genode/doc/challenges.txt

424 lines
21 KiB
Plaintext

=======================================
Future Challenges of the Genode project
=======================================
Abstract
########
This document compiles various ideas to pursue in the context of Genode. It is
meant as source of inspiration for individuals who are interested in getting
involved with the project and for students who want to base their student
research projects on Genode.
Applications and library infrastructure
#######################################
:VNC server implementing Genode's framebuffer session interface:
With 'Input' and 'Framebuffer', Genode provides two low-level interfaces
used by interactive applications. For example, the Nitpicker GUI server uses
these interfaces as a client and, in turn, exports multiple virtual
'Framebuffer' and 'Input' interfaces to its clients. This enables a
highly modular use of applications such as the nesting of GUIs. By
implementing the 'Framebuffer' and 'Input' interfaces with a VNC server
implementation, all graphical workloads of Genode would become available over
the network. One immediate application of this implementation is the remote
testing of graphical Genode applications running on a headless server.
:Interfacing with the SAFE network:
The [https://safenetwork.org/ - SAFE network] is an attempt to fix many
shortcomings of the internet - in particular with respect to privacy and
freedom - at an architectural level. It is a peer-to-peer communication
and storage network that does not depend on single point of
failure or control. It is intriguing to explore the opportunity of
integrating support for the SAFE network not merely as an application but
integrated in the operating system, i.e., in the form of Genode components
or a set of Genode VFS plugins.
:Interactive sound switchbox based on Genode's Audio_out session interface:
Since version 10.05, Genode features a highly flexible configuration concept
that allows the arbitrary routing of session requests throughout the
hierarchic process structure. Even though primarily designed for expressing
mandatory-access control rules, the concept scales far beyond this use case.
For example, it can be used to run an arbitrary number of processes
implementing the same interface and connecting the different interface
implementations. One special case of this scenario is a chain of audio
filters with each using the 'Audio_out' session interface for both roles
client and server. Combined with the Nitpicker GUI server and Genode's
support for real-time priorities, this base techniques enable the creation of
flexible audio mixer / switchboard applications, which require dedicated
frameworks (e.g., Jack audio) on traditional operating systems. The goal of
this project is to create a showcase implementation demonstrating the
feasibility for creating high-quality audio applications on Genode.
Furthermore, we wish for feedback regarding the current design of our bulk
streaming interface when used for low-latency applications.
:Graphical on-target IPC tracing tool using Qt:
Analysing the interaction of components of a multi-server operating system
such as Genode is important to discover bottlenecks of the system and for
debugging highly complex usage scenarios involving many processes. Currently,
Genode handles this problem with two approaches. First, Genode's
recursive structure enables the integration of a subsystem in a basic
OS setup featuring only those drivers and components used for the particular
subsystem. After the successful integration of such a subsystem, it can
be embedded into a far more complex application scenario without any changes.
With this approach, the subject to analyse can be kept at a reasonable level
at integration time. For debugging purposes, the current approach is using
the debugging facilities of the respective base platforms (e.g., using
GDB on Linux, the Fiasco kernel debugger, the OKL4 kernel debugger).
However, in many cases, bottlenecks do not occur when integrating individual
sub systems but after integrating multiple of such subsystems into a large
application scenario. For such scenarios, existing debugging methodologies do
not scale. A tool is desired that is able to capture the relationships
between processes of a potentially large process hierarchy, to display
communication and control flows between those processes, and to visualize the
interaction of threads with the kernel's scheduler.
Since Qt is available natively on Genode, the creation of both offline and
on-target analysis tools has become feasible. The first step of this project
is creating an interactive on-target tool, that displays the interaction
of communicating threads as captured on the running system. The tool should
work on a selected kernel that provides a facility for tracing IPC messages.
The underlying light-weight tracing infrastructure is
[https://genode.org/documentation/release-notes/19.08#Tracinghttps://genode.org/documentation/release-notes/19.08#Tracing - already in place].
The Qt-based tracing tools would complement this infrastructure with
an interactive front end.
:Ports of popular software:
Genode features a ports mechanism to cleanly integrate 3rd-party software.
Thanks to the C runtime, the flexible per-component VFS, the standard
C++ library, and the Noux runtime (for UNIX software), porting software
to Genode is relatively straight forward. The
[http://genode.org/documentation/developer-resources/porting - porting guide]
explains the typical steps. A wish list of software that we'd like to
have available on Genode is available at
[http://usr.sysret.de/jws/genode/porting_wishlist.html].
