Bugfix release, fixing a number of issues. For details, see the
announcement:
https://pythoninsider.blogspot.com/2020/03/python-377-is-now-available.html
Update the hash of the license file for a change in copyright years:
-2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019 Python Software Foundation;
+2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020 Python Software Foundation;
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2020-9273: In ProFTPD 1.3.7, it is possible to corrupt the memory pool
by interrupting the data transfer channel. This triggers a use-after-free
in alloc_pool in pool.c, and possible remote code execution.
And additionally, fixes a number of other issues. For details, see the
release notes:
https://github.com/proftpd/proftpd/blob/1.3.6/RELEASE_NOTES
This also bumps the bundled libcap, so
0001-fix-kernel-header-capability-version.patch can be dropped.
While we are at it, adjust the white space in the .hash function to match
the new agreements.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit a1859b6204)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This bump incorporates the changes found while preparing the 1.3.7
release and drops the CVE 2019-12815 patch which was incorporated in
a73dbfe3b6.
Adjusted the site location to point at github.
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 7f5ea32aed)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Add a patch to correct a typo in the Makefile, so -fno-stack-protector /
-fno-stack-protector-all are really used. With this applied, kvm-unit-tests
will always be built without SSP as intented by upstream. This will fix the
build on ppc64 with SSP that started to fail for an unknown reason since
November 27th.
Moreover, the Arch Linux workaround could also be removed in a follow-up
patch.
Fixes:
- http://autobuild.buildroot.org/results/ad689b08173548af21dd1fb0e827fd561de6dfef
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit dc006056bb)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Call BUSYBOX_INSTALL_INDIVIDUAL_BINARIES in BUSYBOX_INSTALL_TARGET_CMDS,
not in BUSYBOX_INSTALL_INIT_SYSV. This should have been done in commit
b1e07d6d79 but was somehow lost during the
review/aply process.
Signed-off-by: Carlos Santos <unixmania@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 3da205b274)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Unlike <PKG>_DEPENDENCIES, <PKG>_PATCH_DEPENDENCIES only guarantees
extract and patch of listed dependencies, not build. Make this subtlety
more explicit in the documentation.
Cc: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
[yann.morin.1998@free.fr: slight fix]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit d01e808bfe)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Create the staging symlink the same way as the host symlink. This means
using a make dependency rather than recreating it every time.
In coreutils versions below 8.27, re-creation of symbolic links was not
atomic. This means that there is a period in time where the existing link is
removed, before the new one is created. In coreutils 8.27 this was fixed,
see [1]. Note that CentOS 7 ships with coreutils 8.22.
In the following scenario, this is a problem:
- an application is compiled using the sysroot prepared by Buildroot and
links against Xenomai userspace libraries, but its build process is steered
from outside of Buildroot
- to know the correct flags, the application makefile uses the 'xeno-config'
file to request them, and passes DESTDIR=/buildroot/output/staging
- the xeno-config responds with flags based on the path
'/buildroot/output/staging/...'
- while the application build is ongoing, a 'make' happens in Buildroot,
causing the 'staging' symlink to be recreated (even though it already
existed)
- when exactly at this time, the application calls the compiler with -I
flags pointing to output/staging, the build fails with:
-I/buildroot/output/staging/usr/include/xenomai/mercury: Error: ^ is not a directory
-I/buildroot/output/staging/usr/include/xenomai: Error: ^ is not a directory
-I/buildroot/output/staging/usr/include/xenomai/xenomai: Error: ^ is not a directory
-I/buildroot/output/staging/usr/include/xenomai/psos: Error: ^ is not a directory
Failed: ** ^ *
Work around this problem by only creating the staging symlink once, similar
to how the host symlink (if any) is created.
See also commit d0f4f95e39 which changed the
way these symlinks are made. The reasoning in this commit is to move away
from the 'dirs' target.
[1] 376967889e
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 9b82442314)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The package instrumentation step 'step_pkg_size' is populating the files:
output/build/packages-file-list.txt
output/build/packages-file-list-staging.txt
output/build/packages-file-list-host.txt
by comparing the list of files before and after installation of a package,
with some clever tricks to detect changes to existing files etc.
As an optimization, instead of gathering this list before and after each
package, where the 'after-state' of one package is the same as the
'before-state' of the next package, only the 'after-state' is used and
is shared between packages.
This works fine, except at the end of the build, as explained next.
In the target-finalize step, many files will be touched. For example, files
like /etc/hosts, /etc/os-release, but also all object files that are
stripped, and all files touched by post-build scripts or created by rootfs
overlays. This means that the 'after-state' of the last package does not
reflect the actual situation after target-finalize is run.
