thr3ads.net - openssh bugs - [Bug 3855] New: sshd-auth sandbox [Aug 2025]

If this information is useful, please help other people find it:
Share via:

bugzilla-daemon at mindrot.org

2025-Aug-22 07:37 UTC

[Bug 3855] New: sshd-auth sandbox

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

            Bug ID: 3855
           Summary: sshd-auth sandbox
           Product: Portable OpenSSH
           Version: 10.0p2
          Hardware: ARM
                OS: Linux
            Status: NEW
          Severity: major
          Priority: P5
         Component: sshd
          Assignee: unassigned-bugs at mindrot.org
          Reporter: adrian.jarc at aviatnet.com

We have recently investigated an issue where once we updated to
OpenSSH10.0p2 our integration using wolfProvider
(https://github.com/wolfSSL/wolfProvider) for OpenSSL (3.0.12) and
wolfSSL (built in FIPS mode) didn't want to authenticate user.

During the investigation we have found out that wolfSSLs wolfCrypt
module(FIPS140-3 certified) opens file descriptors for /dev/urandom (or
/dev/random if there is no urandom) when generating random seed for
running KEX algorithms (in our case:
diffie-hellman-group-exchange-sha256,diffie-hellman-group14-sha1,diffie-hellman-group14-sha256,diffie-hellman-group16-sha512,diffie-hellman-group18-sha512,ecdh-sha2-nistp256,ecdh-sha2-nistp384,ecdh-sha2-nistp521).
This function stopped working once we upgraded from 9.9p2 to 10.0p2. We
investigated and found out that seed generation tries to open file
descriptor from the sshd-auth module, which is sandboxed before doing
anything, and because of that 4th attempt to open FD fails.

So we now made a "patch" that changes lines
https://github.com/openssh/openssh-portable/blob/V_10_0/sshd-auth.c#L787-L792
from:
"
        privsep_child_demote();

        /* perform the key exchange */
        /* authenticate user and start session */
        do_ssh2_kex(ssh);
        do_authentication2(ssh);
"
to:
"
        /* perform the key exchange */
        /* authenticate user and start session */
        do_ssh2_kex(ssh);
        do_authentication2(ssh);

        privsep_child_demote();
".
Now we are wondering what security issues this might cause, and if
there is maybe a better way to fix those limitations that are imposed.
WolfSSL team is also investigating this issue but in their case, they
were unable to reproduce. In our case it was always successful.

WolfSSL also can't easily change their wolfCrypt module, as that would
mean they need to recertify everything (this takes a while).

If we noticed correctly in our case once privsep_child_demote()
function calls ssh_sandbox_child(box). It executes the implementation
in:
https://github.com/openssh/openssh-portable/blob/V_10_0/sandbox-rlimit.c.


System info:
uname -a command returns: Linux WTM4800 5.4.134-iproc #1 PREEMPT Tue
Jan 25 01:02:13 UTC 2022 armv7l armv7l armv7l GNU/Linux

This Linux is a custom Yocto build for our devices.

OpenSSH is build with configuration from Yocto repository:
https://git.yoctoproject.org/poky/tree/meta/recipes-connectivity/openssh

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-22 07:38 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

Adrian Jarc <adrian.jarc at aviatnet.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|sshd-auth sandbox           |sshd-auth sandbox
                   |                            |limitations

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-22 07:47 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #1 from Adrian Jarc <adrian.jarc at aviatnet.com> ---
Sorry, it looks like I added the wrong sandbox implementation file. It
seems like it is this one:
https://github.com/openssh/openssh-portable/blob/V_10_0/sandbox-seccomp-filter.c

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-28 23:49 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org

--- Comment #2 from Damien Miller <djm at mindrot.org> ---
Moving the privsep_child_demote() to after do_authentication2() not
only disables the sandbox (sshd-auth exits after authentication
completes), but disables all privilege-reduction that it performs.
You're much better off just disabling sandboxing at compile time.

