In order to support external rules loaders that may extend the falco
rules format with new top level objects, move away from providing
individual filter objects to the filter_ruleset via calls to add().
Instead, pass the entire compile output returned by the compiler to
the ruleset using a new method add_compile_output(). Custom users can
then cast back the compile output to the appropriate derived class for
use in the ruleset.
Move the declaration of the compile output to a standalone class so it
can be used by rulesets without including the entire rules loader
header files, and add a new factory method new_compile_output() to the
compiler so it can create a derived class if necessary.
This change is
backwards-compatible with existing rulesets, as the default
implementation of add_compile_output() simply iterates over rules and
calls add() for each rule.
This change also speeds up rule loading. Previously, each rule
condition was compiled twice:
1. First, in the compiler, to see if it was valid.
2. Second, in the falco engine before providing each rule to the
ruleset.
Add the compiled filter to the falco_rule object instead of throwing
it away in the compiler.
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
In some cases, a user of the falco engine may want to extend the falco
rules format to provide additional objects to the rules file.
To support that, add a new method set_rule_loader() that allows a user
to provide classes that derive from
rule_loader::{reader,collector,compiler} and read those additional
objects from the rules file.
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Add methods that allow looking up the factories provided to
add_source(). This allows not having to keep track of the factories
outside of the engine.
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Inline find_source as it can be called in the event processing path.
Also take the cached variant that assigns/uses m_syscall_source_idx
and put it in find_source() instead of process_event().
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Add alternate enable_* methods that allow enabling rulesets by ruleset
id in addition to name. This might be used by some filter_rulesets to
enable/disable rules on the fly via the falco engine.
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Old gcc versions (e.g. 4.8.3) won't allow move elision
but newer versions (e.g. 10.2.1) would complain about
the redundant move.
Signed-off-by: Jason Dellaluce <jasondellaluce@gmail.com>
- avoiding inspector to be allocated for each rule
- use two boolean values for expecting macros and lists
- move items of lists alongside name, under info
- use snake case for json output, like we do for e.g alerts
- correctly retrieve evt names
- consider two levels of lists for exception operators
Signed-off-by: Lorenzo Susini <susinilorenzo1@gmail.com>
The optimization in https://github.com/falcosecurity/falco/pull/2210
had a bug when the engine uses multiple sources at the same
time--m_syscall_source is a pointer to an entry in the indexed vector
m_sources, but if add_source is called multiple times, the vector is
resized, which copies the structs but invalidates any pointer to the
vector entries.
So instead of caching m_syscall_source in add_source(), cache it in
process_events(). m_sources won't change once processing events starts.
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Instead of using a falco_rule struct on the stack, use a single value
inside the falco_source struct. It's mutable as find_source returns a
const struct.
At very high event volumes (> 1M syscalls/second), even the tiny time
it takes to create/destroy the struct starts to add up, and this
switch has some small cpu savings.
Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
