| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since Barebox v2020.11.0, the build system supports a configurable
version string that is baked into the barebox image, see the following
upstream barebox commits:
* commit 98c221e3fad57f271d60 (2020-09-23, "Makefile: add buildsystem version variable")
* commit c7dcac9de597f006fdc8 (2020-09-23, "common: print buildsystem version in barebox banner")
* commit c600d8812e9f3b98591a (2020-09-23, "common: globalvar: add variable for buildsystem_version_string")
* commit 01b53d3b728073996e7e (2020-09-23, "imd: add buildsystem version to metadata")
Set that version string to the Git hash so we can always find out from
which DistroKit version a barebox-in-the-wild was built. (Even if we
don't use barebox >= v2020.11.0 on all platforms yet, future barebox
updates will benefit from this change.)
For mips, we use the upstream barebox rule, which takes the version
string from the platformconfig. For all other platforms, this change
effectively ports the following commit from PTXdist 2021.04.0:
| commit f5500c922424ee71dd99e80f743ed6d994e5bca6
| Author: Steffen Trumtrar <s.trumtrar@pengutronix.de>
| Date: 2021-04-01 10:36:41
|
| rules: barebox: support barebox buildsystem_version
|
| barebox supports setting a buildsystem version. This is useful to have
| an exact buildsystem version like a commit hash accessible from barebox.
|
| Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
| Message-Id: <20210401083641.9251-2-s.trumtrar@pengutronix.de>
| Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Link: https://git.pengutronix.de/cgit/barebox/commit/?id=98c221e3fad57f271d60
Link: https://git.pengutronix.de/cgit/barebox/commit/?id=c7dcac9de597f006fdc8
Link: https://git.pengutronix.de/cgit/barebox/commit/?id=c600d8812e9f3b98591a
Link: https://git.pengutronix.de/cgit/barebox/commit/?id=01b53d3b728073996e7e
Link: https://git.pengutronix.de/cgit/ptxdist/commit/?id=f5500c922424ee71dd99
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This ports the following commits from PTXdist 2021.04.0:
| commit a22b449946eed634f42b06b0c385b29621c6d02e
| Author: Michael Olbrich <m.olbrich@pengutronix.de>
| Date: 2021-04-01 15:58:54
|
| template: kernel: proper handle signed modules
|
| Same as 565aabf6f36d3ed31544d882394604c4a1b13597 but for the kernel
| template instead of the kernel itself.
|
| Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
| commit 565aabf6f36d3ed31544d882394604c4a1b13597
| Author: Denis Osterland-Heim <denis.osterland@diehl.com>
| Date: 2021-03-31 13:35:25
|
| kernel: proper handle signed modules
|
| If CONFIG_MODULE_SIG_ALL is set in kernelconfig then modules will be
| automatically signed during the modules_install phase of a kernel build.
|
| Signed modules are BRITTLE as the signature is outside of the defined ELF
| container. Thus they MAY NOT be stripped once the signature is computed
| and attached. Note the entire module is the signed payload, including any
| and all debug information present at the time of signing.
|
| See: https://www.kernel.org/doc/html/latest/admin-guide/module-signing.html
| Signed-off-by: Denis Osterland-Heim <denis.osterland@diehl.com>
| Message-Id: <20210331113525.22330-1-denis.osterland@diehl.com>
| Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
We might not sign kernel modules at the time, but apply the principle of
least surprise and port it now to be up to date with the upstream
kernel rule template.
Link: https://git.pengutronix.de/cgit/ptxdist/commit/?id=a22b449946eed634f42b
Link: https://git.pengutronix.de/cgit/ptxdist/commit/?id=565aabf6f36d3ed31544
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Migrate with default settings to the new PTXdist version, which was
released today.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
The git repository has a branch called imx_5.4.24_2.1.0 and a tag called
rel_imx_5.4.24_2.1.0. The branch currently contains a commit on top of
the tag. Make sure to use the tag as the stable reference.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.pengutronix.de/20210317085705.29071-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes:
tar: DistroKit/platform-v8a/packages: Cannot open: No such file or directory
tar: Error is not recoverable: exiting now
make: *** [DistroKit/configs/platform-v8a/rules/firmware-imx.make:36: DistroKit/platform-v8a/state/firmware-imx.extract] Error 2
e.g. when running
ptxdist extract firmware-imx
in an unbuilt tree.
