When static linking some dependency library can be missing
(i.e. -latomic for -lcrypto) on linking libraries list. This is
because when static linking libraries dependencies are not
transparently linked into binary.
To avoid moving libraries before/after one another or add new ones
that are not needed at all in the dynamic linking case, we use `pkg-config --libs
LIBRARY` where LIBRARY is the library we "probe" for its existence and
dependency.
In this commit, we:
- Remove 0005-fix-static-link-zlib.patch where -lcrypto and -lz were
swapped, as it is no longer needed thanks to the following point.
- Replace it with 0005-Use-pkgconf-to-get-libs-deps.patch where
-lcrypto has been substituted with `pkg-config --libs libcrypto`
- Add host-pkgconf to ANDROID_TOOLS_DEPENDENCIES
Fixes:
http://autobuild.buildroot.net/results/d3d6679cfc8afe4467368bd3d31483172c1032de/
Signed-off-by: Giulio Benetti <giulio.benetti@micronovasrl.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 1e4f77a2e4)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
* CVE-2019-3836: It was discovered in gnutls before version 3.6.7 upstream
that there is an uninitialized pointer access in gnutls versions 3.6.3 or
later which can be triggered by certain post-handshake messages
* CVE-2019-3829: A vulnerability was found in gnutls versions from 3.5.8
before 3.6.7. A memory corruption (double free) vulnerability in the
certificate verification API. Any client or server application that
verifies X.509 certificates with GnuTLS 3.5.8 or later is affected.
3.6.7.1 is identical to 3.6.7, but fixes a packaging issue in the release
tarball:
https://lists.gnutls.org/pipermail/gnutls-devel/2019-April/013086.html
HTTP URLs changed to HTTPS in COPYING, so update license hash.
Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 1dd5576ccb)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Containerd is no longer maintained under the docker github project and now
has an official website, so refer to that in the help text.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 638504bcdf)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Both patches are already included (a bit earlier in the file) in version
2.0.12, so drop the patches.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 0fda716432)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Switch site to gitlab
- Remove second patch (already in version)
- Use new --{with,without}-usb option
- Add hash for license file
- Fix CVE-2017-6318
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit a911b7d229)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
fakeroot by default forwards {f,l,}chown calls to libc and ignores
permission issues, which may cause issues when building in restricted
environments like user namespaces as set up with bubblewrap where a chown
call with a uid/gid not mapped in the user namespace instead returns EINVAL.
This error is not masked by fakeroot and returned to the caller, causing
failures.
There is no real reason to really perform the *chown calls in the context of
Buildroot (as the calls will likely just fail and files are not accessed
outside the fakeroot environment any way).
This forwarding can be disabled by setting the FAKEROOTDONTTRYCHOWN
environment variable, so set it when fakeroot is executed.
Reported-by: Esben Nielsen <nielsen.esben@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 655acd1df0)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
I left DATACOM and will unlikely have access to the hardware required to
test the package.
Signed-off-by: Carlos Santos <unixmania@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit c45394c1b0)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Include upstream commit 193f1e8 "glob: Do not assume glibc glob
internals". Without this if building glibc with host-make it will fail
with a segfault in make:
>>> glibc glibc-2.28-69-g1e5c5303a522764d7e9d2302a60e4a32cdb902f1 Building
PATH="/scratch/builds/host-make/host/bin:/scratch/builds/host-make/host/sbin:/home/sam/bin:/home/sam/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games" /scratch/builds/host-make/host/bin/host-make -j25 -C /scratch/builds/host-make/build/glibc-glibc-2.28-69-g1e5c5303a522764d7e9d2302a60e4a32cdb902f1/build
/scratch/builds/host-make/host/bin/host-make -r PARALLELMFLAGS="" -C /scratch/builds/host-make/build/glibc-glibc-2.28-69-g1e5c5303a522764d7e9d2302a60e4a32cdb902f1 objdir=`pwd` all
Segmentation fault (core dumped)
Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit a07f69c817)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The current address will soon become invalid so drop it before the
messages start bouncing.
Change-Id: If631cedcaaa55d927d99b18ff299324e9d439cb0
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit c2387c9604)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Instead of stdout where it gets mixed with the normal output, confusing
software parsing the output (E.G. get-developers -e as git sendemail.ccCmd).
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 83f82bd67a)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
tpm2-tools does not need dbus or libglib2, so remove them and the
corresponding toolchain dependencies.
The confusion may have come from the upstream travis configuration, which
also builds tpm2-abrmd (which uses dbus+libglib2).
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit f63a58c350)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Contains a number of fixes for issues discovered post-2.9.8. From the
release notes:
- Fixed readdir bug when non-zero offsets are given to filler and the
filesystem client, after reading a whole directory, re-reads it from a
non-zero offset e.g. by calling seekdir followed by readdir.
https://github.com/libfuse/libfuse/releases/tag/fuse-2.9.9
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 3b6d842fea)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This fixes some omissions from the installation.
Install the udev rules.
Tell buildroot about the fuse device.
Apply setuid permissions on the fusermount tool.
Signed-off-by: Norbert Lange <norbert.lange@andritz.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
(cherry picked from commit ea62ff85b5)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerability:
CVE-2019-9741: An issue was discovered in net/http in Go 1.11.5. CRLF
injection is possible if the attacker controls a url parameter, as
demonstrated by the second argument to http.NewRequest with \r\n followed by
an HTTP header or a Redis command.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerabilities:
*) SECURITY: CVE-2019-0197 (cve.mitre.org)
mod_http2: fixes a possible crash when HTTP/2 was enabled for a http:
host or H2Upgrade was enabled for h2 on a https: host. An Upgrade
request from http/1.1 to http/2 that was not the first request on a
connection could lead to a misconfiguration and crash. Servers that
never enabled the h2 protocol or only enabled it for https: and
did not set "H2Upgrade on" are unaffected by this issue.
