- l4-hackers - example.com Site

accessing msrs
by Ryan Brown 16 Jan '13

16 Jan '13

How does fiasco.oc handle accessing intel model specific registers? Is there any way to access these from user space or does this have to happen in the kernel? Specifically I'd like to access the paravirtualized clock when running under kvm.

2 1

Kernel unit tests
by Bryce Cronkite-Ratcliff 09 Jan '13

09 Jan '13

Hi Hackers, I've made modifications to several parts of the Fiasco.OC kernel and am interested in any existing tests that particularly stress kernel modules (eg scheduler, mem). I've written a handful of simple ones, but I was wondering whether there were particularly promising folders to check out somewhere in the repo. Thanks! Bryce

2 1

No known ld emulation found
by Torsten Rüger 25 Dec '12

25 Dec '12

Hi, I got the latest fiasco sources, tried to compile on a macbook pro (intel core 2 duo) but i get the above message. I did try to set it to amd64 with intel duo option. This is not my final target, i just wanted to play a bit (possibly on an emulator). Do i need a cross compiler or something , because i only have gcc 4.2.1. Something simple i'm missing ? Torsten hopea:build raisantorsten$ make test -e source || ln -sf /Users/raisantorsten/Desktop/fiasco-2012111623/src/kernel/fiasco/src source Cleaning up build directory Creating .Modules.deps make srcdir=/Users/raisantorsten/Desktop/fiasco-2012111623/src/kernel/fiasco/src objbase=/Users/raisantorsten/Desktop/fiasco-2012111623/src/kernel/build -f /Users/raisantorsten/Desktop/fiasco-2012111623/src/kernel/fiasco/src/Makefile.sub1 rm -f \ touch .Clean-auto -n Checking compiler config... CHANGED. Recompiling. No known ld emulation found make[1]: *** [.Host-config] Error 1 make: *** [create-sources] Error 2

3 3

Raspberry pi ?
by Torsten Rüger 25 Dec '12

25 Dec '12

Hi, sorry I'm not really a os "hacker" and I haven't even got the code yet, but was wondering if someone could help me with my issue: Is there any activity of L4 on raspberry pi (a pocket size computer) ? Or is there elsewhere I can ask. Specifically the plan would be to build a ruby-machine on it. Single user but multi-process. If would only need to be able to access usb,ethernet and gpios. And run and start several ruby processes. I know ruby vm well enough to cut that down to size. But I have no idea of how difficult it would be to remove linux and replace it with those 3 device drivers. Or even where to get those, how to integrate them. So i was hoping somebody here could say how crazy this is, or if I can stamp down some trodden path ? Thanks Torsten

3 2

Page tables apparently don't get completely freed on 64-bit Fiasco.OC
by Christian Prochaska 13 Dec '12

13 Dec '12

When running the test case from https://github.com/genodelabs/genode/issues/559 on 64-bit Fiasco.OC, the kernel runs out of memory after the sequential construction and destruction of tasks. I've added some debug output to the kernel which tracks the kernel memory allocations and deallocations. The attached log files cover the creation and destruction of one task on 64-bit and 32-bit Fiasco.OC. In the 64-bit case the allocations reported in lines (05/06), (07/08), (12/13), (14/15), (20/21) and (24/25) don't have matching calls of 'free()'. These allocations belong to the 'Ptab::Walk::sync()' and 'Ptab::Walk::walk()' functions. It looks like only a part of the page table gets destroyed during 'Mem_space::dir_shutdown()'. In the 32-bit case the amout of memory available after the destruction of the task is the same as before the construction, but what I found odd about the log output is that the allocations reported in lines (10/11) and (14/15) don't have matching calls of 'free()', whereas the allocations reported in lines (09) and (12/13) each have two calls of 'free()' (in lines (22/25) and (28/32)), which makes the amount of memory available balanced again. Christian

2 1

L4linux on virtual box
by FU_LONG 13 Dec '12

13 Dec '12

Hi, I have built L4linux(l4linux-mag-x86.iso) and made it run on qemu successfully. But when I try to run it on virtual box, it stops, please see the following picture: I have put the iso file in Attachment, could you please have a try and give me some advise? ------------------ Best Regard, Long Fu 从QQ邮箱发来的超大附件 l4linux-mag-x86.iso (42.18M, 2013年01月06日 17:05 到期)进入下载页面：http://mail.qq.com/cgi-bin/ftnExs_download?t=exs_ftn_download&k=7…

