Re: [PATCH v2 02/15] userfaultfd: introduce struct mfill_state

[David Hildenbrand (Arm)]

On 3/6/26 18:18, Mike Rapoport wrote:
> From: "Mike Rapoport (Microsoft)" <rppt@xxxxxxxxxx>
> 
> mfill_atomic() passes a lot of parameters down to its callees.
> 
> Aggregate them all into mfill_state structure and pass this structure to
> functions that implement various UFFDIO_ commands.
> 
> Tracking the state in a structure will allow moving the code that retries
> copying of data for UFFDIO_COPY into mfill_atomic_pte_copy() and make the
> loop in mfill_atomic() identical for all UFFDIO operations on PTE-mapped
> memory.
> 
> The mfill_state definition is deliberately local to mm/userfaultfd.c, hence
> shmem_mfill_atomic_pte() is not updated.
> 
> Signed-off-by: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx>
> ---

[...]

>  			if (fatal_signal_pending(current))
> @@ -866,10 +882,10 @@ static __always_inline ssize_t mfill_atomic(struct userfaultfd_ctx *ctx,
>  
>  out_unlock:
>  	up_read(&ctx->map_changing_lock);
> -	uffd_mfill_unlock(dst_vma);
> +	uffd_mfill_unlock(state.vma);
>  out:
> -	if (folio)
> -		folio_put(folio);
> +	if (state.folio)
> +		folio_put(state.folio);
>  	VM_WARN_ON_ONCE(copied < 0);
>  	VM_WARN_ON_ONCE(err > 0);
>  	VM_WARN_ON_ONCE(!copied && !err);

I guess, "copied" could be easily avoided by
comparing state.src_addr with start_addr.

Acked-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>

-- 
Cheers,

David