Re: [PATCH] mm/util.c: add a none zero check of "len"

[David Rientjes <rientjes@google.com>]

On Tue, 20 Jan 2015, Pan Xinhui wrote:

> Although this check should have been done by caller. But as it's exported to
> others,
> It's better to add a none zero check of "len" like other functions.
> 
> Signed-off-by: xinhuix.pan <xinhuix.pan@intel.com>
> ---
>  mm/util.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/mm/util.c b/mm/util.c
> index fec39d4..3dc2873 100644
> --- a/mm/util.c
> +++ b/mm/util.c
> @@ -72,6 +72,9 @@ void *kmemdup(const void *src, size_t len, gfp_t gfp)
>  {
>  	void *p;
>  +	if (unlikely(!len))
> +		return ERR_PTR(-EINVAL);
> +
>  	p = kmalloc_track_caller(len, gfp);
>  	if (p)
>  		memcpy(p, src, len);
> @@ -91,6 +94,8 @@ void *memdup_user(const void __user *src, size_t len)
>  {
>  	void *p;
>  +	if (unlikely(!len))
> +		return ERR_PTR(-EINVAL);
>  	/*
>  	 * Always use GFP_KERNEL, since copy_from_user() can sleep and
>  	 * cause pagefault, which makes it pointless to use GFP_NOFS

Nack, there's no need for this since both slab and slub check for 
ZERO_OR_NULL_PTR() and kmalloc_slab() will return ZERO_SIZE_PTR in these 
cases.  kmemdup() would then return NULL, which is appropriate since it 
doesn't return an ERR_PTR() even when memory cannot be allocated.  
memdup_user() would return -ENOMEM for size == 0, which would arguably be 
the wrong return value, but I don't think we need to slow down either of 
these library functions to check for something as stupid as duplicating 
size == 0.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>