users

users@lists.genode.org

2285 discussions

Questions About VMs and Hardware Access
by Zachary Zollers 19 Mar '24

19 Mar '24

Hello Genodians, The last few days, I have been focused on creating my first Linux VM using Seoul. I am happy to report I was largely successful, but have a few questions after completing the exercise. The answers I get from this list have been very helpful in getting through the initial learning curve, so thank you in advance for your responses. [1] While referencing the example Seoul runscripts in the genode-world repository, I did not see an obvious way to facilitate direct PCI passthrough to a VM. Is this supported by the current version of Seoul? If not, is it supported by the VirtualBox port? [2] Is there a good way to support a situation where only one client needs to display and capture input. For example, in my VM scenario, I only care about the visual output of the VM. I was able to accomplish the desired graphical result by configuring a single nitpicker domain that had the entire screen area and no labels. The only issue here is that the single client does not immediately get input focus (I had to click to be able to type in the VM). Am I on the right track here, or is there a better way? [3] What purpose does the platform driver (os/src/drivers/platform) serve? It appears that device discovery is done by either the ACPI driver or pci_decode (possibly others). Is it responsible for delegating access? I assume access to these is granted through an IO_MEM and/or IO_PORT session, but these seem to be provided by core (not the platform driver). Any clarification here would be helpful. Once again, I appreciate your very helpful responses! Thanks, Zachary Zollers

4 3

Running lighttpd.run
by Pirmin Duss 18 Mar '24

18 Mar '24

Hello Genodians After the update to Genode 24.02 one of our more complex run scripts got stuck when trying to download the test-results from the googletest that it did run before. To narrow down the problem, I tried to run the libports/run/lighttpd.run both on 23.11 and 24.02 releases without any of our patches applied. Unfortunately I can't download index.html on both releases. The output is as follows: !% wget http://10.0.2.55/index.html !--2024-03-18 09:49:24-- http://10.0.2.55/index.html !Connecting to 10.0.2.55:80... failed: No route to host. Now I'm thinking, that my setup for the TAP device might not be correct, and would like to ask, if someone can give me a hint what might be wrong? My distribution is Debian 12, and I currently use the following script, to set up the TAP device (commands taken from libports/run/fetchurl.inc and libports/run/lwip_lx.run): !#!/bin/bash ! !set pipefail -e ! !user=pirmin !netdev=enp1s0 !tapdev=tap0 ! !sudo ip tuntap add dev $tapdev mode tap user $user !sudo ip address flush dev $tapdev !sudo ip address add 10.0.2.1/24 brd 10.0.2.255 dev $tapdev !sudo ip link set dev $tapdev addr 02:00:00:ca:fe:01 !sudo ip link set dev $tapdev up ! !sudo iptables -t nat -A POSTROUTING -o $netdev -j MASQUERADE !sudo iptables -A FORWARD -i tap0 -o $netdev -j ACCEPT !sudo iptables -A FORWARD -i $netdev -o tap0 -m state --state !RELATED,ESTABLISHED -j ACCEPT ! !sudo sysctl -w net.ipv4.ip_forward=1 The output from ip address, regarding the TAP device is as follows: !2: enp1s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000 ! link/ether 58:47:ca:73:7b:fc brd ff:ff:ff:ff:ff:ff ! inet <ip-addr> brd <brd> scope global dynamic noprefixroute enp1s0 ! valid_lft 5956sec preferred_lft 5956sec !6: tap0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc fq_codel state DOWN group default qlen 1000 ! link/ether 02:00:00:ca:fe:01 brd ff:ff:ff:ff:ff:ff ! inet 10.0.2.1/24 brd 10.0.2.255 scope global tap0 ! valid_lft forever preferred_lft forever ! inet6 fe80::ff:feca:fe01/64 scope link ! valid_lft forever preferred_lft forever Regards, Pirmin

2 2

Question about Core's Ability to Delegate Errors
by nsf＠ksu.edu 18 Mar '24

18 Mar '24

Hello Genodians, I was working on a project and noticed that the only kernel fault core allows a component to handle is a page fault. Is there a particular reason that core only allows page faults to be handled and not all kernel faults? If there is a problem with receiving all faults I would love to know the issue! thanks in advance for all your help. Sincerely, Nathan

