Re: [Intel-gfx] [PATCH v6 2/2] drm/i915: add guard page to ggtt->error_capture

From: "Das, Nirmoy" <nirmoy.das@linux.intel.com>
To: Andrzej Hajda <andrzej.hajda@intel.com>,
	Jani Nikula <jani.nikula@linux.intel.com>,
	Joonas Lahtinen <joonas.lahtinen@linux.intel.com>,
	Rodrigo Vivi <rodrigo.vivi@intel.com>,
	Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: intel-gfx@lists.freedesktop.org, linux-kernel@vger.kernel.org,
	dri-devel@lists.freedesktop.org,
	Andi Shyti <andi.shyti@linux.intel.com>,
	Chris Wilson <chris.p.wilson@linux.intel.com>,
	Nirmoy Das <nirmoy.das@intel.com>
Subject: Re: [Intel-gfx] [PATCH v6 2/2] drm/i915: add guard page to ggtt->error_capture
Date: Mon, 13 Mar 2023 13:58:53 +0100	[thread overview]
Message-ID: <a07ac81c-f234-a68b-0d68-225259e534a8@linux.intel.com> (raw)
In-Reply-To: <20230308-guard_error_capture-v6-2-1b5f31422563@intel.com>


On 3/10/2023 10:23 AM, Andrzej Hajda wrote:
> Write-combining memory allows speculative reads by CPU.
> ggtt->error_capture is WC mapped to CPU, so CPU/MMU can try
> to prefetch memory beyond the error_capture, ie it tries
> to read memory pointed by next PTE in GGTT.
> If this PTE points to invalid address DMAR errors will occur.
> This behaviour was observed on ADL and RPL platforms.
> To avoid it, guard scratch page should be added after error_capture.
> The patch fixes the most annoying issue with error capture but
> since WC reads are used also in other places there is a risk similar
> problem can affect them as well.
>
> v2:
>    - modified commit message (I hope the diagnosis is correct),
>    - added bug checks to ensure scratch is initialized on gen3 platforms.
>      CI produces strange stacktrace for it suggesting scratch[0] is NULL,
>      to be removed after resolving the issue with gen3 platforms.
> v3:
>    - removed bug checks, replaced with gen check.
> v4:
>    - change code for scratch page insertion to support all platforms,
>    - add info in commit message there could be more similar issues
> v5:
>    - check for nop_clear_range instead of gen8 (Tvrtko),
>    - re-insert scratch pages on resume (Tvrtko)
> v6:
>    - use scratch_range callback to set scratch pages (Chris)
>
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
> Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Acked-by: Nirmoy Das <nirmoy.das@intel.com>
> ---
>   drivers/gpu/drm/i915/gt/intel_ggtt.c | 20 ++++++++++++++++----
>   1 file changed, 16 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c
> index 38e6f0b207fe0c..5ef7e03b11c8e6 100644
> --- a/drivers/gpu/drm/i915/gt/intel_ggtt.c
> +++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c
> @@ -572,8 +572,12 @@ static int init_ggtt(struct i915_ggtt *ggtt)
>   		 * paths, and we trust that 0 will remain reserved. However,
>   		 * the only likely reason for failure to insert is a driver
>   		 * bug, which we expect to cause other failures...
> +		 *
> +		 * Since CPU can perform speculative reads on error capture
> +		 * (write-combining allows it) add scratch page after error
> +		 * capture to avoid DMAR errors.
>   		 */
> -		ggtt->error_capture.size = I915_GTT_PAGE_SIZE;
> +		ggtt->error_capture.size = 2 * I915_GTT_PAGE_SIZE;
>   		ggtt->error_capture.color = I915_COLOR_UNEVICTABLE;
>   		if (drm_mm_reserve_node(&ggtt->vm.mm, &ggtt->error_capture))
>   			drm_mm_insert_node_in_range(&ggtt->vm.mm,
> @@ -583,11 +587,15 @@ static int init_ggtt(struct i915_ggtt *ggtt)
>   						    0, ggtt->mappable_end,
>   						    DRM_MM_INSERT_LOW);
>   	}
> -	if (drm_mm_node_allocated(&ggtt->error_capture))
> +	if (drm_mm_node_allocated(&ggtt->error_capture)) {
> +		u64 start = ggtt->error_capture.start;
> +		u64 size = ggtt->error_capture.size;
> +
> +		ggtt->vm.scratch_range(&ggtt->vm, start, size);
>   		drm_dbg(&ggtt->vm.i915->drm,
>   			"Reserved GGTT:[%llx, %llx] for use by error capture\n",
> -			ggtt->error_capture.start,
> -			ggtt->error_capture.start + ggtt->error_capture.size);
> +			start, start + size);
> +	}
>   
>   	/*
>   	 * The upper portion of the GuC address space has a sizeable hole
> @@ -1280,6 +1288,10 @@ void i915_ggtt_resume(struct i915_ggtt *ggtt)
>   
>   	flush = i915_ggtt_resume_vm(&ggtt->vm);
>   
> +	if (drm_mm_node_allocated(&ggtt->error_capture))
> +		ggtt->vm.scratch_range(&ggtt->vm, ggtt->error_capture.start,
> +				       ggtt->error_capture.size);
> +
>   	ggtt->invalidate(ggtt);
>   
>   	if (flush)
>