[Kos-cvs] kos-web/contents histo_en.i,1.7,1.8 histo_fr.i,1.13,1.14
thomas at kos.enix.org
thomas at kos.enix.org
Wed Dec 29 22:57:12 CET 2004
Update of /var/cvs/kos/kos-web/contents
In directory the-doors:/tmp/cvs-serv7979/contents
Modified Files:
histo_en.i histo_fr.i
Log Message:
Mise à jour de l'historique en anglais
Index: histo_en.i
===================================================================
RCS file: /var/cvs/kos/kos-web/contents/histo_en.i,v
retrieving revision 1.7
retrieving revision 1.8
diff -u -d -r1.7 -r1.8
--- histo_en.i 29 Dec 2004 21:25:07 -0000 1.7
+++ histo_en.i 29 Dec 2004 21:57:09 -0000 1.8
@@ -3,49 +3,55 @@
?>
<p>The KOS project has been launched in June 1998 by Dimitri Ara and
-Thomas Petazzoni. <br/>Their work consisted in sharing ideas about
-Operating Systems, before they really start programming.</p>
+Thomas Petazzoni. <p/>
-<ul>
+<p>Their work consisted in sharing ideas about Operating Systems,
+before they really start programming.</p>
-<li><b>October 2004</b>
-Work by Thomas on a new mapping, reverse mapping and physical memory
-management system. The new system has been thought with
+<h2>October 2004</h2>
+
+<p>Work by Thomas on a new mapping, reverse mapping and physical
+memory management system. The new system has been thought with
synchronization concerns in mind, particularly the one faced with the
-virtual memory management part.</li>
+virtual memory management part.</p>
-<li><b>July 2004</b>
-Presentation of KOS given by David and Thomas at the <a
+<h2>July 2004</h2>
+
+<p>Presentation of KOS given by David and Thomas at the <a
href="http://www.rencontresmondiales.org">Libre Software Meeting</a>
in Bordeaux, France. This conference took place inside the
-<i>Embedded</i> thema.</li>
+<i>Embedded</i> thema.</p>
-<li><b>June 2004</b>
-David and Thomas started writing articles for Linux Magazine France, a
-french magazine related to Linux. These articles explain step by step
-how to create a simple operating system on PC. It is based on a
-dedicated operating system called <a
+<h2>June 2004</h2>
+
+<p>David and Thomas started writing articles for Linux Magazine
+France, a french magazine related to Linux. These articles explain
+step by step how to create a simple operating system on PC. It is
+based on a dedicated operating system called <a
href="http://sos.enix.org">SOS</a>. Of course, a lot of inspiration
came from the KOS project. There should be about 12 articles published
in more than one year, at the rate of an article per month. The
articles are written in French, but they may be available someday in
-English.</li>
+English.</p>
-<li><b>Tuesday, June 1st 2004</b>
-Presentation of KOS at the <a href="http://www.utbm.fr">UTBM</a>, a
+<h2>Tuesday, June 1st 2004</h2>
+
+<p>Presentation of KOS at the <a href="http://www.utbm.fr">UTBM</a>, a
french engineering school. The presentation was made by Thomas, who
-was student in this school at that time.</li>
+was student in this school at that time.</p>
-<li><b>Sunday, April 25th 2004</b>
-Presentation of KOS at <a
+<h2>Sunday, April 25th 2004</h2>
+
+<p>Presentation of KOS at <a
href="http://idile.org/libreast/index.php">Libr'East</a>, a free
software event that took place near Paris in France. The presentation
-was made by Thomas.</li>
+was made by Thomas.</p>
-<li><b>from September 2004 to January 2004</b>
-Work by Thomas and Mélanie Bats on KOS during their studies at the <a
+<h2>from September 2004 to January 2004</h2>
+
+<p>Work by Thomas and Mélanie Bats on KOS during their studies at the <a
href="http://www.utbm.fr">UTBM</a>, a french engineering school. Their
-work was focused on :
+work was focused on :</p>
<ul>
@@ -58,40 +64,46 @@
</ul>
-This work was ended by a report, in French, available on the website.
+<p>This work was ended by a report, in French, available on the website.</p>
-</li>
+<h2>December 2003</h2>
-<li><b>December 2003</b>
-During a coding week-end, Julien and Thomas improved the support of
-ELF binaries. They also worked on user threads and their stacks, and
-also on virtual memory management. They started to look at the PCI
-bus.</li>
+<p><i>Coding week-end with Julien and Thomas</i></p>
-<li><b>October 2003</b>
-During a coding week-end, Julien and Thomas have implemented a couple
-of important syscalls : <i>fork</i>, <i>exec</i> and <i>brk</i>. This
-was possible thanks to work previously done by Thomas on the virtual
-memory subsystem.</li>
+<p>They improved the support of ELF binaries. They also worked on user
+threads and their stacks, and also on virtual memory management. They
+started to look at the PCI bus.</p>
-<li><b>End of May 2003</b>
-Coding week end with David and Thomas at Montpellier. The scheduler
-has been completely changed, with the integration of waitqueues. The
-only thread that is not on a waitqueue is the currently running
-thread. The synchronization primitives of the kitc module have updated
-so that they use these new wait queues.
