genode/doc/release_notes-08-11.txt

831 lines
37 KiB
Plaintext
Raw Normal View History

2011-12-22 15:19:25 +00:00
==============================================
Release notes for the Genode OS Framework 8.11
==============================================
Genode Labs
Summary
#######
This document presents the new features and major changes introduced
in version 8.11 of the Genode OS Framework. It is geared towards
people interested in closely following the progress of the Genode
project and to developers who want to adopt their software to our
mainline development. The document aggregates important fragments
of the updated documentation such that you won't need to scan existing
documents for the new bits. Furthermore, it attempts to provide our
rationale behind the taken design decisions.
The general theme for the release 8.11 is enabling the use of the
Genode OS framework for real-world applications. Because we regard
the presence of device drivers and a way to reuse existing library
code as fundamental prerequisites for achieving this goal, the major
new additions are an API for device drivers written in C, an API for
handling asynchronous notifications, and a C runtime. Other noteworthy
improvements are the typification of capabilities at the C++-language
level, a way for receiving and handling application faults, the
introduction of managed dataspaces, and a new API for scheduling
timed events.
Base framework
##############
This section documents the new features and changes affecting the
'base' repository, in particular the base API.
New features
============
Connection handling
~~~~~~~~~~~~~~~~~~~
The interaction of a client with a server involves the definition of
session-construction arguments, the request of the session creation via
its parent, the initialization of the matching RPC-client stub code
with the received session capability, the actual use of the session
interface, and the closure of the session. A typical procedure of
using a service looks like this:
!#include <rom_session/client.h>
!...
!
!/* construct session-argument string and create session */
!char *args = "filename=config, ram_quota=4K");
!Capability session_cap = env()->parent()->session("ROM", args);
!
!/* initialize RPC stub code */
!Rom_session_client rsc(session_cap);
!
!/* invoke remote procedures, 'dataspace' is a RPC function */
!Capability ds_csp = rsc.dataspace();
!...
!
!/* call parent to close the session */
!env()->parent()->close(session_cap);
Even though this procedure does not seem to be overly complicated,
is has raised the following questions and criticism:
* The quota-donation argument is specific for each server. Most services
use client-donated RAM quota only for holding little meta data and,
thus, are happy with a donation of 4KB. Other services maintain larger
client-specific state and require higher RAM-quota donations. The
developer of a client has to be aware about the quota requirements for
each service used by his application.
* There exists no formalism for documenting session arguments.
* Because session arguments are passed to the 'session'-call as a plain
string, there are no syntax checks for the assembled string performed
at compile time. For example, a missing comma would go undetected until
a runtime test is performed.
* There are multiple lines of client code needed to open a session to
a service and the session capability must be maintained manually for
closing the session later on.
The new 'Connection' template provides a way to greatly simplify the
handling of session arguments, session creation, and destruction on the
client side. By implementing a service-specific connection class
inherited from 'Connection', session arguments become plain constructor
arguments, session functions can be called directly on the 'Connection'
object, and the session gets properly closed when destructing the
'Connection'. By convention, the 'Connection' class corresponding to a
service resides in a file called 'connection.h' in the directory of the
service's RPC interface. For each service, a corresponding 'Connection'
class becomes the natural place where session arguments and quota
donations are documented. With this new mechanism in place, the example
above becomes as simple as:
!#include <rom_session/connection.h>
!...
!
!/* create connection to the ROM service */
!Rom_connection rom("config");
!
!/* invoke remote procedure */
!Capability ds_csp = rom.dataspace();
[https://genode.org/documentation/api/base_index#Connecting_to_services - See the API documentation for the connection template...]
2011-12-22 15:19:25 +00:00
Typed capabilities
~~~~~~~~~~~~~~~~~~
A plain 'Capability' is an untyped reference to a remote object of any
type. For example, a capability can reference a thread object or a
session to a service. It is loosely similar to a C void pointer, for which
the programmer maintains the knowledge about which data type is actually
referenced. To facilitate the type-safe use of RPC interfaces at the C++
language level, we introduced a template for creating specialized
capability types ('Typed_capability' in 'base/typed_capability.h') and
the convention that each RPC interface declares a dedicated capability
type. Note that type-safety is not maintained across RPC interfaces. As
illustrated in Figure [layered_ipc], typification is done at the
2011-12-22 15:19:25 +00:00
object-framework level on the server side and via in the 'Connection'
classes at the client side.
