Admittedly, this is my own OCD wanting to get rid of this.
Because I tried (a while back to upgrade QEMU to a newer version),
and (during that attempt) I tried to get rid of this.
Tested on Linux & Mac.
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
The final semantic is the same, but this is a bit more correct.
Build tested on Windows 10 (yes, there is some
Ubuntu mode for Windows 10, and I've been also building LEDE
on it for a few weeks).
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
This should hopefully fix build issues where libraries that we ship in
tools/ were accidentally picked up from other places on the system, e.g.
/usr/local
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Patch cmakes cmcurl component to unconditionally link librt when we discover
a clock_gettime() symbol in -lrt.
This is required to sucessfully link LibreSSL's libcrypto which might or
might not reqauire librt.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Since linker argument order is significant on modern host compilers, pass
the required libcrypto libraries through the HOST_LOADLIBES variable in
order to ensure proper order of the final linker command line.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Fixes link errors for host packages like ruby like the following:
/usr/bin/ld: .../staging_dir/host/lib/libcrypto.a(libcrypto_la-md5_dgst.o):
relocation R_X86_64_PC32 against symbol `memcpy@@GLIBC_2.14' can not be used when making a shared object; recompile with -fPIC
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
A lot of packages (especially those using libtool) depend on sed being
available. xz fails on an incremental build pretty early if it's not
there.
Clean up handling if essential core tools (patch, tar, xz), build them
in parallel and make them depend on sed
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Funnel the required LDFLAGS environment through HOST_CONFIGURE_CMD for now
as the HOST_CONFIGURE_VARS semantics are completely broken.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
LibreSSL's libcrypto.so has an indirect dependency to librt for clock_gettime()
on Linux.
Use pkg-config to portably discover the required linker flags.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
LibreSSL's libcrypto.so has an indirect dependency to librt for clock_gettime()
on Linux.
Use pkg-config to portably discover the required linker flags.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Ensure that pkg-config is available before building libressl as we're going
to need it for setting proper link flags in utilities requiring libcrypto.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
graphite is disabled by default and it's not worth carrying ppl and
cloog for supporting this feature on the old gcc 4.8, which is only used
for ARC
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The firmware image that is used in TP-Link RE450 (and some more devices from
the RE series) is tplink-safeloader.
In the kernel partition, the kernel is compressed in a regular tp-link
firmware that is just used for booting. Since it is only used for compressing
and booting, only four fields are filled in the header:
Vendor, version, kernel load address and kernel entry point.
mktplinkfw-kernel is a simpler version of mktpolinkfw that generate such
images. It also specifies the hardware id (as it is in the product info
section), so when doing a sysupgrade - the existing code will check for
hardware compatibility.
Signed-off-by: Tal Keren <kooolk@gmail.com>
[rd@radekdostal.com: add build target to .../image/tp-link.mk]
Signed-off-by: Radek Dostál <rd@radekdostal.com>
The TL-WR1043ND v4 uses a kernel image with a mktplinkfw header inside the
os-image partition of a tplink-safeloader image.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
In order to make it easier to add new boards in the future, refactor the code
to describe the image layouts of supported boards in a struct array and alter
the image generation routines to figure out the sysupgrade partitions
automatically.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Force root/root as names for uid0/gid0 instead of using the system
names. This helps make packed download tarballs more reproducible
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This is intended to be used for a wide array of package sanity checks.
The first check that is implemented is for the hash of downloaded files.
It checks:
- Missing hash
- Use of SHA256 instead of MD5
- dl/<file> hash not matching hash in makefile
- deprecated MD5SUM variable
The deprecated MD5SUM variable check is skipped for feeds/ until OpenWrt
is updated as well
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Adds support for new EU and US variants; removes a few strings that were
never actually used and have been removed from the stock firmwares.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Sadly, LibreSSL claims to be OpenSSL v2.0.0 while not providing the new
interfaces introduced with OpenSSL v1.1.x, so extend the pre-OpenSSL 1.1.x
compat checks to consider LibreSSL as well.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
The OpenSSL 1.1.x version series undergone some major API changes which made
the RSA structure opaque and deprecated a number of methods, so add some
conditional compat code to make the u-boot source build again.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Our current implementation is pretty old and uses some pre-standard/old
ANSI C style that triggers warnings like:
warning: call to function 'MD5_Init' without a real prototype [-Wunprototyped-calls]
This is caused by declarations specified in a following way:
src/md5.h:60:6: note: 'MD5_Init' was declared here
void MD5_Init ();
Having these warnings makes it harded to notice real problems. We could
try hiding them but it makes more sense to just use a cleaner code.
