Thank you very much -    qemu-system-i386 -m 512M -cdrom ./images/l4linux-mag-x86.iso -serial stdio                                  ^^^^^^^this solved the problem. ... but the next problem apeared: l4linux | Invalid IO-Port access at pc = 00654b2d port=0xcf8 l4linux | Die message: Trap: 13 l4linux | panic: going to sleep forever, bye   I understand, that L4Linux needs an IO-Port - namely the PCI-Bus, but io doesn't allows it. What resources should be given to l4linux ?  I guess the configuration itself is in  x86-legacy.devs or l4lx-x86.io, where x86-legacy.devs seems the description  and  l4lx-x86.io the policy ?   http://l4linux.org/use.shtml  describes the modules.list without using the io-server ? Does it grap the whole system ? In the L4Linux - menuconfig I see a CONFIG_VPCI. It sounds very useful to have a virtual PCI bus - for virtualize more than one instance. Is the virtualpci an server component ?   Is there any manual I have to read ?    Rudolf Weber Gesendet: Montag, 15. Juli 2013 um 00:31 Uhr Von: "Adam Lackorzynski" <adam@os.inf.tu-dresden.de> An: "Rudolf Weber" <rwib@web.de> Cc: "l4 l4hackerlist" <l4-hackers@os.inf.tu-dresden.de> Betreff: Re: l4linux: l4x_setup_memory:Cant get main memory On Sun Jul 14, 2013 at 00:25:05 +0200, Rudolf Weber wrote:

The configuration with the modaddr leads to region conflicts, so I got modaddr 0x02000000 default-kernel fiasco -serial_esc default-bootstrap bootstrap entry L4Linux-mag-x86 roottask moe rom/l4lx-gfx.cfg module l4re module ned module l4lx-gfx.cfg module io module fb-drv module mag module x86-legacy.devs module l4lx-x86.io module vmlinuz module ramdisk-x86.rd

to boot.

The .config file for l4linux is the x86-native_defconfig (the only difference is the setting of CONFIG_L4_OBJ_TREE).

Why can't l4linux get the memory ? l4linux | l4x_setup_memory: Can't get main memory of 65536kiB! What must be done to make the l4linux-Configuration working ?

Hm, tried adding more memory to the VM? Adam -- Adam adam@os.inf.tu-dresden.de Lackorzynski http://os.inf.tu-dresden.de/~adam/[http://os.inf.tu-dresden.de/~adam/]

Gesendet: Donnerstag, 18. Juli 2013 um 23:58 Uhr Von: "Adam Lackorzynski" <adam@os.inf.tu-dresden.de> An: "Rudolf Weber" <rwib@web.de> Cc: "l4 l4hackerlist" <l4-hackers@os.inf.tu-dresden.de> Betreff: Re: l4linux: l4x_setup_memory:Cant get main memory On Thu Jul 18, 2013 at 16:42:26 +0200, Rudolf Weber wrote:

...

The Booting-Result is better now, but an new error occurs:

l4linux | Invalid IO-Port access at pc = 0080b476 port=0x110

In the System.map I found 0080b400 t el3_init_module which is in the src/l4linux/drivers/net/ethernet/3com/3c509.c - Module. Here the driver scans the io ports 0x110,0x120,...,0x200

Ok - I will throw it out. Yes, it boots further, loads the RAM-Disk and the booting in the mag-Window shows further errors. I can't see the errors that quickly, and they don't appear in the log of qemu.

Question1: How can I see the error messages of L4linux on serial lines? I tried this configuration without mag (and hope, all l4linux-Output should appear at the seril console) modaddr 0x002000000 entry L4Linux kernel fiasco -serial_esc roottask fiasco -serial_esc roottask moe rom/l4linux.cfg module l4linux.cfg module l4re module ned module io module x86-legacy.devs module l4lx-x86.io module vmlinuz module ramdisk-x86.rd and l4linux.cfg loader = L4.default_loader; -- io - server vbus_l4linux = loader:new_channel(); vbus_input = loader:new_channel(); loader:start( { caps = { sigma0 = L4.cast(L4.Proto.Factory, L4.Env.sigma0):create(L4.Proto.Sigma0); icu = L4.Env.icu; input = vbus_input:svr(); l4linux = vbus_l4linux:svr(); }, }, "rom/io rom/x86-legacy.devs rom/l4lx-x86.io"); loader:start( { caps = { log = L4.Env.log:m("rws"), vbus = vbus_l4linux; }, l4re_dbg = L4.Dbg.Warn, log = { "l4linux", "yellow" }, }, "rom/vmlinuz mem=64M l4x_rd=rom/ramdisk-x86.rd root=1:0 ramdisk_size=4000 init=/bin/sh"); Result: I can't see any severe errors, but the prompt of the /bin/sh is not shown :-( I opened the Fiasco - Debugger, looked the backtrace of vmlinux-thread and the question come: Question2: How do I map the addresses to the System.map? -> Is there any tutorial available ? The l4linux configuration x86-ux_defconfig works with grub-Module-configuration modaddr 0x02000000 default-kernel fiasco -serial_esc default-bootstrap bootstrap entry L4Linux-mag-x86 roottask moe rom/l4lx-gfx.cfg module l4re module ned module l4lx-gfx.cfg module io module fb-drv module mag module x86-legacy.devs module l4lx-x86.io module vmlinuz module ramdisk-x86.rd Related to Question1 is Question3:How can I see the messages of l4linux on serialline ? My try: The configuration: modaddr 0x002000000 entry L4Linux kernel fiasco -serial_esc roottask fiasco -serial_esc roottask moe rom/l4linux.cfg module l4linux.cfg module l4re module ned module io module x86-legacy.devs module l4lx-x86.io module vmlinuz module ramdisk-x86.rd with x86-legacy.devs and l4lx-x86.io from l4re-snapshot-2013060718 and l4linux.cfg like above don't lead to the prompt of the /bin/sh but I can't understand what is wrong.