:Native Open-Street-Maps (OSM) client:
When using Sculpt OS, we regularly need to spawn a fully fledged web
browser in a virtual machine for using OSM or Google maps. The goal
of this project would be a native component that makes maps functionality
directly available on Genode, alleviating the urge to reach for a SaaS
product. The work would include a review of existing OSM clients regarding
their feature sets and the feasibility of porting them to Genode.
Depending on the outcome of this review, an existing application could
be ported or a new component could be developed, e.g., leveraging Genode's
Qt support.
Application frameworks and runtime environments
###############################################
:OpenJDK:
[http://openjdk.java.net/ - OpenJDK] is the reference implementation of the
Java programming language and hosts an enormous ecosystem of application
software.
Since
[https://genode.org/documentation/release-notes/19.02#Showcase_of_a_Java-based_network_appliance - version 19.02],
Genode features a port of OpenJDK that allows the use of Java for networking
applications.
The next step would be the creation of Genode-specific native classes that
bridge the gap between the Java world and Genode, in particular the glue
code to run graphical applications as clients of Genode's GUI server. Since
OpenJDK has been ported to numerous platforms (such as Haiku), there exists
a comforting number of implementations that can be taken as reference.
:Android's ART VM natively on Genode:
ART is a Java virtual machine that is used for executing applications on
Android. By running ART directly on Genode, the Linux kernel could be
removed from the trusted computing base of Android, facilitating the use of
this mobile OS in high-assurance settings.
:Go language runtime:
Go is a popular language in particular for web applications. In the past,
there were numerous attempts to make the Go runtime available on Genode
but so far, none of those undertakings have landed in the official
Genode source tree. To goal of this project is the hosting of
Go-written applications - in particular networking applications - as
Genode components. The topic comprises work on the tool-chain
and build-system integration, the porting the runtime libraries, and
the glue between the Go and Genode environments.
:Combination of CAmkES with Genode:
[https://wiki.sel4.systems/CAmkES - CAmkES] is a component framework for
seL4. In contrast to Genode, which is a dynamic system, CAmkES-based systems
are defined at design time and remain fixed at runtime. Hence, CAmkES and
Genode can be seen as the opposite ends of component-based used-land
architectures. The goal of this project is to build a bridge between
both projects with the potential to cross-pollinate the respective communities.
Among the principal approaches are embedding of a single CAmkES
component as a Genode component (e.g., an individual device driver),
the hosting of a dynamic Genode system as a component within a
CAmkES system, or the hosting of a CAmkES system composition as a Genode
subsystem.
:Runtime for the D programming language:
The D systems programming language was designed to overcome many gripes that
exists with C++. In particular, it introduces a sane syntax for meta
programming, supports unit tests, and contract-based programming. These
features make D a compelling language to explore when implementing OS
components. Even though D is a compiled language, it comes with a runtime
providing support for exception handling and garbage collection. The goal of
the project is to explore the use of D for Genode programs, porting the
runtime to Genode, adapting the Genode build system to accommodate D
programs, and interfacing D programs with other Genode components written in
C++.
:Using Haskell as systems-development language:
The goal of this project is the application of functional programming
i.e., Haskell, for the implementation of low-level Genode components.
Implementing critical functionalities in such a high-level language instead
of a classical systems language such as C or C++ would pave the way towards
analyzing such components with formal methods.
The use of Haskell for systems development was pioneered by the
[http://programatica.cs.pdx.edu/House/ - House Project]. A more recent
development is [http://halvm.org - HalVM] - a light-weight OS runtime for
Xen that is based on Haskell.
:Xlib compatibility:
Developments like Wayland notwithstanding, most application software on
GNU/Linux systems is built on top of the Xlib programming interface.
However, only a few parts of this wide interface are actually used today.
I.e., modern applications generally deal with pixel buffers instead of
relying on graphical drawing primitives of the X protocol. Hence, it seems
feasible to reimplement the most important parts of the Xlib interface to
target Genode's native GUI interfaces (nitpicker) directly. This would
allow us to port popular application software to Sculpt OS without
changing the application code.
:Bump-in-the-wire components for visualizing session interfaces:
Genode's session interfaces bear the potential for monitoring and
visualizing their use by plugging a graphical application
in-between any two components. For example, by intercepting block
requests issued by a block-session client to a block-device driver,
such a bump-in-the-wire component could visualize
the access patterns of a block device. Similar ideas could be pursued for
other session interfaces, like the audio-out (sound visualization) or NIC
session (live visualization of network communication).