2 1

Nova+Nul one Execution Context suspend a second Execution Context
by Mehdi AICHOUCH 12 Dec '12

12 Dec '12

Hello Alexander, I wrote a simple example to test the NOVA+NUL. In my example, I created two Execution Contexts. When the the first Execution Context starts, it obtains a portal capability to the second Execution Context, and run a while(true) loop as follow : ... while(1) { Logging::printf("#tsc=%llu\n", now); nova_call(pt); } ... The second Execution Context run also a while(true) loop as follow: ... while(1) { count++; Logging::printf("WorkerThread is working... count= %u\n", count); } ... The first Execution Context call the second Execution Context. When the second Execution Context receive the call in its portal function, it calls the msleep( ) function on a TimerHelper instance that it has already created: unsigned portal_func(Utcb &utcb, Utcb::Frame &input, bool &free_cap, cap_sel pid) { Logging::printf("Thread will sleep... count=%u\n", count); timer->msleep(1000); return ENONE; } The problem is that, it is the first Execution Context who is blocked and not the second Execution Context. Could anyone please give me an example of how can I achieve the inverse case, where the first EC continue running and suspends the second EC. this is my code to obtain the portal: cap_sel pseudonym = alloc_cap(); res = ParentProtocol::get_pseudonym(*utcb, "s0/ec2", 0, pseudonym); cap_sel pt = alloc_cap(); res = ParentProtocol::get_portal(*utcb, pseudonym, pt, true, "s0/worker_thread"); Thank you in advance. Mehdi

2 1

Re: startstop test case of the NOVA-NUL microhypervisor
by Alexander Boettcher 10 Dec '12

10 Dec '12

Hi, try to add "name::/s0/timer" to the hello config. Cheers, Alexander Böttcher. Mehdi AICHOUCH <foxmehdi(a)gmail.com> schrieb: >_______________________________________________ >l4-hackers mailing list >l4-hackers(a)os.inf.tu-dresden.de >http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers

2 1

startstop test case of the NOVA-NUL microhypervisor
by Mehdi AICHOUCH 07 Dec '12

07 Dec '12

Hello everyone, I am trying to test the NOVA+NUL microhypervisor. I played with the test application "nul/michal/apps/tests/startstop". So, I modified this test to try a new a new test case, by modifying the "hello" program that is started by the "startstop" program. I added to the Hello program the following code : ... TimerHelper * timer; timer = new TimerHelper(this); while(1) { Logging::printf("Hello ... \n", now); timer->msleep(1000); } ... This simple code will print a message and suspend the Hello program for 1 second. The problem is that when I tested this program alone, it worked perfectly but when this same program is started by the "startstop" program using start_config(), it does not suspend on the msleep(1000). Is my test correct or allowed by the NOVA-NUL microhypervisor? If it is the case, could anyone please help to debug this issue. Many thanks in advance. Mehdi PS: this is my configuration script. #!/usr/bin/env novaboot # -*-sh-*- HYPERVISOR_PARAMS=serial QEMU_FLAGS=-cpu phenom -smp 1 -serial stdio bin/apps/sigma0.nul tracebuffer_verbose S0_DEFAULT hostserial hostvga hostkeyb:0,0x60,1,12,2 \ script_start:1 script_waitchild service_config bin/apps/startstop.nul bin/apps/startstop.nulconfig <<EOF sigma0::cpu:1 sigma0::mem:16 name::/s0/log name::/s0/config name::/s0/timer name::/s0/fs/rom name::/s0/admission || rom://bin/apps/startstop.nul rom://test.nulconfig EOF bin/apps/hello.nul test.nulconfig <<EOF sigma0::cpu:1 sigma0::mem:16 name::/s0/log name::/s0/admission || rom://bin/apps/hello.nul EOF

1 0

how to use initramfs with l4linux
by Ryan Brown 05 Dec '12

05 Dec '12

How do I use an initramfs with l4linux? I tried using "l4x_rd=rom/initrd.cpio.gz" but it doesn't work. I see linux trying to use it: INITRD: 15000000 - 15114e00 ... Unpacking initramfs... INITRD: Freeing memory. ... But then I get Failed to execute /init Kernel panic - not syncing: No init found. Try passing init= option to kernel. See Linux Documentation/init.txt for guidance. I'm able to use the same initramfs file to boot normal linux kernel. I tried various combinations of cpio formats and compression but it doesn't seem to help. The only thing that works is compiling the initramfs into the kernel which I want to avoid.

1 0