Wrt enabling opening /dev/random, unfortunately it's not possible using
the seccomp sandbox without also allowing open() of any file. The
problem is that the seccomp bpf filters cannot inspect pointer
arguments, including file paths. Therefore the only option would be to
allow all __NR_open syscalls, which would significantly weaken the
sandbox.

AFAIK all other libcrypto libraries have long since moved to use the
getrandom(2) syscall which is much easier to allowlist. AFAIK
BoringSSL's FIPS libcrypto uses getrandom(2).

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-28 23:55 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #3 from Damien Miller <djm at mindrot.org> ---
Some other alternatives:

1. Ask the WolfSSL developers if there is any way to get the library to
preopen the file descriptors before the sandbox is applied.
2. Soft-deny all __NR_open syscalls in the sandbox. This will case
open() to faill with an error but won't result in a process-killing
sandbox violation. You'd need to get a guarantee from the WolfSSL
developers that their library will perform safely in this situation.

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-29 08:51 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #4 from Adrian Jarc <adrian.jarc at aviatnet.com> ---
(In reply to Damien Miller from comment #2)> Moving the privsep_child_demote() to after do_authentication2() not
> only disables the sandbox (sshd-auth exits after authentication
> completes), but disables all privilege-reduction that it performs.
> You're much better off just disabling sandboxing at compile time.
> 
> Wrt enabling opening /dev/random, unfortunately it's not possible
> using the seccomp sandbox without also allowing open() of any file.
> The problem is that the seccomp bpf filters cannot inspect pointer
> arguments, including file paths. Therefore the only option would be
> to allow all __NR_open syscalls, which would significantly weaken
> the sandbox.
> 
> AFAIK all other libcrypto libraries have long since moved to use the
> getrandom(2) syscall which is much easier to allowlist. AFAIK
> BoringSSL's FIPS libcrypto uses getrandom(2).
As far as BoringSSL goes, it is not certified for our processor. And
also we have made some further investigation and found that if we move
privsep_child_demote() after do_ssh2_kex() also works, as the problem
was in kex authentication. Is this any better?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-29 08:53 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #5 from Adrian Jarc <adrian.jarc at aviatnet.com> ---
(In reply to Damien Miller from comment #3)> Some other alternatives:
> 
> 1. Ask the WolfSSL developers if there is any way to get the library
> to preopen the file descriptors before the sandbox is applied.
> 2. Soft-deny all __NR_open syscalls in the sandbox. This will case
> open() to faill with an error but won't result in a process-killing
> sandbox violation. You'd need to get a guarantee from the WolfSSL
> developers that their library will perform safely in this situation.
If WolfSSL changes how that works, their wolfCrypt module won't be FIPS
certified anymore, and that does not help. So this is not an option.

As for 2. point, can we get some pointers as how we could do that?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Aug-31 23:39 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #6 from Damien Miller <djm at mindrot.org> ---
(In reply to Adrian Jarc from comment #5)
> If WolfSSL changes how that works, their wolfCrypt module won't be
> FIPS certified anymore, and that does not help. So this is not an
> option.
No, what I mean is asking if there are any existing WolfSSL API calls
that can be made to prepare it for sandboxing.
> As for 2. point, can we get some pointers as how we could do that?
You'll need to identify the syscall that is failing. If you build
OpenSSH with the SANDBOX_SECCOMP_FILTER_DEBUG define set in
sandbox-seccomp-filter.c (don't use this in production) you'll get an
error message including the syscall number.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Sep-01 07:10 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #7 from Adrian Jarc <adrian.jarc at aviatnet.com> ---
(In reply to Damien Miller from comment #6)> (In reply to Adrian Jarc from comment #5)
> 
> > If WolfSSL changes how that works, their wolfCrypt module won't be
> > FIPS certified anymore, and that does not help. So this is not an
> > option.
> 
> No, what I mean is asking if there are any existing WolfSSL API
> calls that can be made to prepare it for sandboxing.
> 
> > As for 2. point, can we get some pointers as how we could do that?
> 
> You'll need to identify the syscall that is failing. If you build
> OpenSSH with the SANDBOX_SECCOMP_FILTER_DEBUG define set in
> sandbox-seccomp-filter.c (don't use this in production) you'll get
> an error message including the syscall number.
I have enabled debug logging, and this is the debug log I get on server
where it fails:

