l4-hackers December 2006

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

24 participants
30 discussions

Problems compiling NICTA::Pistachio-embedded 1.3.0.
by Jaeheon Yi 01 Jan '07

01 Jan '07

Hello, I'm attempting to compile NICTA::Pistachio-embedded, version 1.3.0. I've downloaded the source using tla, via `tla get ertos(a)nicta.com.au--public/pistachio--release my_pistachio`. When I follow the standard L4 compile procedure and perform make, I get this message: [jaeheon@l4rbed ia32-kernel-build]$ make menuconfig dirname: too few arguments Try `dirname --help' for more information. bash: /gccgcc: No such file or directory make[1]: Entering directory `/src/NICTA-1.3.0/ia32-kernel-build' make[2]: Entering directory `/src/NICTA-1.3.0/ia32-kernel-build/config' Compiling rules, please wait......(0.72 sec) Done. Loading 'config.out' failed, continuing... Rewriting Makeconf.local ... make[2]: Leaving directory `/src/NICTA-1.3.0/ia32-kernel-build/config' make[1]: Leaving directory `/src/NICTA-1.3.0/ia32-kernel-build' [jaeheon@l4rbed ia32-kernel-build]$ make dirname: too few arguments Try `dirname --help' for more information. bash: /gccgcc: No such file or directory make[1]: Entering directory `/src/NICTA-1.3.0/ia32-kernel-build' make[1]: *** No rule to make target `src/generic/mapping_alloc.cc', needed by `.depend'. Stop. make[1]: Leaving directory `/src/NICTA-1.3.0/ia32-kernel-build' make: *** [all] Error 2 [jaeheon@l4rbed ia32-kernel-build]$ I've tried doing this for the ARM architecture ala Ben Leslie's L4 dev slides, and that also ends up not finding a certain file. I get the feeling that I am missing some files, or perhaps I have to tweak some config files? I would appreciate any pointers on this issue. Thank you, Jaeheon Yi

2 1

Re: What is difference among these 3 terms: domain, task and thread
by Adam Lackorzynski 29 Dec '06

29 Dec '06

Hi, On Wed Dec 27, 2006 at 16:16:57 +0800, chenggh wrote: > I want to know what is the difference among these 3 terms : domain, task and > thread in L4/Fiasco documents. If sb could understand please show me clearly. Where did you read 'domain' in a Fiasco document? A task is an address space which contains one or more threads. Pagetables are attached to tasks, threads execute code. Adam -- Adam adam(a)os.inf.tu-dresden.de Lackorzynski http://os.inf.tu-dresden.de/~adam/

1 0

What is difference among these 3 terms: domain, task and thread
by chenggh 27 Dec '06

27 Dec '06

Hi: I want to know what is the difference among these 3 terms : domain, task and thread in L4/Fiasco documents. If sb could understand please show me clearly. thank you. Best wishes. cheng

1 0

Re: failed to booting l4linux from serial console
by chenggh 23 Dec '06

23 Dec '06

> CPU: Intel Pentium III (Coppermine) stepping 03 > 0000 [#1] > Modules linked in: > CPU: 0 > EIP: fd10:[<00501eb0>] Not tainted VLI > EFLAGS: 00213202 (2.6.18-l4 #22) > EIP is at _mmx_memcpy+0x30/0x150 > eax: bff00200 ebx: 00000014 ecx: 00000064 edx: 010cfa84 > esi: 0064d924 edi: 00000200 ebp: 010cdd94 esp: 010cdd84 > ds: fd24 es: 0890 ss: 0000 > Process swapper (pid: 1, ti=010cc000 task=010caa70 task.ti=010cc000) > Stack: 00000014 010ca550 fffffff4 00000200 010cde04 004132e4 010cfa84 > 0064d924 00000500 00000000 00000000 ffffffff ffffffff 010cc038 010ca610 > 010cde5c 00000000 10800711 00000000 010ca550 000000d0 00000000 010c93a8 > 010c93a8 Call Trace: > [<0040c449>] show_stack_log_lvl+0xa9/0xd0 > [<0040c61a>] show_registers+0x15a/0x1c0 > [<0040c795>] die+0x115/0x2d0 > [<00000000>] _start+0xffc00000/0x50 > [<004132e4>] copy_process+0x954/0xfb0 > [<00413b2b>] do_fork+0x6b/0x1c0 > [<004092ba>] kernel_thread+0x7a/0x90 > [<00427707>] keventd_create_kthread+0x27/0x70 > [<0042791a>] kthread_create+0xda/0xe0 > [<0069e226>] cpu_callback+0x76/0xd0 > [<0069e2a2>] spawn_ksoftirqd+0x22/0x50 > [<004001cc>] init+0x1c/0x280 > [<00408960>] kernel_thread_start+0x30/0x40 > [<004084e6>] ret_kernel_thread_start+0x6/0x10 > Code: 53 83 ec 04 8b 75 0c 89 e0 25 00 e0 ff ff f7 40 14 00 ff ff 0f 0f 85 > 11 01 00 00 8b 45 10 c1 e8 06 89 45 f0 89 c3 e8 30 43 f0 ff <0f> 0d 06 0f > 0d 46 40 0f 0d 86 80 00 00 00 0f 0d 86 c0 00 00 00 > EIP: [<00501eb0>] _mmx_memcpy+0x30/0x150 SS:ESP 0023:010cdd7c > <0>Kernel panic - not syncing: Attempted to kill init! > panic: going to sleep forever, bye Right now I forget to tell you my CPU type is Duron. Do you change the cpu type in my configuration? And I suggest in the first time you should use the com1 to display the l4 output information and after you check that the l4 booting is correct you use the com1 to display the l4linux output inforamtion instead of l4. It seems that you have only one serial console. :) it is bad. cheng cheng

