[PATCH] drm/vc4: Fix resource leak in 'vc4_get_hang_state_ioctl()' in error handling path

[PATCH] drm/vc4: Fix resource leak in 'vc4_get_hang_state_ioctl()' in error handling path

[Christophe JAILLET]

If one 'drm_gem_handle_create()' fails, we leak somes handles and some
memory.

In order to fix it:
  - move the 'free(bo_state)' at the end of the function in the error
    handling path. This has the side effect to also try to free it if the
    first 'kcalloc' fails. This is harmless.
  - delete already allocated handles
  - remove the now useless 'err' label

The way the code is now written will also delete the handles if the
'copy_to_user()' call fails.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
---
This patch also add the 'vc4_free_hang_state()' call in the error
handling path. It sounds logical to me, but I'm not sure of it.
---
 drivers/gpu/drm/vc4/vc4_gem.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c
index e9c381c42139..891c7a22cf81 100644
--- a/drivers/gpu/drm/vc4/vc4_gem.c
+++ b/drivers/gpu/drm/vc4/vc4_gem.c
@@ -111,8 +111,8 @@ vc4_get_hang_state_ioctl(struct drm_device *dev, void *data,
 					    &handle);
 
 		if (ret) {
-			state->bo_count = i - 1;
-			goto err;
+			state->bo_count = i;
+			goto err_free;
 		}
 		bo_state[i].handle = handle;
 		bo_state[i].paddr = vc4_bo->base.paddr;
@@ -124,13 +124,16 @@ vc4_get_hang_state_ioctl(struct drm_device *dev, void *data,
 			 state->bo_count * sizeof(*bo_state)))
 		ret = -EFAULT;
 
-	kfree(bo_state);
-
 err_free:
-
 	vc4_free_hang_state(dev, kernel_state);
 
-err:
+	if (ret) {
+		for (i = 0; i < state->bo_count; i++)
+			drm_gem_handle_delete(file_priv, bo_state[i].handle);
+	}
+
+	kfree(bo_state);
+
 	return ret;
 }
 
-- 
2.11.0

Re: [PATCH] drm/vc4: Fix resource leak in 'vc4_get_hang_state_ioctl()' in error handling path

[Eric Anholt]

[-- Attachment #1: Type: text/plain, Size: 1962 bytes --]

Christophe JAILLET <christophe.jaillet@wanadoo.fr> writes:

> If one 'drm_gem_handle_create()' fails, we leak somes handles and some
> memory.
>
> In order to fix it:
>   - move the 'free(bo_state)' at the end of the function in the error
>     handling path. This has the side effect to also try to free it if the
>     first 'kcalloc' fails. This is harmless.
>   - delete already allocated handles
>   - remove the now useless 'err' label
>
> The way the code is now written will also delete the handles if the
> 'copy_to_user()' call fails.
>
> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
> ---
> This patch also add the 'vc4_free_hang_state()' call in the error
> handling path. It sounds logical to me, but I'm not sure of it.
> ---
>  drivers/gpu/drm/vc4/vc4_gem.c | 15 +++++++++------
>  1 file changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c
> index e9c381c42139..891c7a22cf81 100644
> --- a/drivers/gpu/drm/vc4/vc4_gem.c
> +++ b/drivers/gpu/drm/vc4/vc4_gem.c
> @@ -111,8 +111,8 @@ vc4_get_hang_state_ioctl(struct drm_device *dev, void *data,
>  					    &handle);
>  
>  		if (ret) {
> -			state->bo_count = i - 1;
> -			goto err;
> +			state->bo_count = i;
> +			goto err_free;
>  		}
>  		bo_state[i].handle = handle;
>  		bo_state[i].paddr = vc4_bo->base.paddr;
> @@ -124,13 +124,16 @@ vc4_get_hang_state_ioctl(struct drm_device *dev, void *data,
>  			 state->bo_count * sizeof(*bo_state)))
>  		ret = -EFAULT;
>  
> -	kfree(bo_state);
> -
>  err_free:
> -
>  	vc4_free_hang_state(dev, kernel_state);
>  
> -err:
> +	if (ret) {
> +		for (i = 0; i < state->bo_count; i++)
> +			drm_gem_handle_delete(file_priv, bo_state[i].handle);
> +	}

vc4_free_hang_state() has freed the state pointer you're using here.
Also, I think you'd have a NULL deref from the goto err_free in the
!bo_state error path.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]