Another tiny gain from this switch is slightly reduced binary size, on
x86_64 tplink-safeloader's size 48104 became 48003.
The new code is public domain, uses "heavily cut-down BSD license".
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Earlier compilers such as GCC 4 do not like anynomous unions, make it a named
union "u", and update the code where relevant.
Fixes FS#298
Fixes: a72e1692b8 ("firmware-utils: Add support for the Cisco Meraki MX60/MX60W")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
User-visible changes since 0.64 are as follows:
- Translation fixes
- Project settings have priority
- Reject binary files in patches
- Fix a race condition in diff_file
- Performance: Optimizations to the setup command
- Performance: Optimizations to the bash completion script
- Test suite: Improve the edit test case
- Test suite: Make the symlink test more robust
- Test suite: Test backup failure
- Test suite: Test the header command with hard links
- diff: Report diff failures
- edit: Fix a corner case bug
- mail: Fix the help text
- push: Fix the synopsis
- refresh: Do not remove symlinks
- refresh: Break links to read-only patch files
- refresh: Always preserve modification time
- setup: Report failed look-ups in inspect-wrapper
- quilt.el: Fix quilt-editable when patches are stored in subdirs
- bash_completion: Handle spaces in file names
- bash_completion: Update the list of commands
- bash_completion: Add new command options
- bash_completion: Fix handling of mail, new, push options
- guards: Simplify the help text
- guards: Include the file name also in the "Not found" case
- guards: Add support for an external filelist in --check mode
- guards: Report which config file has problem in --check mode
- guards: Documentation update
- guards: Clarify a user message
Note to packagers: the Makefile was modified to take configure's
--sysconfdir into account as other projects do. As a result, setting
--prefix=/usr will no longer put the configuration files under /etc.
You now need to explicitly pass --sysconfdir=/etc. If you don't,
configuration files will go under /usr/etc, which is not what you want.
This is somewhat less intuitive, but also more consistent with what
other projects are doing.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
RELEASE 2.5.1 - Mon, 03 Nov 2016 13:37:42 -0400
- Add scons-configure-cache.py to packaging. It was omitted
- Use memoization to optimize PATH evaluation across
all dependencies per node.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
This patch adds header support for the Cisco Meraki MX60/MX60W, which
are a part of the apm821xx target. Some structure changes were needed
due to the fact this device uses U-Boot (unlike other devices in
mkmerakifw.c) which uses a different header structure to define the load
offsets for the image.
A thanks to Christian for helping implement this properly.
Cc: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Chris Blake <chrisrblake93@gmail.com>
Recent refactoring introduced a regression. It ignored second argument
of make_support_list function which was originally true for C2600. The
new generic build_image function always passes false.
This patch allows specifying trailing char in a device specific info. It
also switches Archer C9 to the \0 char to make it compliant with vendor
images.
I verified generated images to be binary identical to the ones that
were created before whole refactoring.
Reported-by: Jo-Philipp Wich <jo@mein.io>
Fixes: fd924d2068 ("firmware-utils: tplink-safeloader: use one function for generating images")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Looking into /usr/include/stdlib.h, there's a `devname()` function defined
under some #ifdef's which conflicts with the `static char *devname` definition
in `src/mkwrggimg.c`.
Defining `_ANSI_SOURCE` in the `src/mkwrggimg.c` file, omits that part of the
header.
Another more intrusive approach is to rename `devname` to something like
`g_devname` in `src/mkwrggimg.c`. But I think the `_ANSI_SOURCE` define should
be enough.
