April 2010 - l4-hackers - example.com Site

Re: Regd: L4 Reference and User manual
by kolabor＠freenet.de 29 Apr '10

29 Apr '10

Hi, after looking at the slides you pointed to (at http://www.inf.tu-dresden.de/index.php?node_id=1314&ln=en) the Fiasco for download isn't the one discussed in the slides, right ? So it seems that stuff discussed in your lectures are newer than that what is available for download, esp. the capability part. Also in your recent publications you are speaking about Fiasco.OC and L4Re and not about Fiasco 1.2 and/or L4Env. So what does this all mean ? They are related or is that completly new stuff and just rebranded to known/established names. Some enlightenment would be nice. (Intention of my mail: Try to figure out whether it is worth to start porting stuff to old? outdated? Fiasco 1.2 or to wait [how long?] for the newer stuff you discussing in your slides. Don't want to write or to port stuff twice, probably also interesting for the request of the MPI stuff some days ago.) Tim. > -----Ursprüngliche Nachricht----- > Von: Adam Lackorzynski > Gesendet: Mi. 28.04.10 (00:03) > An: l4-hackers(a)os.inf.tu-dresden.de > Betreff: Re: Regd: L4 Reference and User manual > > > On Mon Apr 26, 2010 at 14:32:21 -0400, Karthik Vadambacheri Manian > wrote: > > I am working on Fiasco compiled for AMD64 architecture. Can some > one point > > which L4 reference manual to read? > > Also the user manual by Alan Au and Gernot Heiser seems to be old. > Is it > > still valid? please let me know. > > For the version of Fiasco you are looking at it is still ok. Also I > think you can learn a lot about the concepts etc. > > > > Adam > -- > Adam adam(a)os.inf.tu-dresden.de > Lackorzynski http://os.inf.tu-dresden.de/~adam/ > > _______________________________________________ > l4-hackers mailing list > l4-hackers(a)os.inf.tu-dresden.de > http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers > > > -----Ursprüngliche Nachricht Ende----- freenetMail mobil – Alle E-Mails auf Ihrem Handy versenden und empfangen. Jetzt kinderleicht und kostenlos einrichten. http://tls.freenet.de/tipp/handymail/index.html

2 1

The system call about Periodic task
by Jungwoo Yang 28 Apr '10

28 Apr '10

I tried to use time periodic task designed by Udo. However, the process using the system call does not enter into periodic mode. So I tried to analyze kernel source code. During examining kernel source code, I found the wired point. line 584 in syscalls.cpp regs->old_param (dst->begin_periodic (regs->time(), 0)); I think that the parameter should be "Periodic" instead of 0. Because the point is for entering periodic mode and 0 means that conventional mode, I think 0 should be changed into "Periodic". After modification of this, A task using the system call enters periodic mode ( I check this by using JDB debugger ). I cannot ensure that this is correct code, but It seems to be working. So, if anyone who want to use periodic task and think that it is not working, just try to modify this code. Thank you for Adam Lackorzynski and Cheng Cuanghui Thank you. Regards, Jungwoo Yang.

1 0

Regd: L4 Reference and User manual
by Karthik Vadambacheri Manian 28 Apr '10

28 Apr '10

Hi All, I am working on Fiasco compiled for AMD64 architecture. Can some one point which L4 reference manual to read? Also the user manual by Alan Au and Gernot Heiser seems to be old. Is it still valid? please let me know. Thanks for your time, karthik <l4-hackers(a)os.inf.tu-dresden.de>

2 1

L4 and ARM Cortex-A8 (OMAP3)
by Thomas DuBuisson 26 Apr '10

26 Apr '10

All, I'm wanting to experiment with L4 on a OMAP3 (beagleboard based) system but am having difficulting finding a suitable L4 microkernel. Does anyone on this list have comments/suggestions wrt which kernel to use, building, u-boot configuration, etc? Cheers, Thomas

2 2

Porting an MPI based appilcation on L4 feasible?
by Karthik Vadambacheri Manian 26 Apr '10

26 Apr '10

Hi All, I am new to L4 fiasco and microkernels in general. I could build the l4 fiasco microkernel and run the 'hello' application on top of l4. I also could run the hiworld client server application on l4 based on the available tutorials. Currently I have a pingpong simulation application which uses MPI and Time Warp libraries. For my research I would like to run this application on l4. Is porting it to l4 feasible? I also found a pinpong application in '../kernel/fiasco/src/test/pingpong'. But building fiasco microkernel does not build it. How to build it? and run on top of l4? Thanks for your time, Karthik.V.M. PhD student, Experimental Computing Lab, University of Cincinnati

