On Linux a proxy listening on a Unix socket relayed only the first
bytes of an answer: sockmap() read with recvfrom() into the client's
address, which for a Unix stream socket Linux fills with an empty
AF_UNIX one, and then sent with sendto() to that address, which a
connected Unix stream socket refuses. A TCP connection has no use for
addresses: sockmap() now sends and receives with NULL ones, and
connect-back sends its command with socksend() and reads the answer
with sockrecvfrom() without an address, which a NULL sin now allows.
An abstract socket name (-iunix:@NAME) was bound with the whole size of
struct sockaddr_un, padded with zeros, so a client naming it by its
bytes, as most do, reached another name. SASIZE() now gives an abstract
address its exact size; getsockname() calls that took SASIZE() of the
old contents as the buffer size take the buffer's size instead.
tcppm refused port 0 for a unix: target, though the manual gives
"tcppm 8080 unix:/var/run/app.sock 0" as the example: the port of a
Unix socket target is ignored, so 0 is allowed.
On a reload the new service bound the path anew and the old one,
closing after it, removed the file, the new service's by then, leaving
nothing to connect to. A service now notes the file it bound (device
and inode) and removes the path on closing only if it is still that
file.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
'auth cacheacl ...' is identical to 'auth cache ...' except ACL is not checked for cached authentication. dstaddr, dstport, dsthost and dstoper (operation) are intended to be used with cacheacl. For example
authcache user,ip,password,dstaddr 600
auth cacheacl iponly strong
allows user to access destination ip without ACL/password revalidation if he has cached attempt to the same ip from the same ip with the same username and password.
srvaddr, srvport are useful to only match with cached attempts to the same `internal` address / service port.
- authcache switched to use hashtables, size parameter added
- overflow fixed on hashinit
- hashtable prefers new values on insert if table is full
- hashtable is able to compact/grow
example:
plugin SSLPlugin.ld.so ssl_plugin
allow user1
parent 1000 http 1.1.1.1 1111
allow user2
parent 1000 https 2.2.2.2 2222
ssl_client_mode 3
ssl_client
proxy
With ssl_client_mode 3 TLS is only handshaked for https parent type and is not handshaked for http parent.
Use unix:/path/to/socket, e.g.
tcppm 1234 unix:/path/to/socket 1234
Under linux abstract sockets are supported with '@' prefix, e.g.
parent 1000 http unix:@virtual.3proxy.socket 1111
Destination port numbers are not used in tcppm/parent, but you must specify any positive value to match the syntaxis.
ssl_client_mode
0 (default) - handshake immediately after connect() (with first parent or with destination if there is no parent)
1 - handshake with destination server (handshake after connection via parents is established)
2 - handshake after data channel is established (e.g. after CONNECT)