Description

There are already numerous ways to run Linux and some programs through emulation in a web browser (e.g. x86 and riscv64 on https://bellard.org/jslinux/), but none use WebGL/WebGPU to run the emulation on the GPU.

I already made a PoC of an AArch64 (64-bit Arm) emulator in OpenCL which is unfortunately hindered by a multitude of OpenCL compiler bugs on all platforms (Intel with beignet or the new compute runtime and AMD with Mesa Clover and rusticl). With more widespread and thus less broken GLSL vs. OpenCL and the less complex implementation requirements for RV32 (especially 32bit integers instead of 64bit), that should not be a major problem anymore.

Goals

Write an RISC-V system emulator in GLSL that is capable of booting Linux and run some userspace programs interactively. Ideally it is small enough to work on online test platforms like Shaderoo with a custom texture that contains bootstrap code, kernel and initrd.

Minimum:

riscv32 without FPU (RV32 IMA) and MMU (µClinux), running Linux in M-mode and userspace in U-mode.

Stretch goals:

FPU support, S-Mode support with MMU, SMP. Custom web frontend with more possibilities for I/O (disk image, network?).

Resources

RISC-V ISA Specifications
Shaderoo
OpenGL 4.5 Quick Reference Card

Result as of Hackweek 2024

WebGL turned out to be insufficient, it only supports OpenGL ES 3.0 but imageLoad/imageStore needs ES 3.1. So we switched directions and had to write a native C++ host for the shaders.

As of Hackweek Friday, the kernel attempts to boot and outputs messages, but panics due to missing memory regions.

Since then, some bugs were fixed and enough hardware emulation implemented, so that now Linux boots with framebuffer support and it's possible to log in and run programs!

The repo with a demo video is available at https://github.com/Vogtinator/risky-v

Looking for hackers with the skills:

risc-v opengl kernel c++

This project is part of:

Hack Week 24

Activity

  • 2 months ago: jlp liked this project.
  • 2 months ago: nkrapp liked this project.
  • 2 months ago: favogt added keyword "risc-v" to this project.
  • 2 months ago: favogt added keyword "opengl" to this project.
  • 2 months ago: favogt added keyword "kernel" to this project.
  • 2 months ago: favogt added keyword "c++" to this project.
  • 2 months ago: llansky3 liked this project.
  • 2 months ago: michals liked this project.
  • 3 months ago: jfkw liked this project.
  • 3 months ago: mwilck liked this project.
  • 3 months ago: horon liked this project.
  • 3 months ago: wfrisch liked this project.
  • 3 months ago: mbrugger liked this project.
  • 3 months ago: vliaskovitis liked this project.
  • 3 months ago: mfriedrich liked this project.
  • 3 months ago: mfriedrich joined this project.
  • 3 months ago: favogt started this project.
  • 3 months ago: favogt originated this project.

  • Comments

    • wfrisch
      3 months ago by wfrisch | Reply

      Impressive! Godspeed with that project.

    • michals
      2 months ago by michals | Reply

      Existing art: https://blog.pimaker.at/texts/rvc1/

    • favogt
      2 months ago by favogt | Reply

      Short update: It's able to boot into buildroot graphically and with keyboard input some applications can be started. Some like vi and nano crash though, which still needs some investigation. Once that's fixed I'll link to a git repo with a writeup!

    Similar Projects

    Create openSUSE images for Arm/RISC-V boards by avicenzi

    Project Description

    Create openSUSE images (or test generic EFI images) for Arm and/or RISC-V boards that are not yet supported.

    Goal for this Hackweek

    Create bootable images of Tumbleweed for SBCs that currently have no images available or are untested.

    Consider generic EFI images where possible, as some boards can hold a bootloader.

    Document in the openSUSE Wiki how to flash and use the image for a given board.

    Boards that I have around and there are no images:

    • Rock 3B
    • Nano PC T3 Plus
    • Lichee RV D1
    • StartFive VisionFive (has some image needs testing)

    Hack Week 22

    Hack Week 21

    Resources


    Investigate Milk-V Jupiter board (riscv64) by a_faerber

    Description

    On Monday of Hackweek 24 we received two Milk-V Jupiter mini-ITX boards (JUPITER_V1.1) via the RISC-V International devboards program. The Jupiter board uses a Spacemit M1 System-on-Chip (SoC).

    The Banana Pi BPI-F3 board with related Spacemit K1 SoC had previously already been investigated by Marvin Friedrich.

    Goals

    Investigate the hardware, any shipping boot firmware and its ability to boot openSUSE Tumbleweed.

