A stack-based buffer overflow in res_rtp_asterisk.c in Sangoma Asterisk before 16.16.1, 17.x before 17.9.2, and 18.x before 18.2.1 and Certified Asterisk before 16.8-cert6 allows an authenticated WebRTC client to cause an Asterisk crash by sending multiple hold/unhold requests in quick succession. This is caused by a signedness comparison mismatch.

Incorrect access controls in res_srtp.c in Sangoma Asterisk 13.38.1, 16.16.0, 17.9.1, and 18.2.0 and Certified Asterisk 16.8-cert5 allow a remote unauthenticated attacker to prematurely terminate secure calls by replaying SRTP packets.

An issue was discovered in Sangoma Asterisk 16.x before 16.16.1, 17.x before 17.9.2, and 18.x before 18.2.1 and Certified Asterisk before 16.8-cert6. When re-negotiating for T.38, if the initial remote response was delayed just enough, Asterisk would send both audio and T.38 in the SDP. If this happened, and the remote responded with a declined T.38 stream, then Asterisk would crash.

An issue was discovered in res_pjsip_session.c in Digium Asterisk through 13.38.1; 14.x, 15.x, and 16.x through 16.16.0; 17.x through 17.9.1; and 18.x through 18.2.0, and Certified Asterisk through 16.8-cert5. An SDP negotiation vulnerability in PJSIP allows a remote server to potentially crash Asterisk by sending specific SIP responses that cause an SDP negotiation failure.

A buffer overflow in res_pjsip_diversion.c in Sangoma Asterisk versions 13.38.1, 16.15.1, 17.9.1, and 18.1.1 allows remote attacker to crash Asterisk by deliberately misusing SIP 181 responses.

An issue was discovered in res_pjsip_diversion.c in Sangoma Asterisk before 13.38.0, 14.x through 16.x before 16.15.0, 17.x before 17.9.0, and 18.x before 18.1.0. A crash can occur when a SIP message is received with a History-Info header that contains a tel-uri, or when a SIP 181 response is received that contains a tel-uri in the Diversion header.

A res_pjsip_session crash was discovered in Asterisk Open Source 13.x before 13.37.1, 16.x before 16.14.1, 17.x before 17.8.1, and 18.x before 18.0.1. and Certified Asterisk before 16.8-cert5. Upon receiving a new SIP Invite, Asterisk did not return the created dialog locked or referenced. This caused a gap between the creation of the dialog object, and its next use by the thread that created it. Depending on some off-nominal circumstances and timing, it was possible for another thread to free said dialog in this gap. Asterisk could then crash when the dialog object, or any of its dependent objects, were dereferenced or accessed next by the initial-creation thread. Note, however, that this crash can only occur when using a connection-oriented protocol (e.g., TCP or TLS, but not UDP) for SIP transport. Also, the remote client must be authenticated, or Asterisk must be configured for anonymous calling.

An issue was discovered in Asterisk Open Source 13.x before 13.37.1, 16.x before 16.14.1, 17.x before 17.8.1, and 18.x before 18.0.1 and Certified Asterisk before 16.8-cert5. If Asterisk is challenged on an outbound INVITE and the nonce is changed in each response, Asterisk will continually send INVITEs in a loop. This causes Asterisk to consume more and more memory since the transaction will never terminate (even if the call is hung up), ultimately leading to a restart or shutdown of Asterisk. Outbound authentication must be configured on the endpoint for this to occur.

An issue was discovered in manager.c in Sangoma Asterisk through 13.x, 16.x, 17.x and Certified Asterisk 13.21 through 13.21-cert4. A remote authenticated Asterisk Manager Interface (AMI) user without system authorization could use a specially crafted Originate AMI request to execute arbitrary system commands.

An issue was discovered in channels/chan_sip.c in Sangoma Asterisk 13.x before 13.29.2, 16.x before 16.6.2, and 17.x before 17.0.1, and Certified Asterisk 13.21 before cert5. A SIP request can be sent to Asterisk that can change a SIP peer's IP address. A REGISTER does not need to occur, and calls can be hijacked as a result. The only thing that needs to be known is the peer's name; authentication details such as passwords do not need to be known. This vulnerability is only exploitable when the nat option is set to the default, or auto_force_rport.