- l4-hackers - example.com Site

Re: Fiasco@ARM - supported hardware
by Daniel Vandersee 19 Jun '07

19 Jun '07

Hi Jorge, I just tried it and it doesn't work for me, too. If I break the program execution after a while it always stops at uart_write+72: > 0xb0002e30 <uart_write+64>: ldr r3, [r2, #32] > 0xb0002e34 <uart_write+68>: tst r3, #4 ; 0x4 > 0xb0002e38 <uart_write+72>: beq 0xb0002e30 <uart_write+64> > 0xb0002e3c <uart_write+76>: ldrb r3, [r0, r12] Did you got it running with skyeye? Attached you will find my skyeye configuration. Maybe you have an idea what goes wrong here? Thanks! Cheers, Daniel Jorge Torres schrieb: > Hi Daniel, > > Have you tried PXA borad with skyeye? > > Thanks, > > JOrge > > On 6/18/07, *Daniel Vandersee* < dvandersee(a)escrypt.com > <mailto:dvandersee@escrypt.com>> wrote: > > Hello, > > how do you test L4 on an ARM platform? Do you have specific hardware > for > that purpose? You mentioned that you use ARM Integrator boards. As there > are several integrator boards available: which one exactly do you use? > Do you use other boards as well? I'm asking because I plan to buy an > ARM > board which is supported out of the box without the need to modify > drivers. > > Do you additionally use software simulators to test L4? If so, which one > do you use? I tried Skyeye to simulate an ARM integrator board as well > as a SA1100-board but didn't get any output over the serial line. Qemu > doesn't work either and terminates with the following error: "qemu: > uncaught target signal 11 (Segmentation fault) - exiting". > > Kind regards, > Daniel > > > > _______________________________________________ > l4-hackers mailing list > l4-hackers(a)os.inf.tu-dresden.de <mailto:l4-hackers@os.inf.tu-dresden.de> > http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers > > cpu: xscale #cpu: pxa25x mach: pxa_lubbock #cpu: pxa27x #mach: pxa_mainstone mem_bank: map=I, type=RW, addr=0x40000000, size=0x0c000000 mem_bank: map=M, type=RW, addr=0xb0000000, size=0x10000000 #mem_bank: map=M, type=RW, addr=0xd0000000, size=0x00100000 #log: logon=1, logfile=/tmp/sk2.log, start=500000, end=1500000 #lcd:state=on,type=pxa,mod=gtk uart: mod=stdio

1 0

Re: L4Env Building Problem!
by Bjoern Doebel 14 Jun '07

14 Jun '07

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Marius.bao wrote: >> Please enter l4/pkg/bootstrap/server/src, run "make V=1" and post the >> complete output here. >> > When I run "make V=1", the make process stops. The output is as follows: > ../../../../mk/Makeconf:120: *** need to give builddir with > O=.../builddir. Stop. Sure, then please run make V=1 O=<path_to_your_builddir> and post the complete output here. > Ps:The ../../../../ directory is l4. > I couldn't compile a package separately. When I want to compile a > single package, I have to enter the build directory to run "make", and > compile all the packages. I suffered a lot as the compile process > takes a long time. You can build (parts of) a single package in at least 2 ways: 1. use the O=<builddir> option when running make just as the error message above tells you. 2. Go to the builddir. There will be a subdirectory pkg with build dirs for all the packages. You can enter these and only build a single one. Regards, Bjoern -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGcOTNP5ijxgQLUNkRAvWjAJ9ipIMpC2VA1et+nzf2JtJeqaWhLgCfRnb2 E0j6eVEgzO+AZVvOKDrSHL0= =c4iI -----END PGP SIGNATURE-----

1 0

Re: L4Env Building Problem!
by Marius.bao 13 Jun '07

13 Jun '07

Hi, >Could you find out which dice binary is used (run "make V=1" in one of >the idl directories, for instance tudos/l4/pkg/names/idl)? And then run >that binary with the --version option and post the output here? It turned out that i used an old version of DICE, I update to the latest version, and it works correctly. Thanks a lot :) when I compiled the bootstrap package, another problem occured: l4/pkg/bootstrap/server/src/startup.cc:215: undefined reference to `__cxa_guard_release' maybe my gcc is too old. my gcc version is gcc-3.4.2.6-fc3

