l4-hackers

l4-hackers@os.inf.tu-dresden.de

2 participants
2880 discussions

problem starting the hello server
by Christian Rehn 11 Jan '01

11 Jan '01

Hi, I,ve got a problem starting the hello server on top of fiasco. I checked out l4-fiasco from the cvs repository and compiled it. My menu.lst looks as follows: kernel=(nd)/boot/tftpboot/rehn/rmgr -sigma0 module=(nd)/boot/tftpboot/rehn/main -nojdb module=(nd)/boot/tftpboot/rehn/sigma0 module=(nd)/boot/tftpboot/rehn/hello When I try to boot l4 rmgr starts, loads the l4 kernel and the sigma0 server but does not load the hello server (I'm not sure if it is loaded but not executed or if it is not loaded but I can not find the code that should load the hello server). Where should the hello server be loaded? Why does the rmgr "exec_load" main and the sigma0 server but not the hello server in function startup() in startup.c? Thanks, Christian

1 0

IPC
by Jordan Curzon 03 Jan '01

03 Jan '01

Hello, What is the most common way for apps(or libs) to find the task id for a service(networking for example)? _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com

1 0

L4 benchmarks
by Greg Law 12 Dec '00

12 Dec '00

Hi, I would like to get some benchmarks for L4 on Pentium. Specifically, I am after the times to create a process and execute a getpid (preferably in cycles). Does anyone know where I can get such info before I try and go through the pain of measuring it myself? Cheers, Greg -- Greg Law, email: gel(a)soi.city.ac.uk Dept. Of Computing, web: http://www.soi.city.ac.uk/~gel/ City University, phone: +44 20 7477 8341 London, UK. EC1V 0HB. fax: +44 20 7477 8587

1 0

Missing files in recent cvs versions
by David Welch 16 Nov '00

16 Nov '00

Hi, Some of the Makefiles in fiasco want to use Makeconf.lib.targets and Makeconf.lib.rules which don't seem to be in the cvs repository. Also the Fiasco bug tracking link gives a 404.

2 2

Ideas for some sample L4 root_task(s) [Learning L4].
by Farid Hajji 25 Oct '00

25 Oct '00

Hi Brent, this is what I'm planning to do next to get used to the L4-Ka interface. If you are interested, let's do this together. Basically it's about exercising misc. L4 syscalls in the root_task itself. Once we understand this, we can grab libmom and add L4 sysdeps to glibc and the Hurd. 01. map from sigma0 the kernel info page into the address space of the root_task. printf() its contents, especially its 'clock' field which is updated by the ukernel. sleep some useconds and printf clock again. [-> learn more about sigma0 rpc protocol, how to send/receive a message to/from sigma0, some memory managment issues in general and l4_sleep(). ] 02. activate some threads in root_task, and let each thread continouusly putc() a unique identifying character to the screen. look at the resulting pattern. A control thread could getc() chars from keyboard and use this char to start/stop the thread that prints this char. [-> learn about threads and scheduling.] Add a pager to a thread which just printf()s each page request, before forwarding it to sigma0. Alternatively, trap the request and handle it by mapping an own page to the requester. Finally, reclaim pages that are not needed anymore (how? and how to get notified about this condition?). [-> learn about pagers and the L4 application/pager protocol.] 03. create new (child) tasks and: * map/grant some pages * send/receive messages: short message, long message, with and without string dopes etc.. * synchronize threads via blocking l4-receive syscalls. [-> learn about messages, ipc calls.] 04. understand what rmgr really does, what its interfaces to the root_task and other tasks are (a.k.a. how to request/receive ressources from rmgr) and try to implement Okuji's idea about relocation of modules (glinux) so that we can boot glinux with vanilla GNU Grub (no modaddr). 05. Attach a root_task thread to one interrupt number (say com1, clock, keyboard or whatever) and printf/acknowledge each received interrupt. [-> Handling interrupts. Is it possible to get an interrupt that is already bound by rmgr? Hmmm... ] [-> What about chaining interrupts/events? ] Write a trivial device driver server (like sorta translator), e.g. a clock device, that gets clock interrupts and reads the clock page in the kernel to update its internal notion of time. Add protocol so that other threads/tasks can ask the time via RPC (simple message) from this clock server. [-> generalize this to device driver servers in general. Think about a Hurd abstraction to access such servers. Driver-Translators? Driver-IPC? ] 06. Grab the code of a real device driver from OSKit and try to turn it into a device driver server as in 05. 07. [...] Real world examples are located in RMGR source code, and in the sources of L4-Fiasco apps and L4Linux. I'll continue to use L4Ka right now though, simply because I can compile it without problems on my FreeBSD box ;-). Any suggestions? -Farid. -- Farid Hajji -- Unix Systems and Network Admin | Phone: +49-2131-67-555 Broicherdorfstr. 83, D-41564 Kaarst, Germany | farid.hajji(a)ob.kamp.net - - - - - - - - - - - - - - - - - - - - - - - + - - - - - - - - - - - - Murphy's Law fails only when you try to demonstrate it, and thus succeeds.