Hello, thank you for the good tip with    Gesendet: Donnerstag, 25. Juli 2013 um 23:50 Uhr Von: "Adam Lackorzynski" <adam@os.inf.tu-dresden.de> An: "Rudolf Weber" <rwib@web.de> Cc: "l4 l4hackerlist" <l4-hackers@os.inf.tu-dresden.de> Betreff: Re: l4linux starting - How can I see all Debug messages on serial line ?

...

Question1: How can I see the error messages of L4linux on serial lines? Add console=ttyLv0 to the L4Linux argument list. So I can see the output on all configurations.

The state is now: 1) Linux-conf x86-ux_defconfig without mag : runs 2) Linux-conf x86-ux_defconfig with mag : runs 3) Linux-conf x86_rw_config without mag : runs 4) Linux-conf x86_rw-config with mag : error My x86_rw-config is in the Attachment. The error is: attempt to access beyond end of device ram0: rw=0, want=4764729332, limit=8000 Buffer I/O error on device ram0, logical block 2382364665 attempt to access beyond end of device ram0: rw=0, want=4606413050, limit=8000 Buffer I/O error on device ram0, logical block 2303206524 Attachement log_ioerror ist the whole output. Why this IO-Error ? Does something overwrites the ramdisk ? Or is the ramdisk shortened ? I give in qemu 1024M and Linux in the cmdline 128M,but the phenomenon is still there. Thank you very much Rudolf Weber The configuration is: Modulelist: modaddr 0x02000000 default-kernel fiasco -serial_esc default-bootstrap bootstrap entry L4Linux-mag-x86 roottask moe rom/l4lx-gfx.cfg module l4re module ned module l4lx-gfx.cfg module io module fb-drv module mag module x86-legacy.devs module l4lx-x86.io module vmlinuz module ramdisk-x86.rd -- Start io vbus_l4linux = loader:new_channel(); vbus_input = loader:new_channel(); vbus_fbdrv = loader:new_channel(); loader:start( { caps = { sigma0 = L4.cast(L4.Proto.Factory, L4.Env.sigma0):create(L4.Proto.Sigma0); icu = L4.Env.icu; input = vbus_input:svr(); l4linux = vbus_l4linux:svr(); fbdrv = vbus_fbdrv:svr(); }, }, "rom/io rom/x86-legacy.devs rom/l4lx-x86.io"); -- Start fb-drv (but only if we need to) local fb = L4.Env.vesa; if (not fb) then fb = loader:new_channel(); loader:start({ caps = { fb = fb:svr(), vbus = vbus_fbdrv }}, "rom/fb-drv -m 0x117"); end local mag_mag = loader:new_channel(); local mag_svc = loader:new_channel(); -- Start mag loader:start( { caps = { vbus = vbus_input; mag = mag_mag:svr(); svc = mag_svc:svr(); fb = fb; }, }, "rom/mag"); -- Start Linux loader:start( { caps = { log = L4.Env.log:m("rws"), fb = mag_svc:create(L4.Proto.Goos, "g=640x480"); vbus = vbus_l4linux; }, l4re_dbg = L4.Dbg.Warn, log = { "l4linux", "yellow" }, }, "rom/vmlinuz mem=64M console=ttyLv0 l4x_rd=rom/ramdisk-x86.rd root=1:0 ramdisk_size=4000 init=/bin/sh"); the other files are from l4re-snapshot-2013060718

Gesendet: Samstag, 27. Juli 2013 um 23:56 Uhr Von: "Adam Lackorzynski" <adam@os.inf.tu-dresden.de> An: "Rudolf Weber" <rwib@web.de> Cc: "l4 l4hackerlist" <l4-hackers@os.inf.tu-dresden.de> Betreff: Re: l4linux starting - How can I see all Debug messages on serial line ? Hi,

On Fri Jul 26, 2013 at 14:55:13 +0200, Rudolf Weber wrote:

...

