l4-hackers January 2007

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

23 participants
21 discussions

arm instruction version
by WanKeun Jo 17 Jan '07

17 Jan '07

Hi, In the l4linux-2.6, if set L4ARCH=arm, the processor type is fixed to CPU_SA1100. And the arm instruction version is armv4. I want to use arch-dependent code for pxa, so I changed CPU_SA1100 to CPU_XSCALE config and arm instruction set version to armv5te at --------------------------------------------------- $ l4linux-2.6/arch/l4/Makefile CFLAGS += -D__LINUX_ARM_ARCH__=5 -march=armv5te --------------------------------------------------- and found a code, --------------------------------------------------- $ l4linux-2.6/arch/l4/kernel/arch-arm/setup.c unsigned int processor_id = 0x0001f000; /* ARMv4 */ --------------------------------------------------- The processor_id is fixed to ARMv4 value. My question is 001. How processor_id affect the linux?? Should I also change this value to a armv5te value?? 002. If I have to change it, what is the proper value for armv5te?? I can't find the value anywhere. Thank you!!!

2 1

AW: Documentation for L4Linux/L4Env
by Christian, Martin, OPEE45 16 Jan '07

16 Jan '07

Hi Adam! > maybe the KMB lecture slides are somehow helpful. They are a good starting point to understand the principles. However the mapping to the code is still unclear. Wouldn't it be possible to get a student (HiWi) writing some documentation? > You should have some good understanding of Linux internals though Yes. And that's exactly my problem ... :-( > understand what the important code does. First understand what the important code is. > Besides that you can of course ask questions and I can try to answer them. >From your lecture I understand that a new architecture would require mainly changes to the "Lower Half" of L4Linux. But I suppose there are also changes required to the Syscall Interface. However, since I want to port it to another L4 Interface, I'll also have to rewrite a lot more. So I'm tending more to use either the L4Linux-2.4 (where there is no documentation) or Wombat, which at least builds on the L4.X2 API. Anyways, I fear I won't manage to do it within 6 weeks ... :-( Thanks for the help, Martin.

2 1

Re: Please help me to solve this problem of my application
by chenggh 15 Jan '07

15 Jan '07

Hi: Now the compilation is ok. But when I run this application there are some errors. This is my log: Fiasco-UX on Linux 2.6.15-23-386 (i686) Mapped 64 MB Memory + 0 KB Framebuffer + 0 MB Input Area on FD 3 Loading Module 0x00090000-0x00096964 [sigma0-ux] Loading Module 0x00120000-0x002bc9c4 [roottask] Copying Module 0x039a1000-0x039e0de8 [names] Copying Module 0x03965000-0x039a06d1 [log] Copying Module 0x038de000-0x03964162 [dm_phys] Copying Module 0x03854000-0x038ddee7 [clock] Copying Module 0x037cc000-0x03853750 [clk] Bootstrapping... Welcome to Fiasco(ux)! DD-L4(v2)/x86 microkernel (C) 1998-2006 TU Dresden Rev: Mon Aug 28 08:34:51 2006 compiled with gcc 3.4.6 for AMD Athlon CPU: AuthenticAMD (6:8:1:0) Model: Duron (Applebred) at 1602 MHz 16/256 Entry I TLB (4K pages) 8 Entry I TLB (4M pages) 32/256 Entry D TLB (4K pages) 8 Entry D TLB (4M pages) 64 KB L1 I Cache (2-way associative, 64 bytes per line) 64 KB L1 D Cache (2-way associative, 64 bytes per line) 64 KB L2 U Cache (8-way associative, 64 bytes per line) Freeing init code/data: 24576 bytes (6 pages) SIGMA0: Hello! Found Fiasco: KIP syscalls: yes. Allocated 20kB for maintenance structures. Roottask. Found Fiasco-UX. Command line found: "roottask" 64512kB ( 63MB) total RAM (reported by bootloader) 55080kB ( 54MB) received RAM from Sigma0 3788kB ( 4MB) reserved RAM for RMGR Received no I/O ports Attached irqs = [ <!0> 1 2 3 4 5 6 7 8 9 A B C D E F ] Roottask: Loading 5 modules. #05: loading "names" from [039a1000-039e0de8] to [002d0000-002d614a][002d7000-002e3000] entry at 0000d060 via trampoline page code symbols at [039ee000-039f0000] (8kB), lines at [039e8000-039ee000] (24kB) #06: loading "log" from [03965000-039a06d1] to [00400000-00406c6a][00407000-00439850] entry at 0000e05c via trampoline page code symbols at [039e6000-039e8000] (8kB), lines at [039e0000-039e6000] (24kB) #07: loading "dm_phys" from [038de000-03964162] to [01500000-01511d6a][01512000-0151c000] entry at 0000f060 via trampoline page code symbols at [039db000-039e0000] (20kB), lines at [039ce000-039db000] (52kB) #08: loading "clock" from [03854000-038ddee7] to [01800000-01812bb4][01813000-01833000] entry at 00010060 via trampoline page code symbols at [039ca000-039ce000] (16kB), lines at [039ba000-039ca000] (64kB) #09: loading "clk" from [037cc000-03853750] to [01100000-011124d4][01113000-01133000] entry at 0001105c via trampoline page code symbols at [039b6000-039ba000] (16kB), lines at [039a7000-039b6000] (60kB) clock_s | L4RM: [PF] write at 0x00000000, ip 018001ac, src 8.02 clock_s | [8.0] l4rm/lib/src/pagefault.c:78:__unknown_pf(): clock_s | unhandled page fault The attachment is the source code, too. Thank you. cheng