    Resources

    Progress

    • On Monday Andreas purchased an ATX power supply, to power up the board.
    • According to UART output (note there's two UART pinouts on the board), there appears to be an undocumented U-Boot bootloader in QSPI, in addition to the documented microSD, NVMe and eMMC boot methods.
    • The on-board bootloader does not seem to support USB ("No working controllers found").
    • Manually loading our GRUB from microSD (via load+bootefi commands from SPI-based U-Boot) works, but in GRUB loading our kernel fails ("error: ../../grub-core/loader/efi/linux.c:521:out of memory.").
    • Marvin instead had success flashing a ready-made Bianbu OS image via titantools to microSD card (i.e., a different U-Boot on microSD), that he then modified to load openSUSE kernels instead.


    FizzBuzz OS by mssola

    Project Description

    FizzBuzz OS (or just fbos) is an idea I've had in order to better grasp the fundamentals of the low level of a RISC-V machine. In practice, I'd like to build a small Operating System kernel that is able to launch three processes: one that simply prints "Fizz", another that prints "Buzz", and the third which prints "FizzBuzz". These processes are unaware of each other and it's up to the kernel to schedule them by using the timer interrupts as given on openSBI (fizz on % 3 seconds, buzz on % 5 seconds, and fizzbuzz on % 15 seconds).

    This kernel provides just one system call, write, which allows any program to pass the string to be written into stdout.

    This project is free software and you can find it here.

    Goal for this Hackweek

    • Better understand the RISC-V SBI interface.
    • Better understand RISC-V in privileged mode.
    • Have fun.

    Resources

    Results

    The project was a resounding success add-emoji Lots of learning, and the initial target was met.


    Explore Microchip PIC64GX1000 Curiosity board (riscv64) by a_faerber

    Description

    The Microchip PIC64 family of RISC-V chipsets was announced this summer, with PIC64GX as first subfamily (with SiFive U54 CPU cores, same as Microchip PolarFire). Later families (PIC64-HPSC and PIC64HX) were announced to feature the long-awaited RISC-V Hypervisor Extension.

    My pre-ordered PIC64GX1000 Curiosity board arrived the week before Hackweek 24.

    Goals

    I intend to investigate enabling openSUSE Tumbleweed riscv64 on this new chipset and board.

    Resources

    Progress

    • Created udev rules for Curiosity board's FTDI based debug UARTs on USB-C - UART B has output from HSS / OpenSBI, UART C has output from U-Boot and kernel
    • Modified the image https://download.opensuse.org/ports/riscv/tumbleweed/images/openSUSE-Tumbleweed-RISC-V-JeOS-efi.riscv64-2024.11.10-Build1.13.raw.xz (20241118) with gdisk to prepend a bootloader partition (copying the binary bootloader from the vendor-supplied microSD, ignoring excess partition size) and added the microchip/pic64gx-curiosity-kit.dtb file to the EFI System Partition (again from the vendor-supplied microSD)
    • This boots into Tumbleweed! with ttyS1 by default (UART C)

    Stretch goals

    Investigate also the Microchip PolarFire SoC Discovery board.

    Progress

    The Microchip PolarFire Icicle board and BeagleV-Fire board had previously already been enabled by Torsten Duwe.

    • Additional udev rules needed to be created for its FTDI UART bridge chip.
    • The board ships with a "FIR filter" demo pre-flashed, not exposing any bootloader, only some demo output on UART C. The board will need to be flashed with a Microchip reference design for Linux.
    • The board has an on-board Flash Pro Express JTAG debugger on USB-C. Proprietary Programming and Debug tools software is available from Microchip to download, requiring an account. This should then be able to flash the .job file.
    • An in-progress patchset for OpenOCD is available in Gerrit, not yet in the openSUSE package. Presumably OpenOCD would not support Microchip's .job file format.


    Tumbleweed on Mars-CM (RISC-V board) by ph03nix

    RISC-V is awesome, Tumbleweed is awesome, chocolate cake is awesome. I'm planning to combine all of them in one project.

    Project Description

    I recently purchased a MILK-V Mars CM and managed to setup it up already using Debian Linux. My project for this Hackweek is to see how far I can get to run Tumbleweed on this compute module board.

    Goal for this Hackweek

    • Run Tumbleweed on the Compute Module

    Resources

    • http://milkv.io/mars-cm
    • https://en.opensuse.org/HCL:VisionFive2


    Linux on Cavium CN23XX cards by tsbogend

    Before Cavium switched to ARM64 CPUs they developed quite powerful MIPS based SOCs. The current upstream Linux kernel already supports some Octeon SOCs, but not the latest versions. Goal of this Hack Week project is to use the latest Cavium SDK to update the Linux kernel code to let it running on CN23XX network cards.