Fixes: 2e2c4377dc08 ("i.MX8M: add firmware-imx 8.8")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
| |
This gets rid of all patches, they are included in 2020.09.0.
|
|
|
|
| |
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
PTXdist 2021.03.0 has a bug: without this, the v7a platform cannot be
built.
This basically reverts 510c23705bc999bda9d0b040418dfebe94fc9c68.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
|
|
| |
This is enabled in barebox and also in the kernel for rpi and v8a.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
One new patch is required to fix a regression for USB on rpi3. Also USB
on rpi3 requires CONFIG_USB_NOP_XCEIV=y now which is enabled in the rpi2
config now.
For the beaglebone two patches are added fixing a regression related to
a dts sync from upstream.
All three patches were already sent to the barebox mailing list.
Acked-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
With the upcoming update, our current MLO config will exceed the maximum
size allowed. Work around this by using XZ compression instead, which
reduces size from 115K to 101K, which gives us a further 8K we can
munch.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
| |
SRAM0 on the AM335x is 109K in size. Specify this in the config, so we
get a build error if our MLO exceeds this size.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
So far, the two STM32MP1 boards we support had the same board name
string within the file names of first stage (TF-A) and second stage
(barebox) bootloaders. This will no longer be true when updating
to a more recent barebox version, as the dk2 image's name has been
changed to dkx to indicate support for both dk1 and dk2.
Prepare for this by splitting up STM32MP_BOARD_NAME.
No functional change.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before this change the content of the export list is by default:
/ (rw,no_root_squash),insecure
which is wrong. The closing parenthesis must be after the insecure
keyword.
It seems I'm the first user of unfs3 in the v7a platform since 2016.
Fixes: c17c88e05a88 ("platform: beaglebone: nfs: add insecure option")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
|
|
|
|
|
|
| |
It is recommended to enable these options.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
The kernel config is currently not clean. Run ptxdist oldconfig kernel.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
New version came out today, but the diff is uneventful.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Trusted Firmware-A v2.4 was released last month and adds amongst others,
enhancements to make use of the trusted boot chain possible with
upstream components.
The patch stack has now been posted upstream in hope we can eventually
drop it:
https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/7339
https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/7340
For now, it was rebased with following modifications:
- Sync with upstream device tree and new TF-A requirements
- CLK_ETH_PLL4P is enabled for possible PTP use
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
| |
The patch for lxa-mc1 is mainline, remove it.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new 'ptxdist lint' feature complained about these lines:
Undefined symbol PTXCONF_IMAGE_RPI2_GPU_MEM used in 'next/configs/platform-v7a/rules/image-rpi2.make'.
Undefined symbol PTXCONF_IMAGE_RPI3_GPU_MEM used in 'next/configs/platform-v7a/rules/image-rpi3.make'.
Indeed, these complaints were caused by multiple issues, probably going
back to copy-and-paste from platform-rpi:
* IMAGE_RPI2_GPU_MEM and IMAGE_RPI3_GPU_MEM were never defined in any
kconfig menu since these lines were added (but a similar kconfig item
exists in configs/platform-rpi)
* configs/platform-v7a/rpi-firmware/config.txt didn't have any magic
replacement markers anymore since commit 98626f8441add596486b
* the "replaced" config.txt was never used by anything after writing it
to PTXDIST_TEMPDIR (which is the case in configs/platform-rpi)
Remove the call completely.
