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
Ignore time taken to update EEPROM as otherwise stick commands can ups… #11228
Conversation
Fantastic! Many thanks. |
I think the issue ref has a typo, correct ref is #11226 |
A design fundamental of the previous scheduler is that a task could never be "starved" of it's execution and locked out. That seems to no longer be the case as multiple fixes have had to address this. This is dangerous - particularly when it can affect critical tasks like RX and impact disarming, failsafes, etc. |
The previous scheduler would starve tasks with a priority of idle. The new scheduler does not. The old scheduler would allow ill behaved tasks such as RX and OSD to way overrun and delay time critical tasks. Such tasks are now deterministic and the new scheduler does not allow such errant behaviour except when a task has “aged” in which case it will ensure that the task is called albeit at a reduced rate. This provides the necessary safety net whilst greatly improving scheduling performance. |
Unfortunately, this PR doesn't resolve #11226 issue yet. |
9d21002
ead754c
to
9d21002
Compare
@SunjunKim Updated as per #11226 (comment). Please retest. |
Could you post a compiled image for STM32F7X2 target again, please? I have no IDE for compiling a BF firmware. |
@SunjunKim You don't have to compile it. It gets compiled automatically by Azure. |
Oh I get it. Thanks |
Also included in mashup at #11198 (comment) |
Merged with #11198 |
1 similar comment
Merged with #11198 |
Fixes: #11226
Issue was caused by write of config to FLASH (THR_LO + YAW_LO + PIT_LO + ROL_HI) which took ~20ms and upset the RX task scheduling.
tasks
command run after this showed the issue.