Re: [PATCH] drm/i915: Replace kmalloc() + copy_from_user() with memdup_user()

[Tvrtko Ursulin]

On 02/09/2025 09:10, Thorsten Blum wrote:
> Replace kmalloc() followed by copy_from_user() with memdup_user() to
> improve and simplify set_context_image(), and to silence the following
> Coccinelle/coccicheck warning reported by memdup_user.cocci:
>
>    WARNING opportunity for memdup_user
>
> No functional changes intended.
>
> Signed-off-by: Thorsten Blum <thorsten.blum@xxxxxxxxx>
> ---
> Resending this (with updated patch subject and description) because the
> CI logs [1] from my first submission [2] about a year ago are no longer
> available.
>
> [1] https://patchwork.freedesktop.org/series/139319/
> [2] https://lore.kernel.org/lkml/20240925141750.51198-1-thorsten.blum@xxxxxxxxx/
> ---
>   drivers/gpu/drm/i915/gem/i915_gem_context.c | 12 +++---------
>   1 file changed, 3 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> index 15835952352e..ed6599694835 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> @@ -2158,18 +2158,12 @@ static int set_context_image(struct i915_gem_context *ctx,
>   		goto out_ce;
>   	}
>   
> -	state = kmalloc(ce->engine->context_size, GFP_KERNEL);
> -	if (!state) {
> -		ret = -ENOMEM;
> +	state = memdup_user(u64_to_user_ptr(user.image), ce->engine->context_size);
> +	if (IS_ERR(state)) {
> +		ret = PTR_ERR(state);
>   		goto out_ce;
>   	}
>   
> -	if (copy_from_user(state, u64_to_user_ptr(user.image),
> -			   ce->engine->context_size)) {
> -		ret = -EFAULT;
> -		goto out_state;
> -	}
> -
>   	shmem_state = shmem_create_from_data(ce->engine->name,
>   					     state, ce->engine->context_size);
>   	if (IS_ERR(shmem_state)) {

LGTM.

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxxx>

Note however that CI is not picking up external patches these days, so I 
will need to re-send it for you. Once it passes I will merge it but if I 
forget feel free to ping.

Regards,

Tvrtko