[BSA-099] Security update for libreofice

Rene Engelhard uploaded new packages for libreoffice which fixed the
following security problems:

CVE-2014-3693:
   Use-After-Free in socket manager of Impress Remote

   It was discovered that LibreOffice 4.0.0 and later does not manage the port
   1599 for the LibreOffice Impress correctly. An external attackers with
   access to that port could cause the deleted port manager to continue to
   process attacker supplied data.

Note that this update also disables the remote contol per default as it
listens on port 1599 "to the world" per default. If you want/need it you
need to enable it manally:
   1. Open LibreOffice, go to "Tools -> Options..."                             
   2. Select "LibreOffice Impress -> General"                                   
   3. Check "Presentation -> Enable remote control" 

For the wheezy-backports distribution the problems have been fixed in
version 1:4.3.3~rc2-1~bpo70+1.

CVE-2014-3640

The sosendto function in slirp/udp.c in QEMU before 2.1.2 allows local users to cause a denial of service (NULL pointer dereference) by sending a udp packet with a value of 0 in the source port and address, which triggers access of an uninitialized socket.

CVE-2014-3693

Use-after-free vulnerability in the socket manager of Impress Remote in LibreOffice 4.x before 4.2.7 and 4.3.x before 4.3.3 allows remote attackers to cause a denial of service (crash) or possibly execute arbitrary code via a crafted request to TCP port 1599.

CVE-2014-5037

Eucalyptus 4.0.0 through 4.0.1, when the log level is set to INFO, logs user and system passwords, which allows local users to obtain sensitive information by reading cloud-requests.log.

CVE-2014-5038

Eucalyptus 3.0.0 through 4.0.1, when the log level is set to DEBUG or lower, logs user and system passwords, which allows local users to obtain sensitive information by reading the cloud log files.

CVE-2014-6620

Cross-site scripting (XSS) vulnerability in Aruba Networks ClearPass before 6.3.6 and 6.4.x before 6.4.1 allows remote attackers to inject arbitrary web script or HTML via unspecified vectors.