2 1

Please help me to solve this problem of my application
by chenggh 13 Jan '07

13 Jan '07

Hi: I write a simple service for L4/Fiasco which encapsulate two methods. clock_gettime() and clock_nanosleep(). Of course now it doesn't not have full functions of these two methods. But when I compile the client application for this server. It has some errors like this: PWD=/home/fiasco/L4.17/build/pkg/clock/examples/client/OBJ-x86_586-l4v2 make -C /home/fiasco/L4.17/build/pkg/clock/examples/client/OBJ-x86_586-l4v2 -f Makefile make[1]: Entering directory `/home/fiasco/L4.17/build/pkg/clock/examples/client/OBJ-x86_586-l4v2' ==> Linking clk LD_PRELOAD=/home/fiasco/L4.17/l4/tool/gendep/libgendep.so GENDEP_TARGET=clk GENDEP_BINARY=ld gcc -o clk /home/fiasco/L4.17/build/lib/x86_586/crt0.o main.o -Ttext=0x01100000 -L/home/fiasco/L4.17/build/lib/x86_586/l4v2 -L/home/fiasco/L4.17//lib/x86_586/l4v2 -L/home/fiasco/L4.17/build/lib/x86_586 -L/home/fiasco/L4.17//lib/x86_586 -L/home/fiasco/L4.17/build/lib -L/home/fiasco/L4.17//lib -T/home/fiasco/L4.17/build/lib/x86_586/main_stat.ld -lclock -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/i486-linux-gnu/3.4.6/libgcc.a -luc_c -ll4sys -Ttext=0x01100000 -Wl,--warn-common -Wl,-gc-sections main.o: In function `main':/home/fiasco/L4.17/l4/pkg/clock/examples/client/main.c:11: undefined reference to `clock_gettime' collect2: ld returned 1 exit status make[1]: *** [clk] Error 1 make[1]: Leaving directory `/home/fiasco/L4.17/build/pkg/clock/examples/client/OBJ-x86_586-l4v2' make: *** [/home/fiasco/L4.17/build/pkg/clock/examples/client/OBJ-x86_586-l4v2] Error 2 it says there is no declaration of clock_gettime(). But according to the same method as showing in BID tutorial my hello example runs well. There is attachment of all my source code file. Please help me and I have this problem for many days. Thank you. Cheng

2 1

I/O port protection
by Daniel Vandersee 12 Jan '07

12 Jan '07

Hi there, is it possible to allow I/O page mapping at the RMGR for a newly created task (using l4_task_new) as it were possible using the "allow_cli" flag in the roottask configuration? Usually the pager of the newly created task would handle its I/O pagefaults. Unfortunately this won't work here as this particular task (L4 loader) tries to directly map I/O pages using an IPC call to the ressource manager. So the question is: How can I tell the ressource manager to allow I/O mapping for this particular task? Thanks, Daniel

2 3

Documentation for L4Linux/L4Env
by Christian, Martin, OPEE45 11 Jan '07

11 Jan '07

Hi, is there some documentation on the L4Linux source code? A kind of what-does-what? I've seen L4Linux has been ported to ARM? Is there a porting guide? HowTo? Wiki? Something more than C or assembler? How much has L4Linux still in common with UML? Which parts were used and where are they now? Best, Martin.

