While an image layout based on MBR and 'bootfs' partition may be easy
to understand for users who are very used to the IBM PC and always have
the option to access the SD card outside of the device (and hence don't
really depend on other recovery methods or dual-boot), in my opinion
it's a dead end for many desirable features on embedded systems,
especially when managed remotely (and hence without an easy option to
access the SD card using another device in case things go wrong, for
example).
Let me explain:
* using a MSDOS/VFAT filesystem to store kernel(s) is problematic, as a
single corruption of the bootfs can render the system into a state
that it no longer boots at all. This makes dual-boot useless, or at
least very tedious to setup with then 2 independent boot partitions
to avoid the single point of failure on a "hot" block (the FAT index
of the boot partition, written every time a file is changed in
bootfs). And well: most targets even store the bootloader environment
in a file in that very same FAT filesystem, hence it cannot be used
to script a reliable dual-boot method (as loading the environment
itself will already fail if the filesystem is corrupted).
* loading the kernel uImage from bootfs and using rootfs inside an
additional partition means the bootloader can only validate the
kernel -- if rootfs is broken or corrupted, this can lead to a reboot
loop, which is often a quite costly thing to happen in terms of
hardware lifetime.
* imitating MBR-boot behavior with a FAT-formatted bootfs partition
(like IBM PC in the 80s and 90s) is just one of many choices on
embedded targets. There are much better options with modern U-Boot
(which is what we use and build from source for all targets booting
off SD cards), see examples in mediatek/mt7622 and mediatek/mt7623.
Hence rename the 'sdcard' feature to 'legacy-sdcard', and prefix
functions with 'legacy_sdcard_' instead of 'sdcard_'.
Tested-by: Stijn Tintel <stijn@linux-ipv6.be>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Add a generic sdcard upgrade method instead of duplicating code in yet
another target, and add a feature flag to only install this upgrade
method in targets that set this flag. Copied from mvebu.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
The mkits.sh script help message states hash algorithm can be
specified using the -H command-line option, but it does not work
currently due to a bug in the script.
This patch fixes this problem by changing the option from -S to
-H and specify getopts parameter after it
Signed-off-by: Yonghyu Ban <yonghyu@empo.im>
If the image generation doesn't add any profiles to the output the
*profile merge* will fail. To avoid that set an empty profile as
fallback.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Keep other profiles.json content if the data belongs to the current
build version.
Also useful for the ImageBuilder, which builds for a single model each
time. Without this commit the profiles.json would only contain the
latest build profile information.
Signed-off-by: Moritz Warning <moritzwarning@web.de>
[improve commit message]
Signed-off-by: Paul Spooren <mail@aparcar.org>
This separates index update from feed update. The result is that all
requested feeds are first updated and only then indexed.
The reason for this change is to prevent errors being reported and
potentially invalid index being generated thanks to cross feeds
dependency.
The feeds script pulls in default all feeds as they come and on install
prefers packages from first feeds (unless special feed is requested).
Thus order of feeds in some way specifies preferences. This is handy for
downstream distributions as they can simply override any package from
upstream feeds by placing their feed before them. This removes need to
patch or fork upstream feeds.
The problem is that such feed most likely depends in some way also on
subsequent feeds. The most likely feeds are 'packages' or 'luci'. The
example would be Python package that needs 'python.mk' from 'packages'
feed. Ordering custom feed after dependent feeds is sometimes just not
possible because of preference requirement described before.
The solution is to just first pull all feeds and generate indexes only
after that. In the end this ensures that index is generated correctly at
first try without any error.
In terms of code this removes 'perform_update' argument from
'update_feed' as with index update removal the update is the only action
performed in that subroutine. Thus this moves condition to 'update'
subroutine.
Signed-off-by: Karel Kočí <karel.koci@nic.cz>
This script hasn't seen an update in multiple years, update it to the
latest version provided upstream. Both `config.guess` and `config.sub`
are copied from upstream[1] and not modified.
The full changelog is available within the upstream repository[1].
[1]: https://git.savannah.gnu.org/git/config.git
Signed-off-by: Paul Spooren <mail@aparcar.org>
Before this commit, it was assumed that mkhash is in the PATH. While
this was fine for the normal build workflow, this led to some issues if
make TOPDIR="$(pwd)" -C "$pkgdir" compile
was called manually. In most of the cases, I just saw warnings like this:
make: Entering directory '/home/.../package/gluon-status-page'
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
[...]