1 0

Re: failed to booting l4linux from serial console
by chenggh 23 Dec '06

23 Dec '06

On Saturday 23 December 2006 16:26, halo duan wrote: > Hi all > as my graphic card of my pc can not support l4dope, I boot l4linux from > serial console, my grub configuration looks like following Oh, My l4linux is 2.6.17. I test it works well. In the l4 configuration of l4linux i think there is no special option to be choose except the l4 build directory. and in the linux configuration of l4linux, the options about serial console is like this: # Serial drivers # CONFIG_SERIAL_8250=y CONFIG_SERIAL_8250_CONSOLE=y CONFIG_SERIAL_8250_PCI=y CONFIG_SERIAL_8250_NR_UARTS=4 CONFIG_SERIAL_8250_RUNTIME_UARTS=4 # CONFIG_SERIAL_8250_EXTENDED is not set # # Non-8250 serial port support # CONFIG_SERIAL_CORE=y CONFIG_SERIAL_CORE_CONSOLE=y # CONFIG_SERIAL_JSM is not set CONFIG_UNIX98_PTYS=y CONFIG_LEGACY_PTYS=y CONFIG_LEGACY_PTY_COUNT=256 according to some errors I suggest you close the option loading module automatically. In this letter there is attachment which is my configuation in l4linux-2.6.17. Maybe it could help you. Good luck. Merry Christmas. cheng

2 1

Re: failed to booting l4linux from serial console
by Adam Lackorzynski 23 Dec '06

23 Dec '06

On Sat Dec 23, 2006 at 17:53:46 +0800, chenggh wrote: > On Saturday 23 December 2006 16:26, halo duan wrote: > > Hi all > > as my graphic card of my pc can not support l4dope, I boot l4linux from > > serial console, my grub configuration looks like following > Oh, My l4linux is 2.6.17. I test it works well. > In the l4 configuration of l4linux i think there is no special option to be > choose except the l4 build directory. > and in the linux configuration of l4linux, the options about serial console > is like this: Please note that you are interfering with the uart here which is not recommended. You can do that of course but then you also need to disable its usage in the kernel to be safe. I recommend using the L4_SERIAL driver. > # Serial drivers > # > CONFIG_SERIAL_8250=y > CONFIG_SERIAL_8250_CONSOLE=y > CONFIG_SERIAL_8250_PCI=y > CONFIG_SERIAL_8250_NR_UARTS=4 > CONFIG_SERIAL_8250_RUNTIME_UARTS=4 > # CONFIG_SERIAL_8250_EXTENDED is not set > > # > # Non-8250 serial port support > # > CONFIG_SERIAL_CORE=y > CONFIG_SERIAL_CORE_CONSOLE=y > # CONFIG_SERIAL_JSM is not set > CONFIG_UNIX98_PTYS=y > CONFIG_LEGACY_PTYS=y > CONFIG_LEGACY_PTY_COUNT=256 > > according to some errors I suggest you close the option loading module > automatically. Which errors are you referring to? Adam -- Adam adam(a)os.inf.tu-dresden.de Lackorzynski http://os.inf.tu-dresden.de/~adam/

