mirror of
https://github.com/microsoft/krabsetw
synced 2026-06-06 16:14:32 +00:00
9863c2cb8a
* [poc] inject schema_locator * Apply suggestions from code review Co-Authored-By: Kyle Reed <kareed@kallanreed.com> * inject schema_locator [C++/CLI] * inject schema_locator into predicate * Update krabs/krabs/schema.hpp Co-Authored-By: Kyle Reed <kareed@kallanreed.com> * updated docs * updated nuspec Co-authored-by: Kyle Reed <kareed@kallanreed.com>
108 lines
4.7 KiB
C++
108 lines
4.7 KiB
C++
// Copyright (c) Microsoft. All rights reserved.
|
|
// Licensed under the MIT license. See LICENSE file in the project root for full license information.
|
|
|
|
// This example shows how to use a user_trace and a kernel_trace in the same program.
|
|
|
|
#include <iostream>
|
|
#include <thread>
|
|
#include <condition_variable>
|
|
#include "..\..\krabs\krabs.hpp"
|
|
#include "examples.h"
|
|
|
|
static void setup_ps_provider(krabs::provider<>& provider);
|
|
static void setup_image_load_provider(krabs::kernel::image_load_provider& provider);
|
|
|
|
void kernel_and_user_trace_001::start()
|
|
{
|
|
// user_trace instances should be used for any non-kernel traces that are defined
|
|
// by components or programs in Windows. You can have multiple ETW traces in a given
|
|
// program but each trace object will consume one thread.
|
|
krabs::user_trace user;
|
|
krabs::kernel_trace kernel;
|
|
|
|
// A trace can have any number of providers, which are identified by GUID or
|
|
// a specific trace name.
|
|
//
|
|
// The GUIDs are defined by the components that emit events, and their GUIDs can
|
|
// usually be found with various ETW tools (like wevutil or Microsoft Message Analyzer).
|
|
krabs::provider<> ps_provider(krabs::guid(L"{A0C1853B-5C40-4B15-8766-3CF1C58F985A}"));
|
|
krabs::kernel::image_load_provider image_load_provider;
|
|
|
|
setup_ps_provider(ps_provider);
|
|
setup_image_load_provider(image_load_provider);
|
|
|
|
// The user_trace needs to know about the provider that we've set up.
|
|
// You can assign multiple providers to a single trace.
|
|
user.enable(ps_provider);
|
|
kernel.enable(image_load_provider);
|
|
|
|
// Begin listening for events. This call blocks, so if you want to do other things
|
|
// while this runs, you'll need to call this on another thread.
|
|
//
|
|
// Additionally, if multiple threads are enabling providers with a single trace object,
|
|
// you'll need to synchronize the call to start. Because 'start' is a blocking call,
|
|
// it will prevent any other thread from enabling additional providers.
|
|
std::thread user_thread([&user]() { user.start(); });
|
|
std::thread kernel_thread([&kernel]() { kernel.start(); });
|
|
|
|
// Let the traces process for 30 seconds.
|
|
std::cout << "starting traces..." << std::endl;
|
|
Sleep(10000);
|
|
std::cout << "stopping traces..." << std::endl;
|
|
user.stop();
|
|
kernel.stop();
|
|
|
|
user_thread.join();
|
|
kernel_thread.join();
|
|
}
|
|
|
|
void setup_ps_provider(krabs::provider<>& provider)
|
|
{
|
|
// user_trace providers typically have any and all flags, whose meanings are
|
|
// unique to the specific providers that are being invoked. To understand these
|
|
// flags, you'll need to look to the ETW event producer.
|
|
provider.any(0xf0010000000003ff);
|
|
|
|
// providers should be wired up with functions (or functors) that are called when
|
|
// events from that provider are fired.
|
|
provider.add_on_event_callback([](const EVENT_RECORD &record, const krabs::trace_context &trace_context) {
|
|
|
|
// Once an event is received, if we want krabs to help us analyze it, we need
|
|
// to snap in a schema to ask it for information.
|
|
krabs::schema schema(record, trace_context.schema_locator);
|
|
|
|
// We then have the ability to ask a few questions of the event.
|
|
std::wcout << L"Event " << schema.event_id();
|
|
std::wcout << L"(" << schema.event_name() << L") received." << std::endl;
|
|
|
|
if (schema.event_id() == 7937) {
|
|
// The event we're interested in has a field that contains a bunch of
|
|
// info about what it's doing. We can snap in a parser to help us get
|
|
// the property information out.
|
|
krabs::parser parser(schema);
|
|
|
|
// We have to explicitly name the type that we're parsing in a template
|
|
// argument.
|
|
// We could alternatively use try_parse if we didn't want an exception to
|
|
// be thrown in the case of failure.
|
|
std::wstring context = parser.parse<std::wstring>(L"ContextInfo");
|
|
std::wcout << L"\tContext: " << context << std::endl;
|
|
}
|
|
});
|
|
}
|
|
|
|
void setup_image_load_provider(krabs::kernel::image_load_provider& provider)
|
|
{
|
|
// Kernel providers accept all the typical callback mechanisms.
|
|
provider.add_on_event_callback([](const EVENT_RECORD &record, const krabs::trace_context &trace_context) {
|
|
krabs::schema schema(record, trace_context.schema_locator);
|
|
|
|
// Opcodes can be found on the kernel provider's documentation:
|
|
// https://msdn.microsoft.com/en-us/library/windows/desktop/aa364068(v=vs.85).aspx
|
|
if (schema.event_opcode() == 10) {
|
|
krabs::parser parser(schema);
|
|
std::wstring filename = parser.parse<std::wstring>(L"FileName");
|
|
std::wcout << L"Loaded image from file " << filename << std::endl;
|
|
}
|
|
});
|
|
} |