users

users@lists.genode.org

2357 discussions

Manually set MSI in Genode under base-foc failed
by Jilong Kuang - SISA 15 Jan '13

15 Jan '13

Hi all, I was trying to manually (hack) set MSI in Genode under base-foc but failed. I'd like to describe my approach here and see if anyone can spot any mistakes or missing parts? First, I added the Msi_flag (0x80000000) to the IRQ number in the l4_icu_bind() and setup_irq_mode() functions in genode/base-foc/src/core/irq_session_component.cc. So the kernel will know this is a MSI-type IRQ. In addition, I added a l4_icu_msi_info() function to return msg info for driver msi configuration (In my case it returns the first available vector in IDT, which is 0x21) Second, in my driver code, I do the following: Genode::Irq_connection _irq(80); //random MSI number Dev->config_msi(0x21); //configure the dev (e1000 card in my case) to use MSI mode while (1) { _irq.wait_for_irq(); Dev->interrupt_handler(); } However, I never saw any MSI-type interrupt triggered by Fiasco kernel. After instrumenting kernel, I did see kernel created an MSI-type irq object and the corresponding Irq_sender object and attached itself to the right receiving thread (Interrupt_handler::handler_cap()). But I don't know why Irq_sender::kinvoke was not triggered in the kernel. Thank you very much for your help! Best regards, Jilong

1 0

MSI support in Genode
by Jilong Kuang 14 Jan '13

14 Jan '13

Hi all, Could any briefly point out how to use MSI in Genode for device driver (e.g., e1000 card)? The existing IRQ API seems to only handle normal interrupt case. Is there any sample code I can refer to? I'm using Fiasco.OC based kernel so MSI itself is supported by the kernel. Thank you very much! Best regards, Jilong

3 4

libpulse symbols beginning by pa_ not found while linking fb_sdl (newbie)
by Paul Dufresne 14 Jan '13

