Age | Commit message (Collapse) | Author |
|
This adds a new recipe tdx-updater which provides the files needed by
tdx-image-fstype.inc rather than copying them in a non OE way in
tdx-image-fstype.inc.
This required to restructure the layout in files. An additional update
directory is inserted. Also in each of the overrides directories a
symlink to library is created.
This now allows to build legacy update images with a custom machine
configuration if one adds the relevant machine override, e.g. for a
machine targeting a colibri-imx6 module:
MACHINEOVERRIDES .= ":colibri-imx6"
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Switch to the generic compressed Kernel image type (zImage) instead of
the U-Boot specific uImage format.
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@gmail.com>
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Do no longer hard-code conf_blk_offset now that we reliably determine
it in check_configblock.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Unfortunately as loadaddr is equal to fdt_addr_r on some modules and
equal to kernel_addr_r on others using it together with any of the
others may cause issues. Therefore no longer make use of loadaddr and
consistently use fdt_addr_r instead.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
The config block check aka check_configblock may fail during update
depending on the size of the hardware area boot partition which may
lead to the config block not getting properly migrated or updated if
still using parts of the legacy update procedure. This patch fixes
this by iterating over various sizes.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Improve on the config block locations checked for migration from
WinCE/WEC. Now also takes vendor part of the Boot Configuration Table
aka BCT into account.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
If an image creates a rootfs which is exactly a multiple of 64 MByte then
the scripts never detects that the last chunk has been read and
'run update' never completes.
Fix that by explicitely setting filesize to 0 when a file read fails.
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Our current update procedure reads the rootfs payload in pre
split files of 64 MB size. However due to the workaround for
u-boot counting in hex and split did in decimal, the number of
chunks is limited to 89 resulting in a maximum size of 5.5GB.
Update the u-boot counting procedure to count in 3 digits from
the current 2 digits. This allows more than 800 plus chunks
which should take care of rootfs size and any future eMMC
update requirements.
Signed-off-by: Sanchayan Maity <sanchayan.maity@toradex.com>
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Our current update procedure requires the rootfs payload to fit in
one file on a FAT filesystem. This limits its usable size to 4GB.
Use the process of splitting up the files in 64MB chunks and then
doing the update which is also what the Ethernet update scripts
currently follow.
However due to the workaround for u-boot counting in hex and split
did in decimal, the number of chunks is limited to 89 resulting in
a maximum size of 5.5GB. We will remove the limitation in a follow
up patch.
Signed-off-by: Sanchayan Maity <sanchayan.maity@toradex.com>
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
This requires to adjust the path to include files, e.g.:
-require recipes/images/trdx-image-fstype.inc
+require recipes-images/images/trdx-image-fstype.inc
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Stefan Agner <stefan.agner@toradex.com>
|