Gesendet: Donnerstag, 25. Juli 2013 um 23:50 Uhr Von: "Adam Lackorzynski" <adam@os.inf.tu-dresden.de> An: "Rudolf Weber" <rwib@web.de> Cc: "l4 l4hackerlist" <l4-hackers@os.inf.tu-dresden.de> Betreff: Re: l4linux starting - How can I see all Debug messages on serial line ?

...

...

Question1: How can I see the error messages of L4linux on serial lines? Add console=ttyLv0 to the L4Linux argument list. So I can see the output on all configurations.

The state is now: 1) Linux-conf x86-ux_defconfig without mag : runs 2) Linux-conf x86-ux_defconfig with mag : runs 3) Linux-conf x86_rw_config without mag : runs 4) Linux-conf x86_rw-config with mag : error

My x86_rw-config is in the Attachment. The error is:

attempt to access beyond end of device ram0: rw=0, want=4764729332, limit=8000 Buffer I/O error on device ram0, logical block 2382364665 attempt to access beyond end of device ram0: rw=0, want=4606413050, limit=8000 Buffer I/O error on device ram0, logical block 2303206524

Attachement log_ioerror ist the whole output.

Why this IO-Error ? Does something overwrites the ramdisk ? Or is the ramdisk shortened ?

I give in qemu 1024M and Linux in the cmdline 128M,but the phenomenon is still there.

When the ramdisk is too big for Linux, it will complain. Since the same config works when not starting mag(?) does it work when just not giving the fb cap to L4Linux?

Yes, I reproduced again by modifing the configuratin: * adding fb_drv - in the list - it works * without mag it works, with mag it failes. * If I don't give the fb-cap to L4Linux, I can't see the l4linux-Window - this is what I expect. I now put the mag-module as last module (behind the ramdisk for linux) - I expect, the order of further modules don't matter ->/home/rw/l4exp/l4re-snapshot-2013060718/src/l4/pkg/bootstrap/server/src/startup.cc:526: void try_move_module(const Region&, l4util_mb_info_t*, unsigned int, long unsigned int, unsigned int): Assertion `to' failed. -> unsigned long long to = regions.find_free(module_area, size, L4_PAGESHIFT); returns null May this destroy the ramdisk too ? (Oh, I remember: I applied the patch from http://os.inf.tu-dresden.de/pipermail/l4-hackers/2013/005903.html to fix the bootstrap-problemon x86_64 of the hello world-Application. Since l4linux is not ported to x86_64, I tried x86) Greetings Rudolf

...

(Oh, I remember: I applied the patch from http://os.inf.tu-dresden.de/pipermail/l4-hackers/2013/005903.html to fix the bootstrap-problemon x86_64 of the hello world-Application. Since l4linux is not ported to x86_64, I tried x86)

Could you try without the patch please?

Yes, I did it: It runs ! As modaddr I tried 0x01100000 like in genode, but this collidates with an module. 0x02000000 works. Experience: I put the mag -Servermodule at the end - it worked - i put it behind io -Module - it worked -> So the position is not very important (I think, the first modules bootstrap, fiasco, sigma0, ... should be in that order) The patch is good for the bootstrap-Problem on x86_64, but not with the configured linux-Kern. Thank you very much! I think, the configuration of the root-servers (which must have a different load addresses because the kernel has not started and build the Virtual Address spaces) is the first challenge (Herausfordung) with a L4-Microkernel. And I think, if someone replaces the root-Servers (i.e. Genode,...) he/she will have the same problem again. The above patch seemed tho rearrange the modules, but someone should improve it. To solve the configuration problem for newbies I suggest: * Perhaps the bootloader grub could read out the inital layout. * And then there should be a compiletime tool which adjust the LMA according to the initial address space. ---- My next goal is to have an l4linux-System with a harddisk and a compiler. (Then I want to test my UML toolbox first on l4linux, then native on L4) There is an firm for process visualization and an idea would be asking them, if the real time window system dope may be useful for them. (If they agree, I will contact the TU Dresden and Genode labs to discuss the business terms.) According an mail I read the Dissertation http://www.genode-labs.com/publications/secure-gui-2009.pdf But N.F. writes in his mail, that l4dope is not more in the focus. In 2003, I had a grub-configuration http://www.infnet.verein.de/betriebssysteme/kleinkernarch/l4/l4boot/dope.lst Is mag an successor of nitpicker ? Is it an replacement for l4dope ? Does someone already running configuration ? Rudolf http://www.infnet.verein.de/home/rw/index.html