14 Jan '13

Hi, I am new to Genode and love it (particularly the release notes!). I am having a problem linking fb_sdl with while compiling Genode 12.11 on LuninuxOS (based on Ubuntu Quantal) 64 bits: paul@...162...:~/Bureau/Genode/genode-12.11/build.lx$ make run/demousing run script /home/paul/Bureau/Genode/genode-12.11/base-linux/run/env using run script /home/paul/Bureau/Genode/genode-12.11/os/run/demo.run building targets: core init drivers/timer server/nitpicker server/liquid_framebuffer app/launchpad app/scout test/nitpicker server/nitlog drivers/framebuffer drivers/pci drivers/input spawn make core init drivers/timer server/nitpicker server/liquid_framebuffer app/launchpad app/scout test/nitpicker server/nitlog drivers/framebuffer drivers/pci drivers/input make[1]: Entering directory `/home/paul/Bureau/Genode/genode-12.11/build.lx' checking library dependencies... Skip target drivers/framebuffer/fiasco_ux because it requires fiasco 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/fiasco_ux because it requires fiasco Skip target drivers/input/ps2/pl050 because it requires pl050 Skip target drivers/input/ps2/x86 because it requires ps2 Skip target drivers/timer/codezero because it requires codezero Skip target drivers/timer/fiasco because it requires fiasco Skip target drivers/timer/foc because it requires foc Skip target drivers/timer/hw/omap4 because it requires hw omap4 Skip target drivers/timer/hw/pbxa9 because it requires hw_pbxa9 Skip target drivers/timer/hw/vea9x4 because it requires hw_vea9x4 Skip target drivers/timer/nova because it requires nova Skip target drivers/timer/okl4_arm because it requires okl4 arm Skip target drivers/timer/okl4_x86 because it requires okl4 Skip target drivers/timer/pistachio because it requires pistachio Library platform Library mini_c Library libz_static Library libpng_static Library syscall ... Program drivers/framebuffer/sdl/fb_sdl COMPILE fb_sdl.o COMPILE input.o LINK fb_sdl /usr/local/genode-gcc/bin/../lib/gcc/x86_64-elf/4.7.2/../../../../x86_64-elf/bin/ld: warning: libpulsecommon-2.1.so, needed by /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse-simple.so, not found (try using -rpath or -rpath-link) /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse.so: undefined reference to `pa_tagstruct_get_boolean' /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse.so: undefined reference to `pa_pdispatch_run' /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse.so: undefined reference to `pa_dbus_add_matches' /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse.so: undefined reference to `pa_memblockq_free' /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse.so: undefined reference to `pa_detect_fork' ... /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse-simple.so: undefined reference to `pa_log_level_meta' ... /usr/lib/gcc/x86_64-linux-gnu/4.7/../../../x86_64-linux-gnu/libpulse.so: undefined reference to `pa_tagstruct_putu32' collect2: error: ld returned 1 exit status make[3]: *** [fb_sdl] Error 1 make[2]: *** [fb_sdl.prg] Error 2 make[1]: *** [gen_deps_and_build_targets] Error 2 make[1]: Leaving directory `/home/paul/Bureau/Genode/genode-12.11/build.lx' Error: Genode build failed make: *** [run/demo] Error 252 paul@...162...:~/Bureau/Genode/genode-12.11/build.lx$ pon But libpulse-dev seems intalled: paul@...162...:~/Bureau/Genode/genode-12.11/build.lx$ apt-cache showpkg libpulse-dev Package: libpulse-dev Versions: 1:2.1-0ubuntu4 (/var/lib/apt/lists/ca.archive.ubuntu.com_ubuntu_dists_quantal_main_binary-amd64_Packages) (/var/lib/dpkg/status) Description Language: File: /var/lib/apt/lists/ca.archive.ubuntu.com_ubuntu_dists_quantal_main_binary-amd64_Packages MD5: a36e2b24ab289ab86564c79e92c8e87f Description Language: en File: /var/lib/apt/lists/ca.archive.ubuntu.com_ubuntu_dists_quantal_main_i18n_Translation-en MD5: a36e2b24ab289ab86564c79e92c8e87f Description Language: fr File: /var/lib/apt/lists/ca.archive.ubuntu.com_ubuntu_dists_quantal_main_i18n_Translation-fr MD5: a36e2b24ab289ab86564c79e92c8e87f Reverse Depends: libpulse-dev:i386,libpulse-dev 1:2.1-0ubuntu4 libpulse-dev:i386,libpulse-dev 1:2.1-0ubuntu4 libpulse-ocaml-dev,libpulse-dev libgmtk-dev,libpulse-dev libfluidsynth-dev,libpulse-dev libsdl1.2-dev,libpulse-dev Dependencies: 1:2.1-0ubuntu4 - libpulse0 (5 1:2.1-0ubuntu4) libpulse-mainloop-glib0 (5 1:2.1-0ubuntu4) libglib2.0-dev (0 (null)) libavahi-client-dev (0 (null)) libpulse-dev:i386 (3 1:2.1-0ubuntu4) libpulse-dev:i386 (6 1:2.1-0ubuntu4) Provides: 1:2.1-0ubuntu4 - Reverse Provides: paul@...162...:~/Bureau/Genode/genode-12.11/build.lx$

1 2

OM: A real-world, tangible, desktop implementation of Genode
by John Bessa 14 Jan '13

14 Jan '13

Hello all, The following writing is my attempt to suggest a real-world application for Genode that "philosophically" gives more "responsibility" to the OS/framework at the expense of application-dominance. Starting in the late-90s, I felt that the OS should be closer to the applications (such as web-, data-, and logic-servers in opposition to the movement, which was to disassociate applications from OSs). My job was maintaining both OSs and applications (as an administrator) and I saw efficiency by blurring the differences between them. My OS model (called the Thinman) was a single OS/application that would have libraries/modules that would be "called upon" and brought to windows on the desktop that had "browsed" for material such as pictures and edited the material using tools (libraries and modules). Of course, new material could be initiated with the "New" button. More recently, I have personally concluded that the SVG, or Scalable Vector Graphics page, should be at the heart of the future Web, and the WC3 seems to agree by proscribing full HTML capability within SVG pages. Thus, nearly anyone can easily make an exceedingly attractive page (far more so than FaceBook or Google+), and then paste in the written material as basic html text. Given this, I tend to use Inkscape (IS) SVG editor a lot. Previously, I used GIMP, but GIMP is an image editor, not a page editor. So, today, I use both, primarily to cover IS's weaknesses in graphics manipulation. As you might predict, the interrelation between the two systems is awkward. You can "cut" from GIMP using the mouse or "ctl" key combinations, and "paste" to IS, and the pasted image can be manipulated, but cannot be saved as an SVG page. (I use snipping tool to make web images). In light of my "native" OS beliefs, I "fantasized" a solution that has an OS-based graphic image editor that I call Object Manipulator (OM). OM takes the input stream from the mouse (or ctl-keys), creates from it an object, and then applies the libraries of tools and filters from GIMP (and other sources) directly to the object, including all the cutting/selecting tools. The image, when fully altered, is then selected and "pasted" back into IS, or other format-specific editor, to be made into a presentable product. Given this model, there is no reason the "pasted" object has to be a graphic; it could be a poem, and the libraries brought to it might include a thesaurus. Thus, OM has no specific purpose, except to bring together objects and libraries (which are technically also objects, but that is a different discussion). OM has no connection to "reality" as its space is "an abstraction." It only works in the context of the OS and has the desktop as its "background" in ways no different than a repairman's bench-top is the "background" for fixing/developing some device such as an appliance. It does not even intend to use the filesystem to communicate with applications (though of course it can by converting to established file formats or by creating a native one). The object might not even need a background, or even an application called OM to call it; it may be defined in of itself as a selected object, perhaps defined in the context of, or possessed by, the mouse. I am hoping that this tangible example (that I emulate using IS and GIMP) can help provide a real-world basis, or context, for the Object OS layer that you are attempting with Genode. Implementing this idea involves subtle philosophic/holistic issues that are at the heart of an "agnostic" OS framework; that is to say that given circumstances either technical or political, a kernel can be switched out at the bottom, or, likewise, and object manipulator can be switched out. This suggests that the OS framework must be impartial and democratic --a sort of United Nations of OSs and desktops. I hope that this writing is comprehensible, and I will continue to attempt to simplify it especially for a wider audience. Regards, John -- Photography and sculpture: http://thinan.com/john_bessa/photography Empathy and Emotional Communication http://en.wikiversity.org/wiki/Empathy_Model http://johnbessa.com/empathy Technology and Education: http://thinman.com

