Attachment 787953 Details for Bug 1114090

ssh_config client configuration

ssh_config (text/plain), 7.37 KB, created by James Carter on 2018-10-31 06:13:17 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: James Carter

Created: 2018-10-31 06:13:17 UTC

Size: 7.37 KB

patch

obsolete

># ssh client systemwide configuration file.  See ssh(1) and 
># ssh_config(5) for more information.  While formerly the default for all unset
># parms was shown, this is getting excessive, and only highlights are now shown.
># Revision history:
>#   2017-07-23	openssh-7.2p2	jimc	For Tumbleweed; rearrange ciphers
>#   2016-09-20	openssh-6.6p1	jimc	More Logjam (CVE-2015-4000)
>#   2011-03-27	openssh-5.8p1	jimc	For SuSE 11.4
>#   (numerous revisions not recorded)
>#   2002-05-20	openssh-3.2.2p	jimc	Set up for Xena
>
># With multiple parm sources the first one wins (put host-specific settings
># earlier).  Sources are {command line, ~/.ssh/config , /etc/ssh/ssh_config }
>
>Host *
>
># Network Parameters
>#   Port			22 (dflt)
>#   AddressFamily		any (dflt) (vs. inet or inet6)
>#   IPQoS		lowdelay throughput (autoswitch, DFLT) -- Quality of Svc
>#   Protocol 1 is crackable and there's a widespread virus that attacks it.
>    Protocol			2
>
># Connection Parameters
>#   ServerAliveCountMax		3 (dflt) 
>    ServerAliveInterval		120
>    TCPKeepAlive		no
>#   ConnectionAttempts		1 (dflt) (whether to retry on net outage: no)
>    ConnectTimeout		5
>#   EscapeChar			~ (dflt)
>
># Cryptography (lists are comma separated, in preference order).
># jimc's tests in 2017: with an AES crypto engine, aes256-gcm is equal in speed
># to aes128-gcm, and twice as fast (+more secure) than aes256-ctr.  aes256-gcm
># is 2.5x faster than chacha, but without the engine chacha is 2 to 2.5x faster
># than aes128-gcm.  VMs don't have access to the engine.  So prefer aes256-gcm
># but allow chacha.  MD5 and SHA1 are at end of life.  Where is Keccak?  
>
># Conspiracy theorists believe that NIST/NSA approved elliptic curves got
># approved because the NSA has a way to attack them; jimc discounts that 
># threat.  Aaron Toponce (2nd ref below) has actual crypto complaints about 
># ecdh-sha2-nistp{256,384,521}.  He also mentions that Github will only use
># SHA1 (in 2015) so I'm keeping it.  
># References for openssh-6.6p1 revision (these are all from 2015)
>#   https://jbeekman.nl/blog/2015/05/ssh-logjam/ (for KEX)
>#   https://pthree.org/2015/01/12/reasonable-ssh-security-for-openssh-6-0-or-later/
>#   https://bbs.archlinux.org/viewtopic.php?id=188613
>
># Fly in ointment: The Dropbear ssh daemon I use on Android is SimpleSSHD by
># galexand (dev email: andingers@gmail.com).  The server can do these MAC algos:
># hmac-sha1-96 hmac-sha1 hmac-md5.  Per Aaron Toponce (2nd ref above), the 
># status of the old hashes in HMAC is ambiguous, since the hack resistance
># issues are different, but MD5 is so broken that we'd better stay away from
># it, and spoofing SHA1 is almost feasible too.  hmac-sha1-96 is the same algo 
># but with the hash truncated from 160 to 96 bits, which is not a good sign.  
># I'm going to reluctantly add hmac-sha1 at the end to support SimpleSSHD.  
>
># Ciphers for data transfer.  aes* == Rijndael.  Most preferred first.  
>Ciphers		aes256-gcm@openssh.com,aes128-gcm@openssh.com,chacha20-poly1305@openssh.com,aes256-ctr,aes192-ctr,aes128-ctr,aes256-cbc
># Default for openssh-7.2p2 = chacha20-poly1305@openssh.com, aes128-ctr,
># 	aes192-ctr, aes256-ctr, aes128-gcm@openssh.com,
># 	aes256-gcm@openssh.com, aes128-cbc, aes192-cbc, aes256-cbc, 3des-cbc
>
># Message Authentication Codes.  "etm" means "encrypt then MAC" which is safer.
># Following Aaron Toponce's recommendation (2nd ref. above):
>MACs		hmac-sha2-512-etm@openssh.com,hmac-sha2-256-etm@openssh.com,hmac-sha2-512,hmac-sha2-256,hmac-sha1
>
># KEX (key exchange) algos: elliptic curve Diffie-Hellman is more efficient and
># is not vulnerable to Logjam.  Beekman recommends this preference order.  
># Using diffie-hellman-group-exchange-*, 