2 1

L4Env Building Problem!
by Marius.bao 11 Jun '07

11 Jun '07

Hi, Thanks for previous information. I've checkouted l4env using: svn checkout -N http://svn.tudos.org/repos/tudos/trunk tudos cd l4 ./repomgr co l4env When I tried to compile l4env, errors occured in some idl file. eg: names.idl in names package import <l4/sys/types.h> import <l4/names/libnames.h> /** IDL interface of the name server. */ interface names { long register([in, string] char *name); long register_thread([in, string] char *name, [in] l4_threadid_t id); long unregister_thread([in, string] char *name, [in] l4_threadid_t id); long unregister_task([in] l4_threadid_t id); long query_name([in, string] char *name, [out] l4_threadid_t *id); long query_id([in] l4_threadid_t id, [out, string, prealloc_client, max_is(NAMES_MAX_NAME_LEN)] (*) char **name);modified ...... }; errors occured in the last function at the (*) line, the error information is as follows: unknown attribute dice: 1 Error(s) and 0 Warning(s) occured. when I modified the function into long query_id([in] l4_threadid_t id, [out, string] char **name); It compiles OK. I don't know what's the problem as I know little about the corba idl.

2 1

Re: l4-hackers Digest, Vol 50, Issue 7
by Erik Schierboom 08 Jun '07

08 Jun '07

> Message: 4 > Date: Fri, 8 Jun 2007 00:17:26 +0200 > From: Adam Lackorzynski <adam(a)os.inf.tu-dresden.de> > Subject: Re: Verify Fiasco assertion using pingpong > To: l4-hackers(a)os.inf.tu-dresden.de > Message-ID: <20070607221726.GN14071(a)os.inf.tu-dresden.de> > Content-Type: text/plain; charset=iso-8859-1 > > > On Thu Jun 07, 2007 at 13:17:09 +0200, eschierb(a)sci.kun.nl wrote: > >> Quoting Martin Pohlack <mp26(a)os.inf.tu-dresden.de>: >> >> >>> eschierb(a)sci.kun.nl wrote: >>> >>> [...] >>> >>> >>>> "/bin/bash: did_not_find_dice___please_install_dice: command not found" >>>> >>> Did you build Dice? It got moved recently in the SVN from l4/tools/dice >>> to tools/dice. I think you have to build it manually. >>> >>> Cheers, >>> Martin >>> >>> >> I believe I did, by invoking the following commands: >> >> " >> cd dice >> ./bootstrap >> ./configure >> ./make >> " >> > > When being in the l4 directory, does calling > ../dice/src/dice > work or not? I.e. is there a dice binary? > > > > Adam > Calling ../dice/src/dice does work so there is a DICE binary, but when I call "make O=/tmp/build" in the l4 directory I still get the following error: if [ -r ../dice/Makefile.drops ]; then \ make -C ../dice -f Makefile.drops; \ fi make[1]: Entering directory `/tmp/tudos/dice' ../l4/mk/Makeconf:155: Sorry, your object or source path became garbled. ../l4/mk/Makeconf:156: OBJ_BASE: /tmp/build ../l4/mk/Makeconf:157: SRC_BASE_ABS: /tmp/tudos/l4 ../l4/mk/Makeconf:158: SRC_DIR: /tmp/tudos/dice ../l4/mk/Makeconf:159: OBJ_DIR: /tmp/tudos/dice ../l4/mk/Makeconf:160: PKGDIR: ../l4/mk/Makeconf:161: L4DIR_ABS: /tmp/tudos/l4 ../l4/mk/Makeconf:162: PKGDIR_ABS: ../l4/mk/Makeconf:163: PKGDIR_OBJ: /tmp/tudos/dice ../l4/mk/Makeconf:164: *** Please investigate.. Stop. make[1]: Leaving directory `/tmp/tudos/dice' make: *** [../dice] Error 2