    Kill DMA and DMA32 memory zones by ptesarik

    Description

    Provide a better allocator for DMA-capable buffers, making the DMA and DMA32 zones obsolete.

    Goals

    Make a PoC kernel which can boot a x86 VM and a Raspberry Pi (because early RPi4 boards have some of the weirdest DMA constraints).

    Resources

    • LPC2024 talk:
    • video:


    Model checking the BPF verifier by shunghsiyu

    Project Description

    BPF verifier plays a crucial role in securing the system (though less so now that unprivileged BPF is disabled by default in both upstream and SLES), and bugs in the verifier has lead to privilege escalation vulnerabilities in the past (e.g. CVE-2021-3490).

    One way to check whether the verifer has bugs to use model checking (a formal verification technique), in other words, build a abstract model of how the verifier operates, and then see if certain condition can occur (e.g. incorrect calculation during value tracking of registers) by giving both the model and condition to a solver.

    For the solver I will be using the Z3 SMT solver to do the checking since it provide a Python binding that's relatively easy to use.

    Goal for this Hackweek

    Learn how to use the Z3 Python binding (i.e. Z3Py) to build a model of (part of) the BPF verifier, probably the part that's related to value tracking using tristate numbers (aka tnum), and then check that the algorithm work as intended.

    Resources


    Modularization and Modernization of cifs.ko for Enhanced SMB Protocol Support by hcarvalho

    Creator:
    Enzo Matsumiya ematsumiya@suse.de @ SUSE Samba team
    Members:
    Henrique Carvalho henrique.carvalho@suse.com @ SUSE Samba team

    Description

    Split cifs.ko in 2 separate modules; one for SMB 1.0 and 2.0.x, and another for SMB 2.1, 3.0, and 3.1.1.

    Goals

    Primary

    Start phasing out/deprecation of older SMB versions

    Secondary

    • Clean up of the code (with focus on the newer versions)
    • Update cifs-utils
    • Update documentation
    • Improve backport workflow (see below)

    Technical details

    Ideas for the implementation.

    • fs/smb/client/{old,new}.c to generate the respective modules
      • Maybe don't create separate folders? (re-evaluate as things progresses!)
    • Remove server->{ops,vals} if possible
    • Clean up fs_context.* -- merge duplicate options into one, handle them in userspace utils
    • Reduce code in smb2pdu.c -- tons of functions with very similar init/setup -> send/recv -> handle/free flow
    • Restructure multichannel
      • Treat initial connection as "channel 0" regardless of multichannel enabled/negotiated status, proceed with extra channels accordingly
      • Extra channel just point to "channel 0" as the primary server, no need to allocate an extra TCPServerInfo for each one
    • Authentication mechanisms
      • Modernize algorithms (references: himmelblau, IAKERB/Local KDC, SCRAM, oauth2 (Azure), etc.


    Improve UML page fault handler by ptesarik

    Description

    Improve UML handling of segmentation faults in kernel mode. Although such page faults are generally caused by a kernel bug, it is annoying if they cause an infinite loop, or panic the kernel. More importantly, a robust implementation allows to write KUnit tests for various guard pages, preventing potential kernel self-protection regressions.

    Goals

    Convert the UML page fault handler to use oops_* helpers, go through a few review rounds and finally get my patch series merged in 6.14.

    Resources

    Wrong initial attempt: https://lore.kernel.org/lkml/20231215121431.680-1-petrtesarik@huaweicloud.com/T/


    Port some classic game to Linux by MDoucha

    Let's pick some old classic game, reverse engineer the data formats and game rules and write an open source engine for it from scratch. Some games from 1990s are simple enough that we could have a playable prototype by the end of the week.

    Write which games you'd like to hack on in the comments. Don't forget to check e.g. on Open Source Game Clones, Github and SourceForge whether the game is ported already.

    Hack Week 24 - Master of Orion II: Battle at Antares & Chaos Overlords

    Work on Master of Orion II continues but we can hack more than one game. Chaos Overlords is a dystopian, lighthearted, cyberpunk turn-based strategy game originally released in 1996 for Windows 95 and Mac OS. The player takes on the role of a Chaos Overlord, attempting to control a city. Gameplay involves hiring mercenary gangs and deploying them on an 8-by-8 grid of city sectors to generate income, occupy sectors and take over the city.

    How to ~~install & play~~ observe the decompilation progress:

    • Clone the Git repository
    • A playable reimplementation does not exist yet, but when it does, it will be linked in the repository mentioned above.

    Further work needed:

    • Analyze the remaining unknown data structures, most of which are related to the AI.
    • Decompile the AI completely. The strong AI is part of the appeal of the game. It cannot be left out.
    • Reimplement the game.

