There may be some state mismatch that drive to that?
I'm wondering why In protect picoquic_protect_packet, we use picoquic_aead_get_checksum_length instead of picoquic_get_checksum_length which protect picoquic_aead_get_checksum_length against
null cnx->crypto_context[epoch].aead_encrypt.
Sometime, I have the following crash due to the call of apicoquic_protect_packet called with a null aead_context.
Below the kind of stack we have:
In fact, we did noticed that
There may be some state mismatch that drive to that?
I'm wondering why In protect picoquic_protect_packet, we use picoquic_aead_get_checksum_length instead of picoquic_get_checksum_length which protect picoquic_aead_get_checksum_length against null cnx->crypto_context[epoch].aead_encrypt.