On 2017-08-17 13:53, Dan Carpenter wrote:
> If i2c_mux_alloc() fails then we'd have a NULL derefecrence here.
>
> Fixes: c4aee3e1b0de ("i2c: mux: pinctrl: remove platform_data")
> Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
Indeed, thanks for the patch. Applied.
(IIUC, small memory allocations are not supposed to fail so this is
not really a "real" problem?)
FTR, I have work in progress that further cleans up i2c muxes.
Most .remove functions disappear from drivers/i2c/muxes/ and
most i2c mux users outside of i2c/muxes are also simplified for
a nice net loss in lines-of-code, and this bug would simply not
have happened with that. Oh well, that's not for the this cycle
anyway...
Cheers,
peda
> diff --git a/drivers/i2c/muxes/i2c-mux-pinctrl.c b/drivers/i2c/muxes/i2c-mux-pinctrl.c
> index 20b90a7a1e61..cc6818aabab5 100644
> --- a/drivers/i2c/muxes/i2c-mux-pinctrl.c
> +++ b/drivers/i2c/muxes/i2c-mux-pinctrl.c
> @@ -173,7 +173,7 @@ static int i2c_mux_pinctrl_probe(struct platform_device *pdev)
> err_del_adapter:
> i2c_mux_del_adapters(muxc);
> err_put_parent:
> - i2c_put_adapter(muxc->parent);
> + i2c_put_adapter(parent);
>
> return ret;
> }
>