"
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug3: fd 8 is not O_NONBLOCK
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug2: fd 9 setting O_NONBLOCK
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug1: Forked child 5500.
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.5500
debug3: oom_adjust_restore
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.5500
debug1: Set /proc/self/oom_score_adj to 0
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.5500
debug1: rexec start in 8 out 8 newsock 8 config_s 9/10
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug2: server_accept_loop: child 5500 for connection from <client-ip>
to <server-ip> received config
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: network sockets: 7, 7
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: server_process_channel_timeouts: setting 0
timeouts
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: channel_clear_timeouts: clearing
2025-09-01T09:00:58.519+00:00 Chassis1 auth.info WTM4800
sshd-session.5500 Connection from <client-ip> port 34166 on
<server-ip>
port 22 rdomain ""
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: Local version string SSH-2.0-OpenSSH_10.0
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: Remote protocol version 2.0, remote software
version OpenSSH_9.6p1 Ubuntu-3ubuntu13.13
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: compat_banner: match: OpenSSH_9.6p1
Ubuntu-3ubuntu13.13 pat OpenSSH* compat 0x04000000
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: fd 7 setting O_NONBLOCK
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: Network child is on pid 5502
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: preauth child monitor started
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_receive: entering
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: monitor_read: checking request 51
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: mm_answer_state: config len 3640
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_send: entering, type 52
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_answer_state: done
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: monitor_read: 51 used once, disabling now
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: server_process_channel_timeouts: setting 0
timeouts [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: channel_clear_timeouts: clearing [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: fd 5 is O_NONBLOCK [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_init: preparing seccomp filter
sandbox [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: privsep user:group 997:996 [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: permanently_set_uid: 997/996 [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_child_debugging: installing
SIGSYS handler [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_child: setting
PR_SET_NO_NEW_PRIVS [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_child: attaching seccomp filter
program [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: ssh_sandbox_child: prctl(PR_SET_SECCOMP):
Invalid argument [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: append_hostkey_type: ssh-rsa key not
permitted by HostkeyAlgorithms [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: list_hostkey_types:
rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: send packet: type 20 [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: SSH2_MSG_KEXINIT sent [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: receive packet: type 20 [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: SSH2_MSG_KEXINIT received [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: local server KEXINIT proposal [preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: KEX algorithms:
diffie-hellman-group-exchange-sha256,diffie-hellman-group14-sha1,diffie-hellman-group14-sha256,diffie-hellman-group16-sha512,diffie-hellman-group18-sha512,ecdh-sha2-nistp256,ecdh-sha2-nistp384,ecdh-sha2-nistp521,ext-info-s,kex-strict-s-v00
at openssh.com
[preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: host key algorithms:
rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers ctos:
aes128-ctr,aes128-gcm at openssh.com,aes256-ctr,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers stoc:
aes128-ctr,aes128-gcm at openssh.com,aes256-ctr,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs ctos: hmac-sha2-256,hmac-sha2-512
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs stoc: hmac-sha2-256,hmac-sha2-512
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression ctos: none,zlib at openssh.com
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression stoc: none,zlib at openssh.com
