Skip to content

[STM32H7] Provisioning fs reformat #957

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 5 commits into
base: main
Choose a base branch
from

Conversation

andreagilardoni
Copy link
Contributor

Changing flash partitioning scheme in order to make room for a KV store

@andreagilardoni andreagilardoni changed the title Provisioning fs reformat [STM32H7] Provisioning fs reformat Sep 23, 2024
Copy link
Contributor

@pennam pennam left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There is no need to erase the whole flash before repartitioning; you only need to erase the MBR.

root.init();
root.erase(0x0, root.get_erase_size());

Furthermore if you keep the KVstorage in partition 3 for both schemes you should be able switch between them without erase the MBR at all

mbed::MBRBlockDevice::partition(&root, 1, 0x0B, 0, 1024 * 1024);
if(default_scheme) {
  mbed::MBRBlockDevice::partition(&root, 4, 0x0B, 14 * 1024 * 1024, 14 * 1024 * 1024);
  mbed::MBRBlockDevice::partition(&root, 3, 0x0B, 13 * 1024 * 1024, 14 * 1024 * 1024);
  mbed::MBRBlockDevice::partition(&root, 2, 0x0B, 1024 * 1024, 13 * 1024 * 1024);
  // use space from 15.5MB to 16 MB for another fw, memory mapped
} else {
  mbed::MBRBlockDevice::partition(&root, 2, 0x0B, 1024 * 1024, 6 * 1024 * 1024);
  mbed::MBRBlockDevice::partition(&root, 3, 0x0B, 6 * 1024 * 1024, 7 * 1024 * 1024);
  mbed::MBRBlockDevice::partition(&root, 4, 0x0B, 7 * 1024 * 1024, 14 * 1024 * 1024);
  // use space from 15.5MB to 16 MB for another fw, memory mapped
}

@andreagilardoni
Copy link
Contributor Author

Furthermore if you keep the KVstorage in partition 3 for both schemes you should be able switch between them without erase the MBR at all

I tried switching from this scheme to the older one and something bad happened and the only way to make it work properly was to erase the partitions. I am double checking this.

@andreagilardoni
Copy link
Contributor Author

I confirm that without the MBR erase I cannot use the old partitioning scheme.
@pennam do you think it is better to split this PR into 2? the first one with just the MBR erase command and the other with the new partitioning scheme, so that one can rollback with this fix present.

@andreagilardoni andreagilardoni force-pushed the provisioning-fs-reformat branch 2 times, most recently from 1d22ce2 to 6a309b7 Compare March 11, 2025 16:46
@andreagilardoni
Copy link
Contributor Author

andreagilardoni commented Mar 11, 2025

@pennam

There is no need to erase the whole flash before repartitioning; you only need to erase the MBR.

What do you think about the way I wrote the MBR table clean?

Furthermore if you keep the KVstorage in partition 3 for both schemes you should be able switch between them without erase the MBR at all

Now partition 3 is being used for user data, this would mean that we need either to switch user data partition to 4, or distinguish between the two cases when using kvstore. I prefer to keep it to partition number 4.

@andreagilardoni andreagilardoni force-pushed the provisioning-fs-reformat branch from 6a309b7 to 392cc64 Compare March 11, 2025 16:52
@megacct
Copy link

megacct commented Mar 11, 2025

@andreagilardoni @pennam Hi, a couple of us have tried, unsuccessfully, to scrub the MBR in order to change the partition layout. Detail here https://forum.arduino.cc/t/how-to-partition-giga-flash-similar-to-original-factory-defaults/1349759. I'm sure we were doing something wrong but it may be worth double-double-checking that the erase in this commit is doing what's expected.

@andreagilardoni
Copy link
Contributor Author

I would like to point out that when repartitioning, for the user or ota block devices some files can be corrupted if they lie in the area that will be used for partition 4.

@andreagilardoni andreagilardoni force-pushed the provisioning-fs-reformat branch from 41d9a90 to 9e05f2b Compare March 12, 2025 12:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants