l4-hackers

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

2827 discussions

Portable servers?
by Alexey Mandrookin 20 Sep '02

20 Sep '02

Hi! How to write portable servers for L4 micro-kernel family? I found what the l4_threadid_t type is different in various L4 implementation. Does any wrapper on top of syscalls exist? ==== Fiasco microkernel: ==== typedef struct { unsigned version_low:10; unsigned lthread:7; unsigned task:11; unsigned version_high:4; unsigned site:17; unsigned chief:11; unsigned nest:4; } l4_threadid_struct_t; typedef union { l4_low_high_t lh; l4_threadid_struct_t id; } l4_threadid_t; typedef l4_threadid_t l4_taskid_t; === L4-KA microkernel: ==== #define L4_X0_VERSION_BITS 10 #define L4_X0_THREADID_BITS 6 #define L4_X0_TASKID_BITS 8 typedef union { struct { unsigned version : L4_X0_VERSION_BITS; unsigned thread : (L4_X0_THREADID_BITS + L4_X0_TASKID_BITS); unsigned chief : L4_X0_TASKID_BITS; } id; struct { unsigned version : L4_X0_VERSION_BITS; unsigned thread : L4_X0_THREADID_BITS; unsigned task : L4_X0_TASKID_BITS; unsigned chief : L4_X0_TASKID_BITS; } x0id; dword_t raw; } l4_threadid_t; ========================== Second question is how to distinguish between IRQ messages and ordinary messages. That code makes me confused: irq_th.lh.low = COM_IRQ + 1; irq_th.lh.high = 0; as I understand, it correspond to irq_th.id.version_low = COM_IRQ + 1; irq_th.id.lthread = 0; irq_th.id.task=0; irq_th.id.version_high=0; irq_th.id.site=0; irq_th.id.chief=0; irq_th.id.nest=0; Is lthread = 0 and task = 0 means interrupts? Third question is how to handle memory mapped IO? I found phystokv() function. It is just a simple macro, which placed deep inside of OSKIT Is that right to use that function for MMIO? What will happen if several servers requests same memory range simultaneously? Best regards Alexei

2 1

Clans&Chiefs question
by Jacob Gorm Hansen 16 Sep '02

16 Sep '02

hi, I have the following setup: - rmgr, fiasco, and sigma - a network server 'net' which exports an abstracted ethernet-device, built with oskit10. - a custom pager, 'pager' which is able to spawn new l4linux servers (currently piggybacked on to it). All started via entries in menu.lst, using Grub The spawned l4linuxes would like to talk to 'net' so that they may boot from NFS. However, net and pager are not in the same clan, so that is not allowed (obviously they need to talk to pager as well to get paged). Both net and pager need to talk to rmgr. Is there a way, other than merging net & pager into one task (which works fine, but is not really what I want), in which this can be made to work in fiasco? Thanks, Jacob

2 1

One more problem
by Alexey Mandrookin 13 Sep '02

13 Sep '02

Hi everybody! New problem appeared. It seems to be a make looking for subdirectories and did not found theirs. $gcc -v gcc version 2.95.3 20010315 (release) $make hello make[3]: Entering directory `/usr/src/Fiasco/l4/pkg/l4util/examples/watchdog' set -e ; for d in ; do \ make -C OBJ-$d all ; done /bin/sh: -c: line 1: syntax error near unexpected token `;' /bin/sh: -c: line 1: `set -e ; for d in ; do make -C OBJ-$d all ; done' make[3]: *** [all] Error 2 make[3]: Leaving directory `/usr/src/Fiasco/l4/pkg/l4util/examples/watchdog' make[2]: *** [all] Error 2 make[2]: Leaving directory `/usr/src/Fiasco/l4/pkg/l4util/examples' make[1]: *** [all] Error 2 make[1]: Leaving directory `/usr/src/Fiasco/l4/pkg/l4util' make: *** [hello] Error 2 Best regards Alexei