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages ctos:  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages stoc:  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: first_kex_follows 0  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: reserved 0  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: peer client KEXINIT proposal [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: KEX algorithms:
sntrup761x25519-sha512 at openssh.com,curve25519-sha256,curve25519-sha256 at
libssh.org,ecdh-sha2-nistp256,ecdh-sha2-nistp384,ecdh-sha2-nistp521,diffie-hellman-group-exchange-sha256,diffie-hellman-group16-sha512,diffie-hellman-group18-sha512,diffie-hellman-group14-sha256,ext-info-c,kex-strict-c-v00
at openssh.com
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: host key algorithms:
ssh-ed25519-cert-v01 at openssh.com,ecdsa-sha2-nistp256-cert-v01 at
openssh.com,ecdsa-sha2-nistp384-cert-v01 at
openssh.com,ecdsa-sha2-nistp521-cert-v01 at openssh.com,sk-ssh-ed25519-cert-v01
at openssh.com,sk-ecdsa-sha2-nistp256-cert-v01 at
openssh.com,rsa-sha2-512-cert-v01 at openssh.com,rsa-sha2-256-cert-v01 at
openssh.com,ssh-ed25519,ecdsa-sha2-nistp256,ecdsa-sha2-nistp384,ecdsa-sha2-nistp521,sk-ssh-ed25519
at openssh.com,sk-ecdsa-sha2-nistp256 at openssh.com,rsa-sha2-512,rsa-sha2-256
[preaut
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers ctos:
chacha20-poly1305 at openssh.com,aes128-ctr,aes192-ctr,aes256-ctr,aes128-gcm at
openssh.com,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers stoc:
chacha20-poly1305 at openssh.com,aes128-ctr,aes192-ctr,aes256-ctr,aes128-gcm at
openssh.com,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs ctos:
umac-64-etm at openssh.com,umac-128-etm at openssh.com,hmac-sha2-256-etm at
openssh.com,hmac-sha2-512-etm at openssh.com,hmac-sha1-etm at
openssh.com,umac-64 at openssh.com,umac-128 at
openssh.com,hmac-sha2-256,hmac-sha2-512,hmac-sha1
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs stoc:
umac-64-etm at openssh.com,umac-128-etm at openssh.com,hmac-sha2-256-etm at
openssh.com,hmac-sha2-512-etm at openssh.com,hmac-sha1-etm at
openssh.com,umac-64 at openssh.com,umac-128 at
openssh.com,hmac-sha2-256,hmac-sha2-512,hmac-sha1
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression ctos: none,zlib at openssh.com,zlib
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression stoc: none,zlib at openssh.com,zlib
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages ctos:  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages stoc:  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: first_kex_follows 0  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: reserved 0  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: kex_choose_conf: will use strict KEX ordering
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: algorithm: ecdh-sha2-nistp256 [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: host key algorithm: ecdsa-sha2-nistp256
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: client->server cipher: aes128-ctr MAC:
hmac-sha2-256 compression: none [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: server->client cipher: aes128-ctr MAC:
hmac-sha2-256 compression: none [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: expecting SSH2_MSG_KEX_ECDH_INIT [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: receive packet: type 30 [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: SSH2_MSG_KEX_ECDH_INIT received [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.info WTM4800
sshd-session.5500 ssh_dispatch_run_fatal: Connection from <client-ip>
port 34166: error in libcrypto [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: monitor_read_log: child log fd closed
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_receive: entering
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_receive: monitor fd closed
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: mm_reap: preauth child exited with status 255
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: do_cleanup
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: PAM: sshpam_thread_cleanup entering
2025-09-01T09:00:58.641+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: Killing privsep child 5502
2025-09-01T09:00:58.645+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug1: child_close: enter
2025-09-01T09:00:58.645+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug1: child_reap: preauth child 5500 for connection from <client-ip>
to <server-ip> exited with status 255
2025-09-01T09:00:58.645+00:00 Chassis1 auth.info WTM4800 sshd.29539
srclimit_penalise: ipv4: new <client-ip>/32 deferred penalty of 1
seconds for penalty: connections without attempting authentication
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug3: fd 8 is not O_NONBLOCK
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug2: fd 9 setting O_NONBLOCK
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug1: Forked child 5500.
