Not all Rockchip boards have their console running at 1500000 baud in
U-Boot and the kernel. Such is the case for puma-haikou RK3399-based
SoM+Carrierboard.
In order to prepare for the addition of puma-haikou to meta-rockchip,
let's replace the baudrate in TF-A by the one defined in the machine
conf file in the RK_CONSOLE_BAUD variable.
Cc: Quentin Schulz <foss+yocto@0leil.net>
Signed-off-by: Quentin Schulz <quentin.schulz@theobroma-systems.com>
The _virtual notation is not an override. These syntax "fixes" need to be
reverted.
In the case of the kernel override, when it was added, the rock-pi-e needed
the latest kernel (linux-yocto-dev) but now the default linux-yocto kernel
will suffice. So this mistake actually switched the rock-pi-e from
linux-yocto-dev back to linux-yocto inadvertently but at a time when
linux-yocto-dev was no longer required.
In the case of the bootloader overrides, u-boot was always the default, so
these overrides were always redundant.
Therefore, in the end, simply removing these overrides is the best way
forward (considering these aren't doing anything, and the builds are working
fine regardless).
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
tf-a built with clang is bloated for rk3399 SOCs with 2.6+
it needs looking into, until then use gcc always to build it
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Jon Mason <jon.mason@arm.com>
Cc: Ross Burton <ross.burton@arm.com>
The latest trusted-firmware-a recipe in meta-arm no longer considers
trusted-firmware-a to have potentially multiple providers.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Since the recent patch to switch to UUIDs [0aa5e600: "use uuid
instead of hard-coding root device"] wic fstab-update is not able
to get the correct value for the used device anymore and falls to
the default 'sda'. Thus wrong /dev/sda entries are generated in fstab.
For partitions that should be updated automatically this can be avoided
by either generate entries using uuid or label.
[NOTE: I rearranged the order of the arguments so they line up]
Signed-off-by: MarkusVolk <f_l_k@t-online.de>
There is no "nanopi-m4" defined in any yocto kernel metadata (yet?), therefore
remove this superfluous line.
Build (core-image-base) and run tested (both systemd and sysvinit) on:
- nanopi-m4
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Now that we're booting via UUID, we no longer need these aliases in the DT.
Personally I wasn't able to prove to myself that they actually worked (at
least not with 5.13.y) and fiddling with these aliases didn't seem to affect
the mmc probe order on boot. Additionally it looks like some of these aliases
will be landing upstream shortly.
Build (core-image-base) and run tested (both systemd and sysvinit) on:
- rock64
- rock-pi-e
(i.e. the two rk3328 MACHINEs)
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
For a while we've noticed that /etc/fstab ends up with "junk" appended to the
end e.g.:
/dev/sda1 loader1 vfat defaults 0 0
/dev/sda2 reserved1 vfat defaults 0 0
/dev/sda3 reserved2 vfat defaults 0 0
/dev/sda4 loader2 vfat defaults 0 0
/dev/sda5 atf vfat defaults 0 0
/dev/sda6 /boot vfat defaults 0 0
In most cases this doesn't appear to affect the systems negatively.
However, with the recent patch to switch to UUIDs [0aa5e600: "use uuid
instead of hard-coding root device"] this started becoming an issue on systems
using systemd. Therefore we need to stop wic from adding these junk entries so
that systems continue to boot correctly.
Build tested with core-image-base, nodistro, with both sysvinit and systemd
for:
- marsboard-rk3066
- rock2-square
- firefly-rk3288
- vyasa-rk3288
- nanopi-m4[-2gb]
- tinker-board[-s]
- rock-pi-e
- rock-pi-4[abc]
- rock64
Run tested, core-image-base, both sysvinit and systemd on:
- tinker-board
- rock-pi-e
- rock-pi-4b
- rock64
Commit message updated by Trevor Woerner <twoerner@gmail.com>
Tested-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: MarkusVolk <f_l_k@t-online.de>
Recent upstream kernel changes have made the mmc probing order unpredictable.
Therefore, boards with both an emmc and sdmmc interface aren't guaranteed to
boot with a hard-coded root device selected.
For example, on the rock64, with linux-yocto 5.10.y, using the uSD card (i.e.
the sdmmc interface) about 50% of the time the boot would succeed, and roughly
50% of the time it wouldn't:
...
[ 0.612233] Waiting for root device /dev/mmcblk1p7...
