Move alignment helper to its own module in preparation of HMAC module

[AnthonyGrondin]

Thank you!

Thank you for your contribution.
Please make sure that your submission includes the following:

Must

• The code compiles without errors or warnings.
• All examples work.
• cargo fmt was run.
• Your changes were added to the CHANGELOG.md in the proper section.
• You updated existing examples or added examples (if applicable).
• Added examples are checked in CI

Nice to have

• You add a description of your work to this PR.
• You added proper docs for your newly added features and code.

This PR splits up #465, as suggested in #465 (comment) to move the alignment helper and SHA module refactor into it's own PR to be more easily reviewable.

Code mostly taken from:

• a9807ae
• 7f3b639

I have tested on esp32s3, but not on the other architectures as I don't have them on hand.

[bjoernQ]

Looks good! Just one small thing

[jessebraham]

Something is not quite right here, running on an ESP32 produces the following output:

Took 26941 cycles
SHA512 Hash output [15, 0e, 8d, ee, c8, d8, dc, de, b1, ef, a2, 86, cf, 49, 6a, ac, 65, 67, 3f, d8, ce, 00, b3, 64, ce, 5e, 51, 48, 10, ee, 4b, 5f, c2, 89, 1d, e1, 53, 81, 21, 1c, ab, 31, b2, c3, c9, ed, ca, 77, e5, d7, 24, 6c, 80, 86, 94, 9a, 1a, 00, 51, e1, 80, ec, 8c, e1]
Took 1595586 cycles
SHA512 Hash output [ee, 8d, 0e, 15, de, dc, d8, c8, 86, a2, ef, b1, ac, 6a, 49, cf, d8, 3f, 67, 65, 64, b3, 00, ce, 48, 51, 5e, ce, 5f, 4b, ee, 10, e1, 1d, 89, c2, 1c, 21, 81, 53, c3, b2, 31, ab, 77, ca, ed, c9, 6c, 24, d7, e5, 9a, 94, 86, 80, e1, 51, 00, 1a, e1, 8c, ec, 80]
HW SHA is 59x faster

I confirmed that the hashes match when running against the main branch.

[AnthonyGrondin]

Something is not quite right here, running on an ESP32 produces the following output:

Took 26941 cycles
SHA512 Hash output [15, 0e, 8d, ee, c8, d8, dc, de, b1, ef, a2, 86, cf, 49, 6a, ac, 65, 67, 3f, d8, ce, 00, b3, 64, ce, 5e, 51, 48, 10, ee, 4b, 5f, c2, 89, 1d, e1, 53, 81, 21, 1c, ab, 31, b2, c3, c9, ed, ca, 77, e5, d7, 24, 6c, 80, 86, 94, 9a, 1a, 00, 51, e1, 80, ec, 8c, e1]
Took 1595586 cycles
SHA512 Hash output [ee, 8d, 0e, 15, de, dc, d8, c8, 86, a2, ef, b1, ac, 6a, 49, cf, d8, 3f, 67, 65, 64, b3, 00, ce, 48, 51, 5e, ce, 5f, 4b, ee, 10, e1, 1d, 89, c2, 1c, 21, 81, 53, c3, b2, 31, ab, 77, ca, ed, c9, 6c, 24, d7, e5, 9a, 94, 86, 80, e1, 51, 00, 1a, e1, 8c, ec, 80]
HW SHA is 59x faster

I confirmed that the hashes match when running against the main branch.

I did change the following from the original code, based on what was originally in sha.rs. Did it introduce this bug? I don't have an ESP32 on hand yet to test:
Original:

esp-hal/esp-hal-common/src/reg_access.rs

Line 11 in d744d00

const U32_FROM_BYTES: fn([u8; 4]) -> u32 = u32::from_ne_bytes;

Updated (This PR):

esp-hal/esp-hal-common/src/reg_access.rs

Lines 11 to 15 in ed4d136

	// ESP32 does reversed order
	#[cfg(esp32)]
	const U32_FROM_BYTES: fn([u8; 4]) -> u32 = u32::from_be_bytes;
	#[cfg(not(esp32))]
	const U32_FROM_BYTES: fn([u8; 4]) -> u32 = u32::from_ne_bytes;

[jessebraham]

LGTM, thanks for taking care of this!