Spi loopback adjustments #91194
Merged
Spi loopback adjustments #91194
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The spi_loopback test suite creates three internal threads. The stack sizes for these threads is hardcoded to 512, which is to little for some socs, namely the nrf54h20 cpuapp if pm device runtime is enabled. Instead, determine the stack sizes the same way ztest stack sizes are determined. Signed-off-by: Bjarki Arge Andreasen <[email protected]>
The test which measures and validates transfer times is using the
helper spi_loopback_transceive() to perform the transfer. This
helper internally gets the spi controller, which is useful for
most tests, but for this one, it means the time to get and put
the spi controller is included in the transfer time measurement.
This commit gets the spi controller before calling
spi_loopback_transceive() which results in only the actual
transfer time being measured.
Before this commit, on the nrf54h20:
START - test_spi_complete_multiple_timed
Transfer took 745 us vs theoretical minimum 108 us
Latency measurement: 637 us
PASS - test_spi_complete_multiple_timed in 0.008 seconds
START - test_spi_complete_multiple_timed
Transfer took 700 us vs theoretical minimum 54 us
Latency measurement: 646 us
Assertion failed at ...
Very high latency
FAIL - test_spi_complete_multiple_timed in 0.027 seconds
After this commit:
START - test_spi_complete_multiple_timed
Transfer took 250 us vs theoretical minimum 108 us
Latency measurement: 142 us
PASS - test_spi_complete_multiple_timed in 0.008 seconds
START - test_spi_complete_multiple_timed
Transfer took 204 us vs theoretical minimum 54 us
Latency measurement: 150 us
PASS - test_spi_complete_multiple_timed in 0.008 seconds
Signed-off-by: Bjarki Arge Andreasen <[email protected]>
|
|
@decsny The timing measurements are great! Really puts into perspective how inefficient "inline PM" is :) 500us of get/put for every total transfer time of around 200us is wild |
decsny
approved these changes
Jun 6, 2025
JarmouniA
approved these changes
Jun 6, 2025
tbursztyka
approved these changes
Jun 6, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Adjust the spi_loopback test to use a more optimal method of determining stack sizes used internally, and produce a more accurate transfer time measurement when pm device runtime is enabled.