Hello,
My project is not a real project in sens that it's only for fun.
My main goal is to provide a little (R.T. if possible) Unix like O.S. as QNX or Plan9 on top of L4. I start on top of a micro kernel write in C, but if an Ada implementation exists (and works ;)), I will choose it will be choosen in priority.
For now a study design concept of the O.S. By now, I follow exactly the User Guide Manual, with the typical OS example.
I've created a sourceforge site & M.Ls. (search ADX). There is no traffic actually.
Best regards ..
Frédéric.
___________________________________________________________ Do You Yahoo!? -- Un e-mail gratuit @yahoo.fr ! Yahoo! Courrier : http://fr.mail.yahoo.com
On Wed, 26 Sep 2001, Frédéric BOYER wrote:
Hello,
My project is not a real project in sens that it's only for fun.
My main goal is to provide a little (R.T. if possible) Unix like O.S. as QNX or Plan9 on top of L4. I start on top of a micro kernel write in C, but if an Ada implementation exists (and works ;)), I will choose it will be choosen in priority.
I'm afraid we do not have an Ada implementation right now, we are looking into developing one (see below).
Our "great plan" is to create what we call a "partitioned OS": we have customers (mostly in the aerospace industry) who want to run several applications on different APIs in a single machine, where some applications are very safety-critical (plane will crash if they fail), others are not (or to a lesser extent). Any safety-critical code that is going to be used in a plane must be certified (there is a standard called DO-178, maybe you heard of it) and depending on the potential damage caused by a failure, the required level of certification (and thus the effort it takes to certify) differs.
Now, since we want to run code that has to be certified at different levels in a single machine, we need to make sure they they can not interfere with each other. We think that L4 is an ideal foundation for such a system, however, since we have to certify the microkernel at the highest of all levels, we will use our own implementation "P4" which has been designed for good testability.
Ada will be one of several subsystems we want to implement (the most important one for our customers, though).
All this is currently in a planning phase. The actual work on Ada is not going to start before February next year. A Student has agreed to do it as his diploma thesis. So far, all we have is P4 implemented on MIPS machines and an implementation of OSEK OS that runs on top of it (OSEK is a standard API used in car electronics equipment).
For now a study design concept of the O.S. By now, I follow exactly the User Guide Manual, with the typical OS example.
I think it would be a good idea if we could keep in touch, exchange ideas etc. Would you mind if I pass your email address to the student who will be doing the Ada implementation ?
I've created a sourceforge site & M.Ls. (search ADX). There is no traffic actually.
I will keep an eye on it.
Cheers
Rob
---------------------------------------------------------------- Robert Kaiser email: rkaiser@sysgo.de SYSGO RTS GmbH Am Pfaffenstein 14 D-55270 Klein-Winternheim / Germany fax: (49) 6136 9948-10
I'm afraid we do not have an Ada implementation right now, we are looking into developing one (see below).
No pb, just want to create a GNURL version which is a L4 compatible with ABI. When P4 will be ready, we'll switch from fiasco to L4 very easy.
Our "great plan" is to create what we call a "partitioned OS": we have customers (mostly in the aerospace industry) who want to run several applications on different APIs in a single machine, where some applications are very safety-critical (plane will crash if they fail), others are not (or to a lesser extent). Any safety-critical code that is going to be used in a plane must be certified (there is a standard called DO-178, maybe you heard of it) and depending on the potential damage caused by a failure, the required level of certification (and thus the effort it takes to certify) differs.
Now, since we want to run code that has to be certified at different levels in a single machine, we need to make sure they they can not interfere with each other. We think that L4 is an ideal foundation for such a system, however, since we have to certify the microkernel at the highest of all levels, we will use our own implementation "P4" which has been designed for good testability.
Ada will be one of several subsystems we want to implement (the most important one for our customers, though).
All this is currently in a planning phase. The actual work on Ada is not going to start before February next year. A Student has agreed to do it as his diploma thesis. So far, all we have is P4 implemented on MIPS machines and an implementation of OSEK OS that runs on top of it (OSEK is a standard API used in car electronics equipment).
I think it would be a good idea if we could keep in touch, exchange ideas etc. Would you mind if I pass your email address to the student who will be doing the Ada implementation ?
Of course no pb. In fact, instead of read a quick & dirty port of the L4 library, we could port together th e GNURL on top of L4, so we could write servers very easely. With that tools, we can write a Unix like implementation or a partitioned OS.
Which plateform will you use to write P4 ? GNAT, x86 ?
Please fell free to give my email to your student.
best regards,
Frédéric
___________________________________________________________ Do You Yahoo!? -- Un e-mail gratuit @yahoo.fr ! Yahoo! Courrier : http://fr.mail.yahoo.com
Hi Frederic
On Wed, 26 Sep 2001, Frédéric BOYER wrote:
Which plateform will you use to write P4 ? GNAT, x86 ?
Hmm, I'm afraid you misunderstood: P4 is our own microkernel. It implements the L4 API just like Fiasco and L4 (there are some small deviations, mostly because we have tried to keep machine dependencies out of the API). P4 currently exists for MIPS machines (R3000/R4000/R5000). Our plan is to use it as basis for the (GNAT) Ada implementation, so that should run on MIPS initially. We also plan to make a PowerPC version of P4 in parallel. Ideally, it should be possible to port the Ada runtime system to a different architecture with very little effort once the microkernel exists for that architecture, but that remains to be proven...
Please fell free to give my email to your student.
Thanks!
Cheers
Rob
---------------------------------------------------------------- Robert Kaiser email: rkaiser@sysgo.de SYSGO RTS GmbH Am Pfaffenstein 14 D-55270 Klein-Winternheim / Germany fax: (49) 6136 9948-10
l4-hackers@os.inf.tu-dresden.de