Attention is currently required from: osmith, pespin.

dexter has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/34192 )

Change subject: pcuif_proto: check confirm flag in struct gsm_pcu_if_pch ......................................................................

Patch Set 3:

(5 comments)

Commit Message:

https://gerrit.osmocom.org/c/osmo-bsc/+/34192/comment/05323a65_ba1da355 PS2, Line 11: PCH, we have not the same flexibility as we have it in osmo-bsc.

you mean osmo-bts here?

Done

https://gerrit.osmocom.org/c/osmo-bsc/+/34192/comment/97db3a6f_3252b292 PS2, Line 14: osmo-bts (cofirm PAGING messages, send non confirmed IMMEDIATE

confirm

Done

File src/osmo-bsc/pcu_sock.c:

https://gerrit.osmocom.org/c/osmo-bsc/+/34192/comment/4a5c72c2_f079adc7 PS2, Line 559: if (!pch->confirm)

why are you making so many suppositions here? Just do whatever the request asks you to do. […]

The problem is that we can not really chose here if we want a confirmation or not. The Ericsson RBS BTS will send the IMM ASS send message whether we want it or not. And this will generate a confirmation.

There is only one thing we could do, I am not even sure if this works or not but in case no conformation is requested, we could leave the propritary Mobile-ID (that is how Ericssin calls it) out. Then there may be no IMM ASS sent. I didn't try this as I thought that this is a situation that is not relevant in practice anyway.

As far as I understand, you would use the pch->confirm flag to decide if the message has to go through pcu_rx_rr_imm_ass_pch() or pcu_rx_rr_paging_pch(). It would work, but it would logically turn the "confirm" flag into a "is_imm_ass" flag, which I think may be confusing. That is basically why I added this extra logic + error messages.

What do you think?

https://gerrit.osmocom.org/c/osmo-bsc/+/34192/comment/a3b40f5b_32417af2 PS2, Line 561: "PCU sends IMMEDIATE ASSIGNMENT via PCH but requests no conformation, will send confirmation anyway!\n");

confirmation

Done

https://gerrit.osmocom.org/c/osmo-bsc/+/34192/comment/16f42202_e6d9ce85 PS2, Line 566: if (pch->confirm)

this also should go out.

In theory osmo-bts would be capable to confirm paging MAC blocks. In osmo-bsc have this since it would require some RSL message from the BTS. This however, is nor specified, nor is it needed. Someone may still try to get a confirmation. I think we should have an error message in this case to raise some awareness.

What we could do though is to send the confirmation as we receive the paging message (right here). But I do not think that this has any practical value, other then it makes the behavior somewhat more coherent to that of osmo-bts.