2 2

Re: Two questions
by Udo A. Steinberg 13 Sep '02

13 Sep '02

On Fri, 13 Sep 2002 13:33:06 +0900 Alexey Mandrookin (AM) wrote: [ Please reply to the l4-hackers list instead of sending private emails! ] AM> I downloaded latest CVS version again. Now I compile under Red Hat Linux AM> release 7.1 AM> gcc version 2.96 20000731 (Red Hat Linux 7.1 2.96-81) AM> New problem appear: Note that gcc 2.96 is NOT an official gcc release by the gcc team. Redhat got a development CVS gcc snapshot and released it as gcc 2.96 in their distribution. The gcc team states that the object format used by 2.96 is not compatible with neither 2.95 nor 3.x. Eventually Redhat people got a clue and shipped a working compiler (kgcc). I strongly recommend to use gcc 2.95.x or gcc 3.x for building the L4 tree. Alternatively try compiling with kgcc instead of gcc. AM> It is look like gcc version incompatibles, so I fixed that problem by AM> changing 'short' type to 'int'. AM> Guess is Ok, because the stack alignment is 'int'. AM> Next, I start make hello agian and one more problem appear: AM> ld -N -Ttext 0x200000 -o hello main.o AM> ../../../../lib/crt0_getopt.o -L../../../../lib/x86_586/l4v2 -L../../../../l AM> ib/x86_586 -L../../../../lib -L../../../../../oskit/lib -L/home/drops/lib -l AM> mc -loskit_support -lmc -llmm -lkern -ll4util AM> ../../../../lib/x86_586/l4v2/libl4util.a(getopt2.o): In function `arg_init': AM> /home/alman/work/L4/l4/pkg/l4util/lib/src/OBJ-x86_586-l4v2/../getopt2.c:54: AM> undefined reference to `strlen' AM> /home/alman/work/L4/l4/pkg/l4util/lib/src/OBJ-x86_586-l4v2/../getopt2.c:54: AM> undefined reference to `strlen' This also works with gcc 2.95 and gcc 3.2. If you install a decent compiler you'll save yourself a lot of headaches. -Udo.

1 0

Re: Two questions
by Adam Lackorzynski 12 Sep '02

12 Sep '02

Hi Alexey! On Thu Sep 12, 2002 at 18:27:18 +0900, Alexey Mandrookin wrote: > could you explain, where is I should send my mail, into > maillist or personally to you? Please use the mailing list. Also consider subscribing (see http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers). > Latest CVS version complation problem: > > cd l4 > make hello > ---------------------Start > here ---------------------------------------------- > make[2]: Entering directory `/home/alman/work/L4/oskit/examples' > ld -Ttext 100000 -L../lib \ > -o fdev_test ../lib/multiboot.o fdev_test.o \ > -lfdev -lfdev_linux -lkern -lmc -llmm -lfdev -llmm \ > ../lib/crtn.o > ../lib/libfdev.a(buf.o): In function `fdev_default_buf_copyin': > /home/alman/work/L4/oskit/libfdev/buf.c:35: undefined reference to `memcpy' > ../lib/libfdev.a(buf.o): In function `fdev_default_buf_copyout': > /home/alman/work/L4/oskit/libfdev/buf.c:45: undefined reference to `memcpy' > make[2]: *** [fdev_test] Error 1 > make[2]: Leaving directory `/home/alman/work/L4/oskit/examples' > make[1]: *** [examples/all.MAKE] Error 2 > make[1]: Leaving directory `/home/alman/work/L4/oskit' > make: *** [oskit] Error 2 > -----------------End here--------------------------------------------------- > > I was try to fix that problem by myself, but another error appears in other > places. I have no problems using gcc version 2.95.4 20011002 (Debian prerelease) and GNU ld version 2.12.90.0.1 20020307 Debian/GNU Linux, I just tried it and it built successfully. > ---------------------Start > here ---------------------------------------------- > title Hello world > root (hd0,4) > module /boot/fiasco/main -nokdb > module /boot/fiasco/sigma0 > module /boot/fiasco/hello > > title L4Linux > root (hd0,4) > kernel /boot/fiasco/rmgr -configfile -sigma0 > modaddr=0x03000000 > module /boot/fiasco/main -nowait > module /boot/fiasco/sigma0 > module /boot/fiasco/rmgr.cfg > module /boot/fiasco/glinux root=/dev/hda5 > -----------------End here--------------------------------------------------- > > RMGR, which included in source tree: l4/pkg/rmgr/server/examples/rmgr.cfg > > ---------------------Start > here ---------------------------------------------- > #!rmgr > > ### this is an example RMGR config file > > ### first, some global settings > > # small address spaces have a size of 64 MB > small_space_size 0x04000000 > > # wait for key press after initialization > bootwait > > ### configuration for Sigma0 and RMGR follows > > # NOTE: it's generally useless to specify any resource allocation > # constraints for Sigma0 and RMGR here: most will be ignored because > # these tasks statically allocate all of their resources. That's why > # we only specify bootstrap parameters here. > > task sigma0 > boot_priority 100 > > task rmgr > boot_priority 100 > # enable next line if you want RMGR to run in small address space # 3 > #boot_small 3 > > ### configuration for boot modules follows > > task modname "glinux" # make sure this entry > # corresponds to a Linux server > boot_small 2 # small address space # 2 > boot_priority 100 > boot_mcp 0x90 # maximum controlled priority > log_mcp 0x90 # logical maximum controlled priority > # -- useful if "boot_mcp 0" > child max 512 in [10, 1023] # at most 512 child tasks in that range > memory in [0, 0x02000000] # use only lower 32 MB RAM > high_memory in [0xc0000000, 0xffffffff] # use only hi memory in that range > small in [1,3] # allow these small address > # spaces to be allocated > # enable next line if you want to disallow allocation if IRQs 3 and 4 > # (IRQs for serial lines COM2 and COM1) > #irq mask 0xffffffe7 > > end > -----------------End here--------------------------------------------------- > > Linux kernel compiled with default parameters, i.e. "L4Linux options" was > not changed > > The error message is: > > ---------------------Start > here ---------------------------------------------- > SIGMA0: Hello! > RMGR: Hi there! > RMGR: running on FIASCO > RMGR: total RAM size = 130623 KB (reported by bootloader) > received 103476 LB RAM from sigma0 > 460 KB reserver for RMGR > RMGR: received no IO ports > RMGR: attached irqs = [ <!0> 1 <!2> <!3> 4 5 6 7 <!8> 9 a b c d e f ] > RMGR: starting task /boot/fiasco/glinux root=/dev/hda5 from > 0x03065000-0x031897dc > to [ 0x003ff000-0x004f0458 0x004f2000-0x0053db74 ] > RMGR: starting task /boot/fiasco/glinux root=/dev/hda5 (#0x5) > at entry 0x3ff000 via trampoline page code 0x54168 > > RMGR: task 5.3 at 00516d4a is trying to get page 00000000 which is reserved > RDB: double page fault > > --double page fault------------------------------------------EIP: 00122c50 > (4.00) jdb: > -----------------End here--------------------------------------------------- Just try using some simple setup first, something like: title L4Linux kernel (nd)/tftpboot/adam/l4/rmgr -sigma0 task modname "glinux" module modaddr 0x005A0000 module (nd)/tftpboot/adam/l4/kernel -nowait -serial_esc -serial -comspeed 115200 -comport 1 -nokdb module (nd)/tftpboot/adam/l4/sigma0 module (nd)/tftpboot/adam/l4l-2.2/glinux no-scroll root=/dev/hda5 Adam -- Adam adam(a)os.inf.tu-dresden.de Lackorzynski http://os.inf.tu-dresden.de/~adam/

