users

users@lists.genode.org

3 participants
2349 discussions

AW: AW: Block Cache
by Wolfgang Schmidt 30 Apr '14

30 Apr '14

Hi, The last time the build Problem was mit on Hardware, but on building it with the run/Demo for qemu. This now works after the comments. What haven't worked right now is using the live cd on real hardware. ( neither older L4 mir newer Linux live Demo) So I asked to know wether a specific Plattform is known to work as I habe no Nova knowledge so far ( to reduce the risk of asking to many stupid questions). Of I understand You correctly the was of using a filesystem with cache wozld be ahci as driver -> blockcache as server -> rump_fs as single client of the blockcache and as server for -> multiple programs accessing files. Is this correct? Best regards Wolfgang ----- Ursprüngliche Nachricht ----- Von: "Stefan Kalkowski" <stefan.kalkowski@...1...> Gesendet: ‎29.‎04.‎2014 09:37 An: "genode-main(a)lists.sourceforge.net" <genode-main(a)lists.sourceforge.net> Betreff: Re: AW: Block Cache Hi Wolfgang On 04/28/2014 10:25 PM, Wolfgang Schmidt wrote: > Hi Stefan, > Can You give a gibt for compatible x86 hw? I tried The democd on an FSC > Athlon64 which did mit start at all. Does for example a comp list for > nova exist? No, I've no compatibility list available. I know of quite different Intel PCs people are using doing experiments with Genode/NOVA. What do you mean with it didn't start at all on your specific hardware? The last time I've read an email from you regarding problems with NOVA on real hardware, it failed to compile the framebuffer driver, because your third-party-code "libports" repository probably wasn't prepare accordingly. Maybe it's better to smooth build problems out, instead of searching for new hardware? > > What so You mean with: it serves only one client- one client application? I meant, the block cache acts like a block driver. In general a block driver in Genode serves only one client, e.g.: a filesystem server. If you want to share the same device with multiple clients, you might use the "part_blk" server, which provides different partitions, found on the device, as separate block sessions to different clients. Does that clarifies your question? Regards Stefan > > Best regards > Wolfgang > ------------------------------------------------------------------------ > Von: Stefan Kalkowski <mailto:stefan.kalkowski@...1...> > Gesendet: ‎28.‎04.‎2014 21:13 > An: genode-main(a)lists.sourceforge.net > <mailto:genode-main@lists.sourceforge.net> > Betreff: Re: Block Cache > > Hi Wolfgang, > > On 04/28/2014 08:27 PM, w_schmidt@...181... wrote: >> Hello,, >> >> I have a few question regarding genodes block cache in 14.02. >> >> If understand it correctly on x86-based computers genode starts on top of >> linux. > > well, that's not true. Yes, Genode is able to run on top of Linux, > mostly for development purposes. But Genode runs also directly on x86 > hardware using the NOVA kernel, or Fiasco.OC kernel, or the OKL4 kernel, > or Pistachio kernel. > >> Therefore multiple caches will be used before something goes to genode >> a) the cache on a hdd >> b) cache on a drive controller >> c) the linux block cache > > If running on Linux yes, but normally it wouldn't make much sense to use > an additional block cache here. > >> d) qemu > > As already said: you don't need to use Qemu. Most run scripts are > starting Qemu by default, for convenience reasons only. > >> >> The first one is about the scenario in which the block cache is used. >> Is the block cache only thought for embedded devices where genode runs >> nativly? Are all embeded devices usable with the cache? > > The block cache uses the block session of Genode. It uses exactly one > block device, and serves exectly one client. As all block device drivers > of Genode are implementing the block session interface, the block cache > can be used on top of all block drivers, whether its a x86 AHCI driver, > or MMC card driver of an ARM SoC, or whatever. > >> >> Is there a way to mount a partition in genode with qemu without using > caches >> from c) and d) ? >> Or is there a x86 hardware where genode can directly operate on? > > Qemu allows to define the caching policy of a hard drive (cache=none). I > don't know whether it's possible to circumvent Linux block cache. I > assume that's hard. > But as mentioned above, you can use conventional x86 PC hardware. > > Regards > Stefan > >> >> As far as I have seen so far the disc cache of Linux does not seem to be >> disablable. I have only foudn ways to flush linux cache (like sync && > echo 3 >> > /proc/sys/vm/drop_caches). >> >> Best regards, >> Wolfgang >> >> >> > ------------------------------------------------------------------------------ >> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE >> Instantly run your Selenium tests across 300+ browser/OS combos. Get >> unparalleled scalability from the best Selenium testing platform > available. >> Simple to use. Nothing to install. Get started now for free." >> http://p.sf.net/sfu/SauceLabs >> _______________________________________________ >> Genode-main mailing list >> Genode-main(a)lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/genode-main >> > > -- > Stefan Kalkowski > Genode Labs > > http://www.genode-labs.com/ · http://genode.org/ > > ------------------------------------------------------------------------------ > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE > Instantly run your Selenium tests across 300+ browser/OS combos. Get > unparalleled scalability from the best Selenium testing platform available. > Simple to use. Nothing to install. Get started now for free." > http://p.sf.net/sfu/SauceLabs > _______________________________________________ > Genode-main mailing list > Genode-main(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/genode-main > > > ------------------------------------------------------------------------------ > "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE > Instantly run your Selenium tests across 300+ browser/OS combos. Get > unparalleled scalability from the best Selenium testing platform available. > Simple to use. Nothing to install. Get started now for free." > http://p.sf.net/sfu/SauceLabs > > > > _______________________________________________ > Genode-main mailing list > Genode-main(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/genode-main > -- Stefan Kalkowski Genode Labs http://www.genode-labs.com/ · http://genode.org/ ------------------------------------------------------------------------------ "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE Instantly run your Selenium tests across 300+ browser/OS combos. Get unparalleled scalability from the best Selenium testing platform available. Simple to use. Nothing to install. Get started now for free." http://p.sf.net/sfu/SauceLabs _______________________________________________ Genode-main mailing list Genode-main(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main