-</li>
+<h2>October 2003</h2>
-<li><b>Beginning of May 2003</b>
-After a very long time of inactivity, the development of Kos has
+<p><i>Coding week-end, with Julien and Thomas</i></p>
+
+<p>Implementation of important syscalls : <i>fork</i>, <i>exec</i> and
+<i>brk</i>. This was possible thanks to work previously done by Thomas
+on the virtual memory subsystem.</p>
+
+<h2>End of May 2003</h2>
+
+<p><i>Coding week end with David and Thomas at Montpellier.</i></p>
+
+<p>The scheduler has been completely changed, with the integration of
+waitqueues. The only thread that is not on a waitqueue is the
+currently running thread. The synchronization primitives of the kitc
+module have updated so that they use these new wait queues.</p>
+
+<h2>Beginning of May 2003</h2>
+
+<p>After a very long time of inactivity, the development of Kos has
started again last week end. The inactivity was mostly due to the lack
-of free time and the lock around the Babel stuff.
-Indeed, Julien and Thomas had together a big reflexion concerning a
+of free time and the lock around the Babel stuff.</p>
+
+<p>Indeed, Julien and Thomas had together a big reflexion concerning a
solution about the way to implement drivers and to export various
interfaces to the user, in order to avoid ugly things like ioctl(). Of
course, many reflexions, prototypes and tests have been done before.
After a long discussion, two solutions appeared to be eligible for
-implementation :
+implementation :</p>
+
<ul>
<li>The first one, very simple, comes from the simplification of
several previous reflexions. It proposes for each object of the
@@ -103,351 +115,531 @@
of the first solution concerning code factorization. It allows to
extend object capacities by using translators.</li>
</ul>
-Finally, the first solution was choosen for the moment, so that the
+
+<p>Finally, the first solution was choosen for the moment, so that the
development can go on on simple basis. We will be able later to modify
this model with a deeper knowledge of the problem if necessary.
The system available on CVS works with the first model, but the tty,
part, ide and fat drivers have ot been implemented again, but it
-should come soon.</li>
+should come soon.</p>
-<li><b>November 2002</b> Coding Week-End with Julien, David
-and Thomas. They implemented during this WE a new module, called
-"kares" (Kos Abstraction and Resource Export Services), which aims at
-replacing the old Babel. This new version is the combination of the
+<h2>November 2002</h2>
+
+<p><i>Coding Week-End with Julien, David
+and Thomas.</i></p>
+
+<p>They implemented during this WE a new module, called "kares" (Kos
+Abstraction and Resource Export Services), which aims at replacing the
+old Babel. This new version is the combination of the
proof-of-concept realized by David and Thomas during August, and
researches made by Julien concerning the namespace management problem.
The new system rely on an nscache (namespace cache), which give a
virtual representation of part of the namespace, on resources with
multiple interfaces, and also a bit of reflexivity. Part of the code
is in C++ (without exceptions or RTTI), in order to see if we can use
-this language. </li>
+this language. </p>
-<li><b>August 2002</b>
-Coding party for Thomas/d2. Set up a framework for the cpl3/cpl0
-interaction, extending the current Babel implementation. Based on the
-principle that a kernel resource can be accessed through multiple
-interfaces, and on a bit of reflexivity to overcome the shortcomings
-of the COM-style GUIDs. Wrote a C++ test program (linux/x86 only): see
-kos-dev/syscall-test/.</li>
+<h2>August 2002</h2>
-<li><b>Middle of July 2002</b>
-A presentation of the KOS project has been done by Thomas Petazzoni
+<p><i>Coding party for Thomas and David</i></p>
+
+<p>They set up a framework for the cpl3/cpl0 interaction, extending
+the current Babel implementation. Based on the principle that a kernel
+resource can be accessed through multiple interfaces, and on a bit of
+reflexivity to overcome the shortcomings of the COM-style GUIDs. Wrote
+a C++ test program (linux/x86 only): see kos-dev/syscall-test/.</p>
+
+<h2>Middle of July 2002</h2>
+
+<p>A presentation of the KOS project has been done by Thomas Petazzoni
during the LSM (Libre Software Meeting) in Bordeaux on Friday, 12th
-July.<br/>
-The slides will be available online soon.
