- l4-hackers - example.com Site

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

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

21 Dec '06

>Hi, >On Wed Dec 20, 2006 at 14:00:26 +0800, halo.duan wrote: >> and I check the host linux configuration. the page offset is >> CONFIG_PAGE_OFFSET=0xC0000000 which is correct as your >> suggestion.I guss that fiasco-ux must be failed to interpret ELF >> of sigma0/roottask/irq0 and execute it . I put the same "hello" >> bin in grub/ fiasco environment. it works fun. JDB does not help >> me becuase that fiasco-ux enter a stage that CRTL-C does not >> work, I can not bring the system to jdb stage with CRTL-C. ps >> can find three processes which are 'fiasco-ux,[I][irq0] and >> [fiasco-ux]. I can use strace for fiasco-ux and [i][irq0]. but >> failed to trace [fiasco-ux]. >Thats correct, as the fiasco-ux is already ptracing the [fiasco-ux] >binary. I think you haven't shown us the full log of the segfault and >the text before yet, so please post it. We should be able to see more >when we see the log. >Adam following is my screen copy of fisco-ux. it's freezed there. I can not use CTRL-C to bring up the jdb. [root@athena l4]# ./fiasco-ux -l hello 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: Fri Dec 8 18:40:04 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. ^^^^^^ freezed to above ps shows there are three related process 1955 pts/0 R+ 2:18 ./fiasco-ux -l hello 1956 pts/0 S+ 0:00 [I](irq0) 1957 pts/0 T+ 0:00 [fiasco-ux] strace shows segment fault, i think it's normal for execute "CLI" in userland .

3 2

hiworld example
by Shakthi Kannan 20 Dec '06

20 Dec '06

Hi, After checking out hello, l4/pkg/generic_ts, l4/pkg/cxx, l4/kernel/fiasco, l4/tool/preprocess from CVS and doing: make -C l4 O=build oldconfig cd l4 make O=build stops at: l4/pkg/generic_ts/lib/clientlib/server.c:16:26: error: l4/log/l4log.h: No such file or directory l4/pkg/generic_ts/lib/clientlib/server.c:17:26: error: l4/env/errno.h: No such file or directory l4/pkg/generic_ts/lib/clientlib/server.c:18:31: error: l4/names/libnames.h: No such file or directory What do I need to checkout for log, env and names? System is x86 running Ubuntu Dapper with make-3.81-2. Is there a complete tarball that contains the entire sources - L4env, fiasco, L4Linux, and others, that I can simply download and build, rather than having to check-out individual modules from CVS? Thanks, SK -- Shakthi Kannan http://www.shakthimaan.com __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

5 8

AW: L4Linux on L4Ka::Pistachio
by Christian, Martin, OPEE45 20 Dec '06

20 Dec '06

Hi Adam! >> a) Is the L4Env required for running L4Linux? > Yes but that's no strict requirement. How much time would it probably take, to adjust L4Linux without L4Env to the L4 Interface of Pistachio? Did anybody try this? >> This is part of my "Diplomarbeit" at the University of Leipzig. > That isn't too far away from here. Well, actually I'm in Friedrichshafen with EADS to write my thesis. But I'm studying at the University of Leipzig. >> b) How much time would it take to port L4Env to L4Ka::Pistachio? > It would probably take me a month or two That's definitely more time, than I've got left. I also assume you know L4Env much better than I do, so it would take me even longer. >> d) What about the Fiasco port to PowerPC from Mathias Lange? > This didn't come to an end. That's a pity! :-( >> e) If Fiasco would be running on PowerPC, would it be possible >> to run L4Linux without changes? > If one wants to have binary compatibility then some PPC code needs > to be there. And then that would not be the way I'm currently doing > things. As far as I understood Michael Hohmuths Diplomarbeit, all Hardware access from Linux is replaced by sycalls to L4. Is that still the case? And so far only the IA32 part of Linux is changed that way, isn't it? So I would basically need to rewrite the PowerPC architecture part of Linux, basically. Isn't it the way things are supposed to be done? > Yes, L4Linux-2.4 also runs in Pistachio, at least it did when we worked > on it. It's all in the l4ka CVS. I'll try it. :-) > I suppose support on this is basically non-existent. L4Ka::Pistachio for PowerPC is also not really supported, so I'm used to not getting support! :-( Thanks, Martin.

3 2

L4linux compile errors..
by Wankeun Jo 20 Dec '06

20 Dec '06

Hi, I use L4Linux-2.6.18 version on pxa255 board. l4linux boots well and then now tried to expand my work for device drivers. I compiled dice and dm_generic, dm_phys packages as well. I want to use the file in "l4linux/arch/arm/mach-pxa/*", so I add the files to be compiled in "l4linux-2.6/arch/l4/Makefile" And I link all the files of "include/asm-arm/arch-pxa/*.h" in "asm-l4/arch-arm/arch" for use. And then I tried to compile l4linux, but failed with the following messages --------------- CHK include/linux/version.h CHK include/linux/utsrelease.h CC arch/l4/kernel/asm-offsets.s In file included from include/linux/sched.h:49, from arch/l4/kernel/asm-offsets.c:93: include/linux/jiffies.h:211:31: missing binary operator before token "#" In file included from include/linux/sched.h:49, from arch/l4/kernel/asm-offsets.c:93: include/linux/jiffies.h: In function `timespec_to_jiffies': include/linux/jiffies.h:316: error: syntax error at '#' token include/linux/jiffies.h:316: error: parse error before "Current" include/linux/jiffies.h:316: error: syntax error at '#' token include/linux/jiffies.h:316: error: parse error before "Current" include/linux/jiffies.h:316: error: syntax error at '#' token include/linux/jiffies.h:316: error: parse error before "Current" include/linux/jiffies.h:316: error: syntax error at '#' token include/linux/jiffies.h:316: error: parse error before "Current" include/linux/jiffies.h:316: error: syntax error at '#' token include/linux/jiffies.h:316: error: parse error before "Current" . . . . include/linux/jiffies.h:389: error: syntax error at '#' token include/linux/jiffies.h:389: error: parse error before "Current" include/linux/jiffies.h:417:6: missing binary operator before token "#" --------------- The error messages are so long. What's the problem?? Is it a compiler problem?? Thank you.

2 1