Opened 11 years ago
Closed 11 years ago
#6880 closed enhancement (fixed)
libvdpau-1.1.1
| Reported by: | Fernando de Oliveira | Owned by: | Fernando de Oliveira |
|---|---|---|---|
| Priority: | high | Milestone: | 7.8 |
| Component: | BOOK | Version: | SVN |
| Severity: | medium | Keywords: | |
| Cc: |
Description
http://people.freedesktop.org/~aplattner/vdpau/libvdpau-1.1.1.tar.bz2
http://people.freedesktop.org/~aplattner/vdpau/libvdpau-1.1.1.tar.bz2.sig
CVE-2015-5198, CVE-2015-5199, and CVE-2015-5200
If you use the NVIDIA .run installer packages, please see https://devtalk.nvidia.com/default/topic/873035 for additional information.
http://lists.freedesktop.org/archives/vdpau/2015-August/000376.html
and
http://lists.freedesktop.org/archives/vdpau/2015-August/000377.html
MD5: 2fa0b05a4f4d06791eec83bc9c854d14 libvdpau-1.1.1.tar.bz2
[VDPAU] libvdpau 1.1.1
Aaron Plattner aplattner at nvidia.com
Mon Aug 31 14:16:33 PDT 2015
libvdpau versions 1.1 and earlier, when used in setuid or setgid
applications, contain vulnerabilities related to environment
variable handling that could allow an attacker to execute
arbitrary code or overwrite arbitrary files. See CVE-2015-5198,
CVE-2015-5199, and CVE-2015-5200 for more details.
This release uses the secure_getenv() function, when available,
to fix these problems. On platforms where secure_getenv() is not
available, the VDPAU environment variables will not be honored by
the library.
If you use the NVIDIA .run installer packages, please see
https://devtalk.nvidia.com/default/topic/873035 for additional
information.
This release also adds tracing of HEVC picture structures to
libvdpau_trace.
Aaron Plattner (2):
Remove pdflatex dependency
Bump version to 1.1.1
Bibhuti Prusty (1):
Add tracing for HEVC picture info
Emil Velikov (3):
.gitignore: add compile
configure.ac: remove AM_MAINTAINER_MODE
autogen.sh: update/copy over from mesa
José Hiram Soltren (2):
Fix Doxygen typo: s/no queued/not queued/
Use secure_getenv(3) to improve security
Errata:
[VDPAU] libvdpau 1.1.1
Aaron Plattner aplattner at nvidia.com
Mon Aug 31 14:28:25 PDT 2015
On 08/31/2015 02:16 PM, Aaron Plattner wrote:
> On platforms where secure_getenv() is not available, the VDPAU
> environment variables will not be honored by the library.
Whoops, this is incorrect, sorry. The updated libvdpau will
instead use a fallback implementation of secure_getenv() when the
platform doesn't provide one.
Change History (8)
comment:1 by , 11 years ago
| Owner: | changed from to |
|---|---|
| Status: | new → assigned |
comment:2 by , 11 years ago
comment:3 by , 11 years ago
| Summary: | libvdpau-1.1.1.tar.bz2 → libvdpau-1.1.1 |
|---|
comment:4 by , 11 years ago
The test passes for me with the current book instructions:
make dlclose make[2]: Entering directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' gcc -DHAVE_CONFIG_H -I. -I.. -I../include -g -O2 -MT dlclose.o -MD -MP -MF .deps/dlclose.Tpo -c -o dlclose.o dlclose.c mv -f .deps/dlclose.Tpo .deps/dlclose.Po /bin/sh ../libtool --tag=CC --mode=link gcc -g -O2 -o dlclose dlclose.o -ldl -L/opt/xorg/lib -lX11 -lpthread libtool: link: gcc -g -O2 -o dlclose dlclose.o -L/opt/xorg/lib /opt/xorg/lib/libX11.so /opt/xorg/lib/libxcb.so /opt/xorg/lib/libXau.so /opt/xorg/lib/libXdmcp.so -ldl -lpthread make[2]: Leaving directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' make check-TESTS make[2]: Entering directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' make[3]: Entering directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' PASS: dlclose ============================================================================ Testsuite summary for libvdpau 1.1.1 ============================================================================ # TOTAL: 1 # make dlclose make[2]: Entering directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' gcc -DHAVE_CONFIG_H -I. -I.. -I../include -g -O2 -MT dlclose.o -MD -MP -MF .deps/dlclose.Tpo -c -o dlclose.o dlclose.c mv -f .deps/dlclose.Tpo .deps/dlclose.Po /bin/sh ../libtool --tag=CC --mode=link gcc -g -O2 -o dlclose dlclose.o -ldl -L/opt/xorg/lib -lX11 -lpthread libtool: link: gcc -g -O2 -o dlclose dlclose.o -L/opt/xorg/lib /opt/xorg/lib/libX11.so /opt/xorg/lib/libxcb.so /opt/xorg/lib/libXau.so /opt/xorg/lib/libXdmcp.so -ldl -lpthread make[2]: Leaving directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' make check-TESTS make[2]: Entering directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' make[3]: Entering directory '/tmp/xdrivers/libvdpau/libvdpau-1.1.1/test' PASS: dlclose ============================================================================ Testsuite summary for libvdpau 1.1.1 ============================================================================ # TOTAL: 1 # PASS: 1 # SKIP: 0 # XFAIL: 0 # FAIL: 0 # XPASS: 0 # ERROR: 0
comment:5 by , 11 years ago
OK, thanks, hence I will keep the test part as is. Actually, just some statistics change, in the update.
comment:6 by , 11 years ago
| Owner: | changed from to |
|---|---|
| Status: | assigned → new |
I'm sorry, don't know what is happening. Earlier, got Performing 4.4-subscription-ops.test: FAIL. Trying to solve it, changed remote server, configurations, now it is worse:
Performing 4.2-cups-printer-ops.test: FAIL Performing 4.3-job-ops.test: FAIL Performing 4.4-subscription-ops.test: FAIL }} I am giving back all tickets, until I feel sure about what I am doing again. Sorry for having delayed these updates.
comment:7 by , 11 years ago
| Owner: | changed from to |
|---|---|
| Status: | new → assigned |
After reboot, cups-2.1.0 restored the initial behavior with
Performing 4.2-cups-printer-ops.test: PASS Performing 4.3-job-ops.test: PASS Performing 4.4-subscription-ops.test: FAIL Performing ipp-2.1.test: PASS
Taking back the tickets. Hope to close them tomorrow.

I'm stuck since yesterday with the test (yes, just one test):
The tests dowsn't do much, just compiles, links and runs dlclose. I couldn't identify what (if any) are the calling parameteres, but did everything by hand and executed ./dlclose and got exactly the same error message.
What is curious is that my first log of previous version 1.1 skipped that test, but now it does execute it with failure instead of skipping it:
I'm tending to remove the test from the page, but problem could be just here. Please, can someone comment or perform the test, please? Or, if wished, take the ticket.