OsmoSGSN - "Cannot handle SM for unknown MM CTX"

Hello, I got question about this issue. I updated my comment on this issue here: https://osmocom.org/issues/6197 Original my post: I observe same situation in my setup. Cannot activate PDP context because of missing RA-CAPABILTIY-UPDATE and RA-CAPABILITY-ACK support in osmoSGSN. My setup is Samsung Galaxy S9 and S23, real BTS/BSC from well-known vendor and Osmo core (SGSN/GGSN/HLR/MSC/STP) (Important thing that I'm not using osmo-pcu). Problem description: S9 attach procedure (CS+PS): 1.After UE reboot I see MM procedure Location request and Location update processed good (on DMtool). Then if needed CS calling is working properly(connection to osmoMSC). 2.After that UE is trying GMM Attach request and it got Attach accept and attach complete. 3.Then UE send Activate PDP context, but SGSN is reject this with cause: "Cannot handle SM for unknown MM CTX" 4.After that UE is trying activate PDP context for second time (After timer expired) and finally it got pdp accept (ping/traffic work) same as in this thread. After deep inspection I found that reason of first pdp activation reject is that TLLI is not updated (SGSN dont know it and causing Cannot handle SM for unknown MM CTX) because when I checked pcap from bssgp I see that BSC send RA-CAPABILTIY-UPDATE with new tlli and osmoSGSN send response Unknown PDU with protocol unspecified but should send RA-CAPABILTIY-UPDATE-ACK. Other thing I dont understand that in msg comes from bsc on bvci 127 but respond goes to bvci 0 (but I didnt define it as bsc-sgsn link use only bvci 127 not bvci 0).Also I think sig bvci 0 is not changeable to bvci 127 on osmoSGSN. To summarize attach procedure is long it takes 30-50 seconds(because of this first reject). S23 attach Same as S9 but after first fail it not doing second attempt and stuck without PS forever in CS domain. Its not problem with MS stack. I tried a lot of workarounds to make s23 attach but no success. Changing to PSonly not work, AT commands not work. Also tried with Huawei E3372,E3131 but same behaviour. I connect this setup to other 2Gcore-simulator(not opensource) to compare and its processed normally for both UEs and modems,it got pdp contexts after 6-8 seconds. Difference is that here I got message RA-CAPABILTIY-UPDATE/RA-CAPABILITY-ACK because got real BSC not osmo-pcu. When I look into code I cannot see RA-CAPABILTIY-UPDATE and RA-CAPABILITY-ACK support so my questions are: 1.Is this RA-CAPABILTIY-UPDATE and RA-CAPABILITY-ACK supported in osmoSGSN and is it plan to do that? 2.Is any other workaround (changing timers,tmsi assignment or whatever to change in bsc) in configuration to make S23/modems attached?Theoretically if UE can set PSonly and AT command to activate PDP context it should work but I never make it happened this AT commands on my terminals. 3.I can provide both pcaps and UE logs to see whats missing if it helps to make add it. 4.Is it possible to add this two messages to support to osmoSGSN code?I can test patch if you provide it. Regards, Jan Kosiński