[image layered_ipc]
2011-12-22 15:19:25 +00:00
From the application-developer's perspective, working with capabilities
has now become type-safe, making the produced code more readable and robust.
[https://genode.org/documentation/api/base_index#Capability_representation - See the updated API documentation for the capability representation...]
2011-12-22 15:19:25 +00:00
Fifo data structure
~~~~~~~~~~~~~~~~~~~
Because the 'List' data type inserts new list elements at the list head,
it cannot be used for implementing wait queues requiring first-in
first-out semantics. For such use cases, we introduced a dedicated
'Fifo' template. The main motivation for introducing 'Fifo' into the
base API is the new semaphore described below.
[https://genode.org/documentation/api/base_index#Structured_data_types - See the new API documentation for the fifo template...]
2011-12-22 15:19:25 +00:00
Semaphore
~~~~~~~~~
Alongside lock-based mutual exclusion of entering critical sections,
organizing threads in a producer-consumer relationship via a semaphore
is a common design pattern for thread synchronization. Prior versions
of Genode provided a preliminary semaphore implementation as part of
the 'os' repository. This implementation, however, supported only one
consumer thread (caller of the semaphore's 'down' function). We have
now enhanced our implementation to support multiple consumer threads
and added the semaphore to Genode's official base API. We have made
the wake-up policy in the presence of multiple consumers configurable
via a template argument. The default policy is first-in-first-out.
[https://genode.org/documentation/api/base_index#Synchronization - See the new API documentation for the semaphore...]
2011-12-22 15:19:25 +00:00
Thanks to Christian Prochaska for his valuable contributions to the new
semaphore design.
Asynchronous notifications
~~~~~~~~~~~~~~~~~~~~~~~~~~
Inter-process communication via remote procedure calls requires both
communication partners to operate in a synchronous fashion. The caller
of an RPC blocks as long as the RPC is not answered by the called
server. In order to receive the call, the server has to explicitly
wait for incoming messages. There are a number of situations where
synchronous communication is not suited.
For example, a GUI server wants to deliver a notification to one of its
clients about new input events being available. It does not want to
block on a RPC to one specific client because it has work to do for
other clients. Instead, the GUI server wants to deliver this
_notification_ with _fire-and-forget_ semantics and continue with
its operation immediately, regardless of whether the client received
the notification or not. The client, in turn, does not want to poll
for new input events at the GUI server but it wants to be _waken_up_
when something interesting happens. Another example is a block-device
driver that accepts many requests for read/write operations at once.
The operations may be processed out of order and may take a long time.
When having only synchronous communication available, the client and
the block device driver would have to employ one distinct thread for
each request, which is complicated and a waste of resources. Instead,
the block device driver just wants to acknowledge the completeness of
an operation _asynchronously_.
Because there are many more use cases for asynchronous inter-process
communication, we introduced a new signalling framework that complements
the existing synchronous RPC mode of communication with an interface for
issuing and receiving asynchronous notifications. It defines interfaces
for signal transmitters and signal receivers. A signal receiver can
receive signals from multiple sources, whereas the sources of incoming
signals are clearly distinguishable. One or multiple threads can either
poll or block for incoming signals. Each signal receiver is addressable
via a capability. The signal transmitter provides fire-and-forget
semantics for submitting signals to exactly one signal receiver. Signals
are communicated in a reliable fashion, which means that the exact number
of signals submitted to a signal transmitter is communicated to the
corresponding signal receiver. If notifications are generated at a higher
rate than as they can be processed at the receiver, the transmitter
counts the notifications and delivers the total amount with the next
signal transmission. This way, the total number of notifications gets
properly communicated to the receiver even if the receiver is not highly
responsive. Notifications do not carry any payload because this payload
would have to be queued at the transmitter.