While these were only warnings and the package still compiled sucessfully,
I also observed that some package even fail to build because of this.
After applying this commit, the variable $(MKHASH) is introduced. This
variable points to $(STAGING_DIR_HOST)/bin/mkhash, which is always the
correct path.
Signed-off-by: Leonardo Mörlein <me@irrelefant.net>
This became a bit of a tragedy, caused by a corner cases which wasn't
put into account during testing. DEFAULT_PACKAGES are defined in
target/linux/<target>/Makefile but also in
target/linux/<target>/<subtarget>/target.mk.
The latter was no longer imported when using DUMP=1, however not using
DUMP=1 while running the Makefile in target/linux/<target>/ caused duplicate
packages in the list.
As a solution, which should have been used from day 0, `make` runs in
target/linux/ without DUMP=1, resulting in no duplicate packages and all
inclusions from include/target.mk, linux/target/<target>/{Makefile,
<subtarget>/target.mk}
While at it, sort the list of default packages.
Signed-off-by: Paul Spooren <mail@aparcar.org>
The commit "1bf2b3fe90 build,json: fixup missing arch_packages" fixes
the missing package architecture locally but runs $(TOPDIR)/Makefile
rather than a target specific one. While this works on local builds just
fine, it causes the buildbots to add garbage to the `arch_packages`
variable:
cd \"/builder/shared-workdir/build\"; git log --format=%h -1
toolchain > /builder/shared-workdir/build/tmp/.ver_check\ncmp -s
/builder/shared-workdir/build/tmp/.ver_check
/builder/shared-workdir/build/staging_dir/toolchain-x86_64_gcc-8.4.0_musl/stamp/.ver_check
|| { \\\n\trm -rf
/builder/shared-workdir/build/build_dir/target-x86_64_musl
/builder/shared-workdir/build/staging_dir/target-x86_64_musl
/builder/shared-workdir/build/staging_dir/toolchain-x86_64_gcc-8.4.0_musl
/builder/shared-workdir/build/build_dir/toolchain-x86_64_gcc-8.4.0_musl;
\\\n\tmkdir -p
/builder/shared-workdir/build/staging_dir/toolchain-x86_64_gcc-8.4.0_musl/stamp;
\\\n\tmv /builder/shared-workdir/build/tmp/.ver_check
/builder/shared-workdir/build/staging_dir/toolchain-x86_64_gcc-8.4.0_musl/stamp/.ver_check;
\\\n}\nx86_64
Only the last line contains the desired string.
Future investigation should check why the build system prints this to
stdout rather than stderr.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Fix 7f4c2b1 "build,json: fix duplicates in default_packages" which
removed duplicate default packages but also removed the package
architecture from the profiles.json.
If DUMP=1 is set, the `ARCH_PACKAGES` is no longer exported and
therefore empty. Fix this by running make twice, once with DUMP=1 and
once without.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Calling without the DUMP=1 argument causes the target specific Makefile
to be "included" again which adds the target specific packages twice,
once on the actual run and once included from `include/target.mk`.
This led to duplicate package entries, causing confusion in downstream
projects using the generated JSON files.
While at it, apply `black` style to Python script.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Add new target feature 'dt-overlay' which makes DTC keep the symbol
names in the generated dtb.
Make sure additional DT overlay sources specified by the new device
variable DEVICE_DTS_OVERLAY get compiled together with the main DTS
(currently overlays got to be in the same folder). Let Build/fit pass
the generated DT overlay blobs to mkits.sh.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Allow adding multiple device tree overlay blobs to an image and
generate configurations for each of them.
This is useful on boards with modern U-Boot which allow e.g. user-
configurable peripherals ("shields") in that way.
Note that currently, each generated configuration adds exactly one
overlay on top of the base image, ie. adding multiple overlays at the
same time is not yet supported.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This commit cleans the `ipkg-build` script via changes suggested by
shellcheck. These are mostly word splitting issues.
Remove the definition of GZIP, this adds three "lookups" of the `gzip`
binary but the rest of the build system doesn't seem to use such
improvements neither.
Signed-off-by: Paul Spooren <mail@aparcar.org>
U-boot will reject the nodes with @ for the address since
commit:
79af75f777
This in turn will cause the failure to boot with OpenWrt
generated images.
