Qt5 has predefined optimization flags depending if you're building for
size, for debug etc. These flags are defined in
mkspecs/common/gcc-base.conf:
QMAKE_CFLAGS_OPTIMIZE = -O2
QMAKE_CFLAGS_OPTIMIZE_FULL = -O3
QMAKE_CFLAGS_OPTIMIZE_DEBUG = -Og
QMAKE_CFLAGS_OPTIMIZE_SIZE = -Os
Then, in the same file, they use them to set
QMAKE_CFLAGS_RELEASE/QMAKE_CXXFLAGS_RELEASE:
QMAKE_CFLAGS_RELEASE += $$QMAKE_CFLAGS_OPTIMIZE
QMAKE_CXXFLAGS_RELEASE += $$QMAKE_CFLAGS_RELEASE
At this point there is our chance to override QMAKE_CFLAGS_OPTIMIZE_* in
qmake.conf, but it's too late, because QMAKE_CFLAGS_RELEASE is already
set (i.e. -O2) so trying to add or remove QMAKE_CFLAGS_OPTIMIZE (that is
reset now on) from QMAKE_CLAGS_RELEASE in
common/features/default_post.prf won't work:
optimize_size {
!isEmpty(QMAKE_CFLAGS_OPTIMIZE):!isEmpty(QMAKE_CFLAGS_OPTIMIZE_SIZE) {
QMAKE_CFLAGS_RELEASE -= $$QMAKE_CFLAGS_OPTIMIZE
QMAKE_CXXFLAGS_RELEASE -= $$QMAKE_CFLAGS_OPTIMIZE
QMAKE_CFLAGS_RELEASE += $$QMAKE_CFLAGS_OPTIMIZE_SIZE
QMAKE_CXXFLAGS_RELEASE += $$QMAKE_CFLAGS_OPTIMIZE_SIZE
}
} else: optimize_full {
!isEmpty(QMAKE_CFLAGS_OPTIMIZE):!isEmpty(QMAKE_CFLAGS_OPTIMIZE_FULL) {
QMAKE_CFLAGS_RELEASE -= $$QMAKE_CFLAGS_OPTIMIZE
QMAKE_CXXFLAGS_RELEASE -= $$QMAKE_CFLAGS_OPTIMIZE
QMAKE_CFLAGS_RELEASE += $$QMAKE_CFLAGS_OPTIMIZE_FULL
QMAKE_CXXFLAGS_RELEASE += $$QMAKE_CFLAGS_OPTIMIZE_FULL
}
}
So let's reset:
QMAKE_CFLAGS_RELEASE
QMAKE_CFLAGS_DEBUG
QMAKE_CXXFLAGS_RELEASE
QMAKE_CXXFLAGS_DEBUG
in our qmake.conf since the only assignment done in
mkspecs/common/gcc-base.conf only regards optimization.
This package is also affected by BR2_TOOLCHAIN_HAS_GCC_BUG_90620 and
it's been worked around by appending -O0 to CFLAGS/CXXFLAGS. This bug
prevented workaround to work overriding optimization flags, so solving
this also solves workaround problem.
Fixes:
http://autobuild.buildroot.net/results/ada/adaa9b4bcc6f9d2b5e82c479859a07e8abf5cf13/http://autobuild.buildroot.net/results/a83/a83bdd1f3bf309c07abebe871b017c331ed36e67/
Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Tested-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
[Arnout: add a comment to qmake.conf.in]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
(cherry picked from commit c4a6f974b1)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Fix CVE-2018-11813: libjpeg 9c has a large loop because read_pixel in
rdtarga.c mishandles EOF.
- Update hash of README (small updates such as authors, year ...)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 75a14ec067)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues (4.10.12):
CVE-2019-14902: Replication of ACLs set to inherit down a subtree on AD
Directory not automatic.
CVE-2019-14907: Crash after failed character conversion at log level 3
or above.
CVE-2019-19344: Use after free during DNS zone scavenging in Samba AD
DC.
