DataFlash: Change erase size to pages to reduce memory usage.

[chrissnow]

Description

DataFlash supports page erases, this means that FATFS can use smaller sector sizes and therefore less memory for buffers since read-modify-write can be done in smaller chunks (there are 8 pages to a block)

It would be possible to optimise erases by erasing blocks or sectors where possible but this is not implemented.

This also means that when using non binary pages sizes that you do not exceed FF_MAX_SS of 4096 with 4224 on 16 and 32Mb devices. (DataFlash actually has 264b/528b per page)

Pull request type

[ ] Fix
[ ] Refactor
[ ] Target update
[X] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

Release Notes

Previously the DataFlash driver exposed the minimum erase size as blocks which are typically 4K, DataFlash however supports page erases as small as 256B.

This change implements page erases which will lower RAM requirements for buffers.

Since this has no relation to any file system and lowers the minimum erase size it should not cause issues with devices with existing data stored on them, however a format of FatFs in future may end up with smaller sectors by default, which will further reduce RAM requirements.

[ciarmcom]

@chrissnow, thank you for your changes.
@ARMmbed/mbed-os-storage @ARMmbed/mbed-os-maintainers please review.

[davidsaada]

This change is surprising, as I wasn't aware of the fact you could erase a page. Is this supported on all dataflash components?

[chrissnow]

Yes. All devices support page erases.

[davidsaada]

Yes. All devices support page erases.

All right then, so this is definitely a good change. Approving.

[0xc0170]

Yes. All devices support page erases.

@chrissnow Where can I verify this detail?

[chrissnow]

@0xc0170
I went through all of these for a start
https://www.adestotech.com/products/data-flash/

Also https://www.adestotech.com/wp-content/uploads/Adesto-Memory-Products-Brochure.pdf

The Adesto AT45DBxxx DataFlash devices offer uniform page erase size as small as 256 bytes

It seems the docs need an update too, it states it's I2C for a start!
https://os.mbed.com/docs/mbed-os/v5.11/apis/dataflash-block-device.html
it also mentions that there a page erases of 512b, it should really state as low as 256b

Though I can't go through every revision of every device, I'm pretty confident that any AT45 will be fine, small page erases are the main reason to choose DataFlash !

[adbridge]

CI started

[mbed-ci]

Test run: FAILED

Summary: 2 of 11 test jobs failed
Build number : 1
Build artifacts

Failed test jobs:

• jenkins-ci/mbed-os-ci_dynamic-memory-usage
• jenkins-ci/mbed-os-ci_greentea-test

[chrissnow]

Are those CI issues related to this PR?

[0xc0170]

Are those CI issues related to this PR?

I don't think so. CI job restarted

[0xc0170]

@chrissnow can you rebase this PR? I am seeing unrelated error in the dynamic memory usage (new to me), I would restart testing

[chrissnow]

@0xc0170 rebased, made a bit of a mess with git but should be ok now.

[adbridge]

CI restarted

[0xc0170]

Internal CI fault, restarted

[0xc0170]

It again failed to checkout the branch, I restarted

[mbed-ci]

Test run: SUCCESS

Summary: 11 of 11 test jobs passed
Build number : 5
Build artifacts

[0xc0170]

@chrissnow Can you add release notes for this functionality change? (section in your first comment here).

[chrissnow]

@0xc0170 Release notes added, Is this roughly the sort of thing you were after?

[0xc0170]

Thanks , yes