users May 2015

users@lists.genode.org

27 participants
28 discussions

Port of base-hw to RISC-V / lowRISC ?
by Rolf Sommerhalder 22 May '15

22 May '15

Dear list, GSoC has accepted a project to port seL4 to RISC-V (64 bit) architecture, which is used in the lowRISC System-on-a-Chip (SoC) Open Source project. Further, lowRISC supports another project to port rump kernels to minimal RISC-V cores which lowRISC plans to use as I/O co-processors, for example as TCP/IP offload engines [1]. There are tool chains and ports of Linux kernels to RISC-V / lowRISC to run them in simulators and emulators (qemu), on FPGA (Xilinx Zynq), and ASIC [2, 3]. Although my HW/SW co-design skills are a bit rusty, I am tempted to reverse that and try such a port. In order to get a better idea about complexity and effort required, I currently study the sources of base-hw for ARMv7, and its recent port to x86_64 in the staging branch. Is anyone aware of such a porting effort, or has anyone else an interest to run Genode base-hw on RISC-V64, and might want to chime in eventually? Thanks, Rolf [1] http://www.lowrisc.org/blog/2015/05/summer-of-code-students-for-lowrisc/ [2] http://riscv.org/ [3] https://github.com/ucb-bar/rocket-chip

2 1

Building Genode for Linux/ARM
by Alexander Senier 19 May '15

19 May '15

Hi all, I'm trying to build Genode for the Linux base platform for ARM without success so far. Despite having the ARM toolchain installed, my builds always are x86_64. Here is what I did after checking out the source tree: $ cd <genode_dir>/tool $ ./toolchain arm After prompting for my super user password the toolchain gets installed to /usr/local/genode-gcc. The resulting tools (genode-arm-*) appear to work OK. $ ./create_builddir linux_arm $ cd ../build/linux_arm $ make run/demo However, when done an SDL window running the *x86_64* version of the demo pops up. $ file init/init init/init: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped I'm building a current Genode master branch on Debian 8 / x86_64. What am I missing? Cheers, Alex

2 6

An improved Trace tool for the Genode framework
by Duss Pirmin 18 May '15

18 May '15

Hello I have written an improved trace tool for the Genode OS framework. The idea is, to redirect the trace messages to a terminal session. This terminal session can then be connected to over an Ethernet or serial line depending on the terminal server configured. At the moment only one subject can be traced at the time. At the moment I only tested it on Linux and with the Tcp_terminal as server to the "outside" world. You can download it form : https://github.com/trimpim/TraceTool/ Currently open points: - read the configuration from the parent - pause/resume traces - trace of multiple subjects simultaneously Best regards Pirmin

1 0

Genode's core and init running on seL4
by Norman Feske 18 May '15

18 May '15

Hello, I am happy to report that I got the first little Genode scenario (base/run/printf.run) working on the seL4 kernel. For those of you who are interested in following this line of work, I thoroughly documented the steps I had to take: http://genode.org/documentation/articles/sel4_part_3 You can find the corresponding topic branch at: https://github.com/nfeske/genode/commits/sel4 This preliminary version of the seL4 support will be included in the upcoming Genode release. Cheers Norman -- Dr.-Ing. Norman Feske Genode Labs http://www.genode-labs.com · http://genode.org Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth

1 0

signature of equality operators
by Duss Pirmin 18 May '15

18 May '15

Hello I have a the following code: #include <vector> class myClass { private: std::vector<Genode::Trace::Subject_id> mSubjectList; public: bool hasSubjectId(Genode::Trace::Subject_id aSubjectId) const { return std::find(mSubjectList.begin(), mSubjectList.end(), aSubjectId) != mSubjectList.end(); } } If I try to compile it, I get the following error: /home/pirmin/Documents/genode/genode/contrib/stdcxx-fadf4f6ba436fe4846572f00bf057c5ef290c25b/include/stdcxx/bits/stl_algo.h:168:4: error: passing ‘const Genode::Trace::Subject_id’ as ‘this’ argument of ‘bool Genode::Trace::Subject_id::operator==(const Genode::Trace::Subject_id&)’ discards qualifiers [-fpermissive] In base/trace/types.h the signature of the operator==() for Subject_id is defined as follows: bool operator == (Subject_id const &other) { return id == other.id; } Shouldnt it be? bool operator == (Subject_id const &other) const { return id == other.id; } Most recomendations on operator overloading, I have read, recommend this. Best regards Pirmin