4 5

Questions about Genode on seL4
by Zachary Zollers 14 Mar '24

14 Mar '24

Hello Genodians, I have been playing around with Genode on seL4 for the last few days and have a couple of questions. [1] Did any modifications need to be made to the seL4 source code for Genode to work properly? [2] I encountered a couple of strange warnings when running the "Hello world" scenario from the Genode foundations book. Below is the clipped console log showing the exact messages. . . . Booting all finished, dropped to user space Warning: device memory in range [0000000400000000,0000000800000000) is unavailable (due to limited untyped cnode range) // To which device is this referring? Warning: device memory in range [0000000800000000,0000001000000000) is unavailable (due to limited untyped cnode range) Warning: device memory in range [0000001000000000,0000002000000000) is unavailable (due to limited untyped cnode range) . . . boot module 'config' (469 bytes) boot module 'init' (456304 bytes) Warning: need physical memory, but Platform object not constructed yet // I'm assuming this is a non-issue, but wanted to check Warning: need physical memory, but Platform object not constructed yet . . . Genode 24.02 183 MiB RAM and 523285 caps assigned to init Warning: void Core::Rpc_cap_factory::free(Genode::Native_capability) not implemented - resources leaked: 0x1 // What are the implications of this? Warning: void Core::Rpc_cap_factory::free(Genode::Native_capability) not implemented - resources leaked: 0x2 Warning: void Core::Rpc_cap_factory::free(Genode::Native_capability) not implemented - resources leaked: 0x4 [init -> hello] Hello world I am not yet familiar (still learning) the internals of seL4 and core, so apologies if these questions have obvious answers. Thanks in advance, Zachary Zollers

2 1

enabling iMX.8M Plus based board
by Alice 14 Mar '24

14 Mar '24

Dear Genodians, I started to work on enabling CompuLab's iot-gate-imx8mplus [1]. As the name suggests, this device is based on the i.MX8MP SoC. So far, I have the following things currently working: - u-boot - Linux (6.1) - run/log on top of base-hw This line of work is available on GitHub [2]. I still want to replace the manufacturer's defconfig with tinyconfig, and reduce the size of the Linux kernel. I keep this task for later. Next, I want to draw attention to my work regarding the FEC network driver, eMMC storage, and the USB-host drivers based on the source of the Linux 6.1 port. If someone is already working on these drivers, we could join efforts. [1] https://www.compulab.com/products/iot-gateways/iot-gate-imx8plus-industrial… [2] https://github.com/a-dmg/genode-imx/tree/imx8mp_iot_gate Cheers, Alice

1 0

Mailing List Migration on Thursday, February 8, 2024
by Christian Helmuth 12 Mar '24

12 Mar '24

Hello Genodians, after thorough planning, we finally migrate this mailing list to Mailman 3. Consequently, the list will not be operational tomorrow Thursday, February 8, 2024, all day All current list subscribers stay subscribed and will receive list postings latest starting Friday. Other changes coming with Mailman 3 are as follows. - Registration URL for new list members https://lists.genode.org/mailman3/postorius/lists/users@lists.genode.org - Account Management URL https://lists.genode.org/mailman3/accounts/login/ - Mailing list archive https://lists.genode.org/mailman3/hyperkitty/list/users@lists.genode.org The old archive at https://lists.genode.org/pipermail/users/ will be frozen and preserved. All previous and future postings are available in the new archive. See you on Friday after the migration -- Christian Helmuth Genode Labs https://www.genode-labs.com/ · https://genode.org/ https://floss.social/@genode · https://genodians.org/ Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth

1 2

Sculpt 23.10 .ova file broken with VMware Workstation 17.5.1
by Justin Clift 01 Mar '24

01 Mar '24

Hiyas, Trying out Sculpt 23.10 in VMware Workstation (17.5.1) today, and found some initial hurdles that seem like they should get fixed. :) When looking at options on the Sculpt download page: https://genode.org/download/sculpt ... it seems like a choice of either a .img file, or a .ova file. --- Problem 1: The .img file doesn't seem to work with VMware Workstation, as it just shows "GRUB" (in the top left) like its partially loading... then just hangs there. --- Problem 2: Trying the .OVA file instead (File -> Open -> choose the .ova file), the import process throws up a warning message that the .ova "did not pass OVF specification conformance". (screenshot attached) That warning can be ignored though, in which case VMware keeps going for a few seconds then throws a hard error about "Capacity mismatch for disk2.vmdk". (screenshot for this attached as well) --- It looks like the .ova file format is just a renamed .tar file though. So untarring it works, revealing two VMware disk images (.vmdk). Manually creating a new VM (2 cpu cores, 2GB ram) and assigning those disks to it gets things working. --- Problem 3: VMware doesn't seem totally happy with those disk images. In disk info window (after importing) the "Current Size" for the disk is 16,777,216 TB. Very much not correct. ;) It's also reporting the same figure for "System Free" (also very much not correct), which seems like a VMware bug to me. (screenshot attached too) The field instead has the correct "System Free" value when I check the disks of other VMs on my system. --- Anyway, I thought the above might be useful to mention in case someone wants to look into them. Regards and best wishes, Justin Clift

2 3

Announcement: Genode OS Framework version 24.02 released
by Norman Feske 29 Feb '24