2 1

manually constructing the type list of RPC functions
by Mhamad Hmad 28 Apr '14

28 Apr '14

hi , the RPC function number in each interface is limited by 9 , I have more than 9 functions . Genode have away to do that by manually constructing the type list of RPC functions instead of using the convenience macro typedef Meta::Type_tuple<Rpc_myfunc1, Meta::Type_tuple<Rpc_myfunc2, Meta::Empty> > Rpc_functions; I add that to my intrface session but i get the next error *error: ‘Meta’ does not name a type* and another error * no type named ‘Rpc_functions’ in ‘struct myrpcsession::Session’* I looked to the "cpu_session.h" and tried to do same but i think i miss some thing . any clue? best , mohammad

2 4

make run/demo stop unexpectedly
by Sun Keqin 28 Apr '14

28 Apr '14

Hi, My colleague met a very strange problem today, so I ask for help. There is an error when he run "make run/demo” in build directory, however, when he run "qemu-system-i386 -no-kvm -m 256 -serial mon:stdio -cdrom var/run/demo.iso" directly, it works successfully. Do you know why it happened? What should I do to fix that? Here is the details. genode build completed using fiasco kernel /home/jimmy/work/os/genode/build/kernel/fiasco.oc/fiasco using sigma0/bootstrap at /home/jimmy/work/os/genode/build/l4 creating ISO image... spawn qemu-system-i386 -no-kvm -m 256 -serial mon:stdio -cdrom var/run/demo.iso L4 Bootstrapper Build: #1 2014: spawn id exp6 not open while executing "expect { -i $running_spawn_id -re $wait_for_re { } eof { puts stderr "Error: Spawned process died unexpectedly"; exit -3 } timeout { puts ..." (procedure "wait_for_output" line 11) invoked from within "wait_for_output $wait_for_re $timeout_value $qemu_spawn_id" (procedure "spawn_qemu" line 51) invoked from within "spawn_qemu $wait_for_re $timeout_value" (procedure "run_genode_until" line 18) invoked from within "run_genode_until forever" (file "/home/jimmy/work/os/genode/genode-13.05/os/run/demo.run" line 139) invoked from within "source $include_name" ("foreach" body line 3) invoked from within "foreach include_name [get_cmd_arg --include ""] { puts "using run script $include_name" source $include_name }" (file "/home/jimmy/work/os/genode/genode-13.05/tool/run" line 730) Cheers, keqin

