ssh tools: Split connect and communication phases We can fallback after a connect error, but we cannot easily fall back once we start sending data as we may have consumed part of the buffer so reconnecting and sending what's left would not make sense. Therefore we now fallback on connect errors, but we issue a hard fail if error happens after communication has been established. Resolves: https://pagure.io/SSSD/sssd/issue/1498 Merges: https://pagure.io/SSSD/sssd/pull-request/3383 Signed-off-by: Simo Sorce <simo@redhat.com> Reviewed-by: Lukáš Slebodník <lslebodn@redhat.com>

ssh tools: Fix issues with multiple IP addresses Cycle through all resolved address until one succeed or all fail. This is needed for dual stack systems where either IPv4 or IPv6 are improperly configured or selectively filtered at some point along the route. Resolves: https://pagure.io/SSSD/sssd/issue/1498 Merges: https://pagure.io/SSSD/sssd/pull-request/3383 Signed-off-by: Simo Sorce <simo@redhat.com> Reviewed-by: Lukáš Slebodník <lslebodn@redhat.com>

ssh tools: The ai structure is not an array, This structure is actually a linked list, so do not mislead readers by treating it as an array. Resolves: https://pagure.io/SSSD/sssd/issue/1498 Merges: https://pagure.io/SSSD/sssd/pull-request/3383 Signed-off-by: Simo Sorce <simo@redhat.com> Reviewed-by: Lukáš Slebodník <lslebodn@redhat.com>

Make DEBUG macro invocations variadic Use a script to update DEBUG macro invocations to use it as a variadic macro, supplying format string and its arguments directly, instead of wrapping them in parens. This script was used to update the code: grep -rwl --include '*.[hc]' DEBUG . | while read f; do mv "$f"{,.orig} perl -e \ 'use strict; use File::Slurp; my $text=read_file(\*STDIN); $text=~s#(\bDEBUG\s*\([^(]+)\((.*?)\)\s*\)\s*;#$1$2);#gs; print $text;' < "$f.orig" > "$f" rm "$f.orig" done Reviewed-by: Jakub Hrozek <jhrozek@redhat.com> Reviewed-by: Stephen Gallagher <sgallagh@redhat.com> Reviewed-by: Simo Sorce <simo@redhat.com>

SSH: Use separate field for domain name in client requests Instead of appending @domain to names when the --domain option of sss_ssh_* is used, put domain name in a separate field in client requests.

Include talloc log in our debug facility https://fedorahosted.org/sssd/ticket/1495

SSH: Don't abort connection in sss_ssh_knownhostsproxy when DNS records are missing https://fedorahosted.org/sssd/ticket/1356

SSH: Supress error message output in sss_ssh_knownhostsproxy

Convert read and write operations to sss_atomic_read https://fedorahosted.org/sssd/ticket/1209

SSH: Fix infinite loop in sss_ssh_knownhostsproxy https://fedorahosted.org/sssd/ticket/1268

SSH: Canonicalize host name and do reverse DNS lookup in sss_ssh_knownhostsproxy https://fedorahosted.org/sssd/ticket/1245

SSH: Allow clients to explicitly specify host alias This change removes the need to canonicalize host names on the responder side - the relevant code was removed.

Fixed uninitialized pointer in SSH known host proxy

SSH: Remove unused --file option of sss_ssh_knownhostsproxy

SSH: Add more debugging messages

SSH: Manage global known_hosts file in the responder https://fedorahosted.org/sssd/ticket/1193

SSH: Continue connecting to SSH server even when SSSD is not running in sss_ssh_knownhostsproxy Additionally, don't drop the connection when the sss_ssh_knownhostsproxy process receives a signal. https://fedorahosted.org/sssd/ticket/1179 https://fedorahosted.org/sssd/ticket/1184

SSH: Refactor responder and client common code

SSH: OpenSSH known_hosts client