l4-hackers May 2006

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

26 participants
21 discussions

ore with fiasco-ux
by Jeff Hieb 31 May '06

31 May '06

Hello, I am having trouble getting the ore arping test to work under fiasco or fiasco-ux. When I run under fiasco-ux: ./fiasco -N -s -l names -l dm_phys -l log -l "l4io --noirq" -l ore -l ore_test_arping Roottask. Found Fiasco-UX. 64512kB ( 63MB) total RAM (reported by bootloader) 57208kB ( 56MB) received RAM from Sigma0 7628kB ( 8MB) 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 6 modules. #05: loading "names" from [0399a000-039e0a2a] to [002d0000-002d774a][002d8000-002e4000] entry at 0000c060 via trampoline page code symbols at [039ee000-039f0000] (8kB), lines at [039e7000-039ee000] (28kB) #06: loading "dm_phys" from [0391c000-039996f4] to [01500000-0151304a][01514000-0151e000] entry at 0000d060 via trampoline page code symbols at [039e3000-039e7000] (16kB), lines at [039d5000-039e3000] (56kB) #07: loading "log" from [038d9000-0391b69c] to [00400000-00407eaa][00408000-0043a890] entry at 0000e05c via trampoline page code symbols at [039d3000-039d5000] (8kB), lines at [039cc000-039d3000] (28kB) #08: loading "l4io --noirq" from [036f0000-038d8801] to [00b80000-00ba32ce][00ba4000-00bf0000] entry at 0000f068 via trampoline page code symbols at [03994000-039cc000] (224kB), lines at [0397a000-03994000] (104kB) #09: loading "ore" from [034bf000-036ef46b] to [01800000-0182f3e6][01830000-01865000] entry at 0001005c via trampoline page code symbols at [03972000-0397a000] (32kB), lines at [0394e000-03972000] (144kB) #0a: loading "ore_test_arping" from [03411000-034be85e] to [00980000-0099872e][00999000-009b9000] entry at 00011068 via trampoline page code symbols at [03949000-0394e000] (20kB), lines at [03936000-03949000] (76kB) io | do_args(): Disabling internal IRQ handling. arping | main(): Hello from the ORe arping shared memory client arping | main(): string ipc receive arping | main(): sending with string ipc arping | l4ore_open(): called ore | l4dde_mm_init(): Using ... ore | 16 MB at 0x00100000 (vmem) ore | 16 MB in 1 regions (kmem) ore | device lo entered promiscuous mode ore | main(): Initialized 1 network devices. ore | Device = lo, IRQ = 0, MAC = 00:00:00:00:00:00 ore | main(): Registering at names... ore | main(): Ready for service. arping | ore_lookup_server(): ORe = 9.02 arping | l4ore_open(): descriptor: 0 arping | ore_do_open(): called ore | [9.2] ore/server/src/clients.c:208:setup_connection(): ore | Error: no device found. arping | ore_do_open(): opened. worker = 7FF.7F arping | [A.2] ore/lib/client/lib.c:153:l4ore_open(): arping | Error: ore_open() returned INVALID_ID arping | main(): opened eth0: -1 for 00:00:00:00:00:00 arping | [A.2] ore/examples/arping/main.c:243:main(): arping | Error: could not open eth0! arping | arping | Exiting with 1 Would appreciate any suggestions Thanks Jeff

2 1

NICTA's virtualised Linux faster than native on L4-embedded
by Gernot Heiser 31 May '06

31 May '06

NICTA's ERTOS team is proud to release LMBench performance figures for Wombat, our architecture-independent para-virtualized Linux running on L4-embedded/Iguana. The benchmarks show that Wombat's context-switching overheads are up to thirty times lower than those of native Linux on an XScale processor. This result, which is due to Wombat benefiting from the fast-context-switching support provided by NICTA's L4-embedded, demonstrates that L4-embedded/Iguana is the leading high-performance microkernel and embedded virtualisation technology, particularly on ARM processors. For details see http://ertos.nicta.com.au/news/home.pml#v-gt-r Gernot

1 0

sigma1
by Luke A. Guest 26 May '06

26 May '06

Hi, I've just started to look at L4 and I'm using L4ka::Pistachio as a starting point. I'm trying to read as much as possible and have found a reference to sigma1, but no docs on it. Can anyone elaborate on what this is for? Thanks, Luke.

5 7

test
by Luke A. Guest 20 May '06

20 May '06

please ignore

1 0

DoS problem still existing?
by Robert Kaiser 19 May '06

19 May '06

Hello L4 Hackers, a long time ago (*), Jochen Liedtke described a potential denial-of-service attack against the L4 kernel by -IIRC- requesting a huge number of mappings, thereby exhausting kernel memory. Does this issue still exist in the current L4 implementations or has it been solved (How?). Kind Regards Rob (*) see http://i30www.ira.uka.de/research/documents/l4ka/preventing-denial-of-servi… -- Robert Kaiser <rkaiser(a)sysgo.com> SYSGO AG Tel.: +49-6136-9948-0 Am Pfaffenstein 14 Fax: +49-6136-9948-10 55270 Klein-Winternheim http://www.sysgo.com