3 3

the ipcmon compiling error
by Guanghui, Cheng 21 Apr '10

21 Apr '10

Hello: there is an error when compiling the ipcmon with ubuntu 9.10 with default gcc (Ubuntu 4.4.1-4ubuntu9) 4.4.1 as following: LD_PRELOAD=/home/cheng/work/L4_DEMO/tudos/build/tool/gendep/libgendep.so GENDEP_TARGET=ipcmon-server.o GENDEP_BINARY=cc1 gcc -m32 -c -DLOG_TAG=\"ipcmon\" -DUSE_UCLIBC=y -DRAM_BASE=0x0 -DSYSTEM_x86_586_l4v2 -DARCH_x86 -DCPUTYPE_586 -DL4API_l4v2 -DCONFIG_L4_CALL_SYSCALLS -DCONFIG_L4_ABS_SYSCALLS -I/home/cheng/work/L4_DEMO/tudos/build/pkg/ipcmon/idl/OBJ-x86-l4v2 -I/home/cheng/work/L4_DEMO/tudos/build/include/x86/l4v2 -I/home/cheng/work/L4_DEMO/tudos/build/include/l4v2 -I/home/cheng/work/L4_DEMO/tudos/build/include/x86 -I/home/cheng/work/L4_DEMO/tudos/build/include -I/home/cheng/work/L4_DEMO/tudos/l4/../dice/include -I/home/cheng/work/L4_DEMO/tudos//include/x86/l4v2 -I/home/cheng/work/L4_DEMO/tudos//include/l4v2 -I/home/cheng/work/L4_DEMO/tudos//include/x86 -I/home/cheng/work/L4_DEMO/tudos//include -nostdinc -I/home/cheng/work/L4_DEMO/tudos/build/include/x86/uclibc -I/home/cheng/work/L4_DEMO/tudos/build/include/uclibc -I/home/cheng/work/L4_DEMO/tudos/build/include/uclibc++ -I/home/cheng/work/L4_DEMO/tudos//include/x86/uclibc -I/home/cheng/work/L4_DEMO/tudos//include/uclibc -I/home/cheng/work/L4_DEMO/tudos//include/uclibc++ -I/usr/lib/gcc/i486-linux-gnu/4.4.1/include -I/usr/lib/gcc/i486-linux-gnu/4.4.1/include-fixed -gstabs+ -g -O2 -fno-strict-aliasing -march=i586 -Wall -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations -fno-common -fno-stack-protector -g /home/cheng/work/L4_DEMO/tudos/build/pkg/ipcmon/idl/OBJ-x86-l4v2/ipcmon-server.c -o ipcmon-server.o ==> Linking ipcmon LD_PRELOAD=/home/cheng/work/L4_DEMO/tudos/build/tool/gendep/libgendep.so GENDEP_TARGET=ipcmon GENDEP_BINARY=ld GENDEP_BINARY_ALT1=ld ld -m elf_i386 -o ipcmon /home/cheng/work/L4_DEMO/tudos/build/lib/x86_586/crt0.o components.o capdescriptor.o capmanager.o blacklist.o whitelist.o ipcmon-server.o -Ttext=0x1000000 --hash-style=sysv -L/home/cheng/work/L4_DEMO/tudos/build/lib/x86_586/l4v2 -L/home/cheng/work/L4_DEMO/tudos/build/lib/x86_586 -L/home/cheng/work/L4_DEMO/tudos/build/lib -L/home/cheng/work/L4_DEMO/tudos//lib/x86_586/l4v2 -L/home/cheng/work/L4_DEMO/tudos//lib/x86_586 -L/home/cheng/work/L4_DEMO/tudos//lib -T/home/cheng/work/L4_DEMO/tudos/build/lib/x86_586/main_stat.ld -lparsecmdline -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 -luc_c++ -lc ++start.o /usr/lib/gcc/i486-linux-gnu/4.4.1/libsupc++.a -lmain -lcxx_base -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/i486-linux-gnu/4.4.1/libgcc_eh.a /usr/lib/gcc/i486-linux-gnu/4.4.1/libgcc.a -luc_c -ll4sys -ll4util -luc_be_minimal_io -Ttext=0x1000000 --warn-common -gc-sections /usr/lib/gcc/i486-linux-gnu/4.4.1/libsupc++.a(cp-demangle.o): In function `.L641': (.text+0x32e7): undefined reference to `__sprintf_chk' Thanks. Cheng Guanghui

