[PATCH v26 10/28] riscv/mm: Implement map_shadow_stack() syscall

[Paul Walmsley]

Hi, 

On Mon, 26 Jan 2026, Deepak Gupta wrote:

> Can you apply following diff on `allocate_shadow_stack` function in this
> patch.
> This fixes the bug that I earlier mentioned. We shouldn't be returning
> location
> to token and instead return base address of shadow stack. Userspace consumer
> should be determining token location itself. This matches the ABI of other
> arches. Sorry for being late on this.
> 
> diff --git a/arch/riscv/kernel/usercfi.c b/arch/riscv/kernel/usercfi.c
> index 27b36034ea85..a8530e6afb1e 100644
> --- a/arch/riscv/kernel/usercfi.c
> +++ b/arch/riscv/kernel/usercfi.c
> @@ -232,7 +232,7 @@ static unsigned long allocate_shadow_stack(unsigned long
> addr, unsigned long siz
>  {
>         int flags = MAP_ANONYMOUS | MAP_PRIVATE;
>         struct mm_struct *mm = current->mm;
> -       unsigned long populate, tok_loc = 0;
> +       unsigned long populate;
> 
>         if (addr)
>                 flags |= MAP_FIXED_NOREPLACE;
> @@ -245,13 +245,11 @@ static unsigned long allocate_shadow_stack(unsigned long
> addr, unsigned long siz
>         if (!set_tok || IS_ERR_VALUE(addr))
>                 goto out;
> 
> -       if (create_rstor_token(addr + token_offset, &tok_loc)) {
> +       if (create_rstor_token(addr + token_offset, NULL)) {
>                 vm_munmap(addr, size);
>                 return -EINVAL;
>         }
> 
> -       addr = tok_loc;
> -
>  out:
>         return addr;
>  }

Thanks, this fix has been rolled into the queued patch.  The Reviewed-by:s 
and Tested-by:s have been dropped.


- Paul