1 0

Re: RH6.2 and Fiasco
by Michael Hohmuth 24 Oct '00

24 Oct '00

James, are you still having problems with Fiasco? Ißm just now catching up with my email... James Riordan <rij(a)zurich.ibm.com> writes: > Now that it is built, it crashes at boot time: > > RMGR: loading (hd1,1)/fiasco/sigma0 > RMGR: detected new-style DD-L4 > RMGR: starting (hd1,1)/fiasco/main -nokdb proto=0x10128c > > Welcome to Fiasco! > DD-L4/x86 microkernel (c) 1998 TU Dresden - Jul 13 2000 > KDB: init > L4KD -BP----------------------------------------------EIP: f00011c9 > > (hang) > > Any further suggestions? This is the kernel-debugger prompt. Have you tried typing "g" (mnemonic: Go) here? Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Re: CVS server problems?
by Michael Hohmuth 24 Oct '00

24 Oct '00

suzukis(a)flavor1.ipc.hiroshima-u.ac.jp writes: > >did anyone tried to use flux' oskit instead of oskit10 with l4? > > I think Fiasco is (although slowly) moving from oskit to oskit10. Towards > to the newest Flux OSkit will be after it. In my previous building of Fiasco, > it was still dependent on both of oskit & oskit10. So, replacing oskit by > oskit10 seems to be long way. Is it so difficult work? No, it is not that difficult, but we haven't switched to oskit10 because of two reasons: First, the newer OSKit is GPL'd, and we have to maintain a version of Fiasco that is not under the GPL to be able to sell commercial licenses. Such licenses are one way for us to continue financing Fiasco development. That said, however, I agree that a port of Fiasco to the newer OSKit would be worthwile; the old OSKit has compilation problems with newer compilers and binutils. It just isn't an option to completely depend on it; it must be possible to compile Fiasco with older OSKit's as well. So I would like to see everything that's needed from the OSKit hidden in behind a module interface that can be implemented by either the old or the new OSKit. We just haven't got around to implement this kind of abstraction. Maybe one of you can contribute? Regards, Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Re: CVS server problems?
by Michael Hohmuth 24 Oct '00

24 Oct '00

L4 hackers, I'm slowly working my way up through my email archive... Sorry for not responding for so long -- I've been quite busy and had trouble keeping up with my email. Farid Hajji <farid.hajji(a)ob.kamp.net> writes: > I was able to compile flux' oskit on FreeBSD-5.X out of the box without > any problems (using gmake, gcc-2.95.2 and binutils-2.9.1). So there is > definitely something wrong with the CVS server at Dresden :-( The CVS-server problems have finally been fixed two weeks or so ago. (Our bug-tracking database is still in trouble... The problems are due to a change in our server architectue because of stability problems with our Linux SMP server.) Michael -- hohmuth(a)innocent.com, hohmuth(a)sax.de http://www.sax.de/~hohmuth/

1 0

RE: first experiences with l4ka on a real pentium
by Brent Fulgham 20 Oct '00

20 Oct '00

Farid, Nice work on the BOCHS write-up. I'll have to give it a try at some point. Anyway -- I also get the endless loop when booting from GRUB. I'll have to try the Dresden binary to see if this corrects the problem. If so, we'll have to see if Okuji or Gord would be willing to add whatever changes were necessary to the main GRUB tree. Since Okuji reads this mailing list, I'm hoping he'll do so (hint, hint ;-) Thanks, -Brent

2 1

L4KA mailing list
by Uwe Dannowski 20 Oct '00

20 Oct '00

Hi L4KA users! To take L4KA-specific discussions off l4-hackers, we've set up a dedicated mailing list for L4KA. To subscribe to the list, send mail containing only the word "subscribe" (without the qoutes) in the body to l4ka-request(a)ira.uka.de . To unsubscribe send the word "unsubscribe". Posting to the list is restricted to list members. Cheers, Uwe

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

l4-hackers