1 0

failed to booting l4linux from serial console
by halo duan 23 Dec '06

23 Dec '06

Hi all as my graphic card of my pc can not support l4dope, I boot l4linux from serial console, my grub configuration looks like following title l4linux root (hd0,0) kernel /l4/bootstrap modaddr 0x02000000 # module /l4/fiasco -serial_esc comspeed 115200 -comport 0 module /l4/fiasco module /l4/sigma0 module /l4/roottask task modname "bmodfs" attached 4 modules module /l4/events module /l4/names --events module /l4/log module /l4/dm_phys --isa=0x00800000 -v --events module /l4/simple_ts -t 300 --events module /l4/rtc --events module /l4/l4io --noirq --events module /l4/bmodfs --events module /l4/libld-l4.s.so module /l4/libloader.s.so module /l4/loadercfg module /l4/vmlinuz26 module /l4/loader --fprov=BMODFS loadercfg # vbeset 0x117 506070 and my loadercfg contains following rules modpath "(hd0,0)/l4" task "vmlinuz26 earlyprintk=yes mem=64M console=ttyS0,115200 root=/dev/hda2" all_sects_writable allow_cli the l4linux build config , I check on the CONFIG_L4_EXTERNAL_RTC=y # CONFIG_L4_BLK_DRV is not set # CONFIG_L4_PDSPBLK_DRV is not set # CONFIG_L4_OSHKOSH_DRV is not set # CONFIG_L4_ORE_DRV is not set CONFIG_L4_SERIAL=y CONFIG_L4_SERIAL_CONSOLE=y # CONFIG_L4_NEED_DSI is not set # CONFIG_L4_NEED_INET_NTOA is not set # CONFIG_L4_DROPSCON is not set ...... # CONFIG_SERIAL_CORE=y CONFIG_SERIAL_CORE_CONSOLE=y # CONFIG_UNIX98_PTYS is not set # CONFIG_LEGACY_PTYS is not set but as i boot up the system , it shows that l4linux can not open the initial console. NET: Registered protocol family 17 input: AT Translated Set 2 keyboard as /class/input/input0 kjournald starting. Commit interval 5 seconds EXT3 FS on hda2, internal journal EXT3-fs: recovery complete. EXT3-fs: mounted filesystem with ordered data mode. VFS: Mounted root (ext3 filesystem). Warning: unable to open an initial console. modprobe: FATAL: Could not load /lib/modules/2.6.18-l4/modules.dep: No such file or directory udevd[59]: init_uevent_netlink_sock: bind failed: No such file or directory udevd[59]: main: uevent socket not available modprobe: FATAL: Could not load /lib/modules/2.6.18-l4/modules.dep: No such file or directory modprobe: FATAL: Could not load /lib/modules/2.6.18-l4/modules.dep: No such file or directory modprobe: FATAL: Could not load /lib/modules/2.6.18-l4/modules.dep: No such file or directory Restarting system. anybody give a hand. best regards

2 1

unmap() guaranteed to succeed?
by Marcus Brinkmann 22 Dec '06

22 Dec '06

Hi, I try to understand if (and if yes, why) unmap() in L4 architecture implementations is guaranteed to succeed. I hesitate to say I understood the description in Marcus Völp, Design and Implementation of the Recursive Virtual Address Space Model for Small Scale Multiprocessor Systems, but at least I skimmed through it and have developed sort of a mental picture of the post-order node processing. Bernhard Kauer in L4.sec Implementation also refers to this work. I understand there are substantial variations of L4 implementations, and I don't have a complete picture of which one uses which algorithm. In any case, I am worried about the following scenario, and wonder if you can share with me your insight on the question if this is a valid concern or not: Let's say there is a mapping of the same page from A to B and from A to C. Now, unmap() proceeds in post-order, and deletes the mapping from A to C. Then it is preempted, and a restart point at B is noted. Next B maps the page to B', which updates the restart point to be at B'. Now, unmap() proceeds by deleting the mapping from B to B'. Again, the unmap operation is preempted, with the restart point being B. The cycle continues. There are variations of this, where B' creates a mapping to B'', which creates a mapping to B''' etc, n times in total, and unmap() undos the last k such mappings before it is preempted. As long as n >= k, the B's can create new mappings as fast or faster as unmap can delete them, so there doesn't seem to be a way for unmap() to succeed eventually. I have the feeling that this is an obvious threat scenario and thus is probably addressed already, but I can't find where. I might be missing something obvious. Thanks, Marcus

