2025-09-18 Meeting Notes

2025-09-18 Meeting Notes

Disclosures

Note: The Devboards Meeting is open to members and non-members alike.

 

RVI_Disclosures_Female_Short.mp4

Participants

Name

Affiliation

Name

Affiliation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Presentations

Title

Presenter

File

Title

Presenter

File

1

 

 

 

2

 

 

 

Notes & Action Items

Board Updates:

  • Beagle, SiFive, StarFive, RIOS, Allwinner, Microchip, SpacemIT, Milk-V

  • Ultrarisc pcb layout has been finished, the development board will come back at the end of this month and be tested. Milk-V Titan will ship this board, expected October 15th. They plan to have at least one for NA Summit.

 

Program Updates:

  • oreboot

  • U-Boot

  • Firmware

    • (Wei Fu) Trying to get EDK2 running on Milk-V Jupiter.

  • Operating System

  • Images

    • Development - Gentoo, FedoraVForce (Mirror US)

      • FedoraVForce

        • All images updated to Fedora 42, and recently made live cd for RISC-V machines.

        • Making images for new RISC-V systems

          • P550 and Megrez

        • Making CentOS images as well

          • Making EPEL packages

        • Trying to incorporate with Red Hat to see what the toolchain can be for RVA23

        • Currently building Fedora 43

      • Gentoo

        • Biggest holdup is the compilers.

        • SiFive upstreamed everything except the P550, and work is held off until gcc16 (hasn’t happened for gcc15).

        • Trying to put something together for the SpacemiT K1, but having issues with the GPU stack.

        • Some discussions to get the Imagination GPU support into the kernel.

          • Downside is all based on thead1520

  • Compilers

    • "are we expecting that RVA20 to RVA23 will be a hard ABI break with a different dpkg-architecture and whatever gentoo does, or will you be able to mix and match packages?

      "the most recent CFI patches posted to libc-alpha change the size of jmp_buf depending on compile time __riscv_zicfiss (despite the type having 10+ unused words), which will break every Perl XS module across an incomplete transition (Perl_Interpreter contains a jmp_buf) and probably other things, but I'm not sure if this is actually a problem or if I'm the right person to bring it up if it is"

      "looks like glibc is going with a hard ABI break for CFI by adding a new word to jmp_buf instead of using one of the existing scratch words"

      "both x86_64/SHSTK and arm64/GCS did this without an ABI break, but here we are"

  • CI/CD

  • RISC-V International

What’s cool?:

  •  

Miscellaneous notes:

  •  


Standard_2.png

RISC-V International