[image signals]
2011-12-22 15:19:25 +00:00
Image [signals] illustrates the roles of signaller thread,
2011-12-22 15:19:25 +00:00
transmitter, receiver, and signal-handler thread.
[https://genode.org/documentation/api/base_index#Asynchronous_notifications - See the new API documentation for asynchronous notifications...]
2011-12-22 15:19:25 +00:00
The current generic implementation of the signalling API employs one
thread at each transmitter and one thread at each receiver. Because
the used threads are pretty heavy weight with regard to resource usage,
ports of Genode should replace this implementation with platform-
specific variants, for example by using inter-process semaphores or
native kernel support for signals.
Region-manager faults
~~~~~~~~~~~~~~~~~~~~~
In Genode, region-manager (RM) sessions are used to manage the
address-space layout for processes. A RM session is an address-space
layout that can be populated by attaching (portions of) dataspaces to
(regions of) the RM session. Normally, the RM session of a process is
first configured by the parent when decoding the process' ELF binary.
During the lifetime of the process, the process itself may attach
further dataspaces to its RM session to access the dataspace's content.
Core as the provider of the RM service uses this information for
resolving page faults raised by the process. In prior versions of
Genode, core ignored unresolvable page faults, printed a debug message
and halted the page-faulted thread. However, this condition may be of
interest, in particular to the process' parent for reacting on the
condition of a crashed child process. Therefore, we enhanced the RM
interface by a fault-handling mechanism. For each RM session, a fault
handler can be installed by registering a signal receiver capability.
If an unresolvable page fault occurs, core delivers a signal to the
registered fault handler. The fault handler, in turn, can request the
actual state of the RM session (page-fault address) and react upon
the fault. One possible reaction is attaching a new dataspace at the
fault address and thereby implicitly resolving the fault. If core
detects that a fault is resolved this way, it resumes the operation
of the faulted thread.
This mechanism works analogously to how page faults are handled by
CPUs, but on a more abstract level. A (n-level) page table corresponds
to a RM session, a page-table entry corresponds to a dataspace-
attachment, the RM-fault handler corresponds to a page-fault
exception handler, and the resolution of page-faults (RM fault)
follows the same basic scheme:
# Application accesses memory address with no valid page-table-entry
(RM fault)
# CPU generates page-fault exception (core delivers signal to fault
handler)
# Kernel reads exception-stack frame or special register to determine
fault address (RM-fault handler reads RM state)
# Kernel adds a valid page-table entry and returns from exception
(RM-fault handler attaches dataspace to RM session, core resumes
faulted thread)
The RM-fault mechanism is not only useful for detecting crashing child
processes but it enables a straight-forward implementation of growing
stacks and heap transparently for a child process. An example for
using RM-faults is provided at 'base/src/test/rm_fault'.
Note that this mechanism is only available on platforms on which core
resolves page faults. This is the case for kernels of the L4 family.
On Linux however, the Linux kernel resolves page faults and suspends
processes performing unresolvable memory accesses (segmentation fault).
Managed dataspaces (experimental)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The RM-fault mechanism clears the way for an exciting new feature
of Genode 8.11 called managed dataspaces. In prior versions of Genode,
each dataspace referred to a contiguous area of physical memory (or
memory-mapped I/O) obtained by one of core's RAM, ROM, or IO_MEM
services, hence we call them physical dataspaces. We have now added
a second type of dataspaces called managed dataspaces. In contrast
to a physical dataspace, a managed dataspace is backed by the content
described by an RM session. In fact, each RM session can be used as
dataspace and can thereby be attached to other RM sessions.
Combined with the RM fault mechanism described above, managed
dataspaces enable a new realm of applications such as dataspaces
entirely managed by user-level services, copy-on-write dataspaces,
non-contiguous large memory dataspaces that are immune to physical
memory fragmentation, process-local RM fault handlers (e.g., managing
the own thread-stack area as a sub-RM-session), and sparsely populated
dataspaces.