For a single complete build this poses no problem. But, if one incrementally
rebuilds a package after the initial build, e.g. with 'make foo-rebuild',
then all changes that happened in target-finalize at the end of the initial
build (the 'after-state' of the last package built) will be detected as
changes caused by the rebuild of package foo. As a result, all these files
will incorrectly be treated as 'owned' by package foo.
Correct this situation by capturing a new state at the end of
target-finalize, so that the 'before-state' of an incremental build will be
correct.
Note: the reasoning above talks about packages-file-list.txt and
target-finalize, but also applies to
packages-file-list-staging.txt/staging-finalize and
packages-file-list-host.txt/host-finalize.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 509db3b88a)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2020-1720: ALTER ... DEPENDS ON EXTENSION is missing authorization checks
https://www.postgresql.org/about/news/2011/
Update the license hash for a change in copyright years:
-Portions Copyright (c) 1996-2019, PostgreSQL Global Development Group
+Portions Copyright (c) 1996-2020, PostgreSQL Global Development Group
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2020-7046: Truncated UTF-8 can be used to DoS submission-login and
lmtp processes
lib-smtp doesn't handle truncated command parameters properly, resulting
in infinite loop taking 100% CPU for the process. This happens for LMTP
(where it doesn't matter so much) and also for submission-login where
unauthenticated users can trigger it.
- CVE-2020-7957: Specially crafted mail can crash snippet generation
Snippet generation crashes if:
- message is large enough that message-parser returns multiple body
blocks
- The first block(s) don't contain the full snippet (e.g. full of
whitespace)
- input ends with '>'
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 250535975d)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
When the kernel has CONFIG_SHMEM disabled, /dev is a ramfs (instead of a
tmpfs) and the name_to_handle_at system call is not supported. This
causes eudev's monitor application to exit on startup.
Upstream eudev has added this fix which is not yet part of a release.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 1b81eb7d04)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Use pkg-config to find openssl dependencies such as lz or latomic
Fixes: static build on sparc v8 (even if there are no autobuilder
failures yet)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit c2b6bb17fa)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Fix CVE-2018-14042: In Bootstrap before 4.1.2, XSS is possible in the
data-container property of tooltip.
- Fix an XSS vulnerability (CVE-2019-8331) in our tooltip and popover
plugins by implementing a new HTML sanitizer
- Update indentation of hash file (two spaces)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit bc31029617)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
- Fix CVE-2019-19307: An integer overflow in parse_mqtt in mongoose.c in
Cesanta Mongoose 6.16 allows an attacker to achieve remote DoS
(infinite loop), or possibly cause an out-of-bounds write, by sending
a crafted MQTT protocol packet.
- Update indentation of hash file (two spaces)
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit c18562a82a)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2016-6328: A vulnerability was found in libexif. An integer overflow
when parsing the MNOTE entry data of the input file. This can cause
Denial-of-Service (DoS) and Information Disclosure (disclosing some
critical heap chunk metadata, even other applications' private data).
- CVE-2017-7544: libexif through 0.6.21 is vulnerable to out-of-bounds heap
read vulnerability in exif_data_save_data_entry function in
libexif/exif-data.c caused by improper length computation of the allocated
data of an ExifMnote entry which can cause denial-of-service or possibly
information disclosure.
- CVE-2018-20030: An error when processing the EXIF_IFD_INTEROPERABILITY and
EXIF_IFD_EXIF tags within libexif version 0.6.21 can be exploited to
exhaust available CPU resources.
- CVE-2019-9278: In libexif, there is a possible out of bounds write due to
an integer overflow. This could lead to remote escalation of privilege in
the media content provider with no additional execution privileges needed.
User interaction is needed for exploitation.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 81a4940d25)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
GCC later than 5.x produce _fdti1.so file with an undefined
symbol str2charp_size due to C99 inline semantics change. So
remove this keyword.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
[yann.morin.1998@free.fr: add upstream status]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 135cc97eef)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Add an upstreamed patch that reorders find_package() commands.
This way Python interpreter will be detected first and based on
it the Python libraries can be found.
Fixes the following CMake error:
Could NOT find PythonLibs (missing: PYTHON_LIBRARIES PYTHON_INCLUDE_DIRS)
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit df734533cf)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
We can only know the details of the license files for known versions. For
custom, older or newer versions, the license files may change, or may be
moved around.
So, do for U-Boot as was done for ATF, linux, and linux-headers, and only
define the list of license files for the latest version.
Reported-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Markus Mayer <mmayer@broadcom.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit ff1a03ab28)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This fixes CVE-2019-5188:
A code execution vulnerability exists in the directory rehashing
functionality of E2fsprogs e2fsck 1.45.4. A specially crafted ext4
directory can cause an out-of-bounds write on the stack, resulting
in code execution. An attacker can corrupt a partition to trigger
this vulnerability.