2 1

Two questions
by Alexey Mandrookin 11 Sep '02

11 Sep '02

Dear gentleman, I am sorry for disturb. First of all, what's happen with Fiasco microkernel? Latest CVS version does not compiles, but two week old version is work Ok. Guess it is because Fiasco migrating to the X.2 version API specification. Is it so? Next question is, why L4Linux does not start? After running L4Linux, the microkernel complain with double page fault at page 0. Best regards Alexei

2 1

Problems compiling Fiasco
by Prasad 24 Aug '02

24 Aug '02

hi, i am new to fiasco. When trying to compile it using the sources from the CVS i am getting a lot of errors. I know that the type of errors are required for any guidance, but would like to know some details about the preferred compiler, OS etc: This would prove very helpful to me. Thanking you all for any help, Prasad. -- Failure is not an option

2 1

Usermode Fiasco, X.2 spec
by Andreas Rottmann 23 Aug '02

23 Aug '02

Hi! <disclaimer>Yes, I know I might be annoing</disclaimer> August has come, and I once again ask for the status of the Fiasco usermode port, as I am eager to try it. Furthermore, I'd like to know if there is/will be any effort to implement the X.2 (upcoming version 4) L4 API building on the Fiasco codebase (that would be quite interesting, since the HURD L4 port will use the X.2/4 API). Regards, Andy -- Andreas Rottmann | Dru@ICQ | 118634484@ICQ | a.rottmann(a)gmx.at http://www.8ung.at/rotty | GnuPG Key: http://www.8ung.at/rotty/gpg.asc Fingerprint | DFB4 4EB4 78A4 5EEE 6219 F228 F92F CFC5 01FD 5B62

2 3

Hello, world!
by Nick Name 04 Jul '02

04 Jul '02

Hi all, just wanted to introduce myself: I am an italian computer science student, and a newcomer in the µkernels stuff ... so I am trying fiasco (by now, still trying to understand bochs, to be able to test it). One of my wish is to experiment with functional programming (with Haskell) and µkernels to see if a standalone system could be mainly functional. But... here comes my first wish: User-mode-fiasco ! Is it already useable? thanks, Vincenzo

1 0

user-level port
by Andreas Rottmann 27 Jun '02

27 Jun '02

Hi! I hope this is the right place to ask. (If not, please point me there ;-)) I have read on the Fiasco homepage that Fiasco is beeing/was ported to run as a user-level process under Unix. What is the state of this project? And what is the mentioned CVS branch? However, I am interested in this. A nice way to go would be to port Fiasco to a386 (http://a386.nocrew.org/), IMO. I don't know if I can be of any help, since I haven't (yet) done much kernel-level hacking... Regards, Andy -- Andreas Rottmann | Dru@ICQ | 118634484@ICQ | a.rottmann(a)gmx.at http://www.8ung.at/rotty | GnuPG Key: http://www.8ung.at/rotty/gpg.asc Fingerprint | DFB4 4EB4 78A4 5EEE 6219 F228 F92F CFC5 01FD 5B62

3 5

Jump to page:

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