The server address and log file were hardcoded into the
opensnitchd.service file, making it almost impossible to change.
Soon we'll be able to change it from the UI.
we were not switching between process monitor methods properly, so we're
falling back to proc method in some cases.
Besides, there's seems to be a descriptors leaking problem in ftrace package
when closing resources.
If a rule has the priority flag set, no others rules will be checked.
So if you name the rule as 000-allow-xx and set the priority flag, the
rule wil lbe the only one that will be checked if it matches a
connection.
See #36 to know more on this feature.
Sort rules by name, so they're checked in alphabetical order.
This way, you can place deny rules at the top of the list to get better
performance, since it won't check the rest of the rules.
Discussion: #36
In case we're connecting to an IP directly, or if an IP is not resolved,
leave the DstHost field empty and format it appropiately on the UIs.
Otherwise we can't know (easily) if the field DstHost of a connection is
an IP or a domain.
When a new connection is about to be established and the system performs
a dns resolution, we displayed it like this: 9.9.9.9 (www.opensnitch.io)
It added visibility of what was going on, but if you created a rule to
filter by destination host, you were prompted twice to allow firstly the
DNS query, and secondly the TCP connection, which was a bit annoying.
Some users (#5) also asked to display just the domain, so now we only
display the domain name.
Destination host, process path and process arguments are now case
insensitive by default.
There's a new rule operator field (sensitive), to change this behaviour:
```
{
"type": "simple",
"operand": "dest.host",
"sensitive": false,
"data": "opensnitch.io",
"list": null
}
```
Rules without this field will evaluate as false by default.
closes#45
Some systems has the IPV6 protocol disabled, so we failed starting up
with the error "Address family not supported by protocol" (#52).
Now we don't exist even if we can't insert the needed rules, we'll just
log the error.
We must intercept RELATED packets, not only for intercept protocols like
ftp-data, but also to handle connection errors (ICMP errors), like the
ones originated when dis/connecting from a wifi network.
When the daemon is stopped, we need to close opened netfilter recurses.
Otherwise we can fall into a situation where we leave NFQUEUE queues
opened, which causes opensnitch to not run anymore until system restart
or a manual intervention, because there's a NFQUEUE queue already created
with the same ID.
This is what was happening as a collateral effect of #41.
Current version is stable enough (based on users' feedback) to get it out
of the release candidate state. OpenSnitch works and is usable.
There's a lot of work yet to do, so let's start from a fresh and shiny
release number.
In old libnetfilter_queue libs, we can't know the UID of the packet
because the function nfq_get_uid is not present.
We check it dynamically on run time, but we were doing it wrong, so the
daemon didn't run with old libs.
Thaks to Pain-Patate for reporting it in #18.
Processes paths with spaces were not handled correctly.
If the /proc/<pid>/exe link contained the word "(deleted)",
we deleted that part from the process path. But the way we
did it caused to not handle paths with spaces correctly.
Should fix#24