Current limitations
-------------------
Currently, managed dataspaces still have two major limitations. First,
this mechanism allows for creating cycles of RM sessions. Core must
detect such cycles during page-fault resolution. Although, a design for
an appropriate algorithm exists, cycle-detection is not yet implemented.
The missing cycle detection would enable a malicious process to force
core into an infinite loop. Second, RM faults are implemented using the
new signalling framework. With the current generic implementation, RM
sessions are far more resource-demanding than they should be. Once the
signalling framework is optimized for L4, RM sessions and thereby
managed dataspaces will become cheap. Until then, we do not recommend
to put this mechanism to heavy use.
Because of these current limitations, managed dataspaces are marked as
an experimental feature. When building Genode, experimental features are
disabled by default. To enable them, add a file called 'specs.conf'
with the following content to the 'etc/' subdirectory of your build
directory:
! SPECS += experimental
For an example of how to use the new mechanism to manage a part of a
process' own address space by itself, you may take a look at
'base/src/test/rm_nested'.
Changes
=======
Besides the addition of the new features described above, the following
parts of the base framework underwent changes worth describing.
Consistent use of typed capabilities and connection classes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We applied capability typification to all interfaces of Genode including
the base API and the interfaces defined in the 'os' repository. Figure
[base_cap_types] provides an overview about the capability types
2011-12-22 15:19:25 +00:00
provided by the base API.
[image base_cap_types]
2011-12-22 15:19:25 +00:00
Overview about the capability types provided by the base API
Furthermore, we have complemented all session interfaces with
appropriate 'Connection' classes taking service-specific session
arguments into account.
For session-interface classes, we introduced the convention to declare
the service name as part of the session-interface via a static member
function:
! static const char *service_name();
Allocator refinements
~~~~~~~~~~~~~~~~~~~~~
Throughout Genode, allocators are not only used for allocating memory
but also for managing address-space layouts and ranges of physical
resources such as I/O-port ranges or IRQ ranges. In these cases, the
address '0' may be a valid value. Consequently, this value cannot be
used to signal allocation errors as done in prior versions of Genode.
Furthermore, because managed dataspaces use the RM session interface to
define the dataspace layout, the address-'0' problem applies here as
well. We have now refined our allocator interfaces and the RM-session
interface to make them fit better for problems other than managing
virtual memory.
Misc changes
~~~~~~~~~~~~
We revised all interfaces to consistently use _exceptions_ to signal
error conditions rather than delivering error codes as return values.
This way, error codes become exception types that have a meaningful
name and, in contrast to global 'errno' definitions, an error exception
type can be defined local to the interface it applies to. Furthermore,
the use of exceptions allows for creating much cleaner looking interfaces.
Traditionally, we have provided our custom _printf_ implementation as C
symbol to make this function available from both C and C++ code. However,
we observed that we never called this function from C code and that the
'printf' symbol conflicts with the libc. Hence, we turned 'printf'
into a C++ symbol residing in the 'Genode' namespace.
Operating-system services and libraries
#######################################
This section documents the new features and changes affecting
the 'os' repository.
New Features
============
Device-driver framework for C device drivers
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Genode's base API features everything needed to create user-level device
drivers. For example, the 'os' repository's PS/2 input driver and the
PCI bus driver are using Genode's C++ base API directly. However, most of
today's device drivers are written in C. To ease the reuse of existing
drivers on Genode, we have introduced a C API for device drivers into
Genode's 'os' repository. The API is called DDE kit (DDE is an acronym
for device-driver environment) and it is located at 'os/include/dde_kit'.
The DDE kit API is the result of long-year experiences with porting device
drivers from Linux and FreeBSD to custom OS environments. The following
references are the most significant contributions to the development of
the API.
;
Christian Helmuth created the initial version of the Linux device-driver
environment for L4. He describes his effort of reusing unmodified sound
drivers on the L4 platform in his thesis
[https://os.inf.tu-dresden.de/papers_ps/helmuth-diplom.pdf - Generische Portierung von Linux-Gerätetreibern auf die DROPS-Architektur].