Compilation error is:
src/mkwrggimg.c:64:14: error: redefinition of 'devname' as different kind of symbol
static char *devname;
^
/usr/include/stdlib.h:286:7: note: previous definition is here
char *devname(dev_t, mode_t);
^
src/mkwrggimg.c:147:12: error: non-object type 'char *(dev_t, mode_t)' (aka 'char *(int, unsigned short)') is not assignable
devname = optarg;
~~~~~~~ ^
src/mkwrggimg.c:192:6: warning: comparison of function 'devname' equal to a null pointer is always false [-Wtautological-pointer-compare]
if (devname == NULL) {
^~~~~~~ ~~~~
src/mkwrggimg.c:192:6: note: prefix with the address-of operator to silence this warning
if (devname == NULL) {
^
&
src/mkwrggimg.c:251:27: warning: incompatible pointer types passing 'char *(dev_t, mode_t)' (aka 'char *(int, unsigned short)') to parameter of type 'const char *' [-Wincompatible-pointer-types]
strncpy(header->devname, devname, sizeof(header->devname));
^~~~~~~
/usr/include/secure/_string.h:119:34: note: expanded from macro 'strncpy'
__builtin___strncpy_chk (dest, src, len, __darwin_obsz (dest))
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
[Jo-Philipp Wich: slightly reformat commit message]
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
- use default host install and clean.
- backport compatibility patch for OS X and LEDE (avoids having to force iconv
for OS X).
- use default HOST_BUILD_DIR.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Since XZ is needed to bootstrap building ccache we must not depend on it,
so remove the dependency on ccache to avoid circular dependencies.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Since XZ is required to bootstrap ccache we must not use the ccache compiler
wrapper to avoid circular dependencies.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Make all tools except tar (which is required to bootstrap xz-utils) and XZ
itself depend on XZ, in order to be able to handle .tar.xz downloads.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
The "tar" utility is required to bootstrap XZ which is required to handle
.tar.xz archives, therfore revert to using the bz2 archive.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
In order to build XZ itself we cannot assume that XZ support is available,
so fetch the bz2 archive variant of its sources instead.
Also drop the FreeBSD portability patch and apply it at prepare time using
sed, to avoid a dependency on GNU patch which in turn depends on XZ support.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* Change git packages to xz
* Update mirror checksums in packages where they are used
* Change a few source tarballs to xz if available upstream
* Remove unused lines in packages we're touching, requested by jow- and blogic
* We're relying more on xz-utils so add official mirror as primary source, master site as secondary.
* Add SHA256 checksums to multiple git tarball packages
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
The reason is that ccache does not recognize and process the .incbin
directive, so caching is unreliable.
See https://github.com/ccache/ccache/issues/136 for more information.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Updates expat to 2.2.0
Fixes several CVEs:
CVE-2016-0718
CVE-2016-4472
CVE-2016-5300
CVE-2012-6702
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Update ccache 3.3.2 and refresh patch
Preserving the original patch comments here by Karl Vogel:
"From 90762a9b8d9a50b6176f10bd6c2e2b9501117561 Mon Sep 17 00:00:00 2001
From: Karl Vogel <karl.vogel@gmail.com>
Date: Tue, 14 Jul 2015 11:05:33 +0200
Subject: [PATCH] Include environment variable GCC_HONOUR_COPTS in hash.
The OpenWRT patch, 910-mbsd_multi.patch, to GCC adds an extra
compilation flag, -fhonour-copts, which is influenced by an
environment variable called GCC_HONOUR_COPTS.
Include this environment var in the hash calculation as otherwise
the gcc stdout warning from a previous compilation might be shown
where, even when GCC_HONOUR_COPTS is in 's'ilent mode.
Signed-off-by: Karl Vogel <karl.vogel@gmail.com>"
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Some gcc versions seem to miscompile code using ternary operators,
work around this by just returning the result if exp is 0.