2 1

[Fwd: Re: L4 implementations]
by Sergio Ruocco 08 Jun '07

08 Jun '07

Forwarding my answer below as requested. Sergio -------- Original Message -------- Subject: Re: L4 implementations Date: Fri, 8 Jun 2007 23:15:41 +1200 From: Shams <shams(a)orcon.net.nz> To: <ruocco(a)disco.unimib.it> References: <f3op51$utv$1(a)sea.gmane.org> <4669154F.3020907(a)disco.unimib.it> Hi Sergio Thanks for the info Can you also please post this to the newsgroup so that other newbies can benefit from it too. Thanks Shams ----- Original Message ----- From: "Sergio Ruocco" <ruocco(a)disco.unimib.it> To: "Shams" <shams(a)orcon.net.nz> Sent: Friday, June 08, 2007 8:37 PM Subject: Re: L4 implementations Shams wrote: > Just a few questions which I am having a hard time understanding: For answers to these and other questions, and further information on the L4 microkernel API, its different implementations (Pistachio, Fiasco, OKL4 etc.), and the various OSes running on them (Iguana, DROPS etc.), see also the following web sites: http://en.wikipedia.org/wiki/L4_microkernel_family http://l4hq.org/kernels/ http://l4hq.org/docs/ http://l4hq.org/projects/os/ Pistachio http://l4ka.org/ Fiasco http://os.inf.tu-dresden.de/fiasco/ OKL4 http://www.ok-labs.com/products/ http://portal.ok-labs.com/ and this paper of mine: http://ertos.nicta.com.au/publications/papers/Ruocco_06.abstract.pml http://citeseer.ist.psu.edu/757180.html Best Regards, Sergio -- Sergio Ruocco, PhD Research Fellow sergio.ruocco(a)disco.unimib.it / gmail.com NOMADIS Lab. phone: +39-02-6448-7879 mobile, embedded real-time systems skype: 'sergioruocco' Università degli Studi di Milano-Bicocca, Italy

1 0

L4 implementations
by Shams 08 Jun '07

08 Jun '07

Hi I played with the demo cd and it works great.. Just a few questions which I am having a hard time understanding: 1. L4 is itself a API/ABI reference, right? 2. Are L4.Pistachio and L4.Fiasco are 2 different implementations of L4? 3. Is L4.Pistachio is the current official reference implementation? Do both of them only differ in internal code only? 3. What is the difference between L4.Pistachio and L4.Fiasco? Why are there 2 separate implementations ie. why not just one codebase? Are you able to merge these 2 together or plan to? 4. Is L4.Sec official and already in the process of being implemented? Is L4.Pistachio and L4.Fiasco going to be replaced by L4.Sec? Thanks Shams

6 15

Verify Fiasco assertion using pingpong
by eschierb＠sci.kun.nl 07 Jun '07

07 Jun '07

Hi, I am trying to achieve a very simple thing, which is to test an assertion in the Fiasco ICP path. More specifically, I want to add the following assertion to the sys_ipc() function: " // This assertion is added by us assert(!(state() & Thread_polling)); if (EXPECT_TRUE(have_send_part || have_receive_part)) ret = do_ipc(have_send_part, partner, have_receive_part, sender, t, regs); " As can be seen, I want to assert that the Thread_polling bit is not set before calling do_ipc(). I planned to do this by compiling the modified Fiasco kernel and the pingpong package. If the assertion would not fire after running the pingpong test suite, I would have at least some confidence that the assertion does not hold. I have already done something similar a while ago, but with the newer versions I cannot seem to get it working. The first problem was with the DICE compiler, if I used the latest version of the DICE compiler I got the "collect2: ld returned 1 exit status". I solved this by using the 3.1.1 DICE version. I then used the following commands: " svn co http://svn.tudos.org/repos/tudos/tags/dice/3.1.1 dice ./repomgr co hello cd dice ./bootstrap ./configure ./make cd .. cd l4/pkg svn up pingpong cd .. make O=/tmp/build config make O=/tmp/build " However, this resulted in the following error " if [ -r ../dice/Makefile.drops ]; then \ make -C ../dice -f Makefile.drops; \ fi make[1]: Entering directory `/tmp/tudos/dice' ../l4/mk/Makeconf:155: Sorry, your object or source path became garbled. ../l4/mk/Makeconf:156: OBJ_BASE: /tmp/build ../l4/mk/Makeconf:157: SRC_BASE_ABS: /tmp/tudos/l4 ../l4/mk/Makeconf:158: SRC_DIR: /tmp/tudos/dice ../l4/mk/Makeconf:159: OBJ_DIR: /tmp/tudos/dice ../l4/mk/Makeconf:160: PKGDIR: ../l4/mk/Makeconf:161: L4DIR_ABS: /tmp/tudos/l4 ../l4/mk/Makeconf:162: PKGDIR_ABS: ../l4/mk/Makeconf:163: PKGDIR_OBJ: /tmp/tudos/dice ../l4/mk/Makeconf:164: *** Please investigate.. Stop. make[1]: Leaving directory `/tmp/tudos/dice' make: *** [../dice] Error 2 " As an alternative to the last three lines, I tried to issue the "make O=/tmp/build" command in the l4/pkg directory. This resulted in the following error: "/bin/bash: did_not_find_dice___please_install_dice: command not found" Any idea how this can be remedied? With regards, Erik Schierboom