Also change the hash file to the new spacing convention introduced
by Yann E. Morin.
Signed-off-by: Titouan Christophe <titouan.christophe@railnova.eu>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 31b8b08b47)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Commit a17402e42d has conditionally
enabled NEON and VFPv3 optimizations. However, the VFPv3 logic is
causing issues on some targets such as Cortex-A5 with VFPv4-D16 but
not VFPv4.
Since the ENABLE_VFPV3=ON option only adds CFLAGS, we can always set
it to OFF, and let Buildroot pass appropriate CFLAGS.
However, the ENABLE_NEON option also adds the build of NEON-specific
code, so we keep this logic.
Fixes:
- https://bugs.buildroot.org/show_bug.cgi?id=11996
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 4d0f3dd870)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2018-10754: In ncurses before 6.1.20180414, there is a NULL Pointer
Dereference in the _nc_parse_entry function of tinfo/parse_entry.c. It
could lead to a remote denial of service if the terminfo library code is
used to process untrusted terminfo data in which a use-name is invalid
syntax (REJECTED).
- CVE-2018-19211: In ncurses 6.1, there is a NULL pointer dereference at
function _nc_parse_entry in parse_entry.c that will lead to a denial of
service attack. The product proceeds to the dereference code path even
after a "dubious character `*' in name or alias field" detection.
- CVE-2018-19217: In ncurses, possibly a 6.x version, there is a NULL
pointer dereference at the function _nc_name_match that will lead to a
denial of service attack. NOTE: the original report stated version 6.1,
but the issue did not reproduce for that version according to the
maintainer or a reliable third-party.
- CVE-2019-17594: There is a heap-based buffer over-read in the
_nc_find_entry function in tinfo/comp_hash.c in the terminfo library in
ncurses before 6.1-20191012.
- CVE-2019-17595: There is a heap-based buffer over-read in the fmt_entry
function in tinfo/comp_hash.c in the terminfo library in ncurses before
6.1-20191012.
Ncurses upstream uses a fairly special way of releasing (security) bugfixes.
Approximately once a week an incremental .patch.gz is released, and once in
a while these incremental patches are bundled up to a bigger patch relative
to the current release in .patch.sh.bz2 format (a bzip2 compressed patch
with a small shell script prepended, luckily apply-patches can handle that),
and the relative patch files deleted.
For details of this process, see the upstream FAQ:
https://invisible-island.net/ncurses/ncurses.faq.html#applying_patches
Apply the latest .patch.sh.bz2 and incremental patches up to 20200118 to fix
a number of (security) issues. Notice that these patch files are NOT
available on the GNU mirrors.
The license file COPYING is updated with the new Copyright year (2019 ->
2020), so update the hash accordingly.
While we are at it, adjust the white space in the .hash file to match
sha256sum output for consistency.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[fix whitespace inconsistency after 'sha256' keyword]
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
[yann.morin.1998@free.fr: fix license hash for (C) year]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
(cherry picked from commit 10fae9624b)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security issues:
- CVE-2019-19921: runc volume mount race condition with shared mounts
- 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.
For details, see the announcement:
https://github.com/containerd/containerd/releases/tag/v1.2.12
containerd is now a separate CNCF sponsored project, and is no longer
explicitly associated with docker/moby.
Signed-off-by: Christian Stewart <christian@paral.in>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(cherry picked from commit 8ebb77ac1d)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes CVE-2019-9755: An integer underflow issue exists in ntfs-3g 2017.3.23.
A local attacker could potentially exploit this by running /bin/ntfs-3g with
specially crafted arguments from a specially crafted directory to cause a
heap buffer overflow, resulting in a crash or the ability to execute
arbitrary code. In installations where /bin/ntfs-3g is a setuid-root
binary, this could lead to a local escalation of privileges.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 4fb3c69854)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes CVE-2020-7595: xmlStringLenDecodeEntities in parser.c in libxml2
2.9.10 has an infinite loop in a certain end-of-file situation.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 615b7c4af5)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes the following security vulnerabilities:
- CVE-2014-9638: oggenc in vorbis-tools 1.4.0 allows remote attackers to
cause a denial of service (divide-by-zero error and crash) via a WAV file
with the number of channels set to zero.
- CVE-2014-9639: Integer overflow in oggenc in vorbis-tools 1.4.0 allows
remote attackers to cause a denial of service (crash) via a crafted number
of channels in a WAV file, which triggers an out-of-bounds memory access.
