Attention is currently required from: fixeria.
pespin has posted comments on this change by pespin. ( https://gerrit.osmocom.org/c/osmo-bts/+/41807?usp=email )
Change subject: {bs,ms}_power_control: Move params inside lchan_power_ctrl_state ......................................................................
Patch Set 3:
(1 comment)
Patchset:
PS2:
So far I had a quick look and I don't really understand/see what do we benefit from this refactoring.
As mentioned in the commit message, these params follow the same lifecycle than that the state, ie. they get per-lchan modified over RSL. By having it inside the struct, we reset the whole state together, so the logic is simplified.
And using the pointer allows to avoid memcpy()ing parameters which are identical for all lchans in 99% cases.
You are still holding a copy on each lchan. All I'm doing is getting rid of one extra pointer field.
I don't see how this pointer being NULL can crate problems in static mode: one shall simply never need to access dynamic power control parameters in static mode.
Yes, that's usually how crashes happen: One ends up doing something should never need to do 😊