-During this meeting, I met different Hurd developers (Marcus Brinkmann
-and Neal Walfield), and we have been able to talk about operating
-system programming.</li>
+July.</p>
-<li><b>Beginning of July 2002</b>
-Rewrote the elf32 linker of the loader. Allows to load any relocatable
+<p>During this meeting, Thomas met different Hurd developers (Marcus
+Brinkmann and Neal Walfield), and we had interesting discussions about
+operating system development.</p>
+
+<h2>Beginning of July 2002</h2>
+
+<p>Rewrote the elf32 linker of the loader. Allows to load any relocatable
elf32 object without any assumption on its internal structure (no ld
-script anymore), nor on the number of initlevels.</li>
+script anymore), nor on the number of initlevels.</p>
+
+<h2>June 2002</h2>
+
+<ul>
+
+ <li>Minor improvement of the tty driver : using Alt+Fx to change
+ console, instead of a, b, c, d, e, f, g, h.</li>
-<li><b>June 2002 :</b>
- <ul>
- <li>Minor improvement of the tty driver : using Alt+Fx (1 <= x <= 8)
- to change console, instead of a, b, c, d, e, f, g, h.
<li>The exportation of symbols and the stack change have to been
- corrected in order to make KOS compliant with gcc 3.1.
+ corrected in order to make KOS compliant with gcc 3.1.</li>
+
<li>Huge improvement in the loader : massive use of hash tables to
- lookup symbols. The relocation is 5 times faster under Bochs, and 2
- times faster on a K7 800 Mhz.
+ lookup symbols. The relocation is 5 times faster under Bochs, and 2
+ times faster on a K7 800 Mhz.</li>
+
<li>Complete rewrite of mod_check, which now use the loader code,
- and allow to generate dependency graphs for all modules, and to
- generate symbol table for Bochs debugger.
+ and allow to generate dependency graphs for all modules, and to
+ generate symbol table for Bochs debugger.</li>
+
<li>Minor modifications to allow the compilation of KOS under
- FreeBSD. Be careful to compile Bochs with BX_CPU_LEVEL=6, because
- FreeBSD gcc generate cmov in the compiled code.
- <li>The number of init/cleanup level is now unlimited.
+ FreeBSD. Be careful to compile Bochs with BX_CPU_LEVEL=6, because
+ FreeBSD gcc generate cmov in the compiled code.</li>
+
+ <li>The number of init/cleanup level is now unlimited.</li>
+
<li>Kos doesn't work for the moment on computer with more than 512
- Mb RAM, because of a bug.
- </ul>
-</li>
+ Mb RAM, because of a bug.</li>
+
+</ul>
+
+<h2>May 2002</h2>
+
+<p><i>A coding day with Julien and Thomas.</i></p>
-<li><b>May 2002 :</b>
-A coding day with Julien and Thomas.
<ul>
<li>Improvement of the console driver : multi console support,
- scrolling support, etc..
+ scrolling support, etc..</li>
+
<li>Interaction between the keyboard driver and the console driver is
- working.
- <li>Check function in _tower_mount was desactivated because of a bug
+ working.</li>
+
+ <li>Check function in _tower_mount was desactivated because of a
+ bug</li>
+
<li>Complete reorganization of the kos-sys module : now we got a
- libtty, a libfile and a libc.
+ libtty, a libfile and a libc.</li>
+
<li>Long reflexion concerning the problem of the interfaction between
the libc and the kernel. In the kernel, we provide methods in a
oriented object manner, but the libc (as Unixes implement it) doesn't
work in a object oriented manner. So we need to code libraries in
CPL3 that do understand this OO interfaces provided by the kernel,
and then implement a glue libc to make traditionnal Unix programs
- working.
+ working.</li>
+
<li>The user program currently makes read() on the tty, and then
- write() the result to the console : allmost a shell !
+ write() the result to the console : allmost a shell !</li>
+
<li>We unzipped the gcc and libc sources, just to see what is needed
- to be done... Lot of things thus !
+ to be done... Lot of things thus !</li>
+
</ul>
-</li>
-<li><b>April 2002 :</b>A coding Week-End with David, Julien
+<h2>April 2002</h2>
+
+<p><i>A coding Week-End with David, Julien
and Thomas, followed of a 5 days coding says between Julien and
-Thomas.<br>
-<ul><li>Implementation of the <i>anon</i> Babel interface for the
- mapping of anonymous regions (/dev/null mappings). Tests
- realized from a CPL3 application, with demand paging.
- <li>Clean and definitive integration of the <i>libhash</i> module,
- which allow creation of hash tables.
- <li>Complete rework of the name space management in Babel. Shadow
- resources, that are (and were) listed by translators, are now
- also organized in a tree, representing the arborescence. New
- functions to mount/umount filesystems. creation of a
- <i>fakefs</i> in order to have a root filesystem at startup.
- <li>Securisation of Babel concerning synchronisation problems (use
- of spinlock).