># the server picks a prime field DH group from /etc/ssh/moduli, which you 
># should rebuild yourself with 2048 bit minimum (see ssh-keygen).  The client
># does what the server tells it and does not use /etc/ssh/moduli.  
># diffie-hellman-group14-sha1 uses a fixed 2048 bit prime, for fallback.  
># diffie-hellman-group1-sha1 uses a fixed 1024 bit prime, assumed compromised.
># curve25519-sha256@libssh.org is in >=vers. 6.5 (with bug) and is reliable
># in >=6.7; Beekman prefers it if available.  
>KexAlgorithms	curve25519-sha256@libssh.org,diffie-hellman-group-exchange-sha256,diffie-hellman-group-exchange-sha1,diffie-hellman-group14-sha1
>
>#	Minimum Diffie-Hellman prime field size.  RFC4419 and SSH default is 
>#	1024, which is being exploited in the wild.  
>    KexDHMin	2048
>
># Host Authentication
>#   StrictHostKeyChecking	ask (dflt) (ask user about adding new host keys)
>#   VerifyHostKeyDNS		no (dflt) (check server's key in DNS SSHFP rec)
>#	Do not check the host key if the server is on localhost.  Illegal
>#	parameter on v2.9 (Solaris), but desireable on modern OpenSSH.
>    NoHostAuthenticationForLocalhost	yes
>
># User Authentication
>    PreferredAuthentications	gssapi-with-mic,publickey,keyboard-interactive,password
>#   HostbasedAuthentication	no (dflt) (SSH host check, then honor .rhosts)
>    GSSAPIAuthentication	yes
>#   PasswordAuthentication	yes (dflt)
>#   NumberOfPasswordPrompts	3 (dflt)
>#   PubkeyAuthentication	yes (dflt)
>#   KbdInteractiveAuthentication yes (dflt)
>#   KbdInteractiveDevices (any?) (server typically offers pam,bsdauth,skey)
>#   RhostsRSAAuthentication	no (dflt) (proto 1 only)
>#   RSAAuthentication		yes (dflt, proto 1 only)
>#   ChallengeResponseAuthentication	yes (dflt) (for smart cards and PAM)
>#   SmartcardDevice		(none)
>#   User			(from command line)
>#   EnableSSHKeysign		no (dflt) (improves HostbasedAuthentication)
>
># Login Procedure
>#   CheckHostIP			yes (dflt) (server must be in known_hosts file)
>#   GSSAPIDelegateCredentials	no (dflt) (use -K to turn on)
>#   GSSAPIKeyExchange		no (dflt) (use GSSAPI in place of host key)
>    SendEnv			"LOGNAME LANG LC_*"
>#   NumberOfPasswordPrompts	3 (dflt)
>#	 Do equivalent of ssh-add if key was missing from the agent.  Dflt=no.
>#	 (not avail in openssh-6.6p1)
>    AddKeysToAgent		yes 
>
># Port and X11 Forwarding
>#	The polarity can be confusing.  "Local" (-L P:H:R) means that the 
>#	a process on the local machine can connect to local port P, and the
>#	server will connect to host H port R to serve that connection. 
>#	"Remote" (-R P:H:R) means that processes on the remote machine can 
>#	connect to its port P, and the local client will connect to host H
>#	port R to serve the connection.  
>#   LocalForward		(from command line -L))
>#   RemoteForward		(from command line -R)
>#   DynamicForward		(none) (like Local but SOCKS chooses the port)
>#   Tunnel		no (dflt)
>#   TunnelDevice		any:any (dflt)
>#   GatewayPorts		no (dflt) (only localhost can use fwd ports)
>#   ClearAllForwardings		no (dflt)
>#   ForwardAgent		no (dflt) (specify -A to turn on, upper case)
>#   ForwardX11			no (dflt) (specify -X to turn on, upper case)
>#	Turn on ForwardX11Trusted unless you have configured X11 security right.
>    ForwardX11Trusted		yes
>#   ForwardX11Timeout		20m (dflt) (only if ForwardX11Trusted = no)
>#   ExitOnForwardFailure	no (dflt) (terminate if any forwarding fails)
>
>
># File Paths
>#   GlobalKnownHostsFile	/etc/ssh/ssh_known_hosts (we don't use this)
>#   IdentityFile		~/.ssh/{identity,id_dsa,id_ecdsa,id_rsa} (dflt)
>#   IdentitiesOnly		no (dflt) (yes = decline other than these)
>#   CertificateFile		(none)
>#   UserKnownHostsFile		~/.ssh/known_hosts (dflt)
>#   HashKnownHosts		no (dflt) (obscure identity of partner hosts)
>#   XAuthLocation		/usr/bin/xauth (dflt)
>
># Miscellaneous 
>#   LogLevel			INFO (dflt)
>#   FingerprintHash		sha256 (dflt) (algo to display key fingerprint)

Actions: View

Attachments on bug 1114090: 787952 | 787953 | 787954