5 6

AW: L4Linux on L4Ka::Pistachio
by Christian, Martin, OPEE45 22 Dec '06

22 Dec '06

Thanks Volkmar, that sound promising! However, I'll first try if the L4Linux Kernel 2.4 is running on Pistachio for PowerPC. If not, I will look deeper into the kernel 2.6, so I can estimate how much work it is. I've got to minimize my time to (scientific) market. ;-) ... It just came to my mind, that I might confuse things. Is the L4Linux with Kernel 2.4 architecture neutral? That is, there is only L4 dependent C Code and no "real" architecture dependent code like IA32/PowerPC Assembler. I suppose I should just start to get the force and read the source. ;-) Nice holidays, Martin. -----Ursprüngliche Nachricht----- Von: Volkmar Uhlig [mailto:volkmar@volkmaruhlig.com] Gesendet: Mittwoch, 20. Dezember 2006 16:17 An: Christian, Martin, OPEE45; 'Adam Lackorzynski' Cc: l4-hackers(a)os.inf.tu-dresden.de Betreff: RE: L4Linux on L4Ka::Pistachio According to some people I talked to over here, Linux for PPC has pretty much everything abstracted already. So the job at hand is to fill the architecture specific function table with the L4 specific implementation. Linux on Power is more abstracted than x86 in order to accommodate the various hardware platforms and chip versions (...and to run on IBM's hypervisor). - Volkmar

2 1

Fwd: 答复: ×a・￠ : segment fault for fiasco-ux
by halo duan 21 Dec '06

21 Dec '06

at last , my athena say hello to me. thanks very much. [root@athena l4]# ./fiasco Fiasco-UX on Linux 2.6.15-1.2054_FC5 (i686) Mapped 64 MB Memory + 0 KB Framebuffer + 0 MB Input Area on FD 3 Loading Module 0x00090000-0x00099380 [sigma0] Loading Module 0x00120000-0x002b09e4 [roottask] Copying Module 0x039bb000-0x039e0d4e [hello] Bootstrapping... Welcome to Fiasco(ux)! DD-L4(v2)/x86 microkernel (C) 1998-2006 TU Dresden Rev: Thu Dec 21 14:23:32 2006 compiled with gcc 4.1.0 for Intel Pentium III CPU: GenuineIntel (6:8:3:2) Model: Pentium III (Coppermine) at 797 MHz 32 Entry I TLB (4K pages) 2 Entry I TLB (4M pages) 64 Entry D TLB (4K pages) 8 Entry D TLB (4M pages) 16 KB L1 I Cache (4-way associative, 32 bytes per line) 16 KB L1 D Cache (4-way associative, 32 bytes per line) 256 KB L2 U Cache (8-way associative, 32 bytes per line) Freeing init code/data: 28672 bytes (7 pages) Calibrating timer loop... done. SIGMA0: Hello! KIP @ 39f5000 Found Fiasco: KIP syscalls: yes allocated 4kB for maintainance structures Roottask. Found Fiasco-UX. Command line found: "roottask" 64512kB ( 63MB) total RAM (reported by bootloader) 57500kB ( 57MB) received RAM from Sigma0 1760kB ( 2MB) reserved RAM for RMGR Received I/O ports 0000-ffff Attached irqs = [ <!0> 1 2 3 4 5 6 7 8 9 A B C D E F ] Roottask: Loading 1 module. #05: loading "hello" from [039bb000-039e0d4e] to [01000000-01002eea][01003000-010055c8] entry at 0000d060 via trampoline page code Hello World hello: My thread-id is 5.0 hello: My thread-id is 5.0 hello: My thread-id ps shown following info 6324 tty1 S+ 0:00 ./fiasco-ux -l hello 6325 tty1 S+ 0:00 [I](irq0) 6326 tty1 T+ 0:00 [fiasco-ux] 6327 tty1 T+ 0:00 [fiasco-ux] 6328 tty1 T+ 0:00 [fiasco-ux] 6329 pts/0 R+ 0:00 ps ax On 12/21/06, Carsten Weinhold <weinhold(a)os.inf.tu-dresden.de> wrote: > Am Donnerstag Dezember 21 2006 12:56 schrieb halo duan: > > Hi! > > > [...] > > following is my screen copy of fisco-ux. it's freezed there. I can not use > > CTRL-C to bring up the jdb. > > [...] > > Calibrating timer loop... done. > > > > ^^^^^^ > > freezed to above > > Are you using a System with a recent version of GLibC (you said you were using > Fedorea Core 5 in another email, I think it ships with GLibC 2.4)? If you do, > do you have the most recent Fiasco version from CVS? A TLS-related problem > with new versions of GLibC has been fixed in Fiasco last week, so you should > update your CVS copy. > > Regards, > Carsten > > > [...] >

1 0

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 December 2006