    Hack Week 20, 21, 22 & 23 - Master of Orion II: Battle at Antares

    Master of Orion II is one of the greatest turn-based 4X games of the 1990s. Explore the galaxy, colonize planets, research new technologies, fight space monsters and alien empires and in the end, become the ruler of the galaxy one way or another.

    How to install & play:

    • Clone the Git repository
    • Run ./bootstrap; ./configure; make && make install
    • Copy all *.LBX files from the original Master of Orion II to the installation data directory (/usr/local/share/openorion2 by default)
    • Run openorion2

    Further work needed:

    • Analyze the rest of the original savegame format and a few remaining data files.
    • Implement most of the game. The open source engine currently supports only loading saved games from the original version and viewing the galaxy map, fleet management and list of known planets.

    Hack Week 19 - Signus: The Artifact Wars

    Signus is a Czech turn-based strategy game similar to Panzer General or Battle Isle series. Originally published in 1998 and open-sourced by the original developers in 2003.

    How to install & play:

    • Clone the Git repository
    • Run ./bootstrap; ./configure; make && make install in both signus and signus-data directories.
    • Run signus

    Further work needed:

    • Create openSUSE package
    • Implement full support for original game data (the open source version uses slightly different data file contents but original game data can be converted using a script).


    YQPkg - Bringing the Single Package Selection Back to Life by shundhammer

    tl;dr

    Rip out the high-level YQPackageSelector widget from YaST and make it a standalone Qt program without any YaST dependencies.

    See section "Result" at the bottom for the current status after the hack week.

    Current Status

    See the development status issue at the GitHub repo.

    tl;dr: It's usable now with all the key features.

    It does real package installation / removal / update with reasonable user feedback.

    The Past and the Present

    We used to have and still have a powerful software selection with the YaST sw_single module (and the YaST patterns counterpart): You can select software down to the package level, you can easily select one of many available package versions, you can select entire patterns - or just view them and pick individual packages from patterns.

    You can search packages based on name, description, "requires" or "provides" level, and many more things.

    The Future

    YaST is on its way out, to be replaced by the new Agama installer and Cockpit for system administration. Those tools can do many things, but fine-grained package selection is not among them. And there are also no other Open Source tools available for that purpose that even come close to the YaST package selection.

    Many aspects of YaST have become obsolete over the years; many subsystems now come with a good default configuration, or they can configure themselves automatically. Just think about sound or X11 configuration; when did you last need to touch them?

    For others, the desktops bring their own tools (e.g. printers), or there are FOSS configuration tools (NetworkManager, BlueMan). Most YaST modules are no longer needed, and for many others there is a replacement in tools like Cockpit.

    But no longer having a powerful fine-grained package selection like in YaST sw_single will hurt. Big time. At least until there is an adequate replacement, many users will want to keep it.

    The Idea

    YaST sw_single always revolved around a powerful high-level widget on the abstract UI level. Libyui has low-level widgets like YPushButton, YCheckBox, YInputField, more advanced ones like YTable, YTree; and some few very high-level ones like YPackageSelector and YPatternSelector that do the whole package selection thing alone, working just on the libzypp level and changing the status of packages or patterns there.

    For the YaST Qt UI, the YQPackageSelector / YQPatternSelector widgets work purely on the Qt and libzypp level; no other YaST infrastructure involved, in particular no Ruby (or formerly YCP) interpreter, no libyui-level widgets, no bindings between Qt / C++ and Ruby / YaST-core, nothing. So it's not too hard to rip all that part out of YaST and create a standalone program from it.

    For the NCurses UI, the NCPackageSelector / NCPatternSelector create a lot of libyui widgets (inheriting YWidget / NCWidget) and use a lot of libyui calls to glue them together; and all that of course still needs a lot of YaST / libyui / libyui-ncurses infrastructure. So NCurses is out of scope here.

    Preparatory Work: Initializing the Package Subsystem

    To see if this is feasible at all, the existing UI examples needed some fixing to check what is needed on that level. That was the make-or-break decision: Would it be realistically possible to set the needed environment in libzypp up (without being stranded in the middle of that task alone at the end of the hack week)?

    Yes, it is: That part is already working:

    https://github.com/yast/yast-ycp-ui-bindings/pull/71


    New KDE Plasma notification app/applet by apappas

    Description

    My memory is terrible so I depend a lot on notifications to carry me through the workday. As a plasma user I am ok with the current applet, but I don't love it. It is too small for the centrality it has in my day. Also I dislike how you can not go back to notifications you have dismissed

    Goals

    Develop a plasma app that * must gather notifications without disrupting the existing notification app * must offer the ablity to refer to dismissed/archived/seen notification up to some defined point in the past * must allow deletion of notifications