4 4

Major revision of the C runtime (VFS)
by Norman Feske 28 Apr '14

28 Apr '14

Hello, I have just merged a significant change of the C runtime into our staging branch. Because this change will likely affect users of the framework (hopefully for the better ;-), I'd like to bring up this topic here. Since many years, our C runtime served us quite well, thanks to the plugin architecture that allows us to combine different backends such the lwIP socket API using libc_lwip_nic_dhcp, using LOG as stdout via libc_log, or using a ROM dataspace as a file via libc_rom. However, the original design of the plugin architecture is starting to show its limitations. Examples of the current limitations: * There is a libc_fs plugin that allows a program to access files from a file-system server. But there is no way to allow one program to access two different file-system servers. * There is a libc_block plugin that hands out a block session as a pseudo block device named '/dev/blkdev'. However, when combined with libc_fs, it is not defined which of both plugins will handle the file with this name. As a quick and dirty work-around, the libc_fs plugin explicitly black-lists '/dev/blkdev'. * In general, if multiple plugins are combined, there is no consistent virtual file system structure exposed via getdirentries. * There is no flexible concept for handling stdio. Most programs use libc_log to direct stdout to the LOG service. But what if we want to direct the output of such a program to a terminal? Granted, there exists the 'terminal_log' server to translate a LOG session to a terminal session but it would be much nicer to have this flexibility at the C runtime level. * Many libc plugins look similar. There are quite a few dusty corners where duplicated code has been accumulated over the years. That said, the semantic details (e.g., the quality of error handling) differ from plugin to plugin. Seeing the number of file systems (and thereby the number of added libc plugins) growing, it became apparent that our original design would not scale too well. On the other hand, we have gathered very positive experiences with the virtual file system implementation of Noux. It allows to us have a coherent VFS with support for stacked file systems (similar to union mounts) of various types. The VFS of Noux is stable and complete enough to accommodate our tool chain. Wouldn't it be a good idea to reuse the Noux VFS for the normal libc? This line of thoughts led to the following issue: https://github.com/genodelabs/genode/issues/999 The result of this line of work has been merged into the staging branch now. I adapted most of the run scripts and targets to the changed libc but certainly missed a few bits. I.e., the following run scripts remain untested and are expected to fail in one way of another: * http_srv_tracing_nonblocking_panda * qt_avplay (the qt4 version) * netperf * l4linux_dynamic * noux_gdb_dynamic * debug_nitpicker * noux_gdb * virtualbox * virtualbox_auto I would appreciate your help to rectify them, so that we can merge the staging branch to the master branch soon. How has the libc changed? Each libc-using program can be configured with a program-local virtual file system as illustrated by the following example: <config> ... <libc stdin="/dev/null" stdout="/dev/log" stderr="/dev/log"> <vfs> <dir name="dev"> <log/> <null/> </dir> <dir name="etc"> <dir name="lighttpd"> <inline name="lighttpd.conf"> ... </inline> </dir> </dir> <dir name="website"> <tar name="website.tar"/> </dir> </vfs> </libc> </config> Here you see a lighttpd server that serves a website coming from a TAR archive (which is obtained from a ROM module named "website.tar"). There are two pseudo devices "/dev/log" and "/dev/null", to which the "stdin", "stdout", and "stderr" attributes refer. The "log" file system consists of a single node that represents a LOG session. The web server configuration is supplied inline as part of the config. (Btw, you can try out a very similar scenario using the 'ports/genode_org.run' script) The VFS implementation resides at 'os/include/vfs/'. This is where you can see the file system types that are available (look for '*_file_system.h' files). Because the same code is used by Noux, we have one unified and coherent VFS implementation throughout the framework now. There are two things needed to adapt your work to the change. * Remove the use of the libc_{rom, block, log, fs} plugins from your target description files. Those plugins are no more. As of now, the VFS is still internally a plugin, but it is always included with the libc. * Configure the VFS of your libc-using program in your run script. For most former users of the sole libc_log plugin, this configuration looks like this: <config> <libc stdout="/dev/log" stderr="/dev/log"> <vfs> <dir name="dev"> <log/> </dir> </vfs> </libc> </config> For former users of the other plugins, there are the 'block', 'rom', and 'fs' file-system types available. I hope that you will like the new VFS infrastructure. If you have any questions, bug reports, or critique, feedback is always welcome. 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

