On Wed, Apr 03, 2019 at 11:08:53AM +0200, Felix Fietkau wrote:
> On 2019-03-19 11:37, Stanislaw Gruszka wrote:
> > diff --git a/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c b/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c
> > index ec94d612f53c..75dceeeed059 100644
> > --- a/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c
> > +++ b/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c
> > @@ -152,6 +152,21 @@ static void mt76x02_pre_tbtt_tasklet(unsigned long arg)
> > spin_unlock_bh(&q->lock);
> > }
> >
> > +void mt76x02e_init_beacon_config(struct mt76x02_dev *dev)
> > +{
> > + tasklet_init(&dev->pre_tbtt_tasklet, mt76x02_pre_tbtt_tasklet,
> > + (unsigned long)dev);
> > +
> > + /* Fire a pre-TBTT interrupt 8 ms before TBTT */
> > + mt76_rmw_field(dev, MT_INT_TIMER_CFG, MT_INT_TIMER_CFG_PRE_TBTT, 8 << 4);
> > + mt76_rmw_field(dev, MT_INT_TIMER_CFG, MT_INT_TIMER_CFG_GP_TIMER,
> > + MT_DFS_GP_INTERVAL);
> > + mt76_wr(dev, MT_INT_TIMER_EN, 0);
> > +
> > + mt76x02_init_beacon_config(dev);
> > +}
> > +EXPORT_SYMBOL_GPL(mt76x02e_init_beacon_config);
> > +
> > static int
> > mt76x02_init_tx_queue(struct mt76x02_dev *dev, struct mt76_sw_queue *q,
> > int idx, int n_desc)
> > @@ -230,8 +245,6 @@ int mt76x02_dma_init(struct mt76x02_dev *dev)
> > return -ENOMEM;
> >
> > tasklet_init(&dev->tx_tasklet, mt76x02_tx_tasklet, (unsigned long) dev);
> > - tasklet_init(&dev->pre_tbtt_tasklet, mt76x02_pre_tbtt_tasklet,
> > - (unsigned long)dev);
> Moving the tasklet init to mt76x02e_init_beacon_config can crash the
> kernel, because it leads to the tasklet being re-initialized on every
> reset. I will update that commit and move it back.
Ehh. Some further patches will not apply and will require to be modified.
Perhaps could be easer move mt76x02e_init_beacon_config() to
mt76x2_init_hardware() (or other proper function that is used only
once once during initialization). Or just apply fix on top of the set
as separate patch. Anyway I'm leaving that up to you. Thanks.
Stanislaw
