Am Donnerstag, 15. September 2022, 19:09:00 CEST schrieb Palmer Dabbelt:
> We could make the T-Head CMOs depend on a new-enough assembler to have
> Zicbom, but it's not strictly necessary because the T-Head CMOs
> circumvent the assembler.
>
> Fixes: 8f7e001e0325 ("RISC-V: Clean up the Zicbom block size probing")
> Cc: stable@xxxxxxxxxxxxxxx
> Reported-by: kernel test robot <lkp@xxxxxxxxx>
> Reported-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
> Signed-off-by: Palmer Dabbelt <palmer@xxxxxxxxxxxx>
Reviewed-by: Heiko Stuebner <heiko@xxxxxxxxx>
> ---
> arch/riscv/include/asm/cacheflush.h | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/arch/riscv/include/asm/cacheflush.h b/arch/riscv/include/asm/cacheflush.h
> index a89c005b4bbf..273ece6b622f 100644
> --- a/arch/riscv/include/asm/cacheflush.h
> +++ b/arch/riscv/include/asm/cacheflush.h
> @@ -42,8 +42,12 @@ void flush_icache_mm(struct mm_struct *mm, bool local);
>
> #endif /* CONFIG_SMP */
>
> -#ifdef CONFIG_RISCV_ISA_ZICBOM
> +/*
> + * The T-Head CMO errata internally probe the CBOM block size, but otherwise
> + * don't depend on Zicbom.
> + */
> extern unsigned int riscv_cbom_block_size;
> +#ifdef CONFIG_RISCV_ISA_ZICBOM
> void riscv_init_cbom_blocksize(void);
> #else
> static inline void riscv_init_cbom_blocksize(void) { }
>