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.5500
debug3: oom_adjust_restore
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.5500
debug1: Set /proc/self/oom_score_adj to 0
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.5500
debug1: rexec start in 8 out 8 newsock 8 config_s 9/10
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug2: server_accept_loop: child 5500 for connection from <client-ip>
to <server-ip> received config
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: network sockets: 7, 7
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: server_process_channel_timeouts: setting 0
timeouts
2025-09-01T09:00:58.493+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: channel_clear_timeouts: clearing
2025-09-01T09:00:58.519+00:00 Chassis1 auth.info WTM4800
sshd-session.5500 Connection from <client-ip> port 34166 on
<server-ip>
port 22 rdomain ""
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: Local version string SSH-2.0-OpenSSH_10.0
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: Remote protocol version 2.0, remote software
version OpenSSH_9.6p1 Ubuntu-3ubuntu13.13
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: compat_banner: match: OpenSSH_9.6p1
Ubuntu-3ubuntu13.13 pat OpenSSH* compat 0x04000000
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: fd 7 setting O_NONBLOCK
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: Network child is on pid 5502
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: preauth child monitor started
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_receive: entering
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: monitor_read: checking request 51
2025-09-01T09:00:58.519+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: mm_answer_state: config len 3640
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_send: entering, type 52
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_answer_state: done
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: monitor_read: 51 used once, disabling now
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: server_process_channel_timeouts: setting 0
timeouts [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: channel_clear_timeouts: clearing [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: fd 5 is O_NONBLOCK [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_init: preparing seccomp filter
sandbox [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: privsep user:group 997:996 [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: permanently_set_uid: 997/996 [preauth]
2025-09-01T09:00:58.561+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_child_debugging: installing
SIGSYS handler [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_child: setting
PR_SET_NO_NEW_PRIVS [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: ssh_sandbox_child: attaching seccomp filter
program [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: ssh_sandbox_child: prctl(PR_SET_SECCOMP):
Invalid argument [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: append_hostkey_type: ssh-rsa key not
permitted by HostkeyAlgorithms [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: list_hostkey_types:
rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: send packet: type 20 [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: SSH2_MSG_KEXINIT sent [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: receive packet: type 20 [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: SSH2_MSG_KEXINIT received [preauth]
2025-09-01T09:00:58.565+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: local server KEXINIT proposal [preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: KEX algorithms:
diffie-hellman-group-exchange-sha256,diffie-hellman-group14-sha1,diffie-hellman-group14-sha256,diffie-hellman-group16-sha512,diffie-hellman-group18-sha512,ecdh-sha2-nistp256,ecdh-sha2-nistp384,ecdh-sha2-nistp521,ext-info-s,kex-strict-s-v00
at openssh.com
[preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: host key algorithms:
rsa-sha2-512,rsa-sha2-256,ecdsa-sha2-nistp256 [preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers ctos:
aes128-ctr,aes128-gcm at openssh.com,aes256-ctr,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers stoc:
aes128-ctr,aes128-gcm at openssh.com,aes256-ctr,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.588+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs ctos: hmac-sha2-256,hmac-sha2-512
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs stoc: hmac-sha2-256,hmac-sha2-512
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression ctos: none,zlib at openssh.com
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression stoc: none,zlib at openssh.com
