| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
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>
|
|
|
|
|
|
| |
Migrate to the most current PTXdist version with default settings.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
| |
|
|
|
|
|
|
| |
Update to the latest toolchain, which was released this week.
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
A new ptxdist version is out.
We set PTXCONF_KERNEL_CONFIG_BASE_VERSION=y in all platforms, as
DistroKit is a base layer for other BSPs.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With the recent move to Linux v5.8, we get a new kernel warning
because it skew timings are now specified twice; once from use of
"rgmii-id" and one from the [rt]xc-skew-ps in the device tree.
This patch removes those timings since the board works fine
with the new default settings.
Patch has also been sent[1] to the respective mailing lists for upstream
inclusion.
[1]: https://lore.kernel.org/linux-arm-kernel/20200807150355.6116-1-h.assmann@pengutronix.de/
Signed-off-by: Holger Assmann <h.assmann@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The latest kernel release now contains the device trees for the
Linux Automation MC-1. We can thus drop them from the BSP.
On the Kconfig front, we dropped CONFIG_MMC_OMAP(_HS) in favor of
CONFIG_MMC_SDHCI_OMAP. For a while now, Linux has had a sdhci-omap
driver for TI's OMAP SoCs making use of the SDHCI core library.
This was in addition to the older omap-hsmmc driver, which didn't.
Linux commit 0b4edf11187 ("ARM: dts: Move am33xx and am43xx mmc nodes
tosdhci-omap driver") first included in v5.8 changes the device tree
compatible of the mmc nodes on the am33xx and am43xx SoCs away from
the omap-hsmmc to the SDHCI driver. Add the new compatibles, so we
aren't broken by the change.
Co-authored-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Holger Assmann <h.assmann@pengutronix.de>
|
|
|
|
|
|
| |
A new ptxdist version is out, update.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
| |
With the barebox version bump, we now also have drivers for the
STM32's PWM peripherals. The Linux Automation MC-1 features a PWM RGB
LED, so enable the drivers for it to be usable.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This allows us to drop the patch stack we had for Linux Automation MC-1
support. The series went into barebox with a fix that makes a warning
about the eMMC environment go away. In return we import a single patch
to address a breakage on the rpi3b platform. Being a bug fix, we should
be able to drop the patch along with v2020.07.0.
On the Kconfig front, CONFIG_OF_NET was dropped upstream, because it was
an unreferenced symbol. All Kconfig changes are the result of an
olddefconfig.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
| |
A new ptxdist version is out, so we update our configs.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
partition-type is specific to DOS partitions. The counterpart for GPT
partitions would be partition-type-uuid, which already defaults to "L",
which is ok for our purposes.
So far, this didn't have an impact, but starting with genimage v13,
commit 013b22f26b73 ("Add support for Hybrid MBR partitions"), support
for a partition-type key in a GPT partition was introduced:
Using this option with a GPT partition table will create a hybrid
MBR partition table with a maximum of 3 partition entries(this
limit does not effect the maximum number of GPT partition entries
in the same image).
The ARM TF-A can't cope with this new format and fails to boot barebox
with a "ERROR: Partition ssbl not found" message. Drop partition-type
to prevent the ptxdist 2020.06.0 upgrade that includes host-genimage v13
from breaking the stm32mp157c-dk2 and lxa-mc1 images.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The MC-1 is the only supported board that configures loglevel via
bootspec and not via barebox environment. For symmetry and because
internal CI checks for it, migrate the loglevel to the environment.
This also affects the stm32mp157c-dk2 target, which is fine. No one
minds shorter boot times.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
| |
The barebox images for all v7a boards (e.g. rpi3, bbb) have this already
enabled. For symmetry with them and because internal CI checks for
it, enable the command.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
| |
The barebox images for all v7a boards (e.g. rpi3, bbb) have this already
enabled. For symmetry with them and because internal CI checks for
it, enable state support.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
| |
A new ptxdist version is out, update DistroKit.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
|
| |
With TF-A and barebox support added in previous commits, everything is
now in place for having DistroKit generate a MC-1 rootfs image. Do so.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds barebox support for the Linux Automation GmbH MC-1 board built
around the Octavo Systems OSD32MP157C-512M SiP.
The patches have been posted[1] upstream, but are not yet accepted.
It's however expected for them to be part of barebox v2020.06.0, at which
time we can drop the patchset again.
[1]: https://lists.infradead.org/pipermail/barebox/2020-April/041653.html
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The device tree is based on the one in linux-stm32/stm32-next, which
will probably be merged for Linux v5.8-rc1. Instead of waiting that
long, we import it here with some stuff removed/changed, so it's usable
without the prerequisite patches.
We can drop them again, when we upgrade to v5.8-rc1.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Our current kernel config already supports one STM32MP board, but we
have some HW on the MC-1, that the DK-2 lacks, namely a KSZ9031 PHY,
an EEPROM in the SiP and PWM LEDs.
We also got a display controller, but in interest of keeping kernel
code size down, enabling DRM and associated drivers is left to a
ptxdist layer specific to the MC-1.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The MC-1 can boot from SD-Card with the DK2 TF-A, but for use in
production, it's meant to boot from eMMC. Patch in the MC-1 device tree,
which enables the appropriate SDMMC controller.
RAM timings are those of the DK-2, which have worked so far, but should
be reconsidered prior to upstreaming.
[Note: eMMC boot doesn't yet work, presumably due to hardware issues.]
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
TF-A v2.3 brings in support for multi-image STM32MP1 builds, which means
less boilerplate for us, because we can build images for multiple
STM32MP1 boards at once with the same rule. We want to do this in the
follow-up commit, so bump the version now.