2 1

synchronization between normal and secure world
by Stefan Brenner 14 May '15

14 May '15

Hi, just a short question, I am using the TZ VMM example on i.MX53. How can I syncronize accesses to a shared memory range between secure world and normal world? I would be interested in a mutex that can be acquired both from Genode in secure world, as well as from Linux in normal world. Ciao

4 11

Design review for Native File system support - ext2
by Janani k 13 May '15

13 May '15

Hi, I am trying to port an existing native file-system(ext2) to Genode. So, I started porting this in the same place as rump_ext2 so that this can be set up fast to get quick running of the code. I designed it so as to call our native implementation calls instead of rump calls. To get the block sessions working and basic ext2 functionalities checking, I did quick code work to get it up and running. Sorry didn't follow the Genode specifications in coding. I want get suggestions & feedback's on this design and then I will align code to Genode specifications and move further. I am attaching the patch file. And I will list out major things I did in this work to give an overview. 1. It contains totally 3 parts. one which has all Genode related code which used to deal with rump also and handle registry things, so I kept this one as is so as to utilize it again. Second part File-system interface which serves as an abstraction of all the extended file-system types(ext2,ext3,ext4). Third part ext2 this contains the implementation of ext2. 2. All the calls previously to rump will be replaced by this new ext2 implementation. 3. Till now I could get it up running. In this, I linked creating directory or file code, remove a directory, listing of directory entries works. This you can see in main.cc file comments where this gets called. 4. Since I used ram data space the changes I did to the ext2.raw are not persistent. So I removed 3 lines in rump_ext2.run which create ext2.raw file again and again when this demo is run and instead I made it persistent. And I created some directories before hand in this. 5. For mounting it and checking contents use :: $ sudo mount -t ext2 -o loop,rw ./build.lx/bin/ext2.raw /mnt/ext2/ 6 For checking ext2 structure from outside use :: $ dumpe2fs build.lx/bin/ext2.raw Entire development is done running Genode on linux. For compiling and running ext2 use :: make -C build.lx/ run/rump_ext2 Thanks & regards, Janan

3 6

Improved window management
by Nobody III 13 May '15

13 May '15

Genode's window manager is very simple and lacks the following functionality: minimize/maximize/restore close window force kill It seems like the nitpicker service should be modified to support whichever of these functions it doesn't already. I'm unsure of how exactly force kill could be implemented--especially since not all users can be considered trustworthy. How do we fit users into the Genode security model? Users need to be given reasonable authority--complete authority in some cases. Any ideas?

2 4

Segfaults while calling Rpc_methods from client.h
by Aditya Kousik 12 May '15

12 May '15

Hello, Calling Rpc_<method>(args, ...) from Session_client causes segfault 11 when I try to pass anything other than a default datatype. I'm trying to pass an object of std::string, and it prints the string inside the <method> just fine, but when I make the call<Rpc_method>(st), it breaks. I tried working around this by creating a char* pointer instead. Print inside this method works and the rpc to session_component method also reaches, but whatever pointer I receive in the session_component is garbage data. How do I fix the RPC calls? Thanks, Aditya.

2 4

Dynamic linker error?
by Aditya Kousik 11 May '15

11 May '15

Hello, I'm trying to run a modified version of the hello_tutorial to test some libraries that I've recently ported. While I am able to build the binaries, and the boot image, init gives me the following error:

3 8

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users May 2015