#1199 bad parm to SockSelect segfaults 4.1.3 64-bit debian 7

v4.2.0
closed
Mark Miesfeld
None
none
1
2014-05-05
2013-10-20
Geoff Stevens
No

bad parm to SockSelect segfaults 4.1.3 64-bit debian 7

The issue occurs when the stem containing write sockets is incorrectly set (initially due to a typo). Not tested with the other sockets stems, but I suspect the same problem would occur.

Running udptest2.rex without a parameter sets the socket stem correctly and works. Running udptest2.rex with the parameter 'fail' sets the socket stem incorrectly and segfaults.

[614345.855481] udptest2.rex[12879]: segfault at 18 ip 00007f96fd6caf29 sp 00007fff8f34a160 error 4 in librexx.so.4.0.5[7f96fd587000+179000]

Additionally, there was no such problem in 4.1.0, making the while thing even more mysterious after an upgrade.

1 Attachments

Discussion

  • Mark Miesfeld
    Mark Miesfeld
    2013-10-20

    Geoff, thanks for the test case. Without it this bug would have just sat here.

    Committed revision 9477. [r9477] 4.1 fixes
    Committed revision 9478. [r9478] trunk

    C:\work.ooRexx\socket.segfaults>udptest2.rex fail
    REXX-ooRexx_4.1.4(MT) 6.03 29 Sep 2013
    rc1 1
    rc2 0
    crash

    C:\work.ooRexx\socket.segfaults>udptest2.rex fail
    REXX-ooRexx_4.1.4(MT) 6.03 20 Oct 2013
    rc1 1
    rc2 0
    timeout
    timeout
    timeout
    timeout
    timeout
    timeout
    timeout
    timeout
    timeout
    timeout

    C:\work.ooRexx\socket.segfaults>

     

    Related

    Commit: [r9477]
    Commit: [r9478]

  • Mark Miesfeld
    Mark Miesfeld
    2013-10-20

    • status: open --> pending
    • assigned_to: Mark Miesfeld
    • Group: None --> 4.1.4
     
  • Mark Miesfeld
    Mark Miesfeld
    2014-02-15

    • Group: 4.1.4 --> v4.2.0
     
  • Mark Miesfeld
    Mark Miesfeld
    2014-02-15

    In 4.2.0

     
  • Rick McGuire
    Rick McGuire
    2014-04-15

    • status: pending --> closed
     


Anonymous


Cancel   Add attachments