2011-12-22 15:19:25 +00:00
;
Gerd Griessbach approached the problem of re-using Linux USB drivers
by following the DDE approach in his diploma thesis
[https://os.inf.tu-dresden.de/papers_ps/griessbach-diplom.pdf - USB for DROPS].
2011-12-22 15:19:25 +00:00
;
Marek Menzer adapted Linux DDE to Linux 2.6 and explored the DDE
approach for block-device drivers in his student research project
[https://os.inf.tu-dresden.de/papers_ps/menzer-beleg.pdf - Portierung des DROPS Device Driver Environment (DDE) für Linux 2.6 am Beispiel des IDE-Treibers ]
2011-12-22 15:19:25 +00:00
and his diploma thesis
[https://os.inf.tu-dresden.de/papers_ps/menzer-diplom.pdf - Entwicklung eines Blockgeräte-Frameworks für DROPS].
2011-12-22 15:19:25 +00:00
;
Thomas Friebel generalized the DDE approach and introduced the DDE kit
API to enable the re-use of device driver from other platforms than
Linux. In particular, he experimented with the block-device drivers of
FreeBSD in his diploma thesis
[https://os.inf.tu-dresden.de/papers_ps/friebel-diplom.pdf - Übertragung des Device-Driver-Environment-Ansatzes auf Subsysteme des BSD-Betriebssystemkerns].
2011-12-22 15:19:25 +00:00
;
Dirk Vogt successfully re-approached the port of USB device drivers
from the Linux kernel to L4 in his student research project
[https://os.inf.tu-dresden.de/papers_ps/beleg-vogt.pdf - USB for the L4 Environment].
2011-12-22 15:19:25 +00:00
The current incarnation of the DDE kit API provides the following
features:
* General infrastructure such as init calls, assertions, debug output
* Interrupt handling (attach, detach, disable, enable)
* Locks, semaphores
* Memory management (slabs, malloc)
* PCI access (find device, access device config space)
* Virtual page tables (translation between physical and virtual
addresses)
* Memory-mapped I/O, port I/O
* Multi-threading (create, exit, thread-local storage, sleep)
* Timers, jiffies
For Genode, we have created a complete reimplementation of the DDE kit
API from scratch by fully utilizing the existing Genode infrastructure
such as the available structured data types, core's I/O services,
the synchronization primitives, and the thread API.
[image dde_kit]
2011-12-22 15:19:25 +00:00
Figure [dde_kit] illustrates the role of DDE kit when re-using an
2011-12-22 15:19:25 +00:00
unmodified device driver taken from the Linux kernel. DDE kit translates
Genode's C++ base API to the DDE kit C API. The DDE kit API, in turn, is
used as back end by the Linux driver environment, which translates Linux
kernel interfaces to calls into DDE kit. With this translation in place,
an unmodified Linux device driver can be embedded into the Linux driver
environment. The device API is specific for a class of devices such as
NICs, block devices, or input devices. It can either be used directly as
a function interface by an application that is using the device driver
as a library, or it can be made accessible to external processes via an
RPC interface.
Limitations
-----------
The PCI sub system is not completely implemented, yet.
Alarm API providing a timed event scheduler
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The scheduling of timed events is a recurring pattern found in device
drivers, application frameworks such as Qt4 ('qeventdispatcher'), and
applications. Therefore, we have added a timed event scheduler to the
'os' repository. The new alarm API ('os/include/os/alarm.h') allows
for the scheduling of both one-shot alarms and periodic alarms.
Changes
=======
PS/2 input driver
~~~~~~~~~~~~~~~~~
The original PS/2 driver tried to switch the PS/2 keyboard to
scan-code set 2 and assumed that all modern keyboards support this
mode of operation. However, this assumption was wrong. We observed
that the legacy PS/2 support of some USB keyboards covers only the
emulated (xlate) scan-code set 1 mode. This is also case for the PS/2
emulation in VirtualBox. Therefore, we changed our PS/2 driver to
never touch the keyboard mode but to only detect the current mode
of operation. The driver has now to support both, scan-code set 1 and
scan-code set 2. This change comes along with a slightly more complex
state machine in the driver. Hence, we moved the state machine from
the IRQ handler to a distinct class and changed the control flow of
the driver to fetch only one value from the i8042 PS/2 controller
per received interrupt.