3 3

Creating your first Genode application - Build succesful, but execution fails
by unknown＠example.com 21 Apr '14

21 Apr '14

Hello, I tried to run the genode tutorial. I first installed it on an AMD Athlon x64 with nForce4- Mainboard. run/demo stopped there after init -> nitpicker] create session with args: label="launchpad", ram_quota=12288 I made a second try on a) a virtualBox machine with brand new installation of Unbuntu 12.04 x64 (git) and b) an old Evo N610c Pentium 4 with Ubuntu 12.04 x86 (14.02 from tar.gz of latest releases) I think a few Steps are missing within the Step-By-Step tutorial http://genode.org/documentation/developer-resources/client_server_tutorial As a prestep for “Setting up the build environment” i created a new build directory with './tool/create_builddir linux_x86 BUILD_DIR=build.hello'. As i did not have a lot of files in it, I tried a first ‘make’. In build/etc/build.conf I added REPOSITORIES += $(GENODE_DIR)/hello_testIn the bin directory i created a new config directory with config file. The server section seems to have only one .h file, I put the three listings in one main.cc file (as later in the document it is mentioned that in target.mk SCR_CC = main.cc an obviously no includes to the other files are given). Afterwards i tried ‘make hello’ which returned no error. With ‘cd run’, ‘./core’ only the following output appered: |int main(): --- create local services --- |int main(): --- start init --- |int main(): transferred 17592186044415 MB to init |int main(): --- init created, waiting for exit condition --- |[init] _map_local: lx_mmap failed (addr_in=0,addr_out=ffffffffffffffed/-19) overmap=0 |[init] C++ runtime: Genode::Rm_session::Region_conflict |[init] void* abort(): abort called If i try to run it again with dmesg the follwoing output appears: [ 4774.659901] core[12189]: segfault at 7f2ed162e068 ip 00000000010310c9 sp 00000000400ffce0 error 4 in core[1000000+5b000] (equal if i try git verison or archive version) I also tried to point the repository to the already existing hello_tutorial, but get the same result. What am I doing wrong? Best regards Wolfgang

4 12

Using Lwip with fdopen
by Mhamad Hmad 18 Apr '14

18 Apr '14

Hi, I built small test server application using lwip stack to work like echo server . i tried to associate the socket with a stream by using fdopen but it failed . when i using the mode "r" or "r+" it give invalid parameter error, in the same time when i try to write on the file descriptor returned by lwip_accept it is write to the client . the part of code is : while (1) { magsock = lwip_accept (socket_v , 0, 0 ); if (magsock ==-1) { perror ("lwip_accept"); } else { // this is work char *buf= "hi"; if( write (,magsock,buf , sizeof(buf) )<0) { perror ("write"); } // this is not work FILE *file; file= fdopn(magsock ,"r"); if (file==NULL) { perror ("fdopn"); } int c ; while(c=fgetc(file)) printf(c); fclose(file); } } any clue or suggestion ? best, --

