Repository navigation
Conversation
The 1.11.1 release tarball lacks cipher/simd-common-riscv.h, so the RISC-V vector code fails to build. Fixed upstream in 1.11.2. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The VisionFive2 was wired straight into riscv64_defconfig, and its paths still pointed at board/riscv64/visionfive2/ after the directory was renamed, so the defconfig no longer built. Make it an ix-board package like the aarch64 boards, with the JH7110 kernel options in its .mk, so riscv64 can take more boards and a single-board build still gets a working kernel. The factory password hash moves from a kernel patch to the board's own infix.dtsi. The camera, audio and display drivers are not carried over. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The VisionFive2's U-Boot and OpenSBI were built as part of riscv64_defconfig, which ties the generic riscv64 image to one board. Give it its own boot defconfig, like the aarch64 boards, and let mkimage.sh put the SD card image together. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
riscv64 had fallen behind the other architectures. It was still on Linux 6.8.2, which none of the kernel patches in patches/linux/ apply to, it lacked the test mode the regression tests rely on, and it still built an ext2 root filesystem that nothing uses. Rebase it on aarch64_defconfig so the two only differ in architecture and boards. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The riscv64 kernel config was a StarFive desktop config, with sound, camera, DRM and NFS root, and it never picked up the netfilter, ipset and LSM settings the other architectures have. Derive it from the aarch64 config instead, without the SoC drivers, so it carries the common Infix feature set and leaves hardware to the board packages. On riscv nothing selects SMP, GPIOLIB, PINCTRL, PM and friends the way the arm64 platform options do, so they are set explicitly. Also enable earlycon over SBI for board bring-up. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The 6.18 device tree for the K1 has no SD card controller, nor the I2C bus and P1 PMIC that supply the card, and the 6.18 SD driver never turns on the card clock, so Linux cannot use the SD card the BPi-F3 boots from. Take the nodes and driver fixes from v7.2, the nodes trimmed of the USB3, SPI and PCIe parts they were diffed against. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The K1 BootROM loads a signed first stage loader that brings up DDR before anything else runs. Mainline U-Boot can build that SPL now, but only with SpacemiT's signing tool and DDR firmware blob, so use the vendor SPL for the time being and run mainline U-Boot after it. The FSBL also loads an OpenSBI FIT from its own partition, so wrap the OpenSBI build in the format it expects. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
SpacemiT K1 based board with two GbE ports, eMMC and an SD card slot. This first spin runs from SD card. The SD pads stay at 3.3V because 6.18 cannot switch their IO power domain for UHS. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The vendor FSBL brings up DDR and loads upstream OpenSBI and mainline U-Boot from the SD card. Mainline gained K1 SD and eMMC support after v2026.10, so U-Boot comes from a kkit branch with those changes picked on top of the release. U-Boot has no Ethernet for the K1 yet. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The empty default never made it into the environment, since setting a variable to the empty string deletes it. So boards without a variant script printed '## Error: "ixvariant" not defined' on every boot with a uboot.env on aux. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Neither OpenSBI nor Linux 6.18 can reset the K1, so reboot hangs the board. The P1 PMIC driver for it landed in v6.19, its MFD cell in v7.2. Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
mattiaswal
approved these changes
Oct 12, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Banana Pi BPI-F3 (SpacemiT K1) support, plus a cleanup of the riscv64 BSP so it follows the aarch64 layout: one board package per board.
What works
Limitations, resolved by the upcoming 7.4 kernel upgrade
The 6.18 kernel lacks most K1 support, so this carries backports for the SD card and PMIC reboot. Mainline has these, and more:
With 7.4 the kernel backports go away, and the PDMA and 3.3V workarounds get rechecked.
Other limitations
Checklist
Tick relevant boxes, this PR is-a or has-a: