Once UE release after inactivity timer, will it send CSI report?

Hi Experts,
Once UE release after inactivity timer, will it send CSI report?
And: does NR supports aperiodic CQI also?

NR supports aperiodic CSI reporting.
csi = cqi, pmi, ri

Thanks @RFSpecialist.
Will it send cqi/pmi/ri/li once UE release after inactivity timer or during onduration only it reports csi?

Do you mean RRC inactivty timer here (which is there to compensate the possible RRC state mismatch between network and UE)?

Yes, I mean to say, can we verify, csi report once UE released after rrc inactivity?

CSI is reported only in connected mode.
UE need uplink dedicated resources to send the report.

Right… my understanding is, once inactivitytimer expires at UE, UE will consider that as RRC connection state mismatch, so releases the RRC connection, moves to dereg state and starts fresh attach.

Being able to send csi or not depends on the fact that whether UE is configured with CSI reporting and has the RRC connection maintained.

No, inactivity does not trigger deregistration.
UE will stay attach.
It will only release RRC connection.

Somehow, my understanding is different.
The below info is from “NR in bullets” which makes sense.
I’ll try to look for spec ref too.

The picture above is not from NR in bullets.

This is my notes which i keep whenever I find something interesting :blush:.
But the info is in fact from the bullets book. :slight_smile:

I have this same understanding.

I have this doubt sometime back: Why would UE trigger NAS registration again?

This may apply to SA.
But I don’t think it applies to NSA.

Yes, only SA.

Even for SA I doubt this part highlighted in yellow:

Only way it would is if there’s uplink data pending.

Why every data inactivity timer would trigger registration?
It makes no sense.

Yes, exactly.

From 38.331

5.3.8.5 UE actions upon the expiry of DataInactivityTimer

Upon receiving the expiry of DataInactivityTimer from lower layers while in RRC_CONNECTED, the UE shall:

1> perform the actions upon going to RRC_IDLE as specified in 5.3.11, with release cause ‘RRC connection failure’

This is for SA and is for the case when UE keeps RRC connection longer than gNodeB.
And will inform of RRC connection failure.
But it is weird why it needs to perform registration for each RRC connection failure.

Like when you drop a call you need to perform registration?