- <li>Extension of the <i>kitc</i> module : semaphores and wait
- queues.
- <li>Rework of <i>fat</i> module in order to make it work with the
- new way of managing the name space.
- <li>Automatic detection of the filesystem type.
- <li>Rework of the way to mount/umount a partition.
- <li>Creation of a minimal ELF loader.
- <li>Improvement of the user libc.
- <li>Creation of a real syscall system, with multiple arguments.
- <li>Creation of a <i>libcmdline</i> used to analyse arguments
- given to Grub.
- <li>Bug tracking.
+Thomas.</i></p>
+
+<ul>
+
+ <li>Implementation of the <i>anon</i> Babel interface for the
+ mapping of anonymous regions (/dev/null mappings). Tests realized
+ from a CPL3 application, with demand paging.</li>
+
+ <li>Clean and definitive integration of the <i>libhash</i> module,
+ which allow creation of hash tables.</li>
+
+ <li>Complete rework of the name space management in Babel. Shadow
+ resources, that are (and were) listed by translators, are now also
+ organized in a tree, representing the arborescence. New functions to
+ mount/umount filesystems. creation of a <i>fakefs</i> in order to
+ have a root filesystem at startup.</li>
+
+ <li>Securisation of Babel regarding synchronisation problems (use
+ of spinlock).</li>
+
+ <li>Extension of the <i>kitc</i> module : semaphores and wait
+ queues.</li>
+
+ <li>Rework of <i>fat</i> module in order to make it work with the
+ new way of managing the name space.</li>
+
+ <li>Automatic detection of the filesystem type.</li>
+
+ <li>Rework of the way to mount/umount a partition.</li>
+
+ <li>Creation of a minimal ELF loader.</li>
+
+ <li>Improvement of the user libc.</li>
+
+ <li>Creation of a real syscall system, with multiple arguments.</li>
+
+ <li>Creation of a <i>libcmdline</i> used to analyse arguments given
+ to Grub.</li>
+
+ <li>Bug tracking.</li>
+
</ul>
-</li>
-<li><b>March 2002 :</b>
-<ul><li>Integration of Babel latest version
- <li>Modules <i>vmm</i> and <i>ide</i> are now working with the new
- Babel.
- <li>Creation of the <i>part</i> module (partition management),
- which is now working with the new Babel.
+<h2>March 2002</h2>
+
+<ul>
+
+ <li>Integration of Babel latest version</li>
+
+ <li>Modules <i>vmm</i> and <i>ide</i> are now working with the new
+ Babel.</li>
+
+ <li>Creation of the <i>part</i> module (partition management), which
+ is now working with the new Babel.</li>
+
</ul>
-</li>
-<li><b>End of Februar 2002 :</b>
-<ul><li>Improvement of IDE and FAT drivers.
- <li>Load in memory and execution under KOS of a binary compiled
- and linked under Linux. It's the first real program which is
- executing under KOS.
- <li>New version of <i>rmap</i> to solve deadlock problems.
- <li>It's now possible to map a page in the user space of any team
- from any team (<i>map foreign</i>)
+<h2>End of Februar 2002</h2>
+
+<ul>
+
+ <li>Improvement of IDE and FAT drivers.</li>
+
+ <li>Load in memory and execution under KOS of a binary compiled and
+ linked under Linux. It's the first real program which is executing
+ under KOS.</li>
+
+ <li>New version of <i>rmap</i> to solve deadlock problems.</li>
+
+ <li>It's now possible to map a page in the user space of any team
+ from any team (<i>map foreign</i>)</li>
+
</ul>
-</li>
-<li><b>Februar 2002 :</b>One month of coding for Thomas, one Week-End with Julien, and another Week-End with David.<br>
+<h2>Februar 2002</h2>
+
+<p><i>One month of coding for Thomas, one Week-End with Julien, and
+another Week-End with David.</i></p>
+
+<p>During his coding month, Thomas did the following things :</p>
+
+<ul>
+
+ <li>Implementation of an IDE driver, which detects hard drives and
+ CD-ROM drivers, and which is able to read a hard drive.</li>
+
+ <li>Coding of a very simple FAT driver</li>
+
+ <li>Creation independent teams, with different address spaces.</li>
-During his coding month, Thomas did the following things :
-<ul><li>Implementation of an IDE driver, which detects hard drives and
- CD-ROM drivers, and which is able to read a hard drive.
- <li>Coding of a very simple FAT driver
- <li>Creation independent teams, with different address spaces.
</ul>
-During the Week-End with Julien :
-<ul><li>Integration in the KOS source tree of a new version of Babel.
- <li>Begininng of IDE driver integration to Babel with an interface.
