- l4-hackers - example.com Site

Loading L4 after Uboot on ZynqMP
by mkravchik＠hotmail.com 10 Apr '25

10 Apr '25

Hello, I'm trying to run L4 on a Xilinx ZyncMP board. The board runs Uboot from ROM and then I tried to load L4 as described in https://l4re.org/bsp/zynqmp.html. However, loading a uimage fails, I'm getting an error: ZynqMP> fatload usb 0:1 0x0ffffc0 bootstrap_hello.uimage ZynqMP> bootm 0x0ffffc0 ## Booting kernel from Legacy Image at 00ffffc0 ... Image Name: L4 Image #8 Created: 2025-03-31 14:14:32 UTC Image Type: AArch64 Linux Kernel Image (uncompressed) Data Size: 1530680 Bytes = 1.5 MiB Load Address: 01000000 Entry Point: 01000000 Verifying Checksum ... OK XIP Kernel Image FDT and ATAGS support not compiled in ### ERROR ### Please RESET the board ### I tried to use rawimage, but there I get to the following problem: ZynqMP> fatload usb 0:1 0x1000000 bootstrap_hello-cfg.rawimage 2149176 bytes read in 151 ms (13.6 MiB/s) ZynqMP> go 0x1000000 ## S� L4 Bootstrapper Build: #15 Thu Apr 3 11:39:37 IDT 2025, 9.4.0 Scanning up to 2048 MB RAM, starting at offset 32MB Memory size is 2048MB (00000000 - 7fffffff) RAM: 0000000000000000 - 000000007fffffff: 2097152kB Total RAM: 2048MB Scanning fiasco -serial_esc Scanning sigma0 Scanning moe rom/hello.cfg Compressed modules: cannot decompress at 0x10a8000, try 0x11bc000 Uncompressing modules (modaddr = 0x11bc000 (backwards)): mod06: 0x126c000-0x1296b38: hello mod05: 0x126b000-0x126b080: hello.cfg mod04: 0x11e1000-0x126a118: ned mod03: 0x11bc000-0x11e0c40: l4re Loading fiasco found node 0 kernel info page (via ELF) at 0x3000 L4Re is configured for EL2 mode (virtualization / hypervisor) but got started in EL1 by the boot-loader. Please change your boot-loader to start L4Re in EL2 mode. I'm not sure how to proceed from here. The Uboot is loaded from the board's ROM and is running at EL2. Any help is appreciated, Moshe _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

3 5

L4re RAM/Flash usage and scheduling timing Info
by Eric Miniere via l4-hackers 28 Mar '25

28 Mar '25

Hello Everyone, I am looking for an estimate of the memory footprint to run the hypervisor L4RE on a quad Core ARM Cortex R52. Does it run directly from flash or it needs to be copied to run in RAM ? same question for Guests OS (ie FreeRtos)? What could be the minimal needed L4RE footprint for code and data ? Also if someone could share a rough idea of scheduling time for an ARM R52 running at 400Mhz ? or nb of cycle it takes to schedule ? Thanks _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

2 3

Trouble booting L4Re on PC hardware?
by Richard Clark 14 Mar '25

14 Mar '25

Hi! I've had much success building and running my applications in QEMU. No issues there. But now I've got myself some real hardware. Windows boots on it. Debian Linux boots on it. Linux even boots from the USB I create with rufus. No problems there. What I would like to do, is to take the system.iso image that normally runs under QEMU (x86_64), burn that to my USB stick (same way I put linux boot on there), and then boot the USB stick. The hardware will boot from the USB, and the image copy to the USB appears to work, since I've done that and booted linux. But I cannot seem to get the system.iso file to boot. I've tried UEFI boot, Classic boot, copy as UEFI image, and copy as DD image. It seems to ignore the USB drive and (by default) boot windows from the NVME drive. The device has no serial port. Just an HDMI screen with USB mouse and keyboard. It does have ethernet ports and I could try that, but it doesn't appear that the UEFI bios supports TFTP. How do I take my QEMU work and boot it on a real PC? Thanks! Richard _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

3 2

How to L4Re on the S32G
by qiang xu 27 Feb '25

27 Feb '25

*Hello L4Re experts,* I would like to know how to run L4Re on the S32G. I couldn’t find any related documentation. I’m a beginner in this technical field, so could you provide friendly step-by-step instructions? Regards, Qiang _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

4 25

saving configuration to version control
by Baruch Burstein 09 Feb '25

09 Feb '25

Hello. After running `make config`, I am trying to save my configuration to git and later restore it. This is for both fiasco and l4. I found that the command `make savedefconfig` works (just like for linux kernel) to create a defconfig file for uploading to git, but the command `make defconfig` for restoring the configuration from this file doesn't work. What file should I be saving/restoring, or what is the process to save/restore my configuration? Either a defconfig or a full configuration would be ok. Thank you _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

