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
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