29 Feb '24

I am happy to announce the release of Genode 24.02 today. https://genode.org/news/genode-os-framework-release-24.02 It is safe to say that the release cycle was dominated by infrastructure topics that clear the way for addressing the usability topics on our road map. For example, the new version dissolves a major road block for suspend/resume, namely the suspend-cycle awareness of drivers. It also lays the groundwork for the flexible and robust routing and mixing of audio signals. We are also celebrating the wrapped-up transition to our modern Linux device-driver environment by replacing our dated Linux-based TCP/IP stack to a new port based on DDE-Linux 6.1.20. The wealth of infrastructural topics notwithstanding, the new version also brings exciting news for application developers who can now use the Goa SDK to casually target Sculpt OS as a remote test target. Summary of improvements: - Revised audio infrastructure - Sculpt OS as remote test target for Goa SDK - TCP/IP based on DDE-Linux 6.1.20 - Suspend/resume awareness of GPU, AHCI, NVMe drivers - Improved human-interface device handling - Multi-client use of Vivante GPUs - De-privileged light-weight event tracing The complete story is given by the release documentation: https://genode.org/documentation/release-notes/24.02 Cheers Norman -- Dr.-Ing. Norman Feske Genode Labs https://www.genode-labs.com · https://genode.org Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth

1 0

Rust Project Ideas
by nsf＠ksu.edu 22 Feb '24

22 Feb '24

Hello Genodians, I am new to Genode and am looking into projects I might be able to do to help get practice with Genode and get a better understanding of rust. On the last rust update, it seemed like the Genode team determined one of the next steps would be to explore rust on top of native Genode framework. Would there be a project that would be able to help achieve this goal? I thought of working with the Gnode api to create a transition layer so that components written in rust can directly work with the api. Would this be a project be helpful or would something else be more appropriate? Thanks in advance for the help, and if I am off base, please let me know! Sincerely, Nathan

4 4

Testing pc_linux image
by Pirmin Duss 16 Feb '24

16 Feb '24

Hello Genodians I'm back to working on the integration of the Intel IGC driver in to the pc_nic_drv. I have rebased my work tree to latest master. I can run the driver and it starts up. The log output is as follows: [init -> nic_drv_igc] NR_IRQS: 4352, nr_irqs: 4352, preallocated irqs: 256 [init -> nic_drv_igc] time-clocksource: dde_counter: mask: 0xffffffffffffff max_cycles: 0x1d854df40, max_idle_ns: 3526361616960 ns [init -> nic_drv_igc] time-clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns [init -> nic_drv_igc] time-clocksource: Switched to clocksource dde_counter [init -> nic_drv_igc] e1000-e1000_main: Intel(R) PRO/1000 Network Driver [init -> nic_drv_igc] e1000-e1000_main: Copyright (c) 1999-2006 Intel Corporation. [init -> nic_drv_igc] e1000e-netdev: Intel(R) PRO/1000 Network Driver [init -> nic_drv_igc] e1000e-netdev: Copyright(c) 1999 - 2015 Intel Corporation. [init -> nic_drv_igc] Intel(R) 2.5G Ethernet Linux Driver [init -> nic_drv_igc] Copyright(c) 2018 Intel Corporation. [init -> nic_drv_igc] igc 04:00.0: PCIe PTM not supported by PCIe bus/controller [init -> nic_drv_igc] igc 04:00.0 eth0: MAC: d8:bb:c1:5b:3a:18 [init -> nic_drv_igc] sched_clock: Marking stable (144026000, 246486000)->(405777000, -15265000) [init -> mac_report_rom] report 'nic_drv_igc -> devices' [init -> mac_report_rom] <devices> [init -> mac_report_rom] <nic name="eth0" mac_address="d8:bb:c1:5b:3a:18"/> [init -> mac_report_rom] </devices> [init -> nic_drv_igc] igc 04:00.0 eth0: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX [init -> nic_drv_igc] create uplink for net device eth0 [init -> nic_router] [uplink_igc] NIC sessions: 1 [init -> nic_router] [uplink_igc] snd ETH d8:bb:c1:5b:3a:18 > ff:ff:ff:ff:ff:ff IPV4 0.0.0.0 > 255.255.255.255 UDP 68 > 67 DHCP DISCOVER d8:bb:c1:5b:3a:18 > 0.0.0.0 [init -> nic_router] [uplink_igc] rcv ETH b4:2e:99:f9:18:72 > d8:bb:c1:5b:3a:18 IPV4 0.0.0.0 > 0.0.0.0 The last two lines get repeated. In Wireshark I see that the DHCP server offers an IP to the MAC of the Intel IGC, but nic_router doesn't see that. To be sure, that nothing is missing from the binary I would like to test the Linux kernel Image that get built by run make pc_linux. From what I understand, this wouldn't even need a root filesystem, as the kernel can do DHCP on its own. Before I start piecing something together, I'd like to ask, if someone does have a ready solution which allows to boot that image via iPXE? Regards, Pirmin

2 2

← Newer
1
...
4
5
6
7
8
9
10
...
229
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users