On Mon, Oct 26, 2020 at 08:26:54PM +0800, liush wrote:
> From: Liu Shaohua <li...@allwinnertech.com>
> 
> The argument to pfn_to_virt() should be pfn not the value of CSR_SATP.
> 
> Reviewed-by: Palmer Dabbelt <palmerdabb...@google.com>
> Reviewed-by: Anup Patel <a...@brainfault.org>
> Signed-off-by: liush <li...@allwinnertech.com>

Reviewed-by: Pekka Enberg <penb...@kernel.org>

> ---
>  arch/riscv/mm/fault.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/riscv/mm/fault.c b/arch/riscv/mm/fault.c
> index 1359e21..3c8b9e4 100644
> --- a/arch/riscv/mm/fault.c
> +++ b/arch/riscv/mm/fault.c
> @@ -86,6 +86,7 @@ static inline void vmalloc_fault(struct pt_regs *regs, int 
> code, unsigned long a
>       pmd_t *pmd, *pmd_k;
>       pte_t *pte_k;
>       int index;
> +     unsigned long pfn;
>  
>       /* User mode accesses just cause a SIGSEGV */
>       if (user_mode(regs))
> @@ -100,7 +101,8 @@ static inline void vmalloc_fault(struct pt_regs *regs, 
> int code, unsigned long a
>        * of a task switch.
>        */
>       index = pgd_index(addr);
> -     pgd = (pgd_t *)pfn_to_virt(csr_read(CSR_SATP)) + index;
> +     pfn = csr_read(CSR_SATP) & SATP_PPN;
> +     pgd = (pgd_t *)pfn_to_virt(pfn) + index;
>       pgd_k = init_mm.pgd + index;
>  
>       if (!pgd_present(*pgd_k)) {
> -- 
> 2.7.4
> 

Reply via email to