pcp
[Top] [All Lists]

pcp updates: qa, small libpcp fix

To: PCP <pcp@xxxxxxxxxxx>
Subject: pcp updates: qa, small libpcp fix
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Fri, 31 Jul 2015 00:50:05 -0400 (EDT)
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <2112461694.1470729.1438318139348.JavaMail.zimbra@xxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: mWCSF6qtBdOe2P6IRZ7i9WBKNG9Hww==
Thread-topic: pcp updates: qa, small libpcp fix
Changes committed to git://git.pcp.io/pcp master

Nathan Scott (2):
      libpcp: fix buffer unpinning on securesocket error paths
      qa: reenable test 233 on Linux, passing on all my test boxen


 qa/233                         |   12 +-----------
 src/libpcp/src/secureconnect.c |   10 +++++-----
 src/libpcp/src/secureserver.c  |    6 +++---
 3 files changed, 9 insertions(+), 19 deletions(-)


commit 3b94e7d90efc9eeb0ff390229100950ab2c442a4
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Fri Jul 31 14:45:58 2015 +1000

    qa: reenable test 233 on Linux, passing on all my test boxen

commit 1f696bb0f7d15e53c3e95c73796f72831398d964
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Fri Jul 31 10:51:07 2015 +1000

    libpcp: fix buffer unpinning on securesocket error paths
    
    Thanks to MiloÅ PrchlÃk from Red Hat QE for helping uncover this
    one, showed up on some platforms with QA failures like:
    
    > > +Starting pmcd ... pmcd_wait: pdubuf.c:196: __pmUnpinPDUBuf: Assertion
    > > `((__psint_t) handle % sizeof(int)) == 0' failed.
    
    Turns out to be some incorrect handling of pinned-buffer state,
    and calling __pmUnpinPDUBuf with uninitialised data as a result.
    Another win for QA!

<Prev in Thread] Current Thread [Next in Thread>
  • pcp updates: qa, small libpcp fix, Nathan Scott <=