On 1/27/22 02:44, Jinoh Kang wrote:
What we need to do here is similar to the infrastructure that already exists for device asyncs, namely "unknown_status" etc. It would be nice to use that instead of reinventing it, and although I haven't tried, it seems possible.
That one was on the table, too. In fact it can also help eliminate the initial_status == STATUS_ALERTED check.
One catch is that async_set_unknown_status also sets direct_result to 0, which means to always fire off APC on completion. I wasn't entirely sure of what the effects of { .unknown_status = 1, .direct_result = 1 } would be.
I believe that !direct_result is correct here, actually. Note that we only get an APC when calling async_terminate(), but we shouldn't be introducing any extra terminate calls.
async_add_queue() as it is above is not great. I'm not sure that code actually works in every case;
!pending && terminated && alerted was the condition I was able to deduce to detect this exact condition. It does sound a little arbitrary though, especially since it's testing for three unrelated conditions.
it definitely increases the mental burden even if it does. (Consider for instance that it would be triggered for *every* async).
Instead what I'd suggest is to use the request introduced here in every case, even if the initial status was pending.
You mean, along with use of unknown_status?
In a sense they're orthogonal, but yes, I think doing both would help.