[PATCH v2 18/25] drm/i915/display: Enable RM timeout detection

Matt Roper matthew.d.roper at intel.com
Wed Apr 3 18:02:58 UTC 2024


On Wed, Apr 03, 2024 at 04:52:46PM +0530, Balasubramani Vivekanandan wrote:
> From: Mitul Golani <mitulkumar.ajitkumar.golani at intel.com>
> 
> Enable RM timeout interrupt to detect any hang during display engine
> register access. This interrupt is supported only on Display version 14.

This doesn't seem to be true.  Bit 29 of the IIR register is still there
on Xe2_LPD (version 20) and I see no reason to believe it won't continue
to be present after that.

It doesn't seem like this patch should even be part of the BMG series.
This was a new interrupt bit first added back on MTL/ARL and continuing
onward from there.  It's basically a "hardware is broken" interrupt that
might be useful for debugging, but we hope to never actually see it show
up.  It isn't something that's related to the general enabling of any
specific platform, especially since it's something the hardware already
supports on a few other platforms already present in the Xe driver.

> Current default timeout is 2ms.
> 
> WA: 14012195489

As Jani noted, this doesn't belong as a trailer.  But it's also untrue;
this isn't related to any kind of workaround and the number here doesn't
reference anything in the workaround database.

> Bspec: 50110
> 
> CC: Suraj Kandpal <suraj.kandpal at intel.com>
> Signed-off-by: Mitul Golani <mitulkumar.ajitkumar.golani at intel.com>
> Signed-off-by: Balasubramani Vivekanandan <balasubramani.vivekanandan at intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_display_irq.c | 10 ++++++++++
>  drivers/gpu/drm/i915/i915_reg.h                  |  3 +++
>  2 files changed, 13 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.c b/drivers/gpu/drm/i915/display/intel_display_irq.c
> index f846c5b108b5..3035b50fcad9 100644
> --- a/drivers/gpu/drm/i915/display/intel_display_irq.c
> +++ b/drivers/gpu/drm/i915/display/intel_display_irq.c
> @@ -851,6 +851,13 @@ gen8_de_misc_irq_handler(struct drm_i915_private *dev_priv, u32 iir)
>  {
>  	bool found = false;
>  
> +	if (iir & GEN8_DE_RM_TIMEOUT) {
> +		u32 val = intel_uncore_read(&dev_priv->uncore,
> +				RMTIMEOUTREG_CAPTURE);
> +		drm_warn(&dev_priv->drm, "Register Access Timeout = 0x%x\n", val);
> +		found = true;
> +	}
> +
>  	if (DISPLAY_VER(dev_priv) >= 14) {
>  		if (iir & (XELPDP_PMDEMAND_RSP |
>  			   XELPDP_PMDEMAND_RSPTOUT_ERR)) {
> @@ -1666,6 +1673,9 @@ void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
>  			de_port_masked |= DSI0_TE | DSI1_TE;
>  	}
>  
> +	if (DISPLAY_VER(dev_priv) == 14)
> +		de_misc_masked |= GEN8_DE_RM_TIMEOUT;
> +
>  	de_pipe_enables = de_pipe_masked |
>  		GEN8_PIPE_VBLANK |
>  		gen8_de_pipe_underrun_mask(dev_priv) |
> diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> index 875d76fb8cd0..d1692b32bb8a 100644
> --- a/drivers/gpu/drm/i915/i915_reg.h
> +++ b/drivers/gpu/drm/i915/i915_reg.h
> @@ -4212,6 +4212,8 @@
>  #define RM_TIMEOUT		_MMIO(0x42060)
>  #define  MMIO_TIMEOUT_US(us)	((us) << 0)
>  
> +#define RMTIMEOUTREG_CAPTURE	_MMIO(0x420e0)
> +
>  /* interrupts */
>  #define DE_MASTER_IRQ_CONTROL   (1 << 31)
>  #define DE_SPRITEB_FLIP_DONE    (1 << 29)
> @@ -4398,6 +4400,7 @@
>  #define GEN8_DE_MISC_IMR _MMIO(0x44464)
>  #define GEN8_DE_MISC_IIR _MMIO(0x44468)
>  #define GEN8_DE_MISC_IER _MMIO(0x4446c)
> +#define  GEN8_DE_RM_TIMEOUT		REG_BIT(29)

Given that this was first introduced in Xe_LPD+, the "GEN8" prefix here
is inappropriate.


Matt

>  #define  XELPDP_PMDEMAND_RSPTOUT_ERR	REG_BIT(27)
>  #define  GEN8_DE_MISC_GSE		REG_BIT(27)
>  #define  GEN8_DE_EDP_PSR		REG_BIT(19)
> -- 
> 2.25.1
> 

-- 
Matt Roper
Graphics Software Engineer
Linux GPU Platform Enablement
Intel Corporation


More information about the Intel-xe mailing list