[ 0.634551] mmc_host mmc1: Bus speed (slot 0) = 300000Hz (slot req 300000Hz, actual 300000HZ div = 0)
[ 0.639064] mmc_host mmc0: Bus speed (slot 0) = 50000000Hz (slot req 50000000Hz, actual 50000000HZ di)
[ 0.640007] mmc0: new high speed SDXC card at address 5048
[ 0.641176] mmcblk0: mmc0:5048 SD64G 58.0 GiB
[ 0.647610] random: fast init done
[ 0.648279] GPT:Primary header thinks Alt. header is not at the end of the disk.
[ 0.648941] GPT:376479 != 121634815
[ 0.649252] GPT:Alternate GPT header not at the end of the disk.
[ 0.649796] GPT:376479 != 121634815
[ 0.650106] GPT: Use GNU Parted to correct GPT errors.
[ 0.650598] mmcblk0: p1 p2 p3 p4 p5 p6 p7
NOTE the discrepancy between the kernel waiting for device /dev/mmcblk1p7,
which comes from the hard-coded kernel cmdline, and the kernel probing putting
the sdmmc on mmcblk0.
With linux-yocto 5.13.y on the rock64 using the uSD card the board would never
boot, the sdmmc always appears on mmcblk0.
Instead of simply changing the hard-coded root device (i.e. from mmcblk0 to
mmcblk1) switch to using partition UUIDs instead. Hard-coding the boot device
would work with 5.13.y but would fail 50% of the time with 5.10.y; who knows
what other kernels will do?
In any case, switching to UUIDs works regardless of board, kernel, or
available mmc interfaces.
Boot tested on:
- rock64
- nanopi-m4-2gb
- tinker-board
- rock-pi-e
- rock-pi-4b
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
The rock64 has an ARM Mali 450 MP2 GPU, therefore enable mesa's lima for
accelerated, open-source graphics.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
It looks like I missed a case for the new bitbake override syntax. My tests
weren't done from a fresh build so either a preexisting image was still
available, or the unfixed syntax caused a race.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
With bitbake commit 7dcf317cc141dc980634f8c18bfa84f83e57206a
("bitbake: Switch to using new override syntax"), applied on
Aug 2, 2021, the OVERRIDE separator is now a colon instead of
an underscore. Therefore all builds performed with a bitbake
before this change must use a meta-rockchip commit before this
one, and any builds performed with a bitbake after this change
must use a meta-rockchip starting from this commit onwards.
Build-tested for all meta-rockchip MACHINEs.
Run tested on:
- tinker-board
- nanopi-m4-2gb
- rock64
- rock-pi-4b
- rock-pi-e
The tinker-board and rock-pi-e work fine. The rest of the boards
seem to have a, hopefully unrelated, issue running a
5.13-yocto-standard kernel. However, all boards work with the
5.10-yocto-standard kernel.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Adding "-rockchip" to the Linux kernel name implies, to me anyway, that this
is a vendor kernel. The PREFERRED_PROVIDERs of all kernels specified in this
BSP are upstream linux-yocto kernels, not vendor kernels. Therefore remove the
version name extension to avoid confusion.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
The latest updates to linux-yocto-dev now include support for the rock-pi-e so
do away with our custom recipe and use the one from oe-core.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
By exporting a couple more variables the wks file for every rockchip device
can be built from one template instead of having separate wks files for each
board and platform.
The following BSP variables were checked before and after this change to make
sure they remained valid/sensible:
- WKS_FILE
- UBOOT_SUFFIX
- SPL_BINARY
- IMAGE_FSTYPES
Built-tested for every MACHINE in this BSP.
Run-tested on the following devices to ensure they continue to boot correctly
to a cmdline (core-image-base):
- tinker-board
- rock-pi-e
- rock-pi-4b
- rock64
- nanopi-m4-2gb
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
The ext4 IMAGE_FSTYPES does not need to be mentioned explicitly. It will be
automatically generated in cases where it is needed.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Create a conf/machine/include/rockchip-wic.inc file to contain all the common
wic/wks things for easy inclusion by any MACHINEs that use wic for their image
creation.
NOTE: the wic image type of rock-pi-e changed from "wic.xz" to "wic" which
matches all the other meta-rockchip MACHINEs that use wic
The following variables were checked before and after to make sure they remain
correct/sensible:
- IMAGE_FSTYPES
- WKS_FILE_DEPENDS
- IMAGE_BOOT_FILES
- RK_CONSOLE_BAUD
- RK_CONSOLE_DEVICE
- RK_BOOT_DEVICE
- SERIAL_CONSOLES
- WICVARS
Build-tested for all currently-defined MACHINEs.
Boot-tested on the following boards to make sure they continue to boot to a
console correctly (core-image-base):
- tinker-board
- rock64
- rock-pi-4b
- rock-pi-e
- nanopi-m4-2gb
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Consolidate all the various console definitions to the common
conf/machine/include/rockchip-defaults.inc file and create
RK_CONSOLE_BAUD and RK_CONSOLE_DEVICE variables that can be
reused in the wks files.