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
TP-Link has changed the way the region is stored in the firmware header,
and now provides US- and EU-specific images for the Archer C7. Adding the
new region codes is necessary to make LEDE/OpenWrt flashable on devices
with the new stock firmwares again.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
In commit df4f41261c ("archs38: Introduce images for SD-cards")
we introduced building of SD-card images for ARC HS38-based boards.
While building images mkdosfs utility is used.
On machines I used for testing mentioned change this utility was
already installed so I didn't figure-out that requirement.
But thanks to Lede's autobuilder this missing bit was highlighted,
see failed build job here:
http://phase1.builds.lede-project.org/builders/archs38%2Fsd/builds/0/steps/images/logs/stdio
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: John Crispin <john@phrozen.org>
Cc: Jo-Philipp Wich <jo@mein.io>
Update make_ext4fs to latest git head in order to support creating empty
filesystem images by making the source directory argument optional.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Move the copyright printing code to the usage printing function, to
reduce noise from regular execution but preserve information about
the author.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
This reverts commit 8c68c104ea.
It is used for apm821xx, which needs ext2 (not ext4) images for some
devices.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
- avoid using tools/include/endian.h on OS X to fix compilation.
- remove unneeded Host/Compile definition.
- refresh patches.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
This patch adds support of Mikrotik yaffs2 filesystem image for kernel file
and tools/kernel2minor package.
We neede this to boot kernel through RouterBoot on new Mikrotik NOR flash devices.
Signed-off-by: Sergey Sergeev <adron@yapic.net>
The libmagic shipped with RedHat 5 does not define
MAGIC_NO_CHECK_ELF and MAGIC_NO_CHECK_COMPRESS. e2fsprogs should
check for that, otherwise the build will fail.
Signed-off-by: Hauke Mehrtens <hauke.mehrtens@intel.com>
In the upstream kernel and the upstream squashfs4 tools the xz
compression header looks the following:
struct disk_comp_opts {
__le32 dictionary_size;
__le32 flags;
};
We added some other members and also moved some existing members. Place
the members which are already in upstream header at the same position
as in that kernel and add our own at the end. The kernel should not
have a problem when there are some additional members and just ignore
them.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Added optional command line option for patch-image tool
Default 16KB size is still maintained as this is an optional argument.
if one wants to specify or increase size they can provide this option.
sample usage: patch-dtb <file> <dtb> [dtb max size]
Signed-off-by: Vishnu Swaroop Duddu <duddu.swaroop@intel.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Avoid using the --parallel argument to the CMake bootstrap, as that doesn't
allow us to remove the -j argument. Instead, pass the HOST_JOBS arguments
in MAKEFLAGS.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
omap requires msdos partition to boot from sdcard.
Prepare for upcoming full sdcard image generation.
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
sparse-0.5.0 won't link against lvm 3.5.0 as "llvm-config --libs" does not
include system libs like pthreads. This problem got fixed but not released
yet, so we need to fetch from their git repo.
Signed-off-by: Stefan Sydow <stefan.sydow@tu-berlin.de>
Latest Xcode doesn't include openssl anymore. To compile
mkimage from u-boot source you need SSL headers on your host.
This patch provides libressl host package for any Darwin
compilation. Unfortunately openssl from MacPorts can not be
used, as the installed headers in /opt/local are breaking
GDB compilation. Tested with a RB532 image build and resulting
kernel booted on a device via TFTP.
Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name> [fixes, dependencies]
The CPE210 was still described for the OEM upgrade as compatible,
even the wireless configuration isn't compatible anymore between
both series (2ghz and 5ghz).
Update the CPE210 image profile to use the new profile.
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
Fixes the following error:
gdate.c: In function ‘g_date_strftime’:
gdate.c:2497:7: error: format not a string literal, format string not checked [-Werror=format-nonliteral]
tmplen = strftime (tmpbuf, tmpbufsize, locale_format, &tm);
^~~~~~
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Update to latest make_ext4fs Git HEAD to fix build with recent glibc version
which does not implicitely includes sysmacros.h anymore.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Remove the whole board list from mktplinkfw, as OpenWrt doesn't use it and
it was severely out of sync with the list of built images for ar71xx.