So, to rectify that simply replace @ with -.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cover also newly added rootfs@1 and initrd@1 nodes)
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
We so far had two variables IMG_PREFIX and IMAGE_PREFIX with
different content. Since these names are obviously quite
confusing, this patch renames the latter to DEVICE_IMG_PREFIX,
as it's a device-dependent variable, while IMG_PREFIX is only
(sub)target-dependent.
For consistency, also rename IMAGE_NAME to DEVICE_IMG_NAME, as
that's a device-dependent variable as well.
Cc: Paul Spooren <mail@aparcar.org>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Instead of embedding the initrd cpio archive into the kernel, allow
for having an external ramdisk added to the FIT or uImage.
This is useful to overcome kernel size limitations present in many
stock bootloaders, as the ramdisk is then loaded seperately and doesn't
add to the kernel size. Hence we can have larger ramdisks to host ie.
installers with all binaries to flash included (or a web-based
firmware selector).
In terms of performance and total size the differences are neglectible.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Allow for single (external-data) FIT image to hold kernel, dtb and
squashfs. In that way, the bootloader verifies the system integrity
including the rootfs, because what's the point of checking that the
hash of the kernel is correct if it won't boot in case of squashfs
being corrupted? Better allow bootloader to check everything needed
to make it at least up to failsafe mode. As a positive side effect
this change also makes the sysupgrade process on nand potentially
much easier as it is now.
In short: mkimage has a parameter '-E' which allows generating FIT
images with 'external' data rather than embedding the data into the
device-tree blob itself. In this way, the FIT structure itself remains
small and can be parsed easily (rather than having to page around
megabytes of image content). This patch makes use of that and adds
support for adding sub-images of type 'filesystem' which are used to
store the squashfs. Now U-Boot can verify the whole OS and the new
partition parsers added in the Linux kernel can detect the filesystem
sub-images, create partitions for them, and select the active rootfs
volume based on the configuration in FIT (passing configuration via
device tree could be implemented easily at a later stage).
This new FIT partition parser works for NOR flash (on top of mtdblock),
NAND flash (on top of ubiblock) as well as classic block devices
(ie. eMMC, SDcard, SATA, NVME, ...).
It could even be used to mount such FIT images via `losetup -P` on a
user PC if this patch gets included in Linux upstream one day ;)
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This reverts commit b12288fa69.
The patchelf approach is too fragile, and the only users of this have been
converted to make patching unnecessary
Leave the abi_version_str variable in place in rules.mk
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Preparation for supporting dynamic ABI versions that depend on the runtime
configuration. Read the suffix from the staging dir pkginfo version files.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This makes it possible to declare a package ABI_VERSION independent from the
upstream soname by setting PKG_ABI_VERSION in the package makefile.
The library filename is fixed up for files installed to packages and to the
staging dir. References to the original from executables within the same
package are also fixed up
Signed-off-by: Felix Fietkau <nbd@nbd.name>
As multiple LICENSES are shipped and no longer just LICENSE, modify the
OpenWrt tree detection in checkpatch.pl.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Multiple sources are hosted on OpenWrts source server only. The source
URLs to point to the server vary based on different epochs in OpenWrts
history.
Replace all by @OPENWRT which is an "empty" mirror, therefore using the
fallback servers sources.cdn.openwrt.org and sources.openwrt.org.
Signed-off-by: Paul Spooren <mail@aparcar.org>
In case the default sources for a package fail use the CDN rather than
our own mirror. In case the CDN fails, fallback to our mirror.
Also remove mirror1 which isn't available anymore.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Since 4ee3cf2b5a profiles with alternative vendor names may appear
multiple times in `tmp/.targetinfo` or `.targetinfo` (for
ImageBuilders).
The `target-metadata.pl` script adds these profiles then twice to
`PROFILE_NAMES` and the ImageBuilder show the profile twice when running
`make info`.
This patch removes duplicate profile IDs and only adds them once to
`.profiles.mk`.
Signed-off-by: Paul Spooren <mail@aparcar.org>
The recent 7f285d "scripts/feeds: warn when skipping core package
override" floods SDK output with warning of overwriting "linux" and
"toolchain" core packages. This should be ignored as these are not
regular packages added via feeds.
While at it slightly improve the warning string.