The visualization of system behavior would offer valuable insights,
e.g., new opportunities for optimization. But more importantly, they
would be extremely fun to play with.
Virtualization
##############
:VirtualBox on top of KVM on Linux:
Genode's version of VirtualBox replaces the original in-kernel VirtualBox
hypervisor by the virtualization mechanism of the NOVA hypervisor or the
Muen separation kernel. Those mechanisms look very similar the KVM
interface of the Linux kernel. It should in principle be possible to
re-target Genode's version of VirtualBox to KVM. This way, VirtualBox and
Qemu/KVM-based virtual machines could co-exist on the same system, which
is normally not possible. Also, complex Genode scenarios (like Turmvilla)
could be prototyped on GNU/Linux.
:Xen as kernel for Genode:
Using Xen as kernel for Genode would clear the way to remove the
overly complex Linux OS from the trusted computing base of Xen
guests OSes.
Xen is a hypervisor that can host multiple virtual machines on one physical
machine. For driving physical devices and for virtual-machine management, Xen
relies on a privileged guest OS called Dom0. Currently, Linux is the
predominant choice to be used as Dom0, which implicates a trusted computing
base of millions of lines of code for the other guest OSes.
Even though Xen was designed as hypervisor, a thorough analysis done by Julian
Stecklina concludes that Xen qualifies well as a kernel for Genode. For
example, Julian implemented a version of Genode's IPC framework that utilizes
Xen's communication mechanisms (event channels and shared memory).
:Genode as virtualization layer for Qubes OS:
[https://www.qubes-os.org/ - Qubes OS] is a desktop operating system
that follows the principle of security through compartmentalization.
In spirit, it is closely related to Genode. In contrast Genode's
clean-slate approach of building a fine-grained multi-component system,
Qubes employs Xen-based virtual machines as sandboxing mechanism. In
[https://blog.invisiblethings.org/2015/10/01/qubes-30.html - version 3.0],
Qubes introduced a Hypervisor Abstraction Layer, which decouples Qubes
from the underlying virtualization platform. This exploration project
pursues the goal of replacing Xen by Genode as virtualization layer
for Qubes.
:Qemu:
As we use Qemu as primary testing platform for most of the kernels, a port
of Qemu to Genode is needed in order to move our regular work flows to
Genode as development platform. The basic prerequisites namely libSDL and a
C runtime are already available such that this porting work seems to be
feasible. In our context, the ia32, amd64, and ARM platforms are of most
interest. Note that the project does not have the immediate goal of
using hardware-based virtualization. However, if there is interest,
the project bears the opportunity to explore the provisioning of the
KVM interface based on Genode's VFS plugin concept.
:Hardware-accelerated graphics for virtual machines:
In
[https://genode.org/documentation/release-notes/17.08#Hardware-accelerated_graphics_for_Intel_Gen-8_GPUs - Genode 17.08],
we introduced a GPU multiplexer for Intel Broadwell along with support
for Mesa-based 3D-accelerated applications.
While designing Genode's GPU-session interface, we also aimed at supporting
the hardware-accelerated graphics for Genode's virtual machine monitors like
VirtualBox or Seoul, but until now, we did not took the practical steps of
implementing a virtual GPU device model.
The goal of this project is the offering of a virtual GPU to a Linux guest
OS running on top of Genode's existing virtualization and driver
infrastructure.
Device drivers
##############
:Sound on the Raspberry Pi:
The goal of this project is a component that uses the Raspberry Pi's
PWM device to implement Genode's audio-out-session interface. Since
Genode's version of libSDL already supports this interface as audio
backend, the new driver will make the sound of all SDL-based games
available on the Raspberry Pi.
:Data Plane Development Kit (DPDK):
Genode utilizes the network device drivers of the iPXE project, which
perform reasonably well for everyday use cases but are obviously not
designated for high-performance networking.
The [http://dpdk.org/ - DPDK] is a vendor-supported suite of network device
drivers that is specifically developed for high-performance applications.
It presents an attractive alternative to iPXE-based drivers. This project
has the goal to make DPDK drivers available as a Genode component.
