- l4-hackers - example.com Site

Locks in Base_mappable class
by Yuxin Ren 24 Aug '14

24 Aug '14

Hi All, In the Base_mappable class, in /fiasco/src/kern/mapping_tree.cpp file, what is type of the lock in this class, helping lock, spin lock or others? And usually I find it is hard to identify the type of a lock. Could someone tell me how to figure out the lock's type? Thank you very much. Yuxin

2 2

Re: How can a thread map memory to itself
by Björn Döbel 22 Aug '14

22 Aug '14

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, let us stay on the list - others might be interested in your problems as well. > Thank you very much, I tried what you said, But met a problem. In > my test case, I have multiple threads within the same task, but > running on different cores, doing capability mapping to itself. > Each thread runs in a tight loop, which are 1,000,000 iterations. How does this look like in code? > The problem is after some while, the program did not make any > progress. And this only happens when I have more than 5 threads. If > I only have 2 or 3 threads, they work well. I have no idea how to > debug this. Could you give me some hints? Fiasco has a kernel debugger which you can enter by pressing <ESC> at runtime. Pressing <h> will give you a list of commands. Some interesting ones: lp -> lists all present threads. It will show you thread IDs, on which CPU they are running, what address space they belong to and which state they are in. t<id> -> gives you information about a specific thread with a given debug ID (the one from the lp list). Here you will see at which user or kernel EIP the thread is right now. This may help you figuring out where in the code your threads get stuck. Q -> gives you a list of all known kernel objects. Use this for instance to find the task your program is running in. dt<id> <addr> -> dump the address space of a given task starting at address <addr> Good luck! Bjoern -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 Comment: Using GnuPG with Icedove - http://www.enigmail.net/ iEYEARECAAYFAlP1AeYACgkQP5ijxgQLUNkq7QCeJ5HtstwWw3PAIQTrsXpBh9xT 5m0An04wkCPcYU2d5W6rPzBwCvMBD/1S =0wdI -----END PGP SIGNATURE-----

4 5

Doubt about fiasco-l4re on FastModels
by Maria Soler 21 Aug '14

21 Aug '14

Hello, I want to have a scenario where fiasco-l4re runs in the secure world and Linux runs in the non-secure world. For this I am using the CortexA15x1 model in FastModels. I am using the snapshot 2014053111 for fiasco-l4re. I have already compiled and am able to run fiasco-l4re in the secure side. At the moment I am trying to make the example pkg/examples/sys/vm-tz work (included in the L4 version available at http://svn.tudos.org/repos/oc/tudos/trunk/) but I have been unsuccessful so far. I have added the header files that were missing and have been able to compile the example but it fails when I execute it. Has any of you tried to run this example on FastModels? Do you have any ideas why it may fail? It fails both when running the version of fiasco-l4re where it originally appears (with "terminate called after throwing an instance of 'N2L413Runtime_errorE'") and also in the latest version (with "L4Re[rm]: unhandled write page fault @0 pc=100b868"). Also I would like to know why this example is not included in more recent versions of fiasco-l4re. I am particularly interested in fiasco-l4re running on TrustZone in secure mode and something else running in the non secure mode. I would like to know if there is any other example that does something similar and is more recent (and works on FastModels, ideally) or if there is a better way to do this. Thank you very much for your help, María.

2 1

How can a thread map memory to itself
by Yuxin Ren 19 Aug '14

19 Aug '14

Hi All, I want to do cap/mem mapping without IPC to another thread. In particular, I want a thread to map its own memory or capability to itself. How can I achieve this? Thank you. Yuxin

2 1

Re: Mapping memory into process address space
by Björn Döbel 19 Aug '14

19 Aug '14

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 >> Also, do not call the allocator twice within the function call, >> because then you would obtain two different dataspaces and this >> is not what you want here. Better do >> >> L4:Cap<Dataspace> x = >> L4Re::Util::cap_alloc.alloc<L4::Dataspace>(); long alloc = >> L4Re::Env::env()->mem_alloc.alloc(size, x); if (!x.is_valid()) { >> // handle error } > > Okay, I've gotten the allocation of the dataspace to work correctly > (at least as far as I can tell), but I'm still not sure how to map > the resulting dataspace into the newly created process. I'm trying > something like this: > > L4::Cap<Dataspace> stack = > L4Re::Util::cap_alloc.alloc<L4::Dataspace>(); < allocate memory > > task->map(L4Re::This_task, stack.fpage(), 0); > > but it still freezes. Is the fpage() call still waiting on IPC, or > should I be using a different value for snd_base? The snd_base parameter to the map() call is the address in the target address space. Nevertheless, I'm wondering while this would hang. Are you sure this happens in the map() operation? Bjoern -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 Comment: Using GnuPG with Icedove - http://www.enigmail.net/ iEYEARECAAYFAlPzSo4ACgkQP5ijxgQLUNmVHACfYrqIXq5opXk2e2nzac6fe3AB a10AnitnVNIpOIBPQOhqV1RGMD7j5Z5K =xWcr -----END PGP SIGNATURE-----

