On 12/14/2018 10:49 PM, Stephen Boyd wrote:
> It seems that it may be possible to get to the regmap update call
> without ever initializing this variable, so just set it to 0 to be safe.
>
> Cc: Marek Vasut <marek.vasut+renesas@xxxxxxxxx>
> Cc: Alexey Firago <alexey_firago@xxxxxxxxxx>
> Cc: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> Cc: linux-renesas-soc@xxxxxxxxxxxxxxx
> Signed-off-by: Stephen Boyd <sboyd@xxxxxxxxxx>
> ---
> drivers/clk/clk-versaclock5.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/clk/clk-versaclock5.c b/drivers/clk/clk-versaclock5.c
> index 5b393e711e94..026f76276933 100644
> --- a/drivers/clk/clk-versaclock5.c
> +++ b/drivers/clk/clk-versaclock5.c
> @@ -246,7 +246,7 @@ static int vc5_mux_set_parent(struct clk_hw *hw, u8 index)
> struct vc5_driver_data *vc5 =
> container_of(hw, struct vc5_driver_data, clk_mux);
> const u8 mask = VC5_PRIM_SRC_SHDN_EN_XTAL | VC5_PRIM_SRC_SHDN_EN_CLKIN;
> - u8 src;
> + u8 src = 0;
>
> if ((index > 1) || !vc5->clk_mux_ins)
> return -EINVAL;
Can you rather change the else { } branch in that function such that it does
if (vc5->clk_mux)ins == ...)
...
else if (...)
...
+ else
+ return -EINVAL;
The last bit would assure the function would fail in case the
clk_mux_ins == 0, which can happen if someone builds a setup without any
clock into connected to the VC5 (which would not work hardware-vise).
--
Best regards,
Marek Vasut
