getmntent() thread-unsafe in hwloc_topology_load() on Linux #194
Comments
|
I look at using getmntent_r(), thanks. |
|
By the way, you should use a single topology for both threads whenever possible. |
bgoglin
changed the title
bgoglin
added a commit
that referenced
this issue
Jul 8, 2016
getmntent() isn't thread safe, it uses a static internal struct mntent. Two concurrent hwloc_topology_load() would break if they call getmntent() (for finding the cgroup directory) at the same time. Use getmntent_r() instead. However we have to specify a buffer for storing mntent strings. And getmntent_r() doesn't actually report an error if the buffer is too small, it silently truncates output strings, so we can't dynamically realloc that buffer. The getmntent() that we used before this commit was internally limited to 4kB. And Linux actually limits mount options to 3 pages (during mount, not when reading /proc/mounts). So use 4 pages to be above both. Thanks to Corentin Rossignon for reporting the issue (using gcc -fsanitize=thread) Fixes #194 Refs #142 (cherry picked from commit 2337361)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi,
Thread sanitizer report a data race in getmntent which is called at least by hwloc_topology_load.
How to reproduce
Call the sequence "hwloc_topology_init, hwloc_topology_load and hwloc_get_cpubind" in two threads.
Compile the example
with a recent gcc (> 4.8)
gcc -fsanitize=thread getmntent_tsan.c -lhwloc -lpthread
Possible solutions
Use re-entrant version getmntent_r
OR
Put a static mutex around call of hwloc_find_linux_cpuset_mntpnt.
System informations
Linux 2.6.32-504.el6.x86_64 #1 SMP Wed Oct 15 04:27:16 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
Distributor ID: CentOS
Description: CentOS release 6.6 (Final)
Release: 6.6
Codename: Final
Hwloc 1.11.3
gcc 5.3
tsan.txt
getmntent_tsan.c.txt
The text was updated successfully, but these errors were encountered: