| Summary: | TCP Wrappers support does not log successful connections | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Portable OpenSSH | Reporter: | Joseph <jap3003+opensshbugs> | ||||
| Component: | sshd | Assignee: | OpenSSH Bugzilla mailing list <openssh-bugs> | ||||
| Status: | CLOSED WONTFIX | ||||||
| Severity: | normal | ||||||
| Priority: | P2 | ||||||
| Version: | -current | ||||||
| Hardware: | All | ||||||
| OS: | All | ||||||
| Attachments: |
|
||||||
|
Description
Joseph
2002-01-07 09:49:08 AEDT
Created attachment 8 [details]
Patch to fix this bug (gdiff -u)
shouln't the tcpwrapper do this? we should not call syslog at all "should?" Maybe. "Does?" No. This is the way tcpd does it, too. The call to hosts_access() only determines whether or not a host should pass. refuse() [iirc, don't have the source handy] terminates sshd after calling syslog. There doesn't appear to be a corresponding accept() function. However, the fact that we have to call syslog ourselves is mitigated by the fact that there is a function available to crack the request structure, so we don't really have to go dinking with the tcp wrapper internals. It's the %s being sent into the syslog. Although it's not mentioned on the manpage, it's in the header we include. I'm confused. If I set LogLevel VERBOSE in sshd_config, I get something like this in the syslog: Jan 17 08:12:39 myhost sshd[432]: Connection from somehost port someport What else is desired? You can do this by setting LogLevel verbose Mass change of RESOLVED bugs to CLOSED |