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages ctos:  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages stoc:  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: first_kex_follows 0  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: reserved 0  [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: peer client KEXINIT proposal [preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: KEX algorithms:
sntrup761x25519-sha512 at openssh.com,curve25519-sha256,curve25519-sha256 at
libssh.org,ecdh-sha2-nistp256,ecdh-sha2-nistp384,ecdh-sha2-nistp521,diffie-hellman-group-exchange-sha256,diffie-hellman-group16-sha512,diffie-hellman-group18-sha512,diffie-hellman-group14-sha256,ext-info-c,kex-strict-c-v00
at openssh.com
[preauth]
2025-09-01T09:00:58.607+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: host key algorithms:
ssh-ed25519-cert-v01 at openssh.com,ecdsa-sha2-nistp256-cert-v01 at
openssh.com,ecdsa-sha2-nistp384-cert-v01 at
openssh.com,ecdsa-sha2-nistp521-cert-v01 at openssh.com,sk-ssh-ed25519-cert-v01
at openssh.com,sk-ecdsa-sha2-nistp256-cert-v01 at
openssh.com,rsa-sha2-512-cert-v01 at openssh.com,rsa-sha2-256-cert-v01 at
openssh.com,ssh-ed25519,ecdsa-sha2-nistp256,ecdsa-sha2-nistp384,ecdsa-sha2-nistp521,sk-ssh-ed25519
at openssh.com,sk-ecdsa-sha2-nistp256 at openssh.com,rsa-sha2-512,rsa-sha2-256
[preaut
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers ctos:
chacha20-poly1305 at openssh.com,aes128-ctr,aes192-ctr,aes256-ctr,aes128-gcm at
openssh.com,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: ciphers stoc:
chacha20-poly1305 at openssh.com,aes128-ctr,aes192-ctr,aes256-ctr,aes128-gcm at
openssh.com,aes256-gcm at openssh.com
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs ctos:
umac-64-etm at openssh.com,umac-128-etm at openssh.com,hmac-sha2-256-etm at
openssh.com,hmac-sha2-512-etm at openssh.com,hmac-sha1-etm at
openssh.com,umac-64 at openssh.com,umac-128 at
openssh.com,hmac-sha2-256,hmac-sha2-512,hmac-sha1
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: MACs stoc:
umac-64-etm at openssh.com,umac-128-etm at openssh.com,hmac-sha2-256-etm at
openssh.com,hmac-sha2-512-etm at openssh.com,hmac-sha1-etm at
openssh.com,umac-64 at openssh.com,umac-128 at
openssh.com,hmac-sha2-256,hmac-sha2-512,hmac-sha1
[preauth]
2025-09-01T09:00:58.614+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression ctos: none,zlib at openssh.com,zlib
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: compression stoc: none,zlib at openssh.com,zlib
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages ctos:  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: languages stoc:  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: first_kex_follows 0  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug2: reserved 0  [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: kex_choose_conf: will use strict KEX ordering
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: algorithm: ecdh-sha2-nistp256 [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: host key algorithm: ecdsa-sha2-nistp256
[preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: client->server cipher: aes128-ctr MAC:
hmac-sha2-256 compression: none [preauth]
2025-09-01T09:00:58.632+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: kex: server->client cipher: aes128-ctr MAC:
hmac-sha2-256 compression: none [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: expecting SSH2_MSG_KEX_ECDH_INIT [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: receive packet: type 30 [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: SSH2_MSG_KEX_ECDH_INIT received [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.info WTM4800
sshd-session.5500 ssh_dispatch_run_fatal: Connection from <client-ip>
port 34166: error in libcrypto [preauth]
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: monitor_read_log: child log fd closed
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_receive: entering
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: mm_request_receive: monitor fd closed
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: mm_reap: preauth child exited with status 255
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: do_cleanup
2025-09-01T09:00:58.635+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug3: PAM: sshpam_thread_cleanup entering
2025-09-01T09:00:58.641+00:00 Chassis1 auth.debug WTM4800
sshd-session.5500 debug1: Killing privsep child 5502
2025-09-01T09:00:58.645+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug1: child_close: enter
2025-09-01T09:00:58.645+00:00 Chassis1 auth.debug WTM4800 sshd.29539
debug1: child_reap: preauth child 5500 for connection from <client-ip>
to <server-ip> exited with status 255
2025-09-01T09:00:58.645+00:00 Chassis1 auth.info WTM4800 sshd.29539
srclimit_penalise: ipv4: new <client-ip>/32 deferred penalty of 1
seconds for penalty: connections without attempting authentication
"