+<p>During the Week-End with Julien :</p>
+
+<ul>
+
+ <li>Integration in the KOS source tree of a new version of
+ Babel.</li>
+
+ <li>Begininng of IDE driver integration to Babel with an
+ interface.</li>
+
</ul>
-During the Week-End with David :
-<ul><li>Creation of <i>post_init</i> levels for the
+<p>During the Week-End with David :</p>
+
+<ul>
+
+ <li>Creation of <i>post_init</i> levels for the
initialization. This allows to have init functions in
<i>INIT_LEVELS</i>, that are executed with interrupts
disabled, and init functions in <i>POST_INIT_LEVELS</i>,
executed with interrupts enabled. The initialization is
cleaner, and wolfgang has less things to do. The
<i>boostrap</i> module is now called <i>init</i>, since it is
- not only a bootstrap.
+ not only a bootstrap.</li>
<li>Rework of <i>arch/mm</i> and <i>vmm</i> in order to cleanly
separate archi dependant code, and arch independant code. For
instance, <i>arch/mm</i> contains functions such as
arch_map_virtual_page, arch_unmap_virtual_page, and <i>vmm</i>
contains wrappers to these functions, called map_virtual_page
- et unmap_virtual_page.
+ et unmap_virtual_page.</li>
<li>Tests concerning the teams : user spaces are really separated,
and we make sure the synchronization of the kernel space is
- made.
+ made.</li>
<li>It's now possible to create kernel threads in any team from
any team : a new parameter in the create_kernel_thread
- function indicates the destination team.
+ function indicates the destination team.</li>
<li>Implementation of the <i>reverse mapping</i>, which will speed
up swapping, allow us to move physical pages with no problem
(cool for DMA). We don't need an anonymous pages management
- with reverse mapping.
- <li>Complete rework of the physical memory management.
- <li>Mapping of the kernel area in one region 4k-2G.
+ with reverse mapping.</li>
+ <li>Complete rework of the physical memory management.</li>
+ <li>Mapping of the kernel area in one region 4k-2G.</li>
<li>New flag in slabs to indicate if pages used for the slab are
- swappable or not.
+ swappable or not.</li>
</ul>
-</li>
+<h2>November 2001</h2>
-<li><b>November 2001 :</b>KOS Week-End with David, Julien and
-Thomas
+<p><i>KOS Week-End with David, Julien and Thomas</i></p>
+
+<ul>
+ <li>New version of the VMM model, simpler than what it was : no
+ more virtual regions drivers, but shadow resource
+ drivers.</li>
-<ul><li>New version of the VMM model, simpler than what it was : no
- more virtual regions drivers, but shadow resource drivers.
<li>Improvement of debugging facilities : now, we can
activate/unactivate each debugging method (serial line,
console, bochs) during execution, and not only at compile
- time.
+ time.</li>
+
<li>Reflexion concerning a pseudo compiler <b>K</b> to ease driver
writing in KOS. Our tries with <i>ocamllex</i> and
- <i>ocamlyacc</i> weren't really sucessful.
+ <i>ocamlyacc</i> weren't really sucessful.</li>
+
<li>Modules <i>task-x86</i>,<i>mm-x86</i> and <i>lib-x86</i> have
been moved into a directory called <i>x86</i>. A symbolic link
- <i>arch</i> point to <i>x86</i> by default.
+ <i>arch</i> point to <i>x86</i> by default.</li>
+
</ul>
-</li>
-<li><b>September-October 2001 :</b>
-<ul><li>Kos compiles again ! (we couldn't compile it before, because of VMM work).
- <li>New Web site in PHP
+<h2>September-October 2001</h2>
+
+<ul>
+ <li>Kos compiles again ! (we couldn't compile it before, because of
+ VMM work).</li>
+
+ <li>New Web site in PHP</li>
+
</ul>
-</li>
-<li><b>August 2001 :</b>KOS week with David, Thomas and Julien :
+<h2>August 2001</h2>
+
+<p><i>KOS week with David, Thomas and Julien</i></p>
+
<ul>
- <li>Reorganisation of modules such as scheduler, task-x86 and idt.
- <li>Support of DSR (Deffered Service Routine) and DST (Differed Service Thread).
- <li>Initialization through a new module called bootstrap (multi-level init. and deletion of the identity mapping before launching the kernel).
- <li>New kmem using the Bonwick algorithm.
- <li>Liblist is written to manage linked lists.
- <li>Bugs fixes, SMP is anticipated.
- <li>Huge brainstorming about VMM and Babel, Babel is rewritten in different smaller modules.
