Debian GNU/Linux port for RISC-V 64-bits: Why it matters and roadmap

What is RISC-V used for and why is Debian interested in building this port?

According to the Debian wiki page, “RISC-V (pronounced "risk-five") is an open source instruction set architecture (ISA) based on established reduced instruction set computing (RISC) principles.

In contrast to most ISAs, RISC-V is freely available for all types of use, permitting anyone to design, manufacture and sell RISC-V chips and software. While not the first open ISA, it is significant because it is designed to be useful in modern computerized devices such as warehouse-scale cloud computers, high-end mobile phones and the smallest embedded systems. Such uses demand that the designers consider both performance and power efficiency. The instruction set also has a substantial body of supporting software, which fixes the usual weakness of new instruction sets.

In this project the goal is to have Debian ready to install and run on systems implementing a variant of the RISC-V ISA:

• Software-wise, this port will target the Linux kernel
• Hardware-wise, the port will target the 64-bit variant, little-endian

This ISA variant is the "default flavour" recommended by the designers, and the one that seems to attract more interest for planned implementations that might become available in the next few years (development boards, possible consumer hardware or servers).”

Update on Debian GNU/Linux riscv64 port

Image source: Debian

Let’s have a look at the graph where the percent of arch-dependent packages that are built for riscv64 (grey line) has been around or higher than 80% since mid-2018.

The arch-dependent packages are almost half of Debian's [main, unstable] archive. It means that the arch-independent packages can be used by all the ports, provided that the software is present on which they rely on.

The update also highlights that around 90% of packages from the whole archive has been made available for this architecture.

Image source: Debian

The graph above highlights that the percentages are very stable for all architectures. Montecelo writes, “This is in part due to the freeze for buster, but it usually happens at other times as well (except in the initial bring-up or in the face of severe problems).” Even the second-class ports appear to be stable.

Montecelo writes, “Together, both graphs are also testament that there are people working on ports at all times, keeping things working behind the scenes, and that's why from a high level view it seems that things just work.”

According to him, apart from the work of porters themselves, there are people working on bootstrapping issues that make it easier to bring up ports, better than in the past. They also make coping better when toolchain support or other issues related to ports, blow up.

He further added, “And, of course, all other contributors of Debian help by keeping good tools and building rules that work across architectures, patching the upstream software for the needs of several architectures at the same time (endianness, width of basic types), many upstream projects are generic enough that they don't need specific porting, etc.”

Future scope and improvements yet to come

To get Debian running on RISC-V will not be easy because of various reasons including limited availability of hardware being able to run Debian port and limited options for using bootloaders. According to Montecelo, this is an area of improvement from them.

He further added, “Additionally, it would be nice to have images publicly available and ready to use, for both Qemu and hardware available like the HiFive Unleashed (or others that might show up in time), but although there's been some progress on that, it's still not ready and available for end users.”

Presently, they are beyond 500 packages from the Rust ecosystem in the archive (which is about 4%) which can’t be built and used until Rust gets support for the architecture. Rust requires LLVM and there’s no Rust compiler based on GCC or other toolchains.

Montecelo writes, “Firefox is the main high-level package that depends on Rust, but many packages also depend on librsvg2 to render SVG images, and this library has been converted to Rust. We're still using the C version for that, but it cannot be sustained in the long term."

Apart from Rust, other packages use LLVM to some extent, but currently, it is not fully working for riscv64. The support of LLVM for riscv64 is expected to be completed this year.

While talking about other programming languages, he writes, “There are other programming language ecosystems that need attention, but they represent a really low percentage (only dozens of packages, of more than 12 thousand; and with no dependencies outside that set). And then, of course, there is a long tail of packages that cannot be built due to a missing dependency, lack of support for the architecture or random failures -- together they make a substantial number of the total, but they need to be looked at and solved almost on a case-by-case basis.”

Why are people excited about this?

Many users seem to be excited about the news, one of the reasons being that there won’t be a need to bootstrap from scratch as Rust now will be able to cross-compile easily because of the Riscv64 support. A user commented on HackerNews, “Debian Rust maintainer here. We don't need to bootstrap from scratch, Rust (via LLVM) can cross-compile very easily once riscv64 support is added.” Also, this appears to be a good news for Debian, as cross-compiling has really come a long way on Debian.

Rest are awaiting for more to get incorporated with riscv. Another user commented, “I am waiting until the Bitmanip extension lands to get excited about RISC-V: https://github.com/riscv/riscv-bitmanip”

Few others think that there is a need for LLVM support for riscv64. A user commented, “The lack of LLVM backend surprises me. How much work is it to add a backend with 60 instructions (and few addressing modes)? It's clearly far more than I would have guessed.”

Another comment reads, “Basically LLVM is now a dependency of equal importance to GCC for Debian. Hopefully this will help motivate expanding architecture-support for LLVM, and by proxy Rust.”

According to users, the architecture of this port misses on two major points, one being the support for LLVM compiler and the other one being the support for Rust based on GCC. If the port gets the LLVM support by this year, users will be able to develop a front end for any programming language as well as a backend for any instruction set architecture. Now, if we consider the case of support for Rust based on GCC, then the port will help developers to get support for many language extensions as GCC provides the same.

A user commented on Reddit, “The main blocker to finish the port is having a working Rust toolchain. This is blocked on LLVM support, which only supports RISCV32 right now, and RISCV64 LLVM support is expected to be finished during 2019.”

Another comment reads, “It appears that enough people in academia are working on RISCV for LLVM to accept it as a mainstream backend, but I wish more stakeholders in LLVM would make them reconsider their policy.”

To know more about this news, check out Debian’s official post.

Debian maintainer points out difficulties in Deep Learning Framework Packaging

Debian project leader elections goes without nominations. What now?

Are Debian and Docker slowly losing popularity?