https://bugs.llvm.org/show_bug.cgi?id=40282
Bug ID: 40282
Summary: tsan: false positives on shmat
Product: compiler-rt
Version: unspecified
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: tsan
Assignee: unassignedb...@nondot.org
Reporter: dvyu...@google.com
CC: llvm-bugs@lists.llvm.org
Reported internally, don't have a public repro.
The crux: shmat/shmdt allocate/deallocate memory (like mmap/munmap). We don't
do anything in shmat/shmdt interceptors, so if the memory becomes reused across
shmat calls in different threads, tsan may report a false positive. Accesses to
the new mapping of the memory "race" with accesses to the old mapping at the
same address.
We need shmat/shmdt interceptors to do the same as mmap/munmap interceptors.
Probably the common code to "map" and "unmap" a range can be factored out.
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