2 3

process map information ( like linux pmap )
by Younwoo Kim 14 Jan '13

14 Jan '13

Hello Genode, When running the test program on Fiasco.OC, often a page fault occurs. I assume that, I made a bug by misusing of a shared library. However, I don't know how to get a shared library text. for example, /proc/id/maps on linux: ( or pmap ) root:/proc/1# cat maps b74fa000-b7505000 r-xp 00000000 08:01 10384183 /lib/i386-linux-gnu/ libnss_files-2.15.so b7505000-b7506000 r--p 0000a000 08:01 10384183 /lib/i386-linux-gnu/ libnss_files-2.15.so ... I want to get a section information of shared library for finding a cause of page fault. What should I do for solving this question?

2 1

Genode network performance
by Daniel Waddington 11 Jan '13

11 Jan '13

Hi, On the subject of improving network stack performance, I observe 50% less throughput on Genode+Fiasco.OC+LWIP versus Linux on exactly the same Intel-based hardware. I am sending UDP packets; Linux @ 324K PPS and Genode @ 167 KPPS. Using 64bit kernels. This is using out-of-the-box Genode IPXE DDE with E1000 NIC card. Does this performance difference sound reasonable and expected? Is there some expected significant performance overhead with using DDE as opposed to a native NIC driver? My next step will be to look at raw ethernet packet performance. Daniel

2 1

[genode] Multiple sessions from usb_drv
by jaeyong yoo 11 Jan '13

11 Jan '13

Hello! Is it possible to have multiple sessions from usb_drv? If it is, how does usb_drv multiplex the receiving packets from NIC device to the corresponding session? Best regards, Jaeyong

3 2

rebol language + live CD
by syd bris 10 Jan '13

10 Jan '13

Hi, The Rebol programming language has recently gone open source. It also has an existing open source equivalent of the Red language. Rebol code is designed to accomplish tasks with a minimum of coding. It seems that both Rebol and Genode have similar approaches to computing: a simplified, efficient means of getting things done. Could the adoption of Rebol or Red language help to get goals achieved in Genode? For example, just in getting a usable system up and running, so that more people could use the OS, and further development could then progress faster? Also, have you given up on the live CD demo as a 2013 goal?

2 1

about platform_execute in usb_drv
by jaeyong yoo 09 Jan '13

09 Jan '13

Hello, Could you explain about the intend behind platform_execute in usb_drv? I see it is similar to executing a function but it uses stack allocated by dde_kit_simple_malloc, and I don't understand why. Always thanks for the kind explanation, Jaeyong

3 4

About stack pointers in thread_base
by jaeyong yoo 07 Jan '13

07 Jan '13

Hello, While I was browsing source code, I found two confusing variables: At genode/base/include/base/thread.h Thread_base::Context { long stack[]; addr_t stack_base; } Although, they are commented a bit, it is hard to understand what are they doing and what's the difference. Could you tell me about them more? Best, Jaeyong

2 3

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users