Also:
* fix -Wall warnings
* add const where appropriate
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 49214
Update flex to 2.6.1
* Flex has moved to Github, adjust download link
* Remove patch that has been applied upstream
* Disable building tests to avoid circular dependency to bison
- disable also docs and examples at the same time
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
SVN-Revision: 49025
internal jsoncpp include order leads to multiple build
errors on Alpine Linux which uses musl libc
use include order from upstream jsoncpp
first error was:
In file included from /usr/include/c++/5.3.0/stdexcept:38:0,
from /home/..../openwrt/build_dir/host/cmake-3.4.3/Utilities/cmjsoncpp/include/json/assertions.h:16,
from /home/..../openwrt/build_dir/host/cmake-3.4.3/Utilities/cmjsoncpp/src/lib_json/json_reader.cpp:7:
/usr/include/c++/5.3.0/exception:35:9: error: '#pragma' is not allowed here
#pragma GCC visibility push(default)
Signed-off-by: Dirk Neukirchen <dirkneukirchen@web.de>
SVN-Revision: 48995
Update GNU findutils to the new stable version 4.6.0
Remove the patch for 32-bit buildhosts as the issue is fixed upstream.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
SVN-Revision: 48991
* Bump pkg-config version to 0.29.1
* Use https for the source download (http gets directed there)
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
SVN-Revision: 48988
This tool creates factory images for JCG routers.
Details can be found in the header comment of jcgimage.c.
Signed-off-by: Reinhard Max <reinhard@m4x.de>
Reviewed-by: Torsten Duwe <duwe@lst.de>
SVN-Revision: 48888
The new u-boot version bundled with the 5.6.x firmwares from Ubiquiti gets
confused by the smaller rootfs partition size; this can lead to various
issues:
1. We've gotten reports that flashing from the 5.6.x stock firmware to
OpenWrt will brick devices; I wasn't able to reproduce this myself
2. Flashing from 5.5.x stock firmware to OpenWrt and back to stock (via
TFTP recovery), following by an update to 5.6.x via web interface can
yield a bricked device with the following properties:
- It can't be booted without entering commands over a serial console, as
u-boot supplies the wrong MTD layout
- The web interface won't accept any image with the original flash
layout, so stock firmware upgrades are impossible
- As the TFTP recovery doesn't update u-boot, returning to the old
u-boot from firmware 5.5.x is impossible
To recover from 2., creating an OpenWrt image which doesn't set u-boot as
read-only and flashing a backup of the old u-boot from there is the only
way known to me. (Fixing the mtdparts variable in u-boot-env from OpenWrt
might also work; settings this from u-boot over serial didn't have
any permanent effect.)
Fix all of this by setting the correct flash layout also used by the stock
firmware. Flashing has been tested from both firmware 5.5.x and 5.6.x. The
fixed layout also matches the mtdparts defined by OpenWrt.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 48829
The ubdev01 profile defines its own MTDPARTS with smaller firmware
partition, so give it its own UBNT_BOARD in mkfwimage.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
SVN-Revision: 48828
It has an important feature (compared to seama) of using multiple input
files, aligning them and padding zeroes until reaching a specified
absolute offset. This is needed for a proper flash layout on NAND. We
want kernel partition to be big enough to handle future updates without
a need to resize it and wipe whole "ubi" partition. It's important as
we don't want to lose block counters.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 48601
This pulls in Paul Kocialkowski's SOURCE_DATE_EPOCH support patch for u-boot,
which landed upstream circa July 2015. Note that this "host" u-boot repo is
only used to compile the 'mkimage' utility, and isn't used to actually compile
a bootloader for any target.
This patch could be removed if/when the host u-boot package is updated to a
contemporary version (but there doesn't seem to be any motivation/need to do
so).
Signed-off-by: bryan newbold <bnewbold@robocracy.org>
Signed-off-by: Felix Fietkau <nbd@openwrt.org> [fix portability error, refresh patches]
SVN-Revision: 48542