- CVE-2014-9640: oggenc/oggenc.c in vorbis-tools 1.4.0 allows remote
attackers to cause a denial of service (out-of-bounds read) via a crafted
raw file.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit fd43037c8c)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In gcc 5.1.0, a change was introduced which causes internal search paths
inside the sysroot to be relative to 'lib64' rather than 'lib'. See [1] [2]
and [3].
For example for dtc:
LD convert-dtsv0
/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/../../../../mips64-octeon-linux-gnu/bin/ld: cannot find crt1.o: No such file or directory
/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/../../../../mips64-octeon-linux-gnu/bin/ld: cannot find crti.o: No such file or directory
collect2: error: ld returned 1 exit status
make[1]: *** [Makefile:236: convert-dtsv0] Error 1
make[1]: *** Waiting for unfinished jobs....
make[1]: Leaving directory '/opt/buildroot/output/build/dtc-1.4.7'
make: *** [package/pkg-generic.mk:241: /opt/buildroot/output/build/dtc-1.4.7/.stamp_built] Error 2
In this case, crt1.o was searched for in following locations:
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/n32/octeon3/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/n32/octeon3/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/../../../../mips64-octeon-linux-gnu/lib/mips64-octeon-linux-gnu/7.3.0/n32/octeon3/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/../../../../mips64-octeon-linux-gnu/lib/../lib32-fp/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/lib64/mips64-octeon-linux-gnu/7.3.0/n32/octeon3/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/lib64/../lib32-fp/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/usr/lib64/mips64-octeon-linux-gnu/7.3.0/n32/octeon3/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/usr/lib64/../lib32-fp/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/../../../../mips64-octeon-linux-gnu/lib/mips64-octeon-linux-gnu/7.3.0/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/opt/ext-toolchain/bin/../lib/gcc/mips64-octeon-linux-gnu/7.3.0/../../../../mips64-octeon-linux-gnu/lib/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/lib64/mips64-octeon-linux-gnu/7.3.0/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/lib64/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/usr/lib64/mips64-octeon-linux-gnu/7.3.0/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
16073 access("/opt/buildroot/output/host/mips64-buildroot-linux-gnu/sysroot/usr/lib64/crt1.o", R_OK) = -1 ENOENT (No such file or directory)
As can be seen above, all attempted paths contain 'lib64' as base,
instead of 'lib' or 'lib32', e.g.
.../sysroot/lib64/../lib32-fp/crt1.o
.../sysroot/lib64/crt1.o
This problem was detected on a gcc 7.x toolchain provided by Marvell as part
of their Octeon SDK. For this toolchain, here are the values of the paths
as detected by the Buildroot toolchain logic, for two different Octeon
processors:
- octeon2 (soft-float) (-mabi=n32 -march=octeon2):
SYSROOT_DIR=/opt/buildroot/output/host/opt/ext-toolchain/mips64-octeon-linux-gnu/sys-root/;
ARCH_SYSROOT_DIR=/opt/buildroot/output/host/opt/ext-toolchain/mips64-octeon-linux-gnu/sys-root/;
ARCH_SUBDIR=;
ARCH_LIB_DIR=lib32/octeon2;
SUPPORT_LIB_DIR=/opt/buildroot/output/host/opt/ext-toolchain/mips64-octeon-linux-gnu/lib32/octeon2/
- octeon3 (hard-float) (-mabi=n32 -march=octeon3):
SYSROOT_DIR=/opt/buildroot/output/host/opt/ext-toolchain/mips64-octeon-linux-gnu/sys-root/;
ARCH_SYSROOT_DIR=/opt/buildroot/output/host/opt/ext-toolchain/mips64-octeon-linux-gnu/sys-root/;
ARCH_SUBDIR=;
ARCH_LIB_DIR=lib32-fp;
SUPPORT_LIB_DIR=/opt/buildroot/output/host/opt/ext-toolchain/mips64-octeon-linux-gnu/lib32-fp/
For both cases (MIPS64n32) Buildroot created a symlink 'lib32->lib', from
SYSTEM_LIB_SYMLINK in system/system.mk. Additionally, the function
create_lib_symlinks in
toolchain/toolchain-external/pkg-toolchain-external.mk will use ARCH_LIB_DIR
and create an additional link $(ARCH_LIB_DIR)->lib.
For the Octeon3 case this thus results in the following symlinks (where the
'lib32' one is normally not needed):
lib32 -> lib/
lib32-fp -> lib/
Since the toolchain is searching based on a 'lib64' component, it will fail
to find its internal paths.
To solve the problem, we need to create an additional symlink 'lib64':
lib64 -> lib/
[1] 257ccd463a
[2] https://gcc.gnu.org/ml/gcc-patches/2014-10/msg03377.html
[3] https://gcc.gnu.org/ml/gcc-patches/2014-11/msg00539.html
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
(cherry picked from commit 45fbadb0b7)
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
