Page 2 of 2

Re: nrf52: external file system

Posted: Wed Feb 17, 2021 1:14 am
by ttmetro
jimmo wrote:
Wed Feb 17, 2021 12:59 am
But it's crazy inefficient because you have a task that has to burn a lot of time just polling that queue.
Thanks for the note on ThreadSafeFlag; I'll use that for now.

BTW: no need to poll for interrupt driven events. The latter wake up the CPU, after handling the ISR it proceeds right from where it went to light sleep (presumably after WFI or something). If it decides nothing is ready it goes right back to sleep.

Re: nrf52: external file system

Posted: Wed Feb 17, 2021 3:08 am
by jimmo
ttmetro wrote:
Wed Feb 17, 2021 1:14 am
BTW: no need to poll for interrupt driven events. The latter wake up the CPU, after handling the ISR it proceeds right from where it went to light sleep (presumably after WFI or something). If it decides nothing is ready it goes right back to sleep.
Yes, but the problem is multiplexing all the different wakeup sources and deciding what is ready. This is exactly what uevent will do -- asyncio polls _only_ uevent, and different ISR handlers (pin, timer, network, socket, etc) can signal a given uevent slot.

Re: nrf52: external file system

Posted: Wed Feb 17, 2021 10:42 pm
by ttmetro
jimmo wrote:
Wed Feb 17, 2021 3:08 am
... different ISR handlers (pin, timer, network, socket, etc) can signal a given uevent slot.
Great! I'm looking forward to this, especially together with low power BT (nrf?) this will be fantastic.