2 1

Some error in using Qt on genode
by longjks 18 Apr '14

18 Apr '14

Hello, I just want to development Qt application on Genode. However I am new to Qt. I do not know how to use Qt on Genode. I found there are some example Qt applications on genode-13.08/libports/src/qt5. I try to run qt5.run on Platform of linux_x86, however it can not work. here is my steps: 1) ./tool/create_builddir linux_x86 BUILD_DIR=build.linux_x86 2) cd build.linux_x86 3) make run/qt5 here is the log: android@...115...:~/longjks_new/genode-13.08_longjks/build.linux_x86$ make run/qt5 using run script /home/android/longjks_new/genode-13.08_longjks/base-linux/run/env using run script /home/android/longjks_new/genode-13.08_longjks/libports/run/qt5.run building targets: core init drivers/input/ps2 drivers/pci drivers/framebuffer drivers/timer server/nitpicker server/liquid_framebuffer app/qt5/qt_launchpad app/qt5/examples/calculatorform app/qt5/examples/tetrix spawn make core init drivers/input/ps2 drivers/pci drivers/framebuffer drivers/timer server/nitpicker server/liquid_framebuffer app/qt5/qt_launchpad app/qt5/examples/calculatorform app/qt5/examples/tetrix make[1]: Entering directory `/home/android/longjks_new/genode-13.08_longjks/build.linux_x86' checking library dependencies... Skip target drivers/framebuffer/imx53 because it requires imx53 Skip target drivers/framebuffer/omap4 because it requires omap4 Skip target drivers/framebuffer/pl11x/pbxa9 because it requires pl11x platform_pbxa9 Skip target drivers/framebuffer/pl11x/vea9x4 because it requires pl11x platform_vea9x4 Skip target drivers/framebuffer/pl11x/vpb926 because it requires pl11x platform_vpb926 Skip target drivers/framebuffer/vesa because it requires vesa Skip target drivers/input/ps2/pl050 because it requires pl050 Skip target drivers/input/ps2/x86 because it requires ps2 Skip target drivers/pci/device_pd because it requires nova Library platform Library libc-string Library libc-locale Library libc-stdlib Library libc-stdio Library libc-gen Library libc-gdtoa Library libc-inet Library libc-stdtime Library libc-regex Library libc-compat Library libc-setjmp Library ldso-startup Library syscall Library base-common Library startup Library cxx Library base Library ldso_crt0 Library ldso-arch Library alarm Library timed_semaphore Library ld Library libc Library libc_log Library stdcxx COMPILE allocator-inst.o /home/android/longjks_new/genode-13.08_longjks/libports/contrib/stdcxx-4.7.2/src/c++98/allocator-inst.cc:30:18: fatal error: memory: No such file or directory compilation terminated. make[3]: *** [allocator-inst.o] Error 1 make[2]: *** [stdcxx.lib] Error 2 make[1]: *** [gen_deps_and_build_targets] Error 2 make[1]: Leaving directory `/home/android/longjks_new/genode-13.08_longjks/build.linux_x86' Error: Genode build failed make: *** [run/qt5] Error 252 I can see the error caused by the source code below: #include <memory> what is this "memory"? where can i found this "memory"? thanks.

2 1

Linux Rump Kernel for Genode?
by Peter Lindener 18 Apr '14

18 Apr '14

Hello to those dwelling on on planet Genodia Genode's use of the NetBSD rump Kernel driver<http://genode.org/documentation/release-notes/14.02#NetBSD_file_systems_usi…> seems like it's a real win.. Is there a chance the Genode OS might do the same kind of thing with a "Linux rump kernel" to also support up to date device driver code that is know to work with the latest Linux systems? .... My hunch is it could be a big win! All the best -Peter

2 2

Is there any profiler on Genode like lmbench on Linux?
by Kloisiie 11 Apr '14

11 Apr '14

Dear all, Is there any profiler on Genode like lmbench on Linux? Shuo Wang UCAs

2 1

An Exploration of ARM TrustZone Technology
by Norman Feske 10 Apr '14

10 Apr '14

Hello all, recently, we have seen a growing interest in ARM TrustZone on the mailing list. If you are interested in this topic, you may consider the following article, which we just published today: http://genode.org/documentation/articles/trustzone I hope that it will answer the most common questions about ARM's security extensions and how our work relates to it. Links to the corresponding Git branches are included. You will also find a video of Genode running in the "secure world", and Android running in the "normal world" of an FreeScale i.MX53-based tablet. 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

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users