Re: [PATCH] mtd: Use kasprintf() instead of fixed buffer formatting

[Boris Brezillon]

On Thu, 28 Jun 2018 10:20:11 +0200
Geert Uytterhoeven <geert+renesas@xxxxxxxxx> wrote:

> Using "%4.4X" in the calculation of the buffer size is misleading, as
> the format string literal has no relation to the actual size needed.
> Hence this is fragile w.r.t. future modification.
> 
> As this is not a hot path, fix this by replacing the formatting in a
> fixed buffer by kasprintf().
> 
> Signed-off-by: Geert Uytterhoeven <geert+renesas@xxxxxxxxx>

Applied.

Thanks,

Boris

> ---
> Compile-tested only.
> 
> Interestingly, this decreases code size (by 16 resp. 24 bytes on arm32
> resp. arm64).
> ---
>  drivers/mtd/chips/gen_probe.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/mtd/chips/gen_probe.c b/drivers/mtd/chips/gen_probe.c
> index 879bebf9d8b12038..837b04ab96a9c526 100644
> --- a/drivers/mtd/chips/gen_probe.c
> +++ b/drivers/mtd/chips/gen_probe.c
> @@ -202,16 +202,19 @@ static inline struct mtd_info *cfi_cmdset_unknown(struct map_info *map,
>  	struct cfi_private *cfi = map->fldrv_priv;
>  	__u16 type = primary?cfi->cfiq->P_ID:cfi->cfiq->A_ID;
>  #ifdef CONFIG_MODULES
> -	char probename[sizeof("cfi_cmdset_%4.4X")];
>  	cfi_cmdset_fn_t *probe_function;
> +	char *probename;
>  
> -	sprintf(probename, "cfi_cmdset_%4.4X", type);
> +	probename = kasprintf(GFP_KERNEL, "cfi_cmdset_%4.4X", type);
> +	if (!probename)
> +		return NULL;
>  
>  	probe_function = __symbol_get(probename);
>  	if (!probe_function) {
>  		request_module("cfi_cmdset_%4.4X", type);
>  		probe_function = __symbol_get(probename);
>  	}
> +	kfree(probename);
>  
>  	if (probe_function) {
>  		struct mtd_info *mtd;