1132519 - TSan: multiple data races from libglib/libdbus/etc.

Status: RESOLVED FIXED

(Core :: General, defect)

Description

[Nathan Froyd [:froydnj]]

Attachment: glib-dbus-races.txt

[cribbing from decoder's script, hopefully he won't mind]

The attached logfile shows a thread/data race detected by TSan (ThreadSanitizer).

Typically, races reported by TSan are not false positives, but it is possible that the race is benign. Even in this case though, we should try to come up with a fix unless this would cause unacceptable performance issues. Also note that seemingly benign races can possibly be harmful (also depending on the compiler and the architecture) [1].

If the bug cannot be fixed, then this bug should be used to either make a compile-time annotation for blacklisting or add an entry to the runtime blacklist.

The stacks in the report obviously aren't helpful.  This is running on a Debian 7.0 system, so it's possibly that the races have been fixed in more recent versions of the libraries, but I'm not holding my breath.  Blacklisting these is probably the most reasonable thing to do.

[1] http://software.intel.com/en-us/blogs/2013/01/06/benign-data-races-what-could-possibly-go-wrong

Comment 1

[Christian Holler (:decoder)]

ThreadSanitizer now offers runtime suppressions for calls originating from certain libraries (e.g. if they are uninstrumented). It is reasonable to add the libraries mentioned here to our suppression list. This is also strictly required if you want to run a TSan build locally and actually get useful output. I will make a patch.

Comment 2

[Christian Holler (:decoder)]

Attachment: Bug 1132519 - [TSan] Suppress races in external uninstrumented libraries. r?jseward

Comment 3

[Christian Holler (:decoder)]

I'm skipping the patch here in favor of the patch in bug 1587463. That patch adds the libraries, but also switches things to use the builtin suppression list, which is a lot more robust for CI and automation in general. We should close this bug as FIXED once bug 1587463 landed.

Comment 4

[Christian Holler (:decoder)]

This should be fixed now from the combination of having the builtin suppression list and the use of the new called_from_lib feature in TSan.