Signed-off-by: Paul Spooren <mail@aparcar.org>
clang's gcc emulation does the right thing with -print-file-name now,
drop the wrapper
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
The qemustart script currently picks the ext4 filesystem rather than
squashfs, while the latter is default for nearly all OpenWrt targets.
Change the default behaviour of qemustart to be in line with the rest.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Device specifications:
* QCA IPQ4019
* 256 MB of RAM
* 32 MB of SPI NOR flash (w25q256)
- 2x 15 MB available; but one of the 15 MB regions is the recovery image
* 2T2R 2.4 GHz
- QCA4019 hw1.0 (SoC)
- requires special BDF in QCA4019/hw1.0/board-2.bin with
bus=ahb,bmi-chip-id=0,bmi-board-id=20,variant=PlasmaCloud-PA2200
* 2T2R 5 GHz (channel 36-64)
- QCA9888 hw2.0 (PCI)
- requires special BDF in QCA9888/hw2.0/board-2.bin
bus=pci,bmi-chip-id=0,bmi-board-id=16,variant=PlasmaCloud-PA2200
* 2T2R 5 GHz (channel 100-165)
- QCA4019 hw1.0 (SoC)
- requires special BDF in QCA4019/hw1.0/board-2.bin with
bus=ahb,bmi-chip-id=0,bmi-board-id=21,variant=PlasmaCloud-PA2200
* GPIO-LEDs for 2.4GHz, 5GHz-SoC and 5GHz-PCIE
* GPIO-LEDs for power (orange) and status (blue)
* 1x GPIO-button (reset)
* TTL pins are on board (arrow points to VCC, then follows: GND, TX, RX)
* 2x gigabit ethernet
- phy@mdio3:
+ Label: Ethernet 1
+ gmac0 (ethaddr) in original firmware
+ used as LAN interface
- phy@mdio4:
+ Label: Ethernet 2
+ gmac1 (eth1addr) in original firmware
+ 802.3at POE+
+ used as WAN interface
* 12V 2A DC
Flashing instructions:
The tool ap51-flash (https://github.com/ap51-flash/ap51-flash) should be
used to transfer the factory image to the u-boot when the device boots up.
Signed-off-by: Marek Lindner <marek.lindner@kaiwoo.ai>
[sven@narfation.org: prepare commit message, rebase, use all LEDs, switch
to dualboot_datachk upgrade script, use eth1 as designated WAN interface]
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Device specifications:
* QCA IPQ4018
* 256 MB of RAM
* 32 MB of SPI NOR flash (w25q256)
- 2x 15 MB available; but one of the 15 MB regions is the recovery image
* 2T2R 2.4 GHz
- QCA4019 hw1.0 (SoC)
- requires special BDF in QCA4019/hw1.0/board-2.bin with
bus=ahb,bmi-chip-id=0,bmi-board-id=16,variant=PlasmaCloud-PA1200
* 2T2R 5 GHz
- QCA4019 hw1.0 (SoC)
- requires special BDF in QCA4019/hw1.0/board-2.bin with
bus=ahb,bmi-chip-id=0,bmi-board-id=17,variant=PlasmaCloud-PA1200
* 3x GPIO-LEDs for status (cyan, purple, yellow)
* 1x GPIO-button (reset)
* 1x USB (xHCI)
* TTL pins are on board (arrow points to VCC, then follows: GND, TX, RX)
* 2x gigabit ethernet
- phy@mdio4:
+ Label: Ethernet 1
+ gmac0 (ethaddr) in original firmware
+ used as LAN interface
- phy@mdio3:
+ Label: Ethernet 2
+ gmac1 (eth1addr) in original firmware
+ 802.3af/at POE(+)
+ used as WAN interface
* 12V/24V 1A DC
Flashing instructions:
The tool ap51-flash (https://github.com/ap51-flash/ap51-flash) should be
used to transfer the factory image to the u-boot when the device boots up.