2 1

Indirect IPC or direct IPC
by chenggh 10 Jan '07

10 Jan '07

Hi: Today I read the DICE Manual 3.1.0 and it talks about something about indirect IPC and direct IPC. According to my understanding it seems like this: indirect IPC: sender - > message buffer -> receiver direct IPC: sender ->sender buffer -> receive buffer ->receiver. But I think this understanding is wrong and I don't search something helpful by google, too. If somebody knows please tell me. In this manaul it says ln L4 version 4 all constructed data types will be transmitted by indirect IPC not direct IPC. I also want to know why. Thank you. Best wishes. cheng

3 2

Why I can't get the seconds and nanoseconds like this?
by chenggh 08 Jan '07

08 Jan '07

Hi: I wrote the simple application like this. #include <stdio.h> #include <l4/util/rdtsc.h> #include <l4/sys/l4int.h> #include <l4/thread/thread.h> char LOG_tag[4]="clk"; int main(void) { l4_uint32_t *second; l4_uint32_t *nanosecond; l4_cpu_time_t tsc; tsc = l4_rdtsc(); l4_tsc_to_s_and_ns(tsc, second, nanosecond); printf("%d, %d \n", *second, *nanosecond); } The result is : clk | 0, 0 clk | clk | Exiting with 6 Does these code have some problems? Thank you. cheng

3 2

Re: L4 Demo CD problems
by Dave Haddon 08 Jan '07

08 Jan '07

I am using a PC104 stack with a PIII/1Ghz, 256Mb, CD and 6Gig HD... no floppy..currently running timesys (6.1) (Linux Kernel 2.6.16.16 for i686) I eventually got the source for L4 to Compile by leaving out bits that were causing compiling errors.. RTC for example.. once I got to that point.. my hard drive died and I had to start from scratch.. arrgghh!!!! After much stuffing around to find a workable drive for the PC104, I figured I'd just download the Demo cd and have a whack at that, and it wouldn't work.. I'm starting to get a bit annoyed.. I originally retrieved the L4linux source from CVS, (l4 and linux-2.6) and tried to compile it together from the instructions on the DROPS web and had a lot of missing pieces in timesys, (mainly upto-date PERL and automake).. I haven't tried with QEMU or VMware as I have the PC104 on the same monitor and keyboard.. it is too easy to switch between the PC104 and my desktop (P4 2.8G, SuSe 9.3, KDE 3.4.0b..with no root permissions...) Any info you can give me would be greatly appreciated.. Dave Haddon On Fri, 2007-01-05 at 08:20 +0100, Bjoern Doebel wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hi, > > > I have downloaded and burnt the Demo CD exactly as described on the > > website, but it won't come up with the grub menu as described. All I get > > is the Grub command prompt. Have I done something wrong? I have been > > trying to get L4linux to work for ages now and nothing is working.. > > (Mainly a harddrive crash and a lack of root permissions on my main PC > > causing problems...) Can someone please help.. > > some more information about your system, what exactly you've been doing to > run L4Linux and what failures you encountered could be helpful. > And can you maybe try to run the DemoCD-Image in QEMU or VMWare? > > Bjoern > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.5 (GNU/Linux) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org > > iD8DBQFFnfxIP5ijxgQLUNkRAuKHAKCQBAHJbkFm6/GaBo7WCyMsnSY1LQCfcKbR > 3LKgzTHKSIU3HlfiEoHadIM= > =6eip > -----END PGP SIGNATURE----- -- CSIRO - Queensland Centre for Advanced Technologies Technology Court Pullenvale Qld 4069 AUSTRALIA PO Box 883 Kenmore Qld 4069 AUSTRALIA Telephone +61 7 3327 4444 Facsimile +61 7 3327 4455

2 1

L4 Demo CD problems
by Dave Haddon 05 Jan '07

05 Jan '07

I have downloaded and burnt the Demo CD exactly as described on the website, but it won't come up with the grub menu as described. All I get is the Grub command prompt. Have I done something wrong? I have been trying to get L4linux to work for ages now and nothing is working.. (Mainly a harddrive crash and a lack of root permissions on my main PC causing problems...) Can someone please help.. Thanks -- CSIRO - Queensland Centre for Advanced Technologies Technology Court Pullenvale Qld 4069 AUSTRALIA PO Box 883 Kenmore Qld 4069 AUSTRALIA Telephone +61 7 3327 4444 Facsimile +61 7 3327 4455

2 1

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 January 2007