Platforms
#########
:Microkernelizing Linux:
Thanks to Genode's generic interfaces for I/O access as provided by core, all
Genode device drivers including drivers ported from Linux and gPXE can be
executed as user-level components on all supported microkernels. However, so
far, we have not enabled the use of these device drivers on Linux as base
platform. The goal of this project is the systematic replacement of in-kernel
Linux device drivers by Genode processes running in user space, effectively
reducing the Linux kernel to a runtime for Genode's core process. But moving
drivers to Genode processes is just the beginning. By employing further
Genode functionality such as its native GUI, lwIP, and Noux, many protocol
stacks can effectively be removed from the Linux kernel.
In 2018, Johannes Kliemann pursued this topic to a state where Genode
could be used as init process atop a customized Linux kernel.
[https://lists.genode.org/pipermail/users/2018-May/006066.html - His work]
included the execution of Genode's regular device drivers for VESA and
PS/2 as regular Genode components so that Genode's interactive demo
scenario ran happily on a laptop. At this time, however, only parts of
his results were merged into Genode's mainline.
The goal of this project is to follow up on Johannes' work, bring the
[https://github.com/genodelabs/genode/pull/2829 - remaining parts] into
shape for the inclusion into Genode, and address outstanding topics, in
particular the handling of DMA by user-level device drivers. Further down
the road, it would be tempting to explore the use of
[https://en.wikipedia.org/wiki/Seccomp - seccomp] as sandboxing mechanism
for Genode on Linux and the improvement of the Linux-specific implementation
of Genode's object-capability model.
:Support for the HelenOS/SPARTAN kernel:
[http://www.helenos.org - HelenOS] is a microkernel-based multi-server OS
developed at the university of Prague. It is based on the SPARTAN microkernel,
which runs on a wide variety of CPU architectures including Sparc, MIPS, and
PowerPC. This broad platform support makes SPARTAN an interesting kernel to
look at alone. But a further motivation is the fact that SPARTAN does not
follow the classical L4 road, providing a kernel API that comes with an own
terminology and different kernel primitives. This makes the mapping of
SPARTAN's kernel API to Genode a challenging endeavour and would provide us
with feedback regarding the universality of Genode's internal interfaces.
Finally, this project has the potential to ignite a further collaboration
between the HelenOS and Genode communities.
:Support for the XNU kernel (Darwin):
XNU is the kernel used by Darwin and Mac OS X. It is derived from the
MACH microkernel and extended with a UNIX-like syscall API. Because the
kernel is used for Mac OS X, it could represent an industry-strength
base platform for Genode supporting all CPU features as used by Mac OS X.
:Genode on the Librem5 phone hardware:
Even though there exists a great variety of ARM-based SoCs, Genode
primarily focuses on the NXP i.MX family because it is - in contrast
to most SoCs in the consumer space - very liberal in terms of
good-quality public documentation and reference code, and it scales
from industrial to end-user-facing use cases (multi-media).
The [https://puri.sm/products/librem-5/ - Librem5] project - with its
mission to build a trustworthy mobile phone - has chosen the i.MX family as
the basis for their product for likely the same reasons that attract us.
To goal of this work is bringing Genode to the Librem5 hardware.
For the Librem5 project, Genode could pave the ground towards new use cases
like high-security markets where a regular Linux-based OS would not be
accepted. For the Genode community, the Librem5 hardware could become an
attractive mobile platform for everyday use, similar to how we developers
use our Genode-based [https://genode.org/download/sculpt - Sculpt OS] on our
laptops.
System management
#################
:Remote management of Sculpt OS via Puppet:
[https://en.wikipedia.org/wiki/Puppet_(company)#Puppet - Puppet] is a
software-configuration management tool for administering a large amount
of machines from one central place. Genode's
[https://genode.org/download/sculpt - Sculpt OS] lends itself to such
an approach of remote configuration management by the means of the
"config" file system (for configuring components and deployments) and
the "report" file system (for obtaining the runtime state of components).
The project would explore the application of the Puppet approach and tools
to Sculpt OS.
Optimizations
#############
:De-privileging the VESA graphics driver:
The VESA graphics driver executes the graphics initialization code provided
by the graphics card via an x86 emulator. To initialize a graphics mode, this
code needs to access device hardware. Currently, we permit access to all
device registers requested by the graphics-card's code. These devices include
the system timer, the PCI configuration registers, and the interrupt
controller, which are critical for the proper operating of the kernel. The
goal of this work is to restrict the permissions of the VESA driver to a
minimum by virtualizing all devices but the actual graphics card.