2 2

Use HW random generator to kaslr-seed dtb on uvmm
by Flavio Erlich 04 Feb '25

04 Feb '25

Hi, According to Linux documentation, if I want to use kASLR I should add on my device tree, on the chosen node a kaslr-seed entry ( https://www.kernel.org/doc/Documentation/devicetree/bindings/chosen.txt) I'm running uvmm and I want to use the dt-modify argument to change the value on dtb to a value that comes from an HW random generator at runtime. I have written an application which reads a random sequence from the HW. How can I pass the random value from that application and pass the u64 value to the ned script which I can use as a seed? I can use the application exit code to read 32-bit and run it twice, but seems to be overkill to do that. I was thinking of sending a newer allocatted Dataspace to the application, it will fill it with random data, but I have no idea if it is possible to read this Dataspace and concatenate the result to the arguments to uvmm. Do you have any other suggestions? Thanks for your help, Flavio _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

2 2

USB on L4RE
by Flavio Erlich 25 Jan '25

25 Jan '25

I found this thesis defence post. https://tu-dresden.de/ing/informatik/sya/professur-fuer-betriebssysteme/die… Does anyone know if or where the code is available? Is the thesis paper available? Thanks, Flavio _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

3 2

emmc-driver imx8mp
by Alvaro Gallardo 08 Jan '25

08 Jan '25

Dear L4Re experts, I hope you are doing well and enjoying this time of the year. I would like to ask for your help or insights on how to use the emmc-driver in an imx8mp based board. In this platform, the usdhc nodes are located out of the pci bus, and they require an i2c device to work, as well as gpio and clocks. I am able to start the emmc server, which identifies one usdhc device and starts the server loop. I pass a virtual bus to the emmc driver with all required devices, as they usually work in the Virtual Machine to access the mmc. The driver scans all of them, finding the usdhc node. But, everything stops when I create a client, either using a dynamic way with: create(0, "device=...","ds-max=5); or declaring the client in the argument as: "emmc-drv -v --client cl1 --device=... --ds-max=5". For the second way, it stops when the Virtual Machine starts to configure the related virtio device (I am using a virtio_net node). I will be grateful if you can give me some information about the issue or some of the following questions: 1. how should the node in the device tree of the Virtual Machine be configured to work with the emmc client? 2. is it possible that the board's specific hardware configuration required for the mmc is stopping the emmc client to work? 3. Is there any way to test the driver before linking it to a virtual machine? I found an example with the "virtio-block-chksum-sync" module, but I could not find this module. I attach the configuration files, the output text and the base device tree structure. Thank you in advance for your attention. Kind regards, Alvaro Gallardo _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

2 2

Network failure when high traffic
by Mohamed Dawod 05 Dec '24

05 Dec '24

Hello, I created a new L4 setup with uvmm module and Linux kernel running on it. I added also virtio-net-device and created new tap devices to be able to communicate with the VM from outside of L4. The setup was done successfully and I could ping from my host machine to the Linux VM and vise versa. The problem appeared when I accelerated the ping using -i option : Ping 192.168.7.2 -i 0.01 It continue pings for a few minutes then it gives (Destination Host Unreachable)! I used the wrapper function vmm.start_virtio_switch() in me demo. Why did this issue happen ? And how can I fix it? Thanks, Mohamed _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

2 1

Video Error Reasons?
by Lonnie Cumberland via l4-hackers 27 Nov '24

27 Nov '24

Hello All, I have been experimenting with L4Re and was able to compile the latest snapshot that seems to work well for almost all of the examples in QEMU from the "make qemu" menu which is a very good start, I think. Lately, I have also used the "make grub2iso" which gives me the same menu of examples as the "make qemu" but for which it will generate an ISO of the example. I generated ISO's for all of the examples and tested them on 3 different x64 computer systems which are all Intel with 2 being NUC's and 1 being a Dell XPS 8950 but when I boot any of the generated L4Re ISO's, it seems to boot all of the core modules but then gets to an error message that seems to be related to the video mode. ------------------------------------------- error: no suitable video mode found ------------------------------------------- And then the screen shows a bunch of lines. I have been looking around though the L4Re code in the snapshot release, but I do not yet have a clue where how to change things so that these ISO's will boot completely on some hardware although things seem to boot up in QEMU. I am thinking that there might just be some setting somewhere that I need to adjust, or something. Any thoughts, suggestions, or ideas would be greatly appreciated. Have a great evening, Lonnie _______________________________________________ l4-hackers mailing list -- l4-hackers(a)os.inf.tu-dresden.de To unsubscribe send an email to l4-hackers-leave(a)os.inf.tu-dresden.de

4 4