Also, I can tell you that wolfSSL fails in lines:
https://github.com/wolfSSL/wolfssl/blob/v5.6.4-stable/wolfcrypt/src/random.c#L3690-L3695

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Sep-01 23:40 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #8 from Damien Miller <djm at mindrot.org> ---
Please don't paste logs in the comment box, it makes bugs hard to read.
Use the attachment feature instead.

I don't see a sandbox violation there. If I had to guess what's
happening I'd say that WolfSSL is attempting open(/dev/urandom),
soft-failing with errno==EACCESS because of
https://github.com/openssh/openssh-portable/blob/master/sandbox-seccomp-filter.c#L259
and returning a failure that terminates the sshd-auth process.

Reiterating your options:

1. Ask the WolfSSL developers if you can get it to prepare for
sandboxing before the sandbox is applied. In other libraries, this
usually means making some API call that loads a seed or opens a file
descriptor before the sandbox makes such things impossible.

2. Get WolfSSL to use getrandom() instead of open(/dev/urandom). It
looks like there is already support in the library for this:
https://github.com/wolfSSL/wolfssl/blob/v5.6.4-stable/wolfcrypt/src/random.c#L3595-L3624

3. Change the sandbox to allow the open syscall. This would
significantly weaken the sandbox as it can't be done selectively
per-path, which is why we don't do it in OpenSSH. Practically, this
means replacing "SC_DENY(__NR_open, EACCES)," with
"SC_ALLOW(__NR_open),"

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Sep-02 04:25 UTC

head link

[Bug 3855] sshd-auth sandbox limitations

https://bugzilla.mindrot.org/show_bug.cgi?id=3855

--- Comment #9 from Adrian Jarc <adrian.jarc at aviatnet.com> ---
(In reply to Damien Miller from comment #8)> Please don't paste logs in the comment box, it makes bugs hard to
> read. Use the attachment feature instead.
> 
> I don't see a sandbox violation there. If I had to guess what's
> happening I'd say that WolfSSL is attempting open(/dev/urandom),
> soft-failing with errno==EACCESS because of
> https://github.com/openssh/openssh-portable/blob/master/sandbox-
> seccomp-filter.c#L259 and returning a failure that terminates the
> sshd-auth process.
> 
> Reiterating your options:
> 
> 1. Ask the WolfSSL developers if you can get it to prepare for
> sandboxing before the sandbox is applied. In other libraries, this
> usually means making some API call that loads a seed or opens a file
> descriptor before the sandbox makes such things impossible.
> 
> 2. Get WolfSSL to use getrandom() instead of open(/dev/urandom). It
> looks like there is already support in the library for this:
> https://github.com/wolfSSL/wolfssl/blob/v5.6.4-stable/wolfcrypt/src/
> random.c#L3595-L3624
> 
> 3. Change the sandbox to allow the open syscall. This would
> significantly weaken the sandbox as it can't be done selectively
> per-path, which is why we don't do it in OpenSSH. Practically, this
> means replacing "SC_DENY(__NR_open, EACCES)," with
> "SC_ALLOW(__NR_open),"
I apologise for attaching full logs in comment box, I have completely
missed the add attachment option. Will keep in mind if there will be a
next time.

Since the whole purpose of writing this to you, was to avoid lowering
security of sandbox, I will try to avoid the 3. option. So I have
written to wolfSSL about options 1. and 2.. Hopefully I will get a
reply soon.

I wrote to you, because ever since finding out what issue happens with
wolfSSL they have become silent. Now that you have helped me find some
additional possible solutions, I have written to them again and am
waiting for a reply.

I am leaving this issue open until I get some reply from wolfSSL. I
will close it after they respond.


Thank you for your help. You have been very helpful.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

openssh bugs - Aug 2025 - [Bug 3855] New: sshd-auth sandbox

[Bug 3855] New: sshd-auth sandbox

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations

[Bug 3855] sshd-auth sandbox limitations