[sandbox-systrace.c]
allow shutdown(2) syscall in sandbox - it may be called by packet_close()
from portable
(Id sync only; change is already in portable)
[sandbox-null.c sandbox-rlimit.c sandbox-seccomp-filter.c]
[sandbox-systrace.c ssh-sandbox.h sshd.c] Support preauth sandboxing
using the Capsicum API introduced in FreeBSD 10. Patch by Dag-Erling
Smorgrav, updated by Loganaden Velvindron @ AfriNIC; ok dtucker@
[ssh-agent.c clientloop.c misc.h packet.c progressmeter.c misc.c
channels.c sandbox-systrace.c]
Use clock_gettime(CLOCK_MONOTONIC ...) for ssh timers so that things like
keepalives and rekeying will work properly over clock steps. Suggested by
markus@, "looks good" djm@.
[sandbox-systrace.c sshd.c]
fix a during the load of the sandbox policies (child can still make
the read-syscall and wait forever for systrace-answers) by replacing
the read/write synchronisation with SIGSTOP/SIGCONT;
report and help hshoexer@; ok djm@, dtucker@
[sandbox-systrace.c]
Add mquery to the list of allowed syscalls for "UsePrivilegeSeparation
sandbox" since malloc now uses it. From johnw.mail at gmail com.
[sandbox-systrace.c]
fail open(2) with EPERM rather than SIGKILLing the whole process. libc
will call open() to do strerror() when NLS is enabled;
feedback and ok markus@
[sshd.c ssh-sandbox.h sandbox.h sandbox-rlimit.c sandbox-systrace.c]
[sandbox-null.c]
rename sandbox.h => ssh-sandbox.h to make things easier for portable
[servconf.c servconf.h sshd.c sshd_config.5 sandbox-rlimit.c]
[sandbox-systrace.c sandbox.h configure.ac Makefile.in]
introduce sandboxing of the pre-auth privsep child using systrace(4).
This introduces a new "UsePrivilegeSeparation=sandbox" option for
sshd_config that applies mandatory restrictions on the syscalls the
privsep child can perform. This prevents a compromised privsep child
from being used to attack other hosts (by opening sockets and proxying)
or probing local kernel attack surface.
The sandbox is implemented using systrace(4) in unsupervised "fast-path"
mode, where a list of permitted syscalls is supplied. Any syscall not
on the list results in SIGKILL being sent to the privsep child. Note
that this requires a kernel with the new SYSTR_POLICY_KILL option.
UsePrivilegeSeparation=sandbox will become the default in the future
so please start testing it now.
feedback dtucker@; ok markus@
Damien Miller
Darren Tucker