PCI bus driver
~~~~~~~~~~~~~~
Until now, Genode's PCI bus driver was only used for experimentation
purposes. With the forthcoming driver framework however, the PCI bus
driver will play a central role in the system. Therefore, we adapted
the interface of the PCI driver to these requirements. Specifically,
the scanning of the PCI bus can now be performed without constraining
the results by a specific vendor ID.
Nitpicker GUI server
~~~~~~~~~~~~~~~~~~~~
We improved the _output_latency_ of the Nitpicker GUI server by flushing
pixels eagerly and deferring the next periodically scheduled flush.
This change has a positive effect on the responsiveness of the GUI to
user input.
Misc changes
~~~~~~~~~~~~
Prior versions of the 'os' repository came with a custom 'os/include/base'
directory with interfaces extending the base API. To avoid confusion
between the 'base' repository and the 'os' repository, 'os'-local API
extensions are now located at 'os/include/os'. This way, the folder
prefix of include statements indicates well from which repository the
included header files comes from.
C runtime
#########
Most of existing libraries rely on the presence of a C library. For
making the reuse of this software on Genode possible, we have now
made a complete C library available for Genode. It comes as a separate
source-code repository called 'libc' and is based on the code of FreeBSD.
The original code is available at the official FreeBSD website.
:FreeBSD website:
[https://www.freebsd.org/developers/cvs.html]
2011-12-22 15:19:25 +00:00
Our libc port comprises the libraries 'gdtoa', 'gen', 'locale', 'stdio',
'stdlib', 'stdtime', 'string', and 'msun'. Currently, it supports the
x86 architecture. Support for other architectures is planned as future
addition. At the current stage, our back end is very basic and most of
its functions are dummy stubs. We used Christian Prochaska's forthcoming
Genode port of Qt4 as test case and successfully used the new libc as
foundation for building graphical Qt4 applications. We will further
extend the back end in correspondence to the growing feature set of the
Genode OS framework.
:Usage:
To use the libc in your application, just add 'libc' to the 'LIBS'
declaration in your build-description file. This declaration will make
the libc headers available for the include path of your target and link
the C library. When building, make sure that the 'libc' repository is
included in your build configuration ('etc/build.conf').
:Limitations:
The current version of the C library is not thread-safe. For most
string and math functions, this is not a problem (as these functions
do not modify global state) but be careful with using more complex
functions such as 'malloc' from multiple threads. Also, 'errno' may
become meaningless when calling libc functions from multiple threads.
We have left out the following files from the Genode port of the
FreeBSD libc: gdtoa 'strtodnrp.c' (gdtoa), 'getosreldate.c' (gen),
'strcoll.c', 'strxfrm.c', 'wcscoll.c', 'wcsxfrm.c' (string),
's_exp2l.c' ('msun').
The current back end is quite simplistic and it may help you to revisit
the current state of the implementation in the 'libc/src/lib/libc'
directory. If one of the functions in 'dummies.c' is called, you will
see the debug message:
! "<function-name> called, not yet implemented!"
However, some of the back-end function implemented in the other files
have dummy semantics but have to remain quiet because they are called
from low-level libc code.
Build infrastructure
####################
Build-directory creation tool
=============================
Because we think that each Genode developer benefits from knowing the
basics about the functioning of the build system, the manual creation of
build directories is described in Genode's getting-started document.
However, for regular developers, creating build directories becomes a
repetitive task. Hence, it should be automated. We have now added a
simple build-directory creation tool that creates pre-configured build
directories for some supported platforms. The tool is located at
'tool/builddir/create_builddir'. To print its usage information, just
execute the tool without arguments.
Improved linking of binary files
================================
For linking binary data, binary file have to be converted to object
files. Over the time, we have used different mechanisms for this
purpose. Originally, we used 'ld -r -b binary'. Unfortunately, these
linker options are not portable. Therefore, the mechanism was changed
to a 'hexdump' and 'sed' magic that generated a C array from binary data.