+ <li>Reorganisation of modules such as scheduler, task-x86 and
+ idt.</li>
+
+ <li>Support of DSR (Deffered Service Routine) and DST (Differed
+ Service Thread).</li>
+
+ <li>Initialization through a new module called bootstrap
+ (multi-level init. and deletion of the identity mapping before
+ launching the kernel).</li>
+
+ <li>New kmem using the Bonwick <i>slab</i> algorithm.</li>
+
+ <li>Liblist is written to manage linked lists.</li>
+
+ <li>Bugs fixes, SMP is anticipated.</li>
+
+ <li>Huge brainstorming about VMM and Babel, Babel is rewritten in
+ different smaller modules.</li>
+
</ul>
-</li>
-<li><b>May-July 2001 :</b>
-<ul><li>The kernel memory allocateur <i>kvalloc</i> is ready.
-<li>Minimal serial line driver (ony output) to ease debugging on real computer.
-<li>Bug tracking in physical memory management.
+<h2>May-July 2001</h2>
+
+<ul>
+ <li>The kernel memory allocateur <i>kvalloc</i> is ready.</li>
+
+ <li>Minimal serial line driver (ony output) to ease debugging on
+ real computer.</li>
+
+ <li>Bug tracking in physical memory management.</li>
+
</ul>
-</li>
-<li><b>April 2001 :</b>KOS Week with David, Julien and Thomas :
-<ul><li>Babel syscall implementation
-<li></i>Usleep</i>, semaphores and <i>queues messages</i> implementation.
-<li>Virtual memory kernel execution (not in identity mapping).
-<li>Display of all symbols adresses while loader execution for debugging facilities.
-<li>Double Fault Handler implementation in order to manage kernel stack overflows.
-<li><i>Kgc</i> first appearance : the Kernel Garbage Collector.
-<li>8042 driver in order to manage PS/2 mice and keyboards
-<li>Tiny keyboard driver
-<li>Sources reorganisation and a lot of bug tracking !
+<h2>April 2001</h2>
+
+<p><i>KOS Week with David, Julien and Thomas.</i></p>
+
+<ul>
+
+ <li>Babel syscall implementation</li>
+
+ <li></i>Usleep</i>, semaphores and <i>queues messages</i>
+ implementation.</li>
+
+ <li>Virtual memory kernel execution (not in identity mapping).</li>
+
+ <li>Display of all symbols adresses while loader execution for
+ debugging facilities.</li>
+
+ <li>Double Fault Handler implementation in order to manage kernel
+ stack overflows.</li>
+
+ <li><i>Kgc</i> first appearance : the Kernel Garbage Collector.</li>
+
+ <li>8042 driver in order to manage PS/2 mice and keyboards</li>
+
+ <li>Tiny keyboard driver</li>
+
+ <li>Sources reorganisation and a lot of bug tracking !</li>
+
</ul>
-<li><b>March - April 2001 :</b>Babel integration in a module, double fault ok and ready to go !
-<li><b>February 2001 :</b>KOS Week-End with David, Julien and Thomas
-<ul><li>The Kernel level multithreading works, except stack overflows (double fault).
-<li>Babel developpement is going on.
-<li>Bug tracking
-<li>Double fault experimentation starts
+
+<h2>March - April 2001</h2>
+
+<p>Babel integration in a module, double fault ok and ready to go
+!</p>
+
+<h2>February 2001</h2>
+
+<p><i>KOS Week-End with David, Julien and Thomas</i></p>
+
+<ul>
+ <li>The Kernel level multithreading works, except stack overflows
+ (double fault).</li>
+
+ <li>Babel developpement is going on.</li>
+
+ <li>Bug tracking</li>
+
+ <li>Double fault experimentation starts</li>
+
</ul>
-</li>
-<li><b>January 2001 :</b>Tiny KOS Week-end for the new millenium with David and Thomas
-<ul><li>A kind of virtual memory management (add,suppr and virtual regions modifications).
-<li>Allocation/disallocation of CPL0 stacks
-<li>Kernel thread creations
-<li>Page fault management
-<li>Minor bugs tracking
-<li>The Kernel level multithreading doesn't work yet
+<h2>January 2001</h2>
+
+<p><i>Tiny KOS Week-end for the new millenium with David and
+Thomas</i></p>
+
+<ul>
+
+ <li>A kind of virtual memory management (add,suppr and virtual
+ regions modifications).</li>
+
+ <li>Allocation/disallocation of CPL0 stacks</li>
+
+ <li>Kernel thread creations</li>
+
+ <li>Page fault management</li>
+
+ <li>Minor bugs tracking</li>
+
+ <li>The Kernel level multithreading doesn't work yet</li>
+
</ul>
-</li>
-<li><b>Halloween 2000 :</b>KOS Week-end with David, Julien,
-Jerome and Thomas. Thomas' memory allocator is ready. Intense
-discussions about Babel.
-</li>
+<h2>Halloween 2000</h2>
-<li><b>October 2000 :</b>David and Thomas upgraded the loader
-(multiples modules types management) and optimized it (the link time
-is divided by 3).