[Stefan Eissing]
*) SECURITY: CVE-2019-0196 (cve.mitre.org)
mod_http2: using fuzzed network input, the http/2 request
handling could be made to access freed memory in string
comparision when determining the method of a request and
thus process the request incorrectly. [Stefan Eissing]
*) SECURITY: CVE-2019-0211 (cve.mitre.org)
MPMs unix: Fix a local priviledge escalation vulnerability by not
maintaining each child's listener bucket number in the scoreboard,
preventing unprivileged code like scripts run by/on the server (e.g. via
mod_php) from modifying it persistently to abuse the priviledged main
process. [Charles Fol <folcharles gmail.com>, Yann Ylavic]
*) SECURITY: CVE-2019-0196 (cve.mitre.org)
mod_http2: using fuzzed network input, the http/2 request
handling could be made to access freed memory in string
comparision when determining the method of a request and
thus process the request incorrectly. [Stefan Eissing]
*) SECURITY: CVE-2019-0217 (cve.mitre.org)
mod_auth_digest: Fix a race condition checking user credentials which
could allow a user with valid credentials to impersonate another,
under a threaded MPM. PR 63124. [Simon Kappel <simon.kappel axis.com>]
*) SECURITY: CVE-2019-0215 (cve.mitre.org)
mod_ssl: Fix access control bypass for per-location/per-dir client
certificate verification in TLSv1.3.
*) SECURITY: CVE-2019-0220 (cve.mitre.org)
Merge consecutive slashes in URL's. Opt-out with
`MergeSlashes OFF`. [Eric Covener]
For more details, see the CHANGES file:
https://www.apache.org/dist/httpd/CHANGES_2.4.39
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 556ad6c25b)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Loop relaxation logic in xtensa gas may produce code in which LEND
register doesn't match actual zero overhead loop end. Fix relaxation
code so that it produces a literal or a pair of const16 instructions
with associated relocation record that works correctly in the presence
of other relaxations. This fixes crash in X11 server caused by window
movement.
Loop relaxation has limited of 32K range, this fix removes this
limitation.
Fixes:
http://autobuild.buildroot.net/results/e05522ce540f4ac23f9a3a8fec724694d9a23101/
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Peter: drop 2.32 patch]
(cherry picked from commit 197b5f9d1c)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, host-xz and host-lzip are built as soon as the
corresponding tools are not provided by the system, independently of
whether they are really needed by the Buildroot configuration. This is
particularly annoying for host-lzip, which is only needed for very few
packages.
This commit modifies the generic package infrastructure to only add
host-lzip and host-xz as dependencies when really needed.
Signed-off-by: Norbert Lange <nolange79@gmail.com>
[Thomas:
- improve commit log
- as suggested by Yann E. Morin, make the lzip case similar to the xz
case]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 004960e967)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Commit cc41950950 added the GST_OMX_VARIANT option which gets a default
value that gets overridden by subsequent conditions. check-package
doesn't like that, so instead make the three cases explicitly mutually
exclusive.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 5b217aad9c)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
target defaults to none, which isn't a legal target:
configure: Using none as OpenMAX IL target
configure: error: invalid OpenMAX IL target, you must specify one of --with-omx-target={generic,rpi,bellagio,tizonia,zynqultrascaleplus}
Instead default to 'generic', fixing the build with E.G. nvidia-tegra23.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit cc41950950)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2019-6256: A Denial of Service issue was discovered in the LIVE555
Streaming Media libraries as used in Live555 Media Server 0.93. It can
cause an RTSPServer crash in handleHTTPCmd_TunnelingPOST, when
RTSP-over-HTTP tunneling is supported, via x-sessioncookie HTTP headers in
a GET request and a POST request within the same TCP session. This occurs
because of a call to an incorrect virtual function pointer in the
readSocket function in GroupsockHelper.cpp.
- CVE-2019-7314: liblivemedia in Live555 before 2019.02.03 mishandles the
termination of an RTSP stream after RTP/RTCP-over-RTSP has been set up,
which could lead to a Use-After-Free error that causes the RTSP server to
crash (Segmentation fault) or possibly have unspecified other impact.
- CVE-2019-9215: n Live555 before 2019.02.27, malformed headers lead to
invalid memory access in the parseAuthorizationHeader function.
The normal live555 web site is temporarily unavailable, so use an
alternative _SITE / drop upstream hash.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit ed30a85e5a)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Remove first and second patches (already in version)
- Remove third and fourth patches (not needed since:
245b5a3b4b)
- Add hash for license file
- Drop autoreconf (as configure.ac is not patched anymore)
- Use new --with-crypto option
- Restrict symlink following on installation (CVE-2017-7500,
CVE-2017-7501)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 3b4cc264d9)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Switch site to "real" upstream instead of debian as debian does not
have latest version
- Drop patch (not needed anymore as getline was renamed in my_getline)
- Add hash for license file
- Fix CVE-2013-0348 and CVE-2017-17663
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 48e6230e5f)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issue:
* CVE-2019-7524: Missing input buffer size validation leads into
arbitrary buffer overflow when reading fts or pop3 uidl header
from Dovecot index. Exploiting this requires direct write access to
the index files.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit e3c53aa8a1)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
