On Fri, 2008-10-10 at 11:33 -0400, Steve Huston wrote:
> Hi,
Hi,
http://issues.apache.org/jira/browse/QPID-1317. 
I was planning to submit a patch when I have some time. If you are
planning to fix it go ahead. 

-Rajika

> 
> I run a Linux (RHEL 4) C++ qpid build/test overnight and last night
> turned up 3 instances of the following new diagnostic:
> 
> ==16486== Syscall param futex(uaddr2) contains uninitialised byte(s)
> ==16486== at 0x1C1844F1: __lll_mutex_unlock_wake (in
> /lib/tls/libpthread-2.3.4.so)
> ==16486== by 0x1C1236FD: pthread_mutex_unlock (in
> /lib/tls/libc-2.3.4.so)
> ==16486== by 0x1BA7ED89: qpid::client::SessionImpl::open(unsigned)
> (Mutex.h:120)
> ==16486== by 0x1BA3F6DF:
> qpid::client::ConnectionImpl::newSession(std::string const&, unsigned,
> unsigned short) (shared_ptr.hpp:374)
> ==16486== by 0x1BA65A2A:
> qpid::client::FailoverListener::start(boost::shared_ptr<qpid::client::
> ConnectionImpl> const&) (shared_ptr.hpp:374)
> ==16486== by 0x1BA409EB: qpid::client::ConnectionImpl::open()
> (scoped_ptr.hpp:93)
> ==16486== by 0x1BA36D26:
> qpid::client::Connection::open(qpid::client::ConnectionSettings
> const&) (shared_ptr.hpp:374)
> ...
> 
> And one new memory leak:
> 
> ==16818== 456 bytes in 3 blocks are possibly lost in loss record 1 of
> 1
> ==16818== at 0x1B901B47: calloc (vg_replace_malloc.c:279)
> ==16818== by 0x1B8F2C8A: _dl_allocate_tls (in /lib/ld-2.3.4.so)
> ==16818== by 0x1BE7C9FD: pthread_create@@GLIBC_2.1 (in
> /lib/tls/libpthread-2.3.4.so)
> ==16818== by 0x1BE7CF47: [EMAIL PROTECTED] (in
> /lib/tls/libpthread-2.3.4.so)
> ==16818== by 0x1BB03F77:
> qpid::sys::ThreadPrivate::ThreadPrivate(qpid::sys::Runnable*)
> (Thread.cpp:44)
> ==16818== by 0x1BB03C13:
> qpid::sys::Thread::Thread(qpid::sys::Runnable*) (shared_ptr.hpp:166)
> ==16818== by 0x1B9615B7: qpid::client::TCPConnector::init()
> (Connector.cpp:219)
> ==16818== by 0x1B95A70D: qpid::client::ConnectionImpl::open()
> (scoped_ptr.hpp:93)
> ==16818== by 0x1B950D26:
> qpid::client::Connection::open(qpid::client::ConnectionSettings
> const&) (shared_ptr.hpp:374)
> ==16818== by 0x8054FAD: Client::Client() (TestOptions.h:67)
> ==16818== by 0x805137F: main (perftest.cpp:530)
> 
> The build was run against svn revision 703303.
> 
> Should these be reported in a jira?
> 
> Thanks,
> -Steve
> 
-- 
http://wso2.org
http://llvm.org
http://www.osdev.org

Reply via email to