The reason behind this one is that without a real variable name, the profile objects are immediately desctructed and the profiling only happens for the small window when they were valid. If the intention is to profile the method where they were defined, this doesn't actually happen.
The Zeek code base has very inconsistent #includes. Many sources
included a few headers, and those headers included other headers, and
in the end, nearly everything is included everywhere, so missing
#includes were never noticed. Another side effect was a lot of header
bloat which slows down the build.
First step to fix it: in each source file, its own header should be
included first to verify that each header's includes are correct, and
none is missing.
After adding the missing #includes, I replaced lots of #includes
inside headers with class forward declarations. In most headers,
object pointers are never referenced, so declaring the function
prototypes with forward-declared classes is just fine.
This patch speeds up the build by 19%, because each compilation unit
gets smaller. Here are the "time" numbers for a fresh build (with a
warm page cache but without ccache):
Before this patch:
3144.94user 161.63system 3:02.87elapsed 1808%CPU (0avgtext+0avgdata 2168608maxresident)k
760inputs+12008400outputs (1511major+57747204minor)pagefaults 0swaps
After this patch:
2565.17user 141.83system 2:25.46elapsed 1860%CPU (0avgtext+0avgdata 1489076maxresident)k
72576inputs+9130920outputs (1667major+49400430minor)pagefaults 0swaps
- All timers are now handled by a single global timer manager, which simplifies how they handled by the IOSource manager.
- This change flows down a number of changes to other parts of the code. The timer manager tag field is removed, which means that matching connections to a timer manager is also removed. This removes the ability to tag a connection as internal or external, since that's how the connections where differentiated. This in turn removes the `current_conns_extern` field from the `ConnStats` record type in the script layer.
- Removes entire FindSoonest method that includes all of the calls to select() for checking for ready sources
- Removes FD_Set checking against IOSources
- Adds system for registering and unregistering file descriptors from IOSources. This allows individual sources to mark themselves as ready to be checked by the loop as they become available.
- Adds entirely new loop architecture based on checking the IOSources for when their next timeout is, and then waiting for either that timeout or when the next source is ready. This also implements the polling based on what the OS supports, instead of just calling select() on all platforms. Currently it supports kqueue, epoll, and plain poll.
- Adds system for pinging the loop to force it to wake up
Comparing parent process ID to 1 to detect loss of parent process was
not necessarily portable, so now it stores parent PID pre-fork and then
monitors for any change.
More aspects of the cluster configuration to get fleshed out later,
but a basic cluster like one would use for a live deployment
can now be instantiated and run under supervision. The new
clusterized-pcap-processing supervisor mode is also not done yet.
The full process hierarchy isn't set up yet, but these changes
help prepare by doing two things:
- Add a -j option to enable supervisor-mode. Currently, just a single
"stem" process gets forked early on to be used as the basis for
further forking into real cluster nodes.
- Separates the parsing of command-line options from their consumption.
i.e. need to parse whether we're in -j supervisor-mode before
modifying any global state since that would taint the "stem" process.
The new intermediate structure containing the parsed options may
also serve as a way to pass configuration info from "stem" to its
descendent cluster node processes.
* origin/topic/robin/gh-239:
Undo a change to btest.cfg from a recent commit
Updating submodule.
Fix zeek-wrapper
Update for renaming BroControl to ZeekControl.
Updating submodule.
GH-239: Rename bro to zeek, bro-config to zeek-config, and bro-path-dev to zeek-path-dev.
Note - this compiles, but you cannot run Bro anymore - it crashes
immediately with a 0-pointer access. The reason behind it is that the
required clone functionality does not work anymore.
This also installs symlinks from "zeek" and "bro-config" to a wrapper
script that prints a deprecation warning.
The btests pass, but this is still WIP. broctl renaming is still
missing.
#239
* 'topic/rework-packets' of https://github.com/jsbarber/bro:
One more tinker to Packet -- ensure no uninitialized values
Packet::IP()-created IP_Hdr should not free
Make enums work for non-C++11 config
Refactor to make bro use a common Packet object. Do a better job of parsing layer 2 and keeping track of layer 3 proto. Add support for raw packet event, including Layer2 headers.
Conflicts:
aux/plugins
Works like old enable_communication(), but for new broker communication
mechanism. Scripts have to explicitly call this if they want to use the
broker communication functionality. Saves a decent chunk of Bros'
initialization time when one doesn't need communication features.
We could bring this back, now derived from PktSrc (though strickly
speaking it's of course not *packets). But not sure if we want that,
as the input framework seems the better place to host it. Then it
would turns into a reader.
An unnecessary null pointer check and uninitialized scalar fields.
Don't expect these to be actual problems, but easy enough to fix in
order to silence Coverity.
A bunch of infrastructure work to move IOSource, IOSourceRegistry (now
iosource::Manager) and PktSrc/PktDumper code into iosource/, and over
to a plugin structure.
Other IOSources aren't touched yet, they are still in src/*.
It compiles and does something with a small trace, but that's all I've
tested so far. There are quite certainly a number of problems left, as
well as various TODOs and cleanup; and nothing's cast in stone yet.
Will continue to work on this.
I got rid of the earlier separate InterpreterPlugin class. Instead
Plugin now has a set of virtual methods HookSomething()... that
plugins can override. For efficiency purposes, they however need to
register first that they are interested in a hook, otherwise the
virtual method will never be called. The idea is to extend the set of
hooks over time as we figure out what's useful.
This is a checkpoint commit that's essentially untested and probably
broken. It compiles, though.
