeremysalwen@jeremysalwen-glaptop:~/fun/extended-collections-rs$ valgrind /home/jeremysalwen/fun/extended-collections-rs/target/debug/deps/extended_collections-4040d7db4dc997da --test skiplist::tsplist --test-threads=1
==169673== Memcheck, a memory error detector
==169673== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==169673== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==169673== Command: /home/jeremysalwen/fun/extended-collections-rs/target/debug/deps/extended_collections-4040d7db4dc997da --test skiplist::tsplist --test-threads=1
==169673== 

running 13 tests
test skiplist::tsplist::tests::test_add ... ==169673== Invalid write of size 8
==169673==    at 0x1F0C05: extended_collections::skiplist::tsplist::link (tsplist.rs:76)
==169673==    by 0x1F1D1E: <extended_collections::skiplist::tsplist::TspList<T, A>>::insert (tsplist.rs:199)
==169673==    by 0x274CD2: extended_collections::skiplist::tsplist::tests::test_add (tsplist.rs:853)
==169673==    by 0x225649: extended_collections::skiplist::tsplist::tests::test_add::{{closure}} (tsplist.rs:851)
==169673==    by 0x1635CD: core::ops::function::FnOnce::call_once (function.rs:231)
==169673==    by 0x3347BE: {{closure}} (lib.rs:1473)
==169673==    by 0x3347BE: call_once<closure,()> (function.rs:231)
==169673==    by 0x3347BE: <F as alloc::boxed::FnBox<A>>::call_box (boxed.rs:724)
==169673==    by 0x390AD9: __rust_maybe_catch_panic (lib.rs:92)
==169673==    by 0x351E37: try<(),std::panic::AssertUnwindSafe<alloc::boxed::Box<FnBox<()>>>> (panicking.rs:276)
==169673==    by 0x351E37: catch_unwind<std::panic::AssertUnwindSafe<alloc::boxed::Box<FnBox<()>>>,()> (panic.rs:388)
==169673==    by 0x351E37: test::run_test::run_test_inner::{{closure}} (lib.rs:1428)
==169673==    by 0x35175C: test::run_test::run_test_inner (lib.rs:1450)
==169673==    by 0x34FEDB: test::run_test (lib.rs:1469)
==169673==    by 0x3489CC: run_tests<closure> (lib.rs:1149)
==169673==    by 0x3489CC: test::run_tests_console (lib.rs:956)
==169673==    by 0x3404A3: test::test_main (lib.rs:289)
==169673==  Address 0x20 is not stack'd, malloc'd or (recently) free'd
==169673== 
==169673== 
==169673== Process terminating with default action of signal 11 (SIGSEGV)
==169673==  Access not within mapped region at address 0x20
==169673==    at 0x1F0C05: extended_collections::skiplist::tsplist::link (tsplist.rs:76)
==169673==    by 0x1F1D1E: <extended_collections::skiplist::tsplist::TspList<T, A>>::insert (tsplist.rs:199)
==169673==    by 0x274CD2: extended_collections::skiplist::tsplist::tests::test_add (tsplist.rs:853)
==169673==    by 0x225649: extended_collections::skiplist::tsplist::tests::test_add::{{closure}} (tsplist.rs:851)
==169673==    by 0x1635CD: core::ops::function::FnOnce::call_once (function.rs:231)
==169673==    by 0x3347BE: {{closure}} (lib.rs:1473)
==169673==    by 0x3347BE: call_once<closure,()> (function.rs:231)
==169673==    by 0x3347BE: <F as alloc::boxed::FnBox<A>>::call_box (boxed.rs:724)
==169673==    by 0x390AD9: __rust_maybe_catch_panic (lib.rs:92)
==169673==    by 0x351E37: try<(),std::panic::AssertUnwindSafe<alloc::boxed::Box<FnBox<()>>>> (panicking.rs:276)
==169673==    by 0x351E37: catch_unwind<std::panic::AssertUnwindSafe<alloc::boxed::Box<FnBox<()>>>,()> (panic.rs:388)
==169673==    by 0x351E37: test::run_test::run_test_inner::{{closure}} (lib.rs:1428)
==169673==    by 0x35175C: test::run_test::run_test_inner (lib.rs:1450)
==169673==    by 0x34FEDB: test::run_test (lib.rs:1469)
==169673==    by 0x3489CC: run_tests<closure> (lib.rs:1149)
==169673==    by 0x3489CC: test::run_tests_console (lib.rs:956)
==169673==    by 0x3404A3: test::test_main (lib.rs:289)
==169673==  If you believe this happened as a result of a stack
==169673==  overflow in your program's main thread (unlikely but
==169673==  possible), you can try to increase the size of the
==169673==  main thread stack using the --main-stacksize= flag.
==169673==  The main thread stack size used in this run was 8388608.
==169673== 
==169673== HEAP SUMMARY:
==169673==     in use at exit: 25,083 bytes in 395 blocks
==169673==   total heap usage: 594 allocs, 199 frees, 90,729 bytes allocated
==169673== 
==169673== LEAK SUMMARY:
==169673==    definitely lost: 0 bytes in 0 blocks
==169673==    indirectly lost: 0 bytes in 0 blocks
==169673==      possibly lost: 0 bytes in 0 blocks
==169673==    still reachable: 25,083 bytes in 395 blocks
==169673==         suppressed: 0 bytes in 0 blocks
==169673== Rerun with --leak-check=full to see details of leaked memory
==169673== 
==169673== For counts of detected and suppressed errors, rerun with: -v
==169673== ERROR SUMMARY: 2 errors from 1 contexts (suppressed: 0 from 0)
Segmentation fault