The following variables were checked before and after this patch
to make sure they are sensible:
- SERIAL_CONSOLES
- RK_CONSOLE_DEVICE
- RK_CONSOLE_BAUD
A boot test was performed on the following boards to make sure
they all continue to boot to a cmdline:
- tinker-board
- rock-pi-e
- nanopi-m4-2gb
- rock64
- rock-pi-4b
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
This ensures that DISTRO_FEATURES can translate correctly
for qtbase PACKAGECONFIGs and we can get right platforms enabled.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Now that there is a second rk3328-based MACHINE (rock64) switch rock-pi-e to
use the common rk3328 include.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
This is a RK3328 board from Pine64.
Board details at https://wiki.pine64.org/wiki/ROCK64.
Default image is built to boot from SD-card. Building an image for
eMMC requires to set RK_BOOT_DEVICE="mmcblk0".
Signed-off-by: Yann Dirson <yann@blade-group.com>
This will allow us to define a single set of kernel BSP for all
variants of the board (which only need to differ in u-boot dts).
Signed-off-by: Yann Dirson <yann@blade-group.com>
TF-A runs between two u-boot stages which both uses 1500000 baud, it
just makes no sense to use the same UART at a different rate.
Here is a sample session with the successive stages, with TF-A artificially
separated for emphasis:
[20210406-175438.135934] U-Boot TPL 2021.01 (Jan 11 2021 - 18:11:43)
[20210406-175438.135956] Channel 0: DDR3, 933MHz
[20210406-175438.236974] BW=32 Col=10 Bk=8 CS0 Row=15 CS=1 Die BW=16 Size=1024MB
[20210406-175438.237000] Channel 1: DDR3, 933MHz
[20210406-175438.237004] BW=32 Col=10 Bk=8 CS0 Row=15 CS=1 Die BW=16 Size=1024MB
[20210406-175438.237008] 256B stride
[20210406-175438.237012] Trying to boot from BOOTROM
[20210406-175438.237015] Returning to boot ROM...
[20210406-175438.237018]
[20210406-175438.573394] U-Boot SPL 2021.01 (Jan 11 2021 - 18:11:43 +0000)
[20210406-175438.573431] Trying to boot from MMC1
[20210406-175438.589254] NOTICE: BL31: v2.3():v2.3-dirty
[20210406-175440.534055] NOTICE: BL31: Built : 15:56:43, Apr 20 2020
[20210406-175441.393423] U-Boot 2021.01 (Jan 11 2021 - 18:11:43 +0000)
[20210406-175441.393429]
[20210406-175441.393433] SoC: Rockchip rk3399
Signed-off-by: Yann Dirson <yann@blade-group.com>
The Yocto Project mailing lists migrated to a new system and email address a
while back. Update the README with the up-to-date information.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Create a common conf/machine/include/tinker.inc and re-spin
- conf/machine/tinker-board.conf
- conf/machine-tinker-board-s.conf
to just contain the differences.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
The COMPATIBLE_MACHINE strings were getting unwieldy, so switch to the
MACHINEOVERRIDE notation so they're neater.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Add support for Radxa's ROCK Pi E device
https://wiki.radxa.com/RockpiE
It's a great surprise to find upstream U-Boot and the Linux kernel already
provide support for this board! On the kernel side this support was
added in 5.11. However, that support is so new that even linux-yocto-dev
(which is based on 5.11) doesn't include the commits that add support
for this board yet. As a result I've added a custom Linux kernel recipe
(linux-stable-bleeding) which should, in time, become unnecessary.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Splits out the three variants of the rock-pi-4 (A, B & C) into their own
machines. Unfortunately, it is not possible to have a single machine
that works for all three, as there isn't any known ways for the
bootloader to distinguish them. The old rock-pi-4 machine is kept around
for use with older kernels
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
linux-yocto_5.8 has been removed from the list of kernels found in
openembedded-core, therefore this bbappend is no longer needed.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Fixes the serial port output stopping mid way through the boot process
by reverting the kernel commit that caused it.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
We already have a dependency on meta-arm/meta-arm in order to build
tf-a (there's no point carrying our own recipe when there's a common,
consolidated one to use in meta-arm).
meta-arm/meta-arm now has a dependency on meta-arm/meta-arm-toolchain.
meta-arm-toolchain has a recipe for virtual/arm-none-eabi-gcc, so we might
as well use that too, and remove our own version. Note that using the
meta-arm-toolchain recipe required a small change to how the dependency is
specified.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