-</li>
+<p><i>KOS Week-end with David, Julien, Jerome and Thomas.</i></p>
-<li><b>August 2000 :</b>Two programming weeks with Julien,
-David and Thomas which bear their fruits : the loader is complete. We
-do not use our own bootsecteur, but instead we use the <i>Grub</i>
-<i>bootloader</i>. Indeed, <i>Grub</i> allows to load a kernel (in our
-case, it's a <i>loader</i>) and moules from file system located either
-on a floppy disk or on an hard driver, or from the network. Moreover,
-<i>Grub</i> switches directly to protected mode before executing the
-kernel. The boot phase now go through the following steps : loading of
-the <i>loader</i> by Grub, loading of the modules by <i>Grub</i>,
-switch to protected mode by <i>Grub</i>, start of the loder, switch to
-pagined mode by the loader, modules relocation, and finally execution
-of the real kernel itself.
-</li>
+<p>Thomas' memory allocator is ready. Intense discussions about
+Babel.</p>
-<li><b>Spring 2000 :</b>Restart from scratch of the project,
-on saner bases, with a better architecture. Kos will be a modular
-system : the kernel itself will be split in different modules
-(scheduler, task, vmm, pmm ...) which will be linked together at boot
-time by a <i>loader</i></li>
+<h2>October 2000</h2>
-<li><b>12 - 14 Feb 2000 :</b> Second KOS Meeting, still at
-Chateau de Meridon. One more time, three intense days, less
-programming than in the first meeting, but a lot of
-chat. Unfortunately, we do not have pictures to show you ... :(</li>
+<p>David and Thomas upgraded the loader (multiples modules types
+management) and optimized it (the link time is divided by 3).</p>
-<li><b>October, 29th - November, 1st 99 :</b> First KOS
-meeting at Chateau de Meridon. A wonderful moment for us. Pictures
-right <a href="./photos.html">there</a> !</li>
+<h2>August 2000</h2>
-<li><b>October, 5th 1999 :</b> New design for KOS-web !</li>
+<p><i>Two programming weeks with Julien, David and Thomas</i></p>
-<li><b>September, 29th 1999 :</b> A new page, as important as
-useful for the KOS-cvs use has been created <a
-href="/cgi-bin/cvsweb">ici</a>.</li>
+<p>The loader is complete. We do not use our own bootsecteur, but
+instead we use the <i>Grub</i> <i>bootloader</i>. Indeed, <i>Grub</i>
+allows to load a kernel (in our case, it's a <i>loader</i>) and moules
+from file system located either on a floppy disk or on an hard driver,
+or from the network. Moreover, <i>Grub</i> switches directly to
+protected mode before executing the kernel. The boot phase now go
+through the following steps : loading of the <i>loader</i> by Grub,
+loading of the modules by <i>Grub</i>, switch to protected mode by
+<i>Grub</i>, start of the loder, switch to pagined mode by the loader,
+modules relocation, and finally execution of the real kernel itself.</p>
-<li><b>September, 20th 1999 :</b> A new version available on
-KOS-cvs ! You can compile it under DOS or Linux from the same
-Makefile.<br> Keyboard driver, a conio.c version wihtout bugs and a
-reactualized bootsector (bigger kernel suport).</li>
+<h2>Spring 2000</h2>
-<li><b>September, 15th 1999 :</b> A new record for KOS : 59
-persons registered on KOS-ml !</li>
+<p>Restart from scratch of the project, on saner bases, with a better
+architecture. Kos will be a modular system : the kernel itself will be
+split in different modules (scheduler, task, vmm, pmm ...) which will
+be linked together at boot time by a <i>loader</i></p>
-<li><b>Summer 1999 :</b> Work interruption because of the
-holidays.</li>
+<h2>12 - 14 Feb 2000</h2>
-<li><b>May, June 1999 :</b> Project reorganisation, we are now
-30 to participate to KOS.</li>
+<p><i>Second KOS Meeting, still at <a
+href="http://www.meridon.com">Chateau de Meridon</a></i></p>
-<li><b>18/04/1999 :</b> Coffeeman and Emmanuel Marty succeded
-in making a 32bits kernel works from the KOS bootsector. DJGPP is now
-used for the KOS compilation.</li>
+<p>One more time, three intense days, less programming than in the
+first meeting, but a lot of chat. Unfortunately, we do not have