For more details, see the release notes:
https://www.samba.org/samba/history/samba-4.10.12.html
In addition, 4.10.13 fixes a number of bugs. For details, see the release
notes:
https://www.samba.org/samba/history/samba-4.10.13.html
Drop now upstreamed
0006-heimdal_build-wscript_build-do-not-add-host-include-.patch
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As recently reported to the list:
http://lists.busybox.net/pipermail/buildroot/2020-January/271937.html
The hardcoded rootfs partition size can lead to hard to understand build
failures if more packages are added.
So drop the hardcoded partition size. Genimage will then size the partition
to match the size of the rootfs image (which by default is also 60MB for ext4).
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit f1d1967422)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The 4.12.2 release brings a large number of fixes:
https://xenproject.org/downloads/xen-project-archives/xen-project-4-12-series/xen-project-4-12-2/
Including a number of security fixes:
XSA-296: VCPUOP_initialise DoS (CVE-2019-18420)
XSA-298: missing descriptor table limit checking in x86 PV emulation
(CVE-2019-18425)
XSA-299: Issues with restartable PV type change operations (CVE-2019-18421)
XSA-301: add-to-physmap can be abused to DoS Arm hosts (CVE-2019-18423)
XSA-302: passed through PCI devices may corrupt host memory after
deassignment (CVE-2019-18424)
XSA-303: ARM: Interrupts are unconditionally unmasked in exception handlers
(CVE-2019-18422)
XSA-304: x86: Machine Check Error on Page Size Change DoS (CVE-2018-12207)
XSA-305: TSX Asynchronous Abort speculative side channel (CVE-2019-11135)
XSA-306: Device quarantine for alternate pci assignment methods
(CVE-2019-19579)
XSA-307: find_next_bit() issues (CVE-2019-19581 CVE-2019-19582)
XSA-308: VMX: VMentry failure with debug exceptions and blocked states
(CVE-2019-19583)
XSA-309: Linear pagetable use / entry miscounts (CVE-2019-19578)
XSA-310: Further issues with restartable PV type change operations
(CVE-2019-19580)
XSA-311: Bugs in dynamic height handling for AMD IOMMU pagetables
(CVE-2019-19577)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 636df7ffcd)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes:
http://autobuild.buildroot.net/results/14af2dc3219847a92c6ec2db14ba387159b61fde
The Xen build system builds and embeds a default XSM FLASK (Flux Advanced
Security Kernel) security policy if it detects SELinux checkpolicy on the
build machine.
If enabled, a gen-policy.py python script is used to convert the binary
FLASK policy to a C array initialization list to embed it in the Xen binary.
Depending on the python version and locale available on the host, this fails
with byte values outside the 0..255 range:
policy.c:7:10: error: unsigned conversion from 'int' to 'unsigned char' changes value from '56575' to '255' [-Werror=overflow]
0xdc8c, 0xdcff, 0x7c, 0xdcf9, 0x08, 0x00, 0x00, 0x00, 0x58, 0x65, 0x6e, 0x46, 0x6c,
To fix this and ensure a consistent build, pass XEN_HAS_CHECKPOLICY=n to
disable the checkpolicy detection.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit b60f3e2ae6)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When building with path containing "m4/" occurence(i.e. make
O=output-m4) gettext-tiny install recipe copies files to wrong place and
later some package using autotools fail to autoreconf(i.e. minicom).
This is due to buggy gettext-tiny Makefile install recipe where they
substitute every "m4/" in INSTALL destination path, including the "m4/"
part of our build folder. Add patch to fix this by using $(patsubst ...)
instead of $(subst m4/,,$@) to substitute only last "m4/" occurence in
path.
Fixes:
https://bugs.busybox.net/show_bug.cgi?id=12481
Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 41b9a64526)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The README file saved by legal-info does not mention the host package
variant of the saved material. Add them.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit ec78068972)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This list dates back to 2012. Since a long time now Buildroot saves the
patches applied as well as the actual source code for some external
toolchains. Update the manual accordingly.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit a74e57c932)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fix CVE-2019-18222: Our bignum implementation is not constant
time/constant trace, so side channel attacks can retrieve the blinded
value, factor it (as it is smaller than RSA keys and not guaranteed to
have only large prime factors), and then, by brute force, recover the
key. Reported by Alejandro Cabrera Aldaya and Billy Brumley.