Signed-off-by: Marek Lindner <marek.lindner@kaiwoo.ai>
[sven@narfation.org: prepare commit message, rebase, use all LEDs, switch
to dualboot_datachk upgrade script, use eth1 as designated WAN interface]
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Device specifications:
* Qualcomm/Atheros QCA9533 v2
* 650/600/217 MHz (CPU/DDR/AHB)
* 64 MB of RAM
* 16 MB of SPI NOR flash (mx25l12805d)
- 2x 7 MB available; but one of the 7 MB regions is the recovery image
* 2x 10/100 Mbps Ethernet
* 2T2R 2.4 GHz Wi-Fi
* multi-color LED (controlled via red/green/blue GPIOs)
* 1x GPIO-button (reset)
* external h/w watchdog (enabled by default)
* TTL pins are on board (arrow points to VCC, then follows: GND, TX, RX)
* 2x fast ethernet
- eth0
+ Label: Ethernet 1
+ 24V passive POE (mode B)
+ used as WAN interface
- eth1
+ Label: Ethernet 2
+ 802.3af POE
+ builtin switch port 2
+ used as LAN interface
* 12-24V 1A DC
* internal antennas
Flashing instructions:
The tool ap51-flash (https://github.com/ap51-flash/ap51-flash) should be
used to transfer the factory image to the u-boot when the device boots up.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
The NOR flash rootfs images stored in a sysupgrade.tar must end with the
JFFS2 marker. Otherwise, devices like OpenMesh A42/A62 are not able to
calculate the md5sum of the fixed squashfs part and store it inside the
u-boot-env.
But the commit ee76bd11bb ("images: fix boot failures on NAND with small
sub pages") adds up to 1020 0x00 bytes after the 0xdead0de EOF marker. The
calculated md5sum will be wrong due do this change and u-boot will fail to
boot the newly flashed device with a message like:
Validating MD5Sum of 'vmlinux'...
Passed!
Validating MD5Sum of 'rootfs'...
Failed!
583a1b7b54b8601efa64ade42742459b != 8850ee812dfd7638e94083329d5d2781
Data validation failed!
and boot the old image again.
Since the original change should not change the behavior of NOR images,
just check for the deadc0de marker at the end of the squashfs-jffs2 image
do avoid the problematic behavior for these images.
Fixes: ee76bd11bb ("images: fix boot failures on NAND with small sub pages")
Signed-off-by: Sven Eckelmann <sven@narfation.org>
The rootfs is padded to the full block size by padjffs2 and a 4 byte magic
value ("deadc0de") is added to the end. On first boot, the JFFS2 is
replacing the "deadc0de" marker when the rootfs_data is initialized.
The static part of the rootfs is therefore $rootfs_size - 4 and not
$rootfs_size - 262144 - 4.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
The padding and block alignment is handled by the image build script and
doesn't need to be duplicated in the fwupgrade.cfg build script.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Some images are created using different filesystems, most popular
squashfs and ext4. To allow downstream projects to distinguesh between
those, add the `filesystem` information to created json files.
Signed-off-by: Paul Spooren <mail@aparcar.org>
The previous behavior prefered same feed for dependent packages as
initial package. This caused inconsitency in installation of packages.
The difference was if two feeds provide same package (different version)
there was different result if you executed install for that specific
version compared to install for package depending on it from different
feed.
This ensures that preferred feed is propagated without change and
selected feed is used only really for package it was selected for.
Signed-off-by: Karel Kočí <karel.koci@nic.cz>
Otherwise, a n00b like myself can get quite confused when moving a
package from core to feeds, for example.
(Hint: one *really* needs to clear out the tmp/info/.packageinfo...
entries for the stale package, but '-f' works as well.)
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
It seems like after a build the /dl dir seems to now contain a .hash
file for each source file due to inproper cleanup so fix it by removing
those intermediate files before leaving the download action.
Fixes: 4e19cbc553 ("download: handle possibly invalid local tarballs")
Reported-by: Hannu Nyman <hannu.nyman@iki.fi>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Currently it's assumed, that already downloaded tarballs are always
fine, so no checksum checking is performed and the tarball is used even
if it might be corrupted.
From now on, we're going to always check the downloaded tarballs before
considering them valid.
Steps to reproduce:
1. Remove cached tarball
rm dl/libubox-2020-08-06-9e52171d.tar.xz
2. Download valid tarball again
make package/libubox/download
3. Invalidate the tarball
sed -i 's/PKG_MIRROR_HASH:=../PKG_MIRROR_HASH:=ff/' package/libs/libubox/Makefile
4. Now compile with corrupt tarball source
make package/libubox/{clean,compile}
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Silence the warning in git 2.27 about undefined fast-forward style
in git pull. Define "ff-only" as the style.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
Daniel Golle
Adrian Schmutzler