On Thu, Jun 14, 2018 at 08:56:25PM +0300, Ville Syrjala wrote:
> From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
>
> Just like with PIPESTAT, the edge triggered IIR on i965/g4x
> also causes problems for hotplug interrupts. To make sure
> we don't get the IIR port interrupt bit stuck low with the
> ISR bit high we must force an edge in ISR. Unfortunately
> we can't borrow the PIPESTAT trick and toggle the enable
> bits in PORT_HOTPLUG_EN as that act itself generates hotplug
> interrupts. Instead we just have to loop until we've cleared
> PORT_HOTPLUG_STAT, or we just give up and WARN.
>
> v2: Don't frob with PORT_HOTPLUG_EN
>
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
Ok, not so great that we have to loop, but looks like there's no better
way to go about it. Good that the problem was root caused:
Reviewed-by: Imre Deak <imre.deak@xxxxxxxxx>
> ---
> drivers/gpu/drm/i915/i915_irq.c | 32 ++++++++++++++++++++++++++++++--
> 1 file changed, 30 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index d5aee0b74f4b..55a4d4f3fbb2 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -1998,10 +1998,38 @@ static void valleyview_pipestat_irq_handler(struct drm_i915_private *dev_priv,
>
> static u32 i9xx_hpd_irq_ack(struct drm_i915_private *dev_priv)
> {
> - u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
> + u32 hotplug_status = 0, hotplug_status_mask;
> + int i;
> +
> + if (IS_G4X(dev_priv) ||
> + IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
> + hotplug_status_mask = HOTPLUG_INT_STATUS_G4X |
> + DP_AUX_CHANNEL_MASK_INT_STATUS_G4X;
> + else
> + hotplug_status_mask = HOTPLUG_INT_STATUS_I915;
>
> - if (hotplug_status)
> + /*
> + * We absolutely have to clear all the pending interrupt
> + * bits in PORT_HOTPLUG_STAT. Otherwise the ISR port
> + * interrupt bit won't have an edge, and the i965/g4x
> + * edge triggered IIR will not notice that an interrupt
> + * is still pending. We can't use PORT_HOTPLUG_EN to
> + * guarantee the edge as the act of toggling the enable
> + * bits can itself generate a new hotplug interrupt :(
> + */
> + for (i = 0; i < 10; i++) {
> + u32 tmp = I915_READ(PORT_HOTPLUG_STAT) & hotplug_status_mask;
> +
> + if (tmp == 0)
> + return hotplug_status;
> +
> + hotplug_status |= tmp;
> I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
> + }
> +
> + WARN_ONCE(1,
> + "PORT_HOTPLUG_STAT did not clear (0x%08x)\n",
> + I915_READ(PORT_HOTPLUG_STAT));
>
> return hotplug_status;
> }
> --
> 2.16.4
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
