Re: [PATCH] riscv: Add support for early boot errata application on MIPS chips
From: Paul Walmsley
Date: Thu Aug 20 2026 - 21:53:39 EST
On Mon, 10 Aug 2026, Aleksa Paunovic via B4 Relay wrote:
> From: Aleksa Paunovic <aleksa.paunovic@xxxxxxxxxxxxx>
>
> MIPS errata implementation previously skipped early boot application
> entirely. Although the only currently existing MIPS erratum does not
> require this, amending this now should make any future addition easier to
> implement.
>
> This commit was based on the existing T-Head implementation.
>
> Suggested-by: Jesse Taube <jtaubepe@xxxxxxxxxx>
> Link: https://lore.kernel.org/linux-riscv/CADRr4bcbD57hmR0XGgo8BjgNp4shEADOondZEtwscCJ9-nxXRQ@xxxxxxxxxxxxxx/
> Signed-off-by: Aleksa Paunovic <aleksa.paunovic@xxxxxxxxxxxxx>
> ---
> The patch was motivated by Jesse's suggestion to implement [2] as an early boot erratum.
> However, since it's in conflict with [1] as well, I decided against sending this one and the ZALRSC patch as a series,
> opting for rebasing [1] and [2] once it's clear we should move forward with this approach.
>
> I tested this on an eight-hart QEMU config. The testing was done by rebasing [1] and [2] onto the patch
> and running most of the same tests used to verify those patches. This includes: coremark testing, futex kselftests
> and GDB to check errata application. The same tests were run on the Boston board, configured with a single P8700 core.
>
> Links:
> [1] https://lore.kernel.org/linux-riscv/20260608-p8700-wfi-v2-1-77efc7459f3d@xxxxxxxxxxxxx/
> [2] https://lore.kernel.org/linux-riscv/20260723-p8700-zalrsc-v4-1-59c21b252e29@xxxxxxxxxxxxx/
Thanks, queued for v7.4.
- Paul