syzkaller: divide error in __tcp_select_window
#414
Comments
matttbe
pushed a commit
that referenced
this issue
Jun 29, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jun 29, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jun 30, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jun 30, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jun 30, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jul 2, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jul 2, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jul 3, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jul 3, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jul 3, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jul 3, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jul 4, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
jenkins-tessares
pushed a commit
that referenced
this issue
Jul 4, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jul 4, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jul 4, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
matttbe
pushed a commit
that referenced
this issue
Jul 5, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: #414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
kernel-patches-daemon-bpf bot
pushed a commit
to kernel-patches/bpf
that referenced
this issue
Jul 5, 2023
Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net>
gregkh
pushed a commit
to gregkh/linux
that referenced
this issue
Jul 23, 2023
commit 0226436 upstream. Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mj22226
pushed a commit
to mj22226/linux
that referenced
this issue
Aug 1, 2023
[ Upstream commit 0226436 ] Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
mj22226
pushed a commit
to mj22226/linux
that referenced
this issue
Aug 2, 2023
[ Upstream commit 0226436 ] Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Whissi
pushed a commit
to Whissi/linux-stable
that referenced
this issue
Aug 3, 2023
[ Upstream commit 0226436 ] Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
sileshn
pushed a commit
to sileshn/ubuntu-kernel-lunar
that referenced
this issue
Sep 21, 2023
BugLink: https://bugs.launchpad.net/bugs/2036075 commit 0226436acf2495cde4b93e7400e5a87305c26054 upstream. Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
tuxedo-bot
pushed a commit
to tuxedocomputers/linux
that referenced
this issue
Oct 4, 2023
BugLink: https://bugs.launchpad.net/bugs/2036075 commit 0226436 upstream. Since the blamed commit, closing the first subflow resets the first subflow socket state to SS_UNCONNECTED. The current mptcp listen implementation relies only on such state to prevent touching not-fully-disconnected sockets. Incoming mptcp fastclose (or paired endpoint removal) unconditionally closes the first subflow. All the above allows an incoming fastclose followed by a listen() call to successfully race with a blocking recvmsg(), potentially causing the latter to hit a divide by zero bug in cleanup_rbuf/__tcp_select_window(). Address the issue explicitly checking the msk socket state in mptcp_listen(). An alternative solution would be moving the first subflow socket state update into mptcp_disconnect(), but in the long term the first subflow socket should be removed: better avoid relaying on it for internal consistency check. Fixes: b29fcfb ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Reported-by: Christoph Paasch <cpaasch@apple.com> Closes: multipath-tcp/mptcp_net-next#414 Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
syzkaller-id: e4d528e2abf5f9ab3d895e5021c40a119902b080
HEAD: 74a1b71
Trace:
Kconfig:
Kconfig_k9_kasan.txt
No reproducers.
The text was updated successfully, but these errors were encountered: