On Mon, Aug 11, 2014 at 03:09:35AM +0100, Filipe Manana wrote:
> +             if (name_len + data_len > buf_len) {
> +                     buf_len = name_len + data_len;
> +                     if (is_vmalloc_addr(buf)) {
> +                             vfree(buf);
> +                             buf = NULL;
> +                     } else {
> +                             char *tmp = krealloc(buf, buf_len, GFP_NOFS);

This could fail with a warning (high order allocation) so I suggest to
add __GFP_NOWARN, the vmalloc fallback will catch fragmented memory case
and fail if theres no memory.

> +
> +                             if (!tmp)
> +                                     kfree(buf);
> +                             buf = tmp;
> +                     }
> +                     if (!buf) {
> +                             buf = vmalloc(buf_len);
> +                             if (!buf) {
> +                                     ret = -ENOMEM;
> +                                     goto out;
> +                             }
> +                     }
> +             }
> +
>               read_extent_buffer(eb, buf, (unsigned long)(di + 1),
>                               name_len + data_len);
>  
> @@ -1071,7 +1094,10 @@ static int iterate_dir_item(struct btrfs_root *root, 
> struct btrfs_path *path,
>       }
>  
>  out:
> -     kfree(buf);
> +     if (is_vmalloc_addr(buf))
> +             vfree(buf);
> +     else
> +             kfree(buf);

There's even kvfree to do this.

>       return ret;
>  }
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to