[omniORB] Uninitialized memory read errors reported in omniORB

Duncan Grisby duncan at grisby.org
Mon Feb 21 11:08:39 GMT 2005


On Friday 18 February, Mike Wyatt wrote:

> UMR: Uninitialized memory read (2 times):
>    * This is occurring while in:
> 	void omni::unix_get_ifinfo(omnivector<const char*>&) [tcpTransportImpl.cc:242]

If you look into this, you'll see that it's inside your operating
system's ioctl implementation. Nothing to do with omniORB.

[...]
> UMR: Uninitialized memory read (3 times):
>    * This is occurring while in:
> 	_so_send       [libc.so.1]
> 	int omni::tcpConnection::Send(void*,unsigned,unsigned

This is unused padding bytes in GIOP messages. It's not a problem.

[...]
> ****  Purify instrumented ClipsES_Srv (pid 26643)  ****
> UMR: Uninitialized memory read (79 times):
>    * This is occurring while in thread 8:
> 	mutex_unlock   [libthread.so.1]
> 	cond_timedwait [libthread.so.1]
> 	pthread_cond_timedwait [libthread.so.1]
> 	int omni_condition::timedwait(unsigned long,unsigned long) [posix.cc:143]
> 	void omni::Scavenger::execute() [giopStrand.cc:647]
> 	void omniAsyncWorker::real_run() [omniAsyncInvoker.h:118]
>    * Reading 4 bytes from 0x1c1fa4 in the heap (2 bytes at 0x1c1fa5 uninit).

This is inside the Solaris mutex implementation. The rest of the UMRs
are the same thing. Nothing to do with omniORB.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



More information about the omniORB-list mailing list