This solution however, is complicated and slow. Now, we have adopted
an idea of Ludwig Hähne to use the 'incbin' directive of the GNU
assembler, which is a very clean, flexible, and fast solution.
Lib-import mechanism
====================
Libraries often require specific include files to be available at the
default include search location. For example, users of a C library
expect 'stdio.h' to be available at the root of the include search
location. Placing the library's include files in the root of the
default search location would pollute the include name space for
all applications, regardless if they use the library or not. To
keep library-include files well separated from each other, we have
enhanced our build system by a new mechanism called lib-import.
For each library specified in the 'LIBS' declaration of a build
description file, the build system incorporates a corresponding
'import-<libname>.mk' file into the build process. Such as file
defines library-specific compiler options, in particular additional
include-search locations. The build system searches for lib-import
files in the 'lib/import/' subdirectories of all used repositories.
Using 'ar' for creating libraries
=================================
The previous versions of Genode relied on incremental linking ('ld -r')
for building libraries. This approach is convenient because the linker
resolves all cross-dependencies between libraries regardless of the
order of how libraries are specified at the linker's command line.
However, incremental linking prevents the linker from effectively
detecting dead code. In contrast, when linking '.a' files, the linker
detects unneeded object files. Traditionally, we have only linked our
own framework containing no dead code. This changed with the new 'libc'
support. When linking the 'libc', the presence of dead code becomes
the normal case rather than the exception. Consequently, our old
incremental-linking approach produced exceedingly large binaries
including all functions that come with the 'libc'. We have now adopted
the classic 'ar' mechanism for assembling libraries and use the linker's
'start-group' 'end-group' feature to resolve inter-library-dependencies.
This way, dead code gets eliminated at the granularity of object files.
In the future, we will possible look into the '-ffunction-sections' and
'-gc-sections' features of the GNU tool chain to further improve the
granularity to function level.
If your build-description files rely on custom rules referring to
'lib.o' files, these rules must be adapted to refer to 'lib.a' files
instead.
Misc changes
============
* Added sanity check for build-description files overriding 'INC_DIR'
instead of extending it.
* Restrict inclusion of dependency files to those that actually matter
when building libraries within 'var/libcache'. This change significantly
speeds up the build process in the presence of large libraries such as
Qt4 and libc.
* Added rule for building 'cpp' files analogously to the 'cc' rule.
Within Genode, we name all C++ implementation files with the 'cc'
suffix. However, Qt4 uses 'cpp' as file extension so we have to
support both.
* Build-description files do no longer need the declaration
'REQUIRES = genode'. Genode's include search locations are now
incorporated into the build process by default.
Applications
############
This section refers to the example applications contained in Genode's
'demo' repository.
We have enhanced the _Scout_widgets_ as used by the launchpad and the
Scout tutorial browser to perform all graphical output double-buffered,
which effectively eliminates drawing artifacts that could occur when
exposing intermediate drawing states via direct (unbuffered) output.
Furthermore, we have added a way to constrain the maximum size of
windows to perform pixel-buffer allocations on realistic window sizes.
Both launchpad and Scout can now start child applications. In Scout
this functionality is realized by special "execute" links. We have
generalized the underlying application logic for creating and
maintaining child processes between both applications and placed
the unification into a separate 'launchpad' library.
We have replaced the default document presented in Scout with an
_interactive_walk-through_guide_ explaining the basic features of Genode.
The document uses the new "execute" link facility to let the user start
a launchpad instance by clicking on a link.
Platform-specific changes
#########################
Genode used to define _fixed-width_integer_types_ in a file 'stdint.h'
placed in a directory corresponding to bit-width of the platform, for
example 'include/32bit/stdint.h'. When building for a 32bit platform,
the build system included the appropriate directory into the
include-search path and thereby made 'stdint.h' available at the root
of the include location. Unfortunately, this clashes with the 'stdint.h'
file that comes with the C library. To avoid conflict with libc header
files, we moved the definition of fixed-width integer types to
'32bit/base/fixed_stdint.h'.