With this update, all boot device drivers are disabled by default.
We can afford continuing to waste a few kilobytes in the first stage
bootloader and in the future, we might want to support boards that
don't boot from SDMMC anyway, so just (re-)enable all of them.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
With this release, the TF-A rule went upstream in revised form, so it's
dropped as part of the migration.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
[adapted to current next]
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
tested on RIoTBoard
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
|
|
| |
entry
All other boards have it. And booting fails without it.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
This lets us drop the am335x USB fixes, because they were already in the
v2020.03.0 release. Only new enabled config option is
CONFIG_USB_GADGET_FASTBOOT_CMD_OEM, which was non-configurable before.
Marking it as enabled avoid a regression for people using `fastboot oem exec`
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
| |
PTXdist can now provide the necessary metadata files for
security_model=mapped-file. With this enabled, the rootfs on 9p has the
correct ownership and permissions.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
NEON was switched off, while ARMv7 has support for it.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
Disable cpu isolation and cpu hotplug.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
Disable thumbee extension.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
A new kernel is out, update all platforms.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
|
|
| |
A new ptxdist version is out, update the configs.
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
| |
this PHY is used on BeagleBone Black
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
| |
Fixes: 3dd5c0fe34cee8ebbcdd ("rules: remove CREDIT lines")
Fixes: 3bce747e47170b219334 ("v7a: add stm32mp157c-dk2 SD-Card image")
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
it is useful only for development of kernel gpio controller and produces
lots of dbug messages on each kernel boot:
[ 0.210511] gpio gpiochip0: (209c000.gpio): created GPIO range 18->18
==> 20e0000.iomuxc PIN 184->184
[ 0.210527] gpio gpiochip0: (209c000.gpio): created GPIO range 19->19
==> 20e0000.iomuxc PIN 187->187
[ 0.210545] gpio gpiochip0: (209c000.gpio): created GPIO range 20->20
==> 20e0000.iomuxc PIN 183->183
[ 0.210561] gpio gpiochip0: (209c000.gpio): created GPIO range 21->21
==> 20e0000.iomuxc PIN 188->188
[ 0.210578] gpio gpiochip0: (209c000.gpio): created GPIO range 22->24
==> 20e0000.iomuxc PIN 123->125
[ 0.210597] gpio gpiochip0: (209c000.gpio): created GPIO range 25->25
==> 20e0000.iomuxc PIN 121->121
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
| |
It is used on RIoTBoard and MarS Board
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
|
|
|
|
|
|
|
|
| |
This file has never existed here, and the respective lines were also
removed from the upstream PTXdist rules and templates in commit
cbd6bda680c65d7e01cf ("CREDITS: remove unmaintained file").
Signed-off-by: Roland Hieber <rhi@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Both STM32MP BootROM and TF-A expect the subsequent bootloader stage to
come from a GPT partition when booting from SD-Card.
Add a genimage config that describes this.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
| |
The STM32MP uses TF-A as first stage boot loader. Configure it
appropriately for use.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Trusted Firmware-A (TF-A) is a reference implementation of secure world
software for Arm A-Profile architectures (Armv8-A and Armv7-A).
TF-A is used as first-stage bootloader on the STM32MP1. We'll use
barebox for second-stage only for now, thus add a rule for TF-A.
Tested-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Rouven Czerwinski <rouven@czerwinskis.de>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
[oldconfiged rpi an v8a platforms while applying]
Signed-off-by: Robert Schwebel <r.schwebel@pengutronix.de>
Link: https://www.mail-archive.com/ptxdist@pengutronix.de/msg15828.html
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since v2019.12.0, barebox now has SD/MMC and Ethernet support for the
STM32MP1. Thus there is nothing holding us back from extending
DistroKit's v7a platform to support the SoC. Add a barebox-stm32mp
rule as first step.
The barebox rule add here is only meant as second-stage bootloader
(SSBL). First-stage (FSBL) will be TF-A introduced in a later patch.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
The STM32MP157 is an ARM SoC from STMicro with 2x Cortex-A7 cores
as well a Cortex-M MCU. Amend kernel configuration to support this
SoC as well as main peripherals used on the STM32MP157C-DK2 SBC.
For now, only driver for the non-video HW components are enabled.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Kernel commit 0782e8572c ("ARM: dts: Probe am335x musb with ti-sysc")
which barebox pulled in during the v2020.02.0 dts/ sync moved the USB nodes
to be under a ti-sysc bus instead of ti,am33xx-usb.
This broke am335x USB under barebox, because the MUSB drivers couldn't
cope with the now different device probe order.
Import the two patches fixing this out of barebox master branch.
These can be dropped when moving to barebox v2020.03.0.
Fixes: 1d84e5419f50 ("v7a: barebox: version bump 2020.01.0 -> 2020.02.0")
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
barebox v2020.02.0 has just been released. Drop the now upstream
rpi patches, oldconfig the configs and use it.
As seen in the config diffs, two of our configuration options have
changed with this release:
CONFIG_CLOCKSOURCE_ARMV8_TIMER=y:
barebox support for ARMv7 (and ARMv8) architected timers on >= ARMv7
has now been reenabled. This should improve clocksource accuracy on
the rpi3 and vexpress. The name will be changed to indicate that
it's not ARMv8 specific in v2020.03.0.
CONFIG_DEFAULT_ENVIRONMENT=n in am335x-mlo:
This symbol became selectable even without CONFIG_ENV_HANDLING.
We didn't use the environment before for this configuration, so we
probably don't want to start now, so this is left disabled.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
|