Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Remove check for openssl so that we can build basic components without
it installed.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The CANOPi board comes with a VIA VL805 4 Port USB controller on the
PCIe bus, which requires an initialization sequence (e.g. firmware
loaded) to be run during boot.

In order to properly initialize this device, it needs to be declared in
the device tree. Otherwise, the PCIe bus may communicate with USB
controller during initialization, which hangs the board on a Synchronous
Abort, e.g.:

```
U-Boot menu
1:     Apertis v2023dev3 5.18.0-0.deb11.4-arm64
2:     Apertis v2023dev3 5.18.0-0.deb11.4-arm64 (rescue target)
Enter choice: 1:       Apertis v2023dev3 5.18.0-0.deb11.4-arm64
Retrieving file: /initrd.img-5.18.0-0.deb11.4-arm64
"Synchronous Abort" handler, esr 0x96000004
elr: 000000000009cc90 lr : 00000000000a8e50 (reloc)
elr: 000000003b370c90 lr : 000000003b37ce50
x0 : d519b040aa010000 x1 : 0000000000000058
x2 : 0000000000400000 x3 : 000000003b3d29c0
x4 : b900080152b00000 x5 : 000000003b3d2450
x6 : 000000000000005b x7 : 000000003b3d29f0
x8 : b900080152affdf0 x9 : 0000000000000008
x10: fffffffffffffff0 x11: 0000000000000006
x12: 000000003af58730 x13: 000000000001ea57
x14: 0000000000000000 x15: 0000000000000000
x16: 000000003b37fac4 x17: 656e2f74656e7265
x18: 000000003af4fd90 x19: 0000000000000210
x20: 000000003b3d2440 x21: 0000000000000001
x22: 0000000000000200 x23: 000000003af43e80
x24: 0000000000000200 x25: 000000000001ea56
x26: 0000000000000000 x27: 0000000000000006
x28: 0000000000000009 x29: 000000003af43d60

Code: eb02007f 54ffff48 f9400cc4 17ffffe0 (f9400404)
Resetting CPU ...
```

The RPi loader loads dtbs and overlays for u-boot from the firmware
partition. In the CANOPi case, it will try to load the vl805.dtbo
overlay, so let's provide it.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Currently, the u-boot binary is going beyond the 1MiB limit on R-Car
Gen3. Upstream u-boot doesn't have this problem as the DISTRO_DEFAULTS
config option is disabled.

Let's disable the CMD_I2C config option to shrink the final image size
between the boundaries.

NOTE: Ideally, other unused options should be shrinked as well. In fact,
if a feature is not supported nor used across all Apertis platforms, it
could be globally disabled.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Merge remote-tracking branch 'origin/apertis/v2023pre' into proposed-updates/debian/unstable/5dfdedcc

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The iMX8MN Variscite Board is not being able to boot from eMMC due to a
regression introduced on v2022.04. This is fixed by the following
upstream commit, so pick it:

  commit 2349ecffec5fc91d3c8964725865132c73d68f15
  Author: Marek Vasut <marex@denx.de>
  Date:   Fri Jul 15 01:58:24 2022 +0200

      mmc: Do not send status of send_status is false

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

arm: dts: imx8mn_var_som: Set atf-bl31 blob entry type

Configure binman ATF blob entry type to use the path from the BL31
environment variable, if defined.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

After commit 395110284b56 ("nitrogen6x: Populate FDTFILE at build-time
for all platforms") iMX.6Q Sabrelite fails to boot due to a bad fdtfile
string:

  Retrieving file: /dtbs/5.18.0-0.deb11.4-armmp/"imx6q-sabrelite".dtb
  ** File not found /dtbs/5.18.0-0.deb11.4-armmp/"imx6q-sabrelite".dtb **

CONFIG_DEFAULT_DEVICE_TREE option is string typed, so __stringify() is
adding extra quotes. Remove this.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Add MMC aliases to ensure that the /dev/mmcblk ID for the MMC device for
bootcount won't change depending on the probe order of the MMC drivers.

This is required as the configs/rpi_arm64_defconfig is currently setting
the bootcount dev/part to mmc:0:2 for all the rpi platforms using that
defconfig, so it must be consistent.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

This commit adds a basic device tree to boot RPi Compute Module 4 (CM4)
boards.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The RPi loader tries to load a dtb for u-boot from the firmware
partition. Currently, the apertis-image-recipes recipes for RPi are
copying these dtb files from the kernel.

It makes more sense to copy these dtb blobs from u-boot rather than the
kernel, as the bootloader is the one using them. Moreover, these dtbs
are built supporting overlays, which we might use for the CANOPi
auto-detection support.

For the above reasons, let's ship the built dtbs for arm64 rpi
platforms.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

In the RPI compute Module 4 (CM4) case, try to detect if the current
board variant is CANOPi, to load the proper device-tree.

The CANOPi board comes with a VIA VL805 4 Port USB controller on the
PCIe BDF 1.0.0. Detect it to differentiate this board from the CM4
I/O board, which doesn't have this module.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

For the RPi CM4 (Compute Module 4), we currently try to load the dtb
file bcm2711-rpi-cm4.dtb, which is not built by the upstream kernel.

Instead, the only CM4 dtb file provided by linux upstream is the
bcm2711-rpi-cm4-io.dtb, so let's use that.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Apertis is currently shipping arm-trusted-firmware version 2.5+dfsg-1.
Let's stick to that version to avoid having to update other pkgs to
bookworm as well.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Merge remote-tracking branch 'origin/apertis/v2023dev3' into proposed-updates/debian/bookworm/146c8e68

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The boot procedure on LAVA is having problem with the u-boot prompt
being set to ">" on iMX8MN VAR-SOM-MX8M-NANO board. Let's set it to
"U-Boot>" so it matches what other supported boards do, e.g. RPi4.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The IVT table is located right after the FIT image. If the FIT image is
generated using external data, the total size needs to be computed
including the FIT struct and external data lengths.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Since u-boot commit 92055e138f2 ("image: Drop if/elseif hash selection
in calculate_hash()") if the iMX6 CAAM module is enabled, it will be
used to calculate the SHA1 hash value.

There's an issue where the iMX6 CAAM module fails due to a bad alignment
on the images:

```
  Verifying Hash Integrity ... sha1Error: Address arguments are not
  aligned CAAM was not setup properly or it is faulty
```

The reason is that both the image and the buffer must be aligned for the
DMA engine (64-byte boundary).

See drivers/crypto/fsl/fsl_hash.c:caam_hash() function:
```
        if (!IS_ALIGNED((uintptr_t)pbuf, ARCH_DMA_MINALIGN) ||
            !IS_ALIGNED((uintptr_t)pout, ARCH_DMA_MINALIGN)) {
                puts("Error: Address arguments are not aligned\n");
                return -EINVAL;
        }
```

The Apertis image can be easily fixed to be 64-byte aligned, however the
buffer requires a little bit more work. Let's make the SHA hash hw
accelleration optional, even if CAAM module is enabled, so it can be
disabled temporarily until this issue is fixed.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The boot procedure on LAVA is having problem with the u-boot prompt
being set to ">" on BSH SMM S2 devices. Let's set it to "U-Boot>" so it
matches what other supported boards do, e.g. RPi4.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>