2 1

About Periodic Tasks
by Jungwoo Yang 21 Apr '10

21 Apr '10

Hello, I'm a student who are studying periodic task on L4, so I'm trying to use your cpu_reserved example. However, what I tried to use res_rt_sched or periodic. I've gotten some errors. title Fiasco Periodic Test kernel /bootstrap -serial -modaddr=0x06000000 module /fiasco -serial_esc -comport 1 module /sigma0 module /roottask task modname "cpu_reserved" allow_cli task modname "res_rt_sched" allow_cli module /events module /names --events module /log module /dm_phys module /cpu_reserved module /res_rt_sched This is my configuration of grub. but I have gotten some error like that. res_rt_s| Error: l4_rt_next_period(): Unknown (unregistered) error res_rt_s| [A.3] cpu_reserve/examples/res_rt_sched/main.c:77:worker(): res_rt_s| Error: l4_rt_next_period(): Unknown (unregistered) error res_rt_s| [A.3] cpu_reserve/examples/res_rt_sched/main.c:77:worker(): When I tried to figure out the error. The problem is from "l4_rt_next_period()". It returns 0x20 error "L4_IPC_RETIMEOUT". I've found the error from "l4cpu_reserve_wait_periodic()" in cpu_reserved server. The function returns 0 if the IPC returns "L4_IPC_RETIMEOUT", so I thought that "L4_IPC_RETIMEOUT" is not real error at that time. But when I execute the compoent "res_rt_sched", it look like that it is not a periodic task. I've spent about 3 weeks because of this.. please help me! Thank you.

2 1

bug report and fixed in the dice compiling
by Guanghui, Cheng 19 Apr '10

19 Apr '10

Hello: In the ubuntu 9.10 with default bison version 2.4.1 the dice configure script failed to check the valid bison version: the old one: 20627 if test "(" "${1}" = "2" -a "(" "${2}" = "1a" -o "${2}" = "1b" -o "${2}" = "2" -o "${2}" = "3" ")" ")"; then 20628 dice_bison_version="${1}.${2} (ok)" 20629 else 20630 dice_bison_version=invalid 20631 fi this verification is only available for bison version 2.1a, 2.2 and 2.3. it will be invalid in the bison 2.4 and later. the new one: 20626 set `$YACC --version|grep 'GNU Bison'|cut -d ' ' -f 4|sed -e 's/\./ /'` 20627 if test "(" "${1}" = "2" -a "(" "${2}" = "1a" -o "${2}" = "1b" -o "${2}" = "2" -o "${2}" = "3" -o "${2}" = "4" -o "${2}" = "4.1" -o "${2}" = "4.2" ")" ")"; then 20628 dice_bison_version="${1}.${2} (ok)" 20629 else 20630 dice_bison_version=invalid That's all. Cheng Guanghui

3 2

Post Bison version fix error
by Harish G. Naik 18 Apr '10

18 Apr '10

So I was advised to download bison 2.3 and install it and use it instead of the system default 2.4.1. I did that and I now get a different, much shorter: c-scanner.ll: In function ‘yy::c_parser::token::yytokentype clex(yy::c_parser::semantic_type*, yy::location*, c_parser_driver&)’: c-scanner.ll:523: error: ‘strlen’ was not declared in this scope make[6]: *** [c-scanner.o] Error 1 make[5]: *** [all-recursive] Error 1 make[4]: *** [all-recursive] Error 1 make[3]: *** [all-recursive] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all] Error 2 make: *** [../dice] Error 2 Can someone tell me what I might still be missing? -Harish

2 3

Re: pci_set_master in dde linux26
by Björn Döbel 07 Apr '10

07 Apr '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, > On 10-4-7 下午4:30, Björn Döbel wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> Hi, >> >>> We tried to run the e100 driver with the hurd porting of dde linux26, but got >>> one problem, i.e., pci_set_master() hasn't been implemented. In >>> linux26/lib/src/arch/l4/pci.c, >>> void pci_set_master(struct pci_dev *dev) >>> { >>> CHECK_INITVAR(dde26_pci); >>> WARN_UNIMPL; >>> } >>> >>> Were there any reasons that pci_set_master() was implemented? It seems >>> pci_set_master() only needs to read or write the PCI configuration space. >> >> obviously we didn't need this feature yet. Feel free to add it. > OK, I see. Does DDE Linux26 works with drivers using MMIO? I'm trying to make > DDE Linux26 with e1000 driver, which uses MMIO, but have no luck. Yes, works here. Bjoern -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAku8ZS8ACgkQP5ijxgQLUNlyqACfWgQyom6YnQBrwNTGp86eu5tp dWkAnjYgM/Pcm0HX5wGECZ0OacYCV2eT =8sro -----END PGP SIGNATURE-----

2 4