net, sunrpc: Remap EPERM in case of connection failure in xs_tcp_setup_socket
authorDaniel Borkmann <daniel@iogearbox.net>
Thu, 4 Jul 2024 06:41:57 +0000 (08:41 +0200)
committerPaolo Abeni <pabeni@redhat.com>
Thu, 11 Jul 2024 10:17:45 +0000 (12:17 +0200)
commit626dfed5fa3bfb41e0dffd796032b555b69f9cde
treed12cf6ffc238526a87a26deca2182788e38763b0
parent26488172b0292bed837b95a006a3f3431d1898c3
net, sunrpc: Remap EPERM in case of connection failure in xs_tcp_setup_socket

When using a BPF program on kernel_connect(), the call can return -EPERM. This
causes xs_tcp_setup_socket() to loop forever, filling up the syslog and causing
the kernel to potentially freeze up.

Neil suggested:

  This will propagate -EPERM up into other layers which might not be ready
  to handle it. It might be safer to map EPERM to an error we would be more
  likely to expect from the network system - such as ECONNREFUSED or ENETDOWN.

ECONNREFUSED as error seems reasonable. For programs setting a different error
can be out of reach (see handling in 4fbac77d2d09) in particular on kernels
which do not have f10d05966196 ("bpf: Make BPF_PROG_RUN_ARRAY return -err
instead of allow boolean"), thus given that it is better to simply remap for
consistent behavior. UDP does handle EPERM in xs_udp_send_request().

Fixes: d74bad4e74ee ("bpf: Hooks for sys_connect")
Fixes: 4fbac77d2d09 ("bpf: Hooks for sys_bind")
Co-developed-by: Lex Siegel <usiegl00@gmail.com>
Signed-off-by: Lex Siegel <usiegl00@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Neil Brown <neilb@suse.de>
Cc: Trond Myklebust <trondmy@kernel.org>
Cc: Anna Schumaker <anna@kernel.org>
Link: https://github.com/cilium/cilium/issues/33395
Link: https://lore.kernel.org/bpf/171374175513.12877.8993642908082014881@noble.neil.brown.name
Link: https://patch.msgid.link/9069ec1d59e4b2129fc23433349fd5580ad43921.1720075070.git.daniel@iogearbox.net
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
net/sunrpc/xprtsock.c