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit a7186d0913)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
With Qt 5.12.x only handwriting/lipi-toolkit needs 3rdparty parts
installation (with Qt 5.6.x although zn_CZ/pinyin and zh_TW tcime).
Fixes:
- https://bugs.busybox.net/show_bug.cgi?id=12456
cp: cannot stat '.../host/arm-buildroot-linux-gnueabihf/sysroot/usr/qtvirtualkeyboard': No such file or directory
Also fix the way we test the variable: we very seldomly use ifdef,
instead we usually test for equality.
Reported-by: Sam Petrocelli <sam.petrocelli@gmail.com>
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
[yann.morin.1998@free.fr: also fix the way we test the variable]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 3645f89922)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This patch adds a new manual section that captures an overview
of the run-tests tool, how to manually run a test and where to
find the test case script.
A brief set of steps is included to go through how to add a new
test case and suggestions on how to test/debug.
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
[yann.morin.1998@free.fr:
- switch the creating and debugging sections
- minor reformatting
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit e2e57d5678)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fix CVE-2019-10155 (IKEv1 information exchange packet's integrity check
value is not verified)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 94c66ece47)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2019-13117: In numbers.c in libxslt 1.1.33, an xsl:number with certain
format strings could lead to a uninitialized read in
xsltNumberFormatInsertNumbers. This could allow an attacker to discern
whether a byte on the stack contains the characters A, a, I, i, or 0, or
any other character.
- CVE-2019-13118: In numbers.c in libxslt 1.1.33, a type holding grouping
characters of an xsl:number instruction was too narrow and an invalid
character/length combination could be passed to xsltNumberFormatDecimal,
leading to a read of uninitialized stack data.
- CVE-2019-18197: In xsltCopyText in transform.c in libxslt 1.1.33, a
pointer variable isn't reset under certain circumstances. If the relevant
memory area happened to be freed and reused in a certain way, a bounds
check could fail and memory outside a buffer could be written to, or
uninitialized data could be disclosed.
Remove patch (already in version)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Peter: mention security impact]
(cherry picked from commit 5645107c39)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Fix CVE-2019-14491: An issue was discovered in OpenCV before 3.4.7
and 4.x before 4.1.1. There is an out of bounds read in the function
cv::predictOrdered<cv::HaarEvaluator> in
modules/objdetect/src/cascadedetect.hpp, which leads to denial of service.
- Fix CVE-2019-14492: An issue was discovered in OpenCV before 3.4.7
and 4.x before 4.1.1. There is an out of bounds read/write in the
function HaarEvaluator::OptFeature::calc in
modules/objdetect/src/cascadedetect.hpp, which leads to denial of service.
- atomic workaround is not needed since version 3.4.8 and
464972855e
- Update hash of license file (Xperience.AI added:
766465ce94)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit f6fb2cae06)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Removed patches 0006 & 0007 which were applied upstream as single
commit on the server-1.20-branch branch:
07efd81b81
Updated upstream URL for patch 0001.
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 5f90daa66f)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Add two patches to fix openssl support:
- 0003-Fix-openssl-detection.patch (suggested by Jonathan Kimmitt)
- 0004-Support-OpenSSL-1.1.0.patch (taken from upstream)
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 62ad96c057)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Commit 5eecaf354c (package/rtl8821au: switch to abperiasamy fork) changed
the upstream location, but didn't update the link in the help text.
Signed-off-by: Christian Stewart <christian@paral.in>
[Peter: extend commit message]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 6d4c2d062e)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Without the device-mapper udev rules, dm devices will not get a proper
symlink like /dev/disk/by-label/LABEL, which in turn causes fstab
LABEL= mounts to fails.
And by extension causes shenanigans with systemd, where it will
unmount a manually mounted disk because it can't resolve the label.
Signed-off-by: Pascal de Bruijn <p.debruijn@unilogic.nl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 51ec0f48ee)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerabilities:
- CVE-2019-14271: In Docker 19.03.x before 19.03.1 linked against the GNU C
Library (aka glibc), code injection can occur when the nsswitch facility
dynamically loads a library inside a chroot that contains the contents of
the container
Signed-off-by: Christian Stewart <christian@paral.in>
[Peter: mention security impact]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 0161899ae5)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerabilities:
- CVE-2019-19221: In Libarchive 3.4.0, archive_wstring_append_from_mbs in archive_string.c
has an out-of-bounds read because of an incorrect mbrtowc or mbtowc call. For example,
bsdtar crashes via a crafted archive.
