boot: zephyr: kconfig: Default to swap using offset for nRF devices #2266
Conversation
Defaults to the new swap using offset mode Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Fixes an issue with swap using offser with serial recovery by changing an #ifdef Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
| @@ -397,14 +397,14 @@ config BOOT_VALIDATE_SLOT0_ONCE | |||
|
|
|||
| config BOOT_PREFER_SWAP_OFFSET | |||
| bool "Prefer the newer swap offset algorithm" | |||
| default y if SOC_FAMILY_NORDIC_NRF | |||
There was a problem hiding this comment.
Can we make it default for all platforms?
By adding: default y if !$(dt_nodelabel_enabled,scratch_partition)
There was a problem hiding this comment.
It's not been made the default for all platforms due to the sector size requirements, if a device has equal sizes for slot 0 and slot 1 then it will work (but is not ideal) but if it has the swap using move ideal setup of slot 0 being larger than slot 1, it will fail, so has been left as swap using move for default for now
There was a problem hiding this comment.
if a device has equal sizes for slot 0 and slot 1 then it will work (but is not ideal)
Using optimized slot sizes is broken with the latest mcuboot: zephyrproject-rtos/zephyr#88850
It seems, I need to switch to equal slot sizes (at least for NXP) to make it work for all modes, even if it's not optimal.
There was a problem hiding this comment.
That has been fixed efa3039 but has not been pulled into zephyr yet
There was a problem hiding this comment.
That has been fixed efa3039 but has not been pulled into zephyr yet
Great! Thank you.
There was a problem hiding this comment.
That has been fixed efa3039 but has not been pulled into zephyr yet
Just tried. Unfortunately, it does not fix the optimized partitions swap move issue.
Still writes "W: Cannot upgrade: not a compatible amount of sectors"
There was a problem hiding this comment.
what board/config is used?
There was a problem hiding this comment.
having a look at frdm_ke17z it looks like it should work, would be just under 2K for swap status fields so not sure why it throws an error, would suggest enabling debug logs for MCUboot and seeing what is printed, then setting the number of sectors for each slot to the same and seeing what the non-optimal message says
There was a problem hiding this comment.
having a look at
frdm_ke17zit looks like it should work, would be just under 2K for swap status fields so not sure why it throws an error, would suggest enabling debug logs for MCUboot and seeing what is printed, then setting the number of sectors for each slot to the same and seeing what the non-optimal message says
For example for frdm_mcxw71, it has Flash total size 1M , 8KB erase and 16B write. Partitioned to:
&flash {
partitions {
compatible = "fixed-partitions";
#address-cells = <1>;
#size-cells = <1>;
/* The MCUBoot swap-move algorithm uses the last 2 sectors
* of the primary slot0 for swap status and move.
*/
boot_partition: partition@0 {
reg = <0x0 DT_SIZE_K(64)>;
};
slot0_partition: partition@10000 {
reg = <0x10000 (DT_SIZE_K(416) + DT_SIZE_K(16))>;
};
slot1_partition: partition@7C000 {
reg = <0x7C000 DT_SIZE_K(416)>;
};
storage_partition: partition@E4000 {
reg = <0xE4000 DT_SIZE_K(112)>;
};
};
};
There was a problem hiding this comment.
BTW: Is see another PR fix is coming "boot: bootutil: Fix max image size computation for swap-move and swap-offset" #2283
Defaults to the new swap using offset mode