2 1

con-ux bug?
by Wei Shen 12 May '06

12 May '06

Hi, I run con-ux with: ./fiasco -G 800x600@16 -m 128 -l names -l dm_phys -l con-ux The output is: >*con | Using fast memcpy. >*con | Character size is 8x14, font has 256 characters >*con | VESA reports 800x600@16 1600bpl (0000) [4096kB] >*con | Color mapping: red=11:5 green=5:6 blue=0:5 res=0:0 > ... .... >*con | Mapping I/O UX video mem c0000000 => 00400000+000000 [4096kB] >*con | Using MMXEXT for colorspace transform >KERNEL: 7.2 (tcb=201c1000) killed: >Unhandled trap >EAX 00000025 EBX 0000122b ECX 0000000a EDX 00000000 >ESI 00000320 EDI 00b0d73c EBP 012ffda8 ESP 012ffda4 >EIP 00b0dc7f EFLAGS 00200206 >CS 0002 SS 007b DS 007b ES 007b FS 0000 GS 0000 >trap 13 (General Protection), error 00000402, from user mode >(internal event regarding IDT gate descriptor no. 0x80) The backtrace: >main // l4con/server/src/main.c >do_switch >vc_clear // vc.c >vc_fill >pslim_fill // pslim.c >sw_fill In "sw_fill", "vc->do_drty" equals 11589436, which seems invalid. Maybe, it should be NULL? Maybe, we should add "new->do_drty = fg_do_drty;" to line 402 in main.c? Regards, Wei

3 8

L4Linux freezes
by Vilmos Nebehaj 12 May '06

12 May '06

Hi, I've compiled and installed Fiasco and L4Linux-2.6 and it works, except that when I start a simple test program that creates lots of pthreads, the whole L4Linux task freezes (simply pthread_create() a thousand threads after each other is enough to reproduce the problem). I can ping the machine from another host, but can't login through ssh, the console becomes totally unresponsive, etc. At the same time, the following message appears in the log: simplets| ERROR: task #23 doesn't match: client 10.03, owner 0.0 l4lx | l4ts_create_task(): failed (server=7.00, ret=-3, exc=0) Other L4 tasks continue running uncluttered. Do you have an idea why this happens? Is this a bug in Fiasco, or is it a misconfiguration problem? -vili

2 2

Re: TCP/IP between L4Linux and L4 task
by Bjoern Doebel 11 May '06

11 May '06

2 1

Re: TCP/IP between L4Linux and L4 task
by Stefan Scheler 11 May '06

11 May '06

Hi, > Sure. Both, L4Linux and your app need to be compiled without native > hardware support. Instead, use a software network switch like the one > provided by l4/pkg/ore. L4Linux needs to compiled with the ORe driver > and your app needs to send/receive packets through ORe. Well, sounds good. So I'm currently trying to integrate the latest e1000 driver (v7.0.33) into ORe as my e1000 adapter is not supported by the current one but encountered the following problem: The updated driver wants to alloc some pages using alloc_page(). This results in a undefined reference to "_alloc_pages" when linking ore: ==> Linking ore LD_PRELOAD=/home/sts/L4_full/l4/tool/gendep/libgendep.so GENDEP_TARGET=ore GENDEP_BINARY=ld gcc -o ore ../../../../../lib/x86_586/crt0.o main.o openclose.o clients.o recv.o send.o netdevice.o irq_handling.o rxtx_string.o debug.o util.o events.o checksum.o xmit_lock.o worker.o local_deliver.o dummy_dsi.o -Wl,--whole-archive ../../../server/lib/net/OBJ-x86_586-l4v2/liblinuxnetdrv.a -Wl,--no-whole-archive -Ttext=0x01800000 -L../../../server/lib/linuxemul/OBJ-x86_586-l4v2 -L../../../../../lib/x86_586/l4v2 -L/home/drops/lib/x86_586/l4v2 -L../../../../../lib/x86_586 -L/home/drops/lib/x86_586 -L../../../../../lib -L/home/drops/lib -T../../../../../lib/x86_586/main_stat.ld -llinuxemul -levents -llinuxnetdrv -ldde_linux -lomega0 -lio -ll4rm -lore_server-worker -lore_server -lore_notify -llist_alloc -lparsecmdline /usr/lib/gcc/i686-pc-linux-gnu/3.4.6/libgcc.a -static -lgeneric_ts -ll4env -ll4rm -ldm_generic -ldm_mem -lthread -lsemaphore -llogserver_capsule -lnames -ll4util_root -ll4util -lsigma0 -lroot.o -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/i686-pc-linux-gnu/3.4.6/libgcc.a -luc_c -ll4sys -Ttext=0x01800000 -Wl,-gc-sections -Wl,--warn-common ../../../server/lib/net/OBJ-x86_586-l4v2/liblinuxnetdrv.a(e1000_main.o): In function `e1000_alloc_rx_buffers_ps': /home/sts/L4_full/l4/pkg/ore/server/lib/net/OBJ-x86_586-l4v2/../../../../server/include/linux_headers/linux/mm.h:454: undefined reference to `_alloc_pages' collect2: ld returned 1 exit status I suppose I have to link something else? Any hints appreciated. BTW is ORe an acronym for something? What does it stand for? Stefan

3 4

[Fwd: Re: TCP/IP between L4Linux and L4 task]
by Bjoern Doebel 11 May '06

11 May '06

3 4

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 May 2006