Re: audit not catching accept call from a 64bit sshd

[Steve Grubb <sgrubb@redhat.com>]

On Tuesday, January 13, 2015 03:16:25 PM hsultan@thefroid.net wrote:
> Hi,
> 
> So I'm a bit dumbfounded by what I'm seeing on my Ubuntu 14.04 Server
> LTS :
> 
> ubuntu@ip-172-31-37-137:~/test$ sudo netstat -anop |grep sshd
> tcp        0      0 0.0.0.0:22              0.0.0.0:*
> LISTEN      1406/sshd        off (0.00/0/0)
> 
> Using file I verified that it's a 64bit process :
> 
> ubuntu@ip-172-31-37-137:~/test$ file /usr/sbin/sshd
> /usr/sbin/sshd: ELF 64-bit LSB  shared object, x86-64, version 1
> (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.24,
> BuildID[sha1]=00e04495db18849d9fa69d05f3626e66c6a6d4e1, stripped
> 
> I then ran strace on it, and launched a telnet to 127.0.0.1 : 22
> 
> ubuntu@ip-172-31-37-137:~/test$ sudo strace -p 1406
> Process 1406 attached
> select(7, [3 4], NULL, NULL, NULL)      = 1 (in [3])
> accept(3, {sa_family=AF_INET, sin_port=htons(58853),
> sin_addr=inet_addr("127.0.0.1")}, [16]) = 5
> fcntl(5, F_GETFL)                       = 0x2 (flags O_RDWR)
> pipe([6, 7])                            = 0
> socketpair(PF_LOCAL, SOCK_STREAM, 0, [8, 9]) = 0
> clone(child_stack=0,
> flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD,
> child_tidptr=0x7f95cba7db10) = 6224
> 
> Clearly it is calling the accept system call.
> 
> Next I look at the audit infrastructure to see what it gives me :
> ubuntu@ip-172-31-37-137:~/test$ sudo auditctl -l
> LIST_RULES: exit,always arch=3221225534 (0xc000003e)
> syscall=accept,accept4
> 
> Then I use my own tool to catch every msg from the netlink (using
> audit_get_reply) port and display each and every msg that has a struct
> audit_reply->len > 0 , as well as audit msgs from netfilter/conntrack.
> Just to make sure that my audit code/settings are correct, I launch a
> telnet to port 22, and another one to port 631 (cups), and here's what I
> see :
> 
> audit(1421190014.387:45784): audit_enabled=1 old=1 auid=1000 ses=16
> res=1
>      [NEW] tcp      6 120 SYN_SENT src=127.0.0.1 dst=127.0.0.1
> sport=58860 dport=22 [UNREPLIED] src=127.0.0.1 dst=127.0.0.1 sport=22
> dport=58860
>   [UPDATE] tcp      6 60 SYN_RECV src=127.0.0.1 dst=127.0.0.1
> sport=58860 dport=22 src=127.0.0.1 dst=127.0.0.1 sport=22 dport=58860
>   [UPDATE] tcp      6 432000 ESTABLISHED src=127.0.0.1 dst=127.0.0.1
> sport=58860 dport=22 src=127.0.0.1 dst=127.0.0.1 sport=22 dport=58860
> [ASSURED]
>   [UPDATE] tcp      6 120 FIN_WAIT src=127.0.0.1 dst=127.0.0.1
> sport=58860 dport=22 src=127.0.0.1 dst=127.0.0.1 sport=22 dport=58860
> [ASSURED]
>   [UPDATE] tcp      6 30 LAST_ACK src=127.0.0.1 dst=127.0.0.1
> sport=58860 dport=22 src=127.0.0.1 dst=127.0.0.1 sport=22 dport=58860
> [ASSURED]
>   [UPDATE] tcp      6 120 TIME_WAIT src=127.0.0.1 dst=127.0.0.1
> sport=58860 dport=22 src=127.0.0.1 dst=127.0.0.1 sport=22 dport=58860
> [ASSURED]
>      [NEW] tcp      6 120 SYN_SENT src=127.0.0.1 dst=127.0.0.1
> sport=40478 dport=6031 [UNREPLIED] src=127.0.0.1 dst=127.0.0.1
> sport=6031 dport=40478
> [DESTROY] tcp      6 src=127.0.0.1 dst=127.0.0.1 sport=40478 dport=6031
> [UNREPLIED] src=127.0.0.1 dst=127.0.0.1 sport=6031 dport=40478
>      [NEW] tcp      6 120 SYN_SENT src=127.0.0.1 dst=127.0.0.1
> sport=36764 dport=631 [UNREPLIED] src=127.0.0.1 dst=127.0.0.1 sport=631
> dport=36764
>   [UPDATE] tcp      6 60 SYN_RECV src=127.0.0.1 dst=127.0.0.1
> sport=36764 dport=631 src=127.0.0.1 dst=127.0.0.1 sport=631 dport=36764
>   [UPDATE] tcp      6 432000 ESTABLISHED src=127.0.0.1 dst=127.0.0.1
> sport=36764 dport=631 src=127.0.0.1 dst=127.0.0.1 sport=631 dport=36764
> [ASSURED]
> audit(1421190031.663:45785): arch=c000003e syscall=43 success=yes
> exit=15 a0=b a1=7f84aee66f7c a2=7fff960572fc a3=7f84ad2b67b8 items=0
> ppid=1 pid=22048 auid=4294967295 uid=0 gid=0 euid=0 suid=0 fsuid=0
> egid=0 sgid=0 fsgid=0 tty=(none) ses=4294967295 comm="cupsd"
> exe="/usr/sbin/cupsd" key=(null)
> audit(1421190031.663:45785): saddr=02008F9C7F0000010000000000000000
> audit(1421190031.663:45785):
> 
> I never see the accept call for sshd ... even though connection is
> established, but I see the one for cups. The machine is pretty much
> idle, so there should be no msgs dropped.
> 
> Any idea what could be the reason ? I just can't see what I could be
> doing wrong here.

Make sure you have both a 32 and 64 bit rule. That's all I can think of. Also, 
check to see if you have audit=1 on your kernel boot parameters. There is a 
chance that sshd is starting before auditd and then its marked inauditable. 
But, it works fine on my system.

-Steve