[PATCH v6 4/8] drm/msm/dp: avoid handling masked interrupts
Sankeerth Billakanti (QUIC)
quic_sbillaka at quicinc.com
Thu Mar 31 05:53:14 UTC 2022
Hi Dmitry,
> On Wed, 30 Mar 2022 at 19:03, Sankeerth Billakanti
> <quic_sbillaka at quicinc.com> wrote:
> >
> > The interrupt register will still reflect the connect and disconnect
> > interrupt status without generating an actual HW interrupt.
> > The controller driver should not handle those masked interrupts.
> >
> > Signed-off-by: Sankeerth Billakanti <quic_sbillaka at quicinc.com>
> > ---
> > drivers/gpu/drm/msm/dp/dp_catalog.c | 5 +++--
> > 1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/msm/dp/dp_catalog.c
> > b/drivers/gpu/drm/msm/dp/dp_catalog.c
> > index 3c16f95..1809ce2 100644
> > --- a/drivers/gpu/drm/msm/dp/dp_catalog.c
> > +++ b/drivers/gpu/drm/msm/dp/dp_catalog.c
> > @@ -608,13 +608,14 @@ u32 dp_catalog_hpd_get_intr_status(struct
> > dp_catalog *dp_catalog) {
> > struct dp_catalog_private *catalog = container_of(dp_catalog,
> > struct dp_catalog_private, dp_catalog);
> > - int isr = 0;
> > + int isr, mask;
> >
> > isr = dp_read_aux(catalog, REG_DP_DP_HPD_INT_STATUS);
> > dp_write_aux(catalog, REG_DP_DP_HPD_INT_ACK,
> > (isr & DP_DP_HPD_INT_MASK));
> > + mask = dp_read_aux(catalog, REG_DP_DP_HPD_INT_MASK);
> >
> > - return isr;
> > + return isr & (DP_DP_HPD_STATE_STATUS_MASK | mask);
>
> I suspect that the logic is inverted here. Shouldn't it be:
>
> return isr & DP_DP_HPD_STATE_STATUS_MASK & mask;
>
> ?
>
The value of DP_DP_HPD_STATE_STATUS_MASK is 0xE0000000 and the value of the read
interrupt mask variable could be is 0xF.
The mask value is indicated via the register, REG_DP_DP_HPD_INT_MASK, bits 3:0.
The HPD status is indicated via a different read-only register REG_DP_DP_HPD_INT_STATUS, bits 31:29.
isr & DP_DP_HPD_STATE_STATUS_MASK & mask, will return 0 always.
> > }
> >
> > int dp_catalog_ctrl_get_interrupt(struct dp_catalog *dp_catalog)
> > --
> > 2.7.4
> >
>
>
> --
> With best wishes
> Dmitry
Thank you,
Sankeerth
More information about the dri-devel
mailing list