For the L4/Fiasco version of Genode, there existed some x86-specific
header files that did not specifically depend on L4/Fiasco, for example
atomic operations. Because these files are not L4/Fiasco-specific and
may become handy for other platforms as well, we moved them to the
generic 'base' repository.
Linux 32bit
===========
:Dissolving Genode's dependency from the glibc:
The port of the C runtime to Genode posed an interesting challenge to
the Linux version of Genode. This version used to rely on certain
functions provided by the underlying glibc:
* For creating and destroying threads, we used to rely on POSIX threads
as provided by the 'pthread' library
* The lock implementation was based on the POSIX semaphore functions
'sem_init', 'sem_wait', and 'sem_post'
* Shared memory was realized by using files ('open', 'close',
'ftruncate') and the 'mmap' interface
* Starting and killing processes was implemented using 'fork' and 'kill'
* Inter-process communication used the glibc's socket functions
For our custom C runtime, we want to override the glibc functionality
with our own implementation. For example, we want to provide the 'mmap'
interface to a Genode application by implementing 'mmap' with
functions of our base API. On Linux, however, this base API, in turn,
used to rely on 'mmap'. This is just an example. The problem applies
also for the other categories mentioned above. We realized that we cannot
rely on the glibc on one hand but at the same time replace it by a custom
C runtime (in fact, we believe that such a thing is possible by using
awkward linker magic but we desire a clean solution). Consequently, we
have to remove the dependency of Genode from the glibc on Linux. Step
by step, we replaced the used glibc functions by custom Linux system-call
bindings. Each binding function has a prefix 'lx_' such that the symbol
won't collide with 'libc' symbols. The new bindings are located at the file
'base-linux/src/platform/linux_syscalls.h'. It consist of 20 functions,
most of them resembling the original interface ('socket', 'connect',
'bind', 'getsockname', 'recvfrom', 'write', 'close', 'open', 'fork',
'execve', 'mmap', 'ftruncate', 'unlink', 'tkill', 'nanosleep').
For other functions, we simplified the semantics for our use case
('sigaction', 'sigpending', 'sigsetmask', 'create_thread'). The most
noteworthy changes are the creation and destruction of threads by
directly using the 'clone' and 'tkill' system calls, and the lock
implementation. Because we cannot anymore rely on the convenience of
using futexes indirectly through the POSIX semaphore interface, we
have adopted the simple locking approach that we already use for the
L4/Fiasco version. This lock implementation is a simple sleeping
spinlock.
:Compromises:
The introduction of custom Linux system-call bindings for Genode has
several pros and cons. With this change, The Linux version of Genode is
not anymore easy to port to other POSIX platforms such as the Darwin
kernel. For each POSIX kernel used as Genode platform, a custom
implementation of our system-call bindings must be created. The
original POSIX variant could still be reanimated, but this version
would inherently lack support for Genode's C runtime, and thus would
have limited value. A positive side effect of this solution, however,
is that 'linux_syscalls.h' documents well the subset of the Linux'
kernel interface that we are actually using.
The replacement of POSIX semaphores with sleeping spinlocks decreases
locking performance quite significantly. In the contention case, the
wakeup from sleeping introduces a high latency of up to one millisecond.
Furthermore, fairness is not guaranteed and the spinning produces a bit
of system load. If this approach turns out to become a serious performance
bottleneck, we will consider creating custom bindings for Linux' futexes.
L4/Fiasco
=========
The concepts of _RM_faults_ and _managed_dataspaces_ as described in
Section [Base framework], had been implemented into the L4/Fiasco
version of core. Although the introduction of these concepts involved
only minimal changes at the API level, the required core-internal
changes had been quite invasive, affecting major parts of the pager
and RM-session implementations.
Prior versions of the L4/Fiasco version of core did not implement
the _cancel-blocking_mechanism_ as specified by the CPU-session API.
The missing implementation resulted in lock-ups when destructing a
thread that blocks for lock. With the new implementation based on
L4/Fiasco's inter-task ex-regs system call, such threads can now
be gracefully destructed.