Blob Blame History Raw
summary: Test for bz522528 (pthread_join() hangs if a thread calls setuid())
description: |
    Bug summary: pthread_join() hangs if a thread calls setuid()
    Bugzilla link: https://bugzilla.redhat.com/show_bug.cgi?id=522528

    Description:

    Created an attachment (id=360523)
    reproducer's source file

    Description of problem:

    When a program creates a thread which calls setuid() and terminates, the other
    thread trying to wait for it hangs.


    Version-Release number of selected component (if applicable):

    RHEL5 (2.5-34), also RHEL4 and the newest upstream sources


    How reproducible:

    always


    Steps to Reproduce:

    1. Compile attached source file of reproducer:

    $ gcc -o reproducer reproducer.c -lpthread

    2. and run it:

    $ ./reproducer

      
    Actual results:

    program hangs


    Expected results:

    program should end without hanging


    Additional info:

    This bug has appeared quite a long ago, you can find it in a sourceware BZ with existing patch at:

    http://sources.redhat.com/bugzilla/show_bug.cgi?id=3270#c2

    The patch solves the problem, but I have no idea why this hasn't been applied in the upstream already. I have attached a slightly modiffied version of the patch to be able to be applied and built on RHEL5 glibc sources and it works and solves the problem.

    There is also existing BZ for RHEL4:

    https://bugzilla.redhat.com/show_bug.cgi?id=248671
contact: Petr Muller <pmuller@redhat.com>
component:
  - glibc
test: ./runtest.sh
framework: beakerlib
recommend:
  - gcc
tag:
  - simple
  - tier1_mfranc
  - mfranc_stable
  - noEWA
  - Tier1
  - not-er15271
  - glibc-buildroot-ready
duration: 75m
link:
  - relates: https://bugzilla.redhat.com/show_bug.cgi?id=522528
extra-summary: /tools/glibc/Regression/bz522528-pthread-join-hangs-if-a-thread-calls-setuid
extra-task: /tools/glibc/Regression/bz522528-pthread-join-hangs-if-a-thread-calls-setuid