3 3

Re: Hiworld Problem
by Bjoern Doebel 06 Jun '07

06 Jun '07

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, >> Did you checkout and build the L4 Environment (as requested in chapter >> 2 of >> the tutorial)? If so, generic_ts should be built and your application >> gets >> linked with it. > I followed http://os.inf.tu-dresden.de/drops/download.html using svn to > check > out the hello module. I think it contains everything necessary. I just > unzipped it and > compiled it. No it doesn't. The hello module includes the minimal stuff necessary to build an L4 application. As stated in the BID tutorial, chapter 1, please checkout the L4Env module and compile it. Thereafter everythin should work fine. Regards, Bjoern -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGZqeyP5ijxgQLUNkRAjA8AJwIMFk71NoJrdjeGEkyBKmA/gTtkACePNgF lXoytPeJXVrx+EJK6qN/kf4= =pOPl -----END PGP SIGNATURE-----

1 0

Hiworld Problem
by Marius.bao 06 Jun '07

06 Jun '07

Hi all, I followed the tutoral--<<Building Infrastructure for DROPS (BID) Tutorial>>(http://os.inf.tu-dresden.de/l4env/doc/html/bid-tut/node3.html) to wirte my fisrt "hiworld" server. I've used the template, ˜/src/l4/pkg/hiworld> ../../mk/tmpl/inst The "hiworld" source file is as follows: #include <stdio.h> int main(int argc, char** argv){ printf("HiWorld"); return 0; } An error occured when I compiled the server. I think it might be a link error. /usr/bin/ld: cannot find -lgeneric_ts ==> Linking hiworld LD_PRELOAD=/home/hector/Develop/L4-2007/l4/tool/gendep/libgendep.so GENDEP_TARGET=hiworld GENDEP_BINARY=ld gcc -o hiworld /home/hector/Develop/L4-2007/l4/build/lib/x86_586/crt0.o main.o -Ttext=0x01800000 -L/home/hector/Develop/L4-2007/l4/build/lib/x86_586/l4v2 -L/home/drops/lib/x86_586/l4v2 -L/home/hector/Develop/L4-2007/l4/build/lib/x86_586 -L/home/drops/lib/x86_586 -L/home/hector/Develop/L4-2007/l4/build/lib -L/home/drops/lib -T/home/hector/Develop/L4-2007/l4/build/lib/x86_586/main_stat.ld -static -lgeneric_ts -ll4env -ll4rm -ldm_generic -ldm_mem -lthread -lsemaphore -llogserver_capsule -lnames -ll4util_root -ll4util -lsigma0 -lroot -ll4env -ll4env_err -ll4rm -ldm_generic -ldm_mem -lthread -lslab -ll4sys -nostdlib -Wl,--nostdlib -luc_c -luclibc_support -u mmap_anon -lc_be_simple_mem -lc_be_mmap -lc_be_mmap_util -lc_be_l4env_start_stop -lc_be_minimal_log_io -luc_be_simple_sleep -ll4env -llogserver_capsule -ll4rm -ldm_generic -ldm_mem -lthread -lgeneric_ts -luc_c /usr/lib/gcc/i386-redhat-linux/3.4.2/libgcc.a -luc_c -ll4sys -Ttext=0x01800000 -Wl,--warn-common -Wl,-gc-sections /usr/bin/ld: cannot find -lgeneric_ts collect2: ld returned 1 exit status make[6]: *** [hiworld] error 1 make[6]: Leaving directory `/home/hector/Develop/L4-2007/l4/build/pkg/hiworld/server/src/OBJ-x86_586-l4v2' make[5]: *** [/home/hector/Develop/L4-2007/l4/build/pkg/hiworld/server/src/OBJ-x86_586-l4v2] error 2 make[5]: Leaving directory `/home/hector/Develop/L4-2007/l4/pkg/hiworld/server/src' make[4]: *** [src] error 2 make[4]: Leaving directory `/home/hector/Develop/L4-2007/l4/pkg/hiworld/server' make[3]: *** [server] error 2 make[3]: Leaving directory `/home/hector/Develop/L4-2007/l4/pkg/hiworld' make[2]: *** [hiworld.bin] error 2 make[2]: Leaving directory `/home/hector/Develop/L4-2007/l4/pkg' make[1]: *** [pkg] error 2 make[1]: Leaving directory `/home/hector/Develop/L4-2007/l4' make: *** [x] error 2 I wonder how to write the makefile myself without using the template. When I compile the "Hiworld" server, what other files(source files,head files) should I need to include.

2 1