Daiki AMINAKA
Daiki AMINAKA
https://github.com/microsoft/msquic/blob/1edd42421a106d35381f4e0a93605860292026c0/src/test/lib/HandshakeTest.cpp#L2203-L2209 --- Expected case 1. Client executes ForceKeyUpdate (SetParam QUIC_PARAM_CONN_FORCE_KEY_UPDATE) 2. Server executes ForceKeyUpdate (SetParam QUIC_PARAM_CONN_FORCE_KEY_UPDATE) --- problematic case 1. Client ForceKeyUpdate 2. Client sends packet with key type of...
Found the cause, https://github.com/microsoft/msquic/blob/290e634880f24440b288b21db7342b1d6f0f52c9/src/test/lib/EventTest.cpp#LL534C1-L537C65 `Listener.Start` set address to `Binding->LocalAddress`, but `Listener->LocalAddress` is empty yet. This should happen not only for XDP, but also conventional datapath if they call `Listener.GetLocalAddr`. `Listener.GetLocalAddr`...
Maybe we are talking different topic? Problem is Listener->LocalAddress is not set appropriately (or is this expected?) even though Binding->LocalAddress is set. This makes GetParam for LocalAddress get empty address.
The call stack is > $6 0x00005555556ba239 in QuicAddrSetToDuoNic (Addr=0x7fffffffcfe0) at /home/daiki/workspace/msquic/src/test/lib/TestHelpers.h:36 > $7 QuicTestPrimeResumption (QuicAddrFamily=, Registration=..., ServerConfiguration=..., ClientConfiguration=..., ResumptionTicket=ResumptionTicket@entry=0x7fffffffd2d8) at /home/daiki/workspace/msquic/src/test/lib/HandshakeTest.cpp:80 > $8 0x00005555556aab99 in QuicTestValidateConnectionEvents3 (Registration=..., Listener=0x555555d36100, ServerLocalAddr=...)...
For some reason gcc-multilib is uninstalled at some point
What! the behavior differs with my local Ubuntu VM. Ubuntu VM can not find gcc-multilib, Ubuntu container cannot find ethertool
The Build Nuget Package job is gotten affected by different PR? https://github.com/microsoft/msquic/pull/3819
Still reproing. I did not touch yet
Next merging to https://github.com/microsoft/xdp-for-windows/pull/622 causes build time warning and missing trace. There is logical conflicts....  Then these trace doesn't appear during functional tests for both user and kernel test...
> Next merging to #622 causes build time warning and missing trace. There is logical conflicts.... > >  > > Then these trace doesn't appear during functional tests for...