1 0

Adding a new scheduling algorithm to Fiasco.OC
by Valentin Hauner 19 Aug '14

19 Aug '14

Hi, I'm planning to add a new scheduling algorithm to Fiasco.OC. I will implement EDF. First of all, is there some handout written by you main developers that copes with the issues of implementing EDF in Fiasco.OC? What I've achieved so far: I've created the files sched_context-edf.cpp and ready_queue_edf.cpp in ./kernel/fiasco/src/kern, using the corresponding source files for your fixed-priority scheduler as templates. I've named the config directive used in ./kernel/fiasco/src/Modules.* files analogous to yours, that is 'sched_edf' in my case. Now, I'm beginning to modify sched_context-edf.cpp and ready_queue_edf.cpp to fit the requirements of EDF, especially the en- and dequeuing as well as the 'dominates' method. Are there any other major source files that I do have to alter? Thanks in advance! Best regards, Valentin

2 17

I share a document to access HW on L4 Linux.
by Taeung 18 Aug '14

18 Aug '14

Thank you L4 Hackers. You have help me in securing Ethernet on L4 Linux. So, I succeeded it. I’m particularly grateful to Martin Schröder, Wei jin, Adam Lackorzynski, Matthias Lange. Although there is already a document 'L4 Runtime Environments : Io, the Io server' a link of which is http://os.inf.tu-dresden.de/L4Re/doc/io.html, I made a guide document for accessing HW on L4 Linux because of ethernet example. (I've used ARM platform Freescale i.MX6 Quad SABRESD ) This document is https://www.dropbox.com/s/abj9ao5uk9r3w72/The_Guide_to_access_HW_on_L4Linux… I don’t know whether this document can help someone in accessing HW on L4Linux. But I want to share this document in order to repay for l4hackers's kindness. - Taeung -

3 5

compiling and running l4android-arm in qemu
by cem akpolat 18 Aug '14

18 Aug '14

Dear all, I have compiled l4android, l4 and fiasco for arm architecture and now I am attempting to run elf image in qemu.However, I see only a black screen in qemu. There is any io messages. What could be the reason? My qemu version is 1.7.1. The images that I have after making elfimages are: bootstrap.elf bootstap_L4LinuxARM bootstap_L4LinuxARM.elf The used command for running image: $qemu-system-arm -M realview-pbx-a9 -m 2048 -kernel bootstrap_L4LinuxARM.elf -serial stdio Thanks in advance, Sincerely, Cem Akpolat

2 1

Mapping memory into process address space
by Noah Zentzis 15 Aug '14

15 Aug '14

------------------------------------------------------------------------ I don't want to use libloader because I'm trying to create a new process with dynamically-generated assembly (and alter it while the process is running), so I can't really output an ELF binary. On 08/18/2014 08:50, Björn Döbel wrote: > The dataspace that prog_attach_ds() is called with is allocated by the > ELF loader in libloader/include/elf. > > reserve_utcb_area() is supposed to not attach a real dataspace but > reserve a virtual memory area that will not be used by the new > application's RM to attach any other memory, because this is where the > UTCBs will later be mapped to by the kernel iirc. I'm pretty sure I need to do this myself, since any calls to map() are causing the program to freeze (could this be because the task doesn't have its UTCB allocated yet?). I'm also not sure what to use for the destination offset of the L4::Task::map() call, so that might be the issue. > At this point your best way to do so is > to use the L4::Task::map() operation on the capability of your newly > created task. How do I specify what point in the target the memory maps to? It looks like the L4::Dataspace::map() would work, but without being able to map the capability, I can't really check. Thanks! - Noah Zentzis

3 3

Build system
by teclis High Elf 14 Aug '14

14 Aug '14

If I execute this command from the l4 build directory: make E=hello S=bootstrap In which Makefile or script is the S= processed (its a bit of a difficult thing to grep for :) thanks

2 3