Fixes: 1db122202276b9f343eb (2017-06-16, "platform-v7a: rpi2: add board support")
Fixes: 98626f8441add596486b (2018-11-29, "add RPI3 Support")
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
| |
A new PTXdist version was released last week, which also includes a
kconfig update from v5.1 to v5.11-rc3.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
The LXA MC-1 can boot from eMMC as well. The BootROM only supports
booting with fast boot acknowledge though, which is a one time
configuration. Include the mmc_extcsd tool needed to configure this:
https://www.barebox.org/doc/latest/boards/stm32mp.html#flashing-barebox
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The images sizes don't correspond to a specific SD-Card or eMMC size,
so software checking for the backup partition table at the end of
the device won't find it.
Dropping it will however reduce the STM32MP1 images size (618M -> 66M).
Do so.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Both forms are accepted, but it looks more natural to not quote
non-string values. This is also what we do for the other genimage
configs (e.g. bootable = true).
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The variables HEADER and START were apparently meant to be used as
replacements in configs/platform-rpi/config/images/boot-vfat.config, but
they are not used nowhere Furthermore, HEADER referenced a non-existent
file, START used the non-existing Kconfig symbol IMAGE_BOOT_VFAT_START,
and was not layer-aware by using PTXDIST_PLATFORMCONFIGDIR, and
start.elf is also already referenced in FILES.
Reviewed-by: Michael Olbrich <m.olbrich@pengutronix.de>
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
Since PTXdist commit 41cb4d9f9287d85ae11b (2018-07-24, Michael Olbrich:
"barebox: support layers with BAREBOX_EXTRA_ENV_PATH"), this path is
layer-aware and resolved relative to the platformconfig directory.
Reviewed-by: Michael Olbrich <m.olbrich@pengutronix.de>
Link: https://git.pengutronix.de/cgit/ptxdist/commit/?id=41cb4d9f9287d85ae11b
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This ports PTXdist commit 4f058f4c37fa9fb1465a:
| commit 4f058f4c37fa9fb1465a77757a35c48e5e4a9179
| Author: Michael Olbrich <m.olbrich@pengutronix.de>
| Date: 2018-07-31 16:12:58
|
| dtc: support layers in DTC_OFTREE_DTS_PATH by default
|
| PTXDIST_PLATFORMCONFIG_SUBDIR is a relative path and will be searched in
| all layers.
|
| Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Reviewed-by: Alexander Dahl <ada@thorsis.com>
Reviewed-by: Michael Olbrich <m.olbrich@pengutronix.de>
Link: https://git.pengutronix.de/cgit/ptxdist/commit/?id=4f058f4c37fa9fb1465a
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
The names of the TF-A files changed with the new PTXdist version. This was
necessary to support multiple platforms in one build: The platforms may
create images with the same name. Without the aditional prefix, the names
clash when copoed to images/.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
| |
Make malta kernel work with qemu-xhci controller. It was needed to
reproduce endiannes bug with candleLight CAN adapter.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
| |
This is required for the i.MX6 platforms supported by platform-v7a.
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
|
|
|
|
|
|
|
| |
Add the same bit of boilerplate we added for the EK1 to get an image
generated for the giant board as well.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The Groboards Giant Board is a ATSAMA5D27C-D1G SiP-based SBC.
The board features a 500MHz ARM Cortex-A5 and 128MB DDR2 SDRAM in the
SiP as well as a MicroSD slot on the PCB.
The device tree is taken from the one in barebox itself based on the
vendor's DTS available at https://github.com/Groboards/giantboard-tools
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
With the previous barebox and kernel changes, we can combine it with the
DistroKit-v7a rootfs to get an image to boot on the sama5d27-som1-ek.