+pictures to show you ... :(</p>
-<li><b>09/04/1999 :</b> Thanks to new developpers, Thomas
-Lorblanches succeded in runing a kernel in protected mode : a giant
-step for KOS, a tiny step for us :)</li>
+<h2>October, 29th - November, 1st 99</h2>
-<li><b>06/04/1999 :</b> Julien Munier created the
-mailing-list. Contribution call on OS-Rings and NG.</li>
+<p><i>First KOS meeting at <a href="http://www.meridon.com">Chateau de
+Meridon</a>.</i></p>
-<li><b>03/04/1999 :</b> Dimitri Ara officialy abandon the
-project. He'll be greatly missed.</li>
+<p>A wonderful moment for us. Pictures right <a
+href="./photos.php">there</a> !</p>
-<li><b>02/03/1999 :</b> Julien Munier is entering the project,
-by a message send on LVI</li>
+<h2>October, 5th 1999</h2>
-<li><b>25/02/1999 :</b> C kernel execution, linked with JLOC</li>
+<p> New design for KOS-web !</p>
-<li><b>15/02/1999 :</b> Stack declaration by the bootsector,
-ASM 16bits kernel execution.</li>
+<h2>September, 29th 1999</h2>
-<li><b>04/01/1999 :</b> Processor detection (386+ is needed)</li>
+<p> A new page, as important as useful for the KOS-cvs use has been
+created <a href="/cgi-bin/cvsweb">ici</a>.</p>
-<li><b>16/12/1998 :</b> First Bootsector displaying "Loading
-KOS ..." (version 0.0.1)</li>
+<h2>September, 20th 1999</h2>
-<li><b>11/10/1998 :</b> First valid Master Boot Record (MBR)</li>
+<p> A new version available on KOS-cvs ! You can compile it under DOS
+or Linux from the same Makefile.</p>
-</ul>
+<p>Keyboard driver, a conio.c version wihtout bugs and a reactualized
+bootsector (bigger kernel suport).</p>
+
+<h2>September, 15th 1999</h2>
+
+<p>A new record for KOS : 59 persons registered on KOS-ml !</p>
+
+<h2>Summer 1999</h2>
+
+<p>Work interruption because of the holidays.</p>
+
+<h2>May, June 1999</h2>
+
+<p>Project reorganisation, there are now 30 persons involved in KOS.</p>
+
+<h2>18/04/1999</h2>
+
+<p>Coffeeman and Emmanuel Marty succeded in making a 32bits kernel
+works from the KOS bootsector. DJGPP is now used for the KOS
+compilation.</p>
+
+<h2>09/04/1999</h2>
+
+<p> Thanks to new developpers, Thomas Lorblanches succeded in runing a
+kernel in protected mode : a giant step for KOS, a tiny step for us
+:)</p>
+
+<h2>06/04/1999</h2>
+
+<p>Julien Munier created the mailing-list. Contribution call on
+OS-Rings and NG.</p>
+
+<h2>03/04/1999</h2>
+
+<p>Dimitri Ara officialy abandon the project. He'll be greatly
+missed.</p>
+
+<h2>02/03/1999</h2>
+
+<p>Julien Munier is entering the project, by a message send on
+LVI</p>
+
+<h2>25/02/1999</h2>
+
+<p>C kernel execution, linked with JLOC</p>
+
+<h2>15/02/1999</h2>
+
+<p>Stack declaration by the bootsector, ASM 16bits kernel
+execution.</p>
+
+<h2>04/01/1999</h2>
+
+<p>Processor detection (386+ is needed)</p>
+
+<h2>16/12/1998</h2>
+
+<p>First Bootsector displaying "Loading KOS ..." (version 0.0.1)</p>
+
+<h2>11/10/1998</h2>
+
+<p>First valid Master Boot Record (MBR)</p>
<?
end_content_section();
Index: histo_fr.i
===================================================================
RCS file: /var/cvs/kos/kos-web/contents/histo_fr.i,v
retrieving revision 1.13
retrieving revision 1.14
diff -u -d -r1.13 -r1.14
--- histo_fr.i 29 Dec 2004 21:27:37 -0000 1.13
+++ histo_fr.i 29 Dec 2004 21:57:09 -0000 1.14
@@ -582,8 +582,8 @@
<h2>12 - 14 Fév 2000</h2>
-<p><i>Seconde Assemblée KOS, toujours
-au Château de Méridon.</i></p>
+<p><i>Seconde Assemblée KOS, toujours au <a
+href="http://www.meridon.com">Château de Méridon</a>.</i></p>
<p>Encore une fois trois jours intenses, moins de programmation mais
beaucoup de débats. Malheureusement pas de photos cette fois ci
@@ -591,10 +591,10 @@
<h2>29 - 1er Nov 99</h2>
-<p><i>Première Assemblée KOS au Château
-de Méridon.</i></p>
+<p><i>Première Assemblée KOS au <a
+href="http://www.meridon.com">Château de Méridon</a>.</i></p>
-<p>Un très bon moment, les photos <a href="./photos.html">ici</a>!</p>
+<p>Un très bon moment, les photos <a href="./photos.php">ici</a>!</p>
<h2>5 Octobre 1999</h2>
More information about the Kos-cvs
mailing list