And adds various security fixes. For details, see :
https://github.com/libarchive/libarchive/releases/tag/v3.4.1
Also remove upstreamed patch.
Signed-off-by: Pierre-Jean Texier <pjtexier@koncepto.io>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit bbc64eae62)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
mender-grubenv currently has 3 problems that prevent an x86_64-efi image from
successfully being made with the genimage.sh script.
- mender-grubenv does not currently depend on Grub2.
While Grub2 is not needed to build the mender-grubenv package, Grub2 needs
to be built first for mender-grubenv to overwrite the default Grub2 files
reliably.
- The MENDER_GRUBENV_ENV_DIR variable points to /boot/efi/EFI/BOOT instead of
/boot/EFI/BOOT, which is where the Grub2 package installs the default files.
This variable now points to the correct location.
- The Grub2 package installs images to $(BINARIES_DIR)/efi-part, which the
mender-grubenv package currently does not do. As such; the default Grub2
configuration file is used instead of the one provided by mender-grubenv.
Adding a MENDER_GRUBENV_INSTALL_IMAGES_CMDS define in mender-grubenv.mk which
copies the installed files from $(TARGET_DIR)/boot/EFI to
$(BINARIES_DIR)/efi-part fixes this issue.
Signed-off-by: Adam Duskett <aduskett@greenlots.com>
[Thomas:
- drop "runtime" on the depends on BR2_TARGET_GRUB2 since we now have
a build-time dependency on it
- explicitly copy the files installed by mender-grubenv in
MENDER_GRUBENV_INSTALL_IMAGES_CMDS instead of blindly copying
everything that is in $(TARGET_DIR)/boot/EFI]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 425f79087a)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerabilities:
- CVE-2019-14271: In Docker 19.03.x before 19.03.1 linked against the GNU C
Library (aka glibc), code injection can occur when the nsswitch facility
dynamically loads a library inside a chroot that contains the contents of
the container
Signed-off-by: Christian Stewart <christian@paral.in>
[Peter: mention security impact]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 39cffd5356)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerabilities:
containerd 1.2.9/gRPC:
- CVE-2019-9512: Some HTTP/2 implementations are vulnerable to ping floods,
potentially leading to a denial of service. The attacker sends continual
pings to an HTTP/2 peer, causing the peer to build an internal queue of
responses. Depending on how efficiently this data is queued, this can
consume excess CPU, memory, or both
- CVE-2019-9514: Some HTTP/2 implementations are vulnerable to a reset
flood, potentially leading to a denial of service. The attacker opens a
number of streams and sends an invalid request over each stream that
should solicit a stream of RST_STREAM frames from the peer. Depending on
how the peer queues the RST_STREAM frames, this can consume excess memory,
CPU, or both
- CVE-2019-9515: Some HTTP/2 implementations are vulnerable to a settings
flood, potentially leading to a denial of service. The attacker sends a
stream of SETTINGS frames to the peer. Since the RFC requires that the
peer reply with one acknowledgement per SETTINGS frame, an empty SETTINGS
frame is almost equivalent in behavior to a ping. Depending on how
efficiently this data is queued, this can consume excess CPU, memory, or
both
containerd 1.2.10/runc:
- CVE-2019-16884: runc through 1.0.0-rc8, as used in Docker through
19.03.2-ce and other products, allows AppArmor restriction bypass because
libcontainer/rootfs_linux.go incorrectly checks mount targets, and thus a
malicious Docker image can mount over a /proc director
Signed-off-by: Christian Stewart <christian@paral.in>
[Peter: mention security impact]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit f40f2bae81)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerability:
- CVE-2019-16884: runc through 1.0.0-rc8, as used in Docker through
19.03.2-ce and other products, allows AppArmor restriction bypass because
libcontainer/rootfs_linux.go incorrectly checks mount targets, and thus a
malicious Docker image can mount over a /proc directory.
Signed-off-by: Christian Stewart <christian@paral.in>
[Peter: mention security impact]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit dbbf08849b)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