Do this.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
[rebased to master]
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
| |
In preparation for generating images for the sama5d27, have our
multiplatform kernel configuration target the sama5d2 as well.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Later commits will have the v7a platform generate an image for Cortex-A5
based sama5d2 as well. Import here a new barebox-at91 rule that extends
the barebox-common rule to provide support for the two sama5d27 boards
supported by barebox:
barebox-sama5d27-som1-ek.img
barebox-sama5d27-som1-ek-xload-mmc.img
barebox-groboards-sama5d27-giantboard.img
barebox-groboards-sama5d27-giantboard-xload-mmc.img
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The previous upgrade to v2020.08.1 gave us first stage boot support on
the sama5d2, but had slight multi-image breakage because the fresh
giantboard support on the sama5d27 messed up the EK1 environment
support: it set nv.boot.default=mmc1 in the built-in environment when
none should've been set. This is fixed in v2020.09.0, which is imported
here.
The config changes are those of olddefconfig.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
Version 1.6.0 includes a fix for building on hosts with gcc 10.
This ports PTXdist commit 7990fbf116f66f82796b (2020-05-24, Michael
Olbrich: "host-dtc: update default version to 1.6.0").
Link: https://lkml.org/lkml/2020/4/1/1206
Link: https://git.pengutronix.de/cgit/ptxdist/commit/?id=7990fbf116f66f82796b
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
| |
Cc: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Update to the most recent kernel version and do an oldconfig with
default values.
Tested-by: Roland Hieber <rhi@pengutronix.de> on qemu-x86_64
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
| |
Update to the most recent kernel version and do an oldconfig with
default values.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
| |
We don't need the whole 1 GiB image, the rootfs is only about 50 MiB in
size. Slim the image down to 64 MiB rootfs + 63 MiB data and 1 MiB of
padding at the end to keep the kernel happy.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
Update to the most recent barebox release, which includes our patch.
CONFIG_USB_GADGET_FASTBOOT_* have been renamed to CONFIG_FASTBOOT_*, for
all other new options use the default values.
Tested-by: Roland Hieber <rhi@pengutronix.de> on qemu-vexpress, rpi3-b
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Roland Hieber <r.hieber@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Update to the most recent kernel version and do an oldconfig with
default values.
Tested-by: Roland Hieber <rhi@pengutronix.de> on qemu-vexpress, rpi3-b
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
All upstream defaultenvs were converted in barebox 2015.01.0 [2]
to use nv variables instead of /env/config, and /env/config is no longer
sourced by the init script since barebox 2019.06.0 [1].
The default defaultenv already contains our variable definitions for
allow_color, user, and autoboot_timeout, and the information in
linux.bootargs.base is detected automatically as linux.bootargs.console
from the port name given by the serial driver and CONFIG_BAUDRATE in the
barebox config, so we don't have to convert those. This leaves
boot.default, and converting this variable should also fix the issue of
barebox trying to boot from net (instead of SD card) by default.
[1]: 2019-02-25, Sascha Hauer: "defaultenv: Convert init script to C",
https://git.pengutronix.de/cgit/barebox/commit/?id=90df2a955e3c66fee2c5
[2]: 2014-11-06, Sascha Hauer: "defaultenv-2: Make use of nonvolatile variables",
https://git.pengutronix.de/cgit/barebox/commit/?id=7962e7a0b423a5dfba25
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Update to the latest barebox version, which also has the side effect
that it can boot the newer Linux kernel (first bad Distrokit commit was
61b5ec910db2fe82a39f, 2019-05-06: "platform-rpi/v7a: kernel version bump
5.0 -> 5.1").
Enable regulator support to get around a build issue in the RPi board
code [1], and allow color by default on the console. Also enable
PINCTRL_BCM283X, since it was renamed from GPIO_BCM283X. For all other
settings, use their default value.
Add a patch to fix a build failure with the bcm2835 sdhci driver.
[1]: "rpi-common.c:462:8: error: implicit declaration of function
'regulator_get_name'; did you mean 'regulator_enable'?
[-Werror=implicit-function-declaration]", fixed upstream via
<https://www.mail-archive.com/barebox@lists.infradead.org/msg36326.html>
Tested-by: Roland Hieber <rhi@pengutronix.de> on rpi1-b, rpi-zero-w
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|