On Fri, Jan 05, 2024 at 12:57:23AM +0100, Francesco Dolcini wrote: > From: Francesco Dolcini <francesco.dolcini@xxxxxxxxxxx> > > m_can_clk_start() already skip starting the clock when > clock support is disabled, remove the redundant check in > m_can_class_register(). > > This also solves the imbalance with m_can_clk_stop() that is called > afterward in the same function before the return. > > Signed-off-by: Francesco Dolcini <francesco.dolcini@xxxxxxxxxxx> Reviewed-by: Markus Schneider-Pargmann <msp@xxxxxxxxxxxx> Best, Markus > --- > I spotted the issue while debugging some other part of the code, > the patch is only compile-tested. > --- > drivers/net/can/m_can/m_can.c | 8 +++----- > 1 file changed, 3 insertions(+), 5 deletions(-) > > diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c > index 16ecc11c7f62..bd1d1626684d 100644 > --- a/drivers/net/can/m_can/m_can.c > +++ b/drivers/net/can/m_can/m_can.c > @@ -2056,11 +2056,9 @@ int m_can_class_register(struct m_can_classdev *cdev) > { > int ret; > > - if (cdev->pm_clock_support) { > - ret = m_can_clk_start(cdev); > - if (ret) > - return ret; > - } > + ret = m_can_clk_start(cdev); > + if (ret) > + return ret; > > if (cdev->is_peripheral) { > ret = can_rx_offload_add_manual(cdev->net, &cdev->offload, > -- > 2.39.2 >