Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Nirmoy Das <nirmoy.das@intel.com>
To: Andi Shyti <andi.shyti@linux.intel.com>,
	Jonathan Cavitt <jonathan.cavitt@intel.com>,
	Matt Roper <matthew.d.roper@intel.com>,
	"Chris Wilson" <chris@chris-wilson.co.uk>,
	Mika Kuoppala <mika.kuoppala@linux.intel.com>,
	Andrzej Hajda <andrzej.hajda@intel.com>
Cc: Intel GFX <intel-gfx@lists.freedesktop.org>,
	DRI Devel <dri-devel@lists.freedesktop.org>
Subject: Re: [Intel-gfx] [PATCH v5 7/9] drm/i915/gt: Ensure memory quiesced before invalidation for all engines
Date: Tue, 18 Jul 2023 16:12:58 +0200	[thread overview]
Message-ID: <00d7515b-8e02-e3e1-2018-67fe0073b427@intel.com> (raw)
In-Reply-To: <20230718133836.574781-8-andi.shyti@linux.intel.com>

Hi Andi,

On 7/18/2023 3:38 PM, Andi Shyti wrote:
> Commit af9e423a8aae ("drm/i915/gt: Ensure memory quiesced before
> invalidation") has made sure that the memory is quiesced before
> invalidating the AUX CCS table. Do it for all the other engines
> and not just RCS.
>
> Signed-off-by: Andi Shyti <andi.shyti@linux.intel.com>
> Cc: Jonathan Cavitt <jonathan.cavitt@intel.com>
> Cc: Matt Roper <matthew.d.roper@intel.com>
> ---
>   drivers/gpu/drm/i915/gt/gen8_engine_cs.c     | 71 +++++++++++++++++---
>   drivers/gpu/drm/i915/gt/intel_gpu_commands.h |  1 +
>   2 files changed, 62 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/gen8_engine_cs.c b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
> index 3275e55b18d90..2f40cd515cc78 100644
> --- a/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
> +++ b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
> @@ -225,6 +225,13 @@ int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode)
>   
>   		bit_group_0 |= PIPE_CONTROL0_HDC_PIPELINE_FLUSH;
>   
> +		/*
> +		 * When required, in MTL+ platforms we need to
> +		 * set the CCS_FLUSH bit in the pipe control
> +		 */
> +		if (GRAPHICS_VER_FULL(rq->i915) >= IP_VER(12, 70))
> +			bit_group_0 |= PIPE_CONTROL_CCS_FLUSH;
> +
>   		bit_group_1 |= PIPE_CONTROL_TILE_CACHE_FLUSH;
>   		bit_group_1 |= PIPE_CONTROL_FLUSH_L3;
>   		bit_group_1 |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
> @@ -309,20 +316,64 @@ int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode)
>   int gen12_emit_flush_xcs(struct i915_request *rq, u32 mode)
>   {
>   	intel_engine_mask_t aux_inv = 0;
> -	u32 cmd, *cs;
> +	u32 cmd = 4;
> +	u32 *cs;
>   
> -	cmd = 4;
> -	if (mode & EMIT_INVALIDATE) {
> +	if (mode & EMIT_INVALIDATE)
>   		cmd += 2;
>   
> -		if (HAS_AUX_CCS(rq->engine->i915) &&
> -		    (rq->engine->class == VIDEO_DECODE_CLASS ||
> -		     rq->engine->class == VIDEO_ENHANCEMENT_CLASS)) {
> -			aux_inv = rq->engine->mask &
> -				~GENMASK(_BCS(I915_MAX_BCS - 1), BCS0);
> -			if (aux_inv)
> -				cmd += 4;
> +	if (HAS_AUX_CCS(rq->engine->i915))
> +		aux_inv = rq->engine->mask &
> +			  ~GENMASK(_BCS(I915_MAX_BCS - 1), BCS0);
> +
> +	/*
> +	 * Aux invalidations on Aux CCS platforms require
> +	 * memory traffic is quiesced prior.
> +	 */
> +	if (aux_inv) {
> +		u32 bit_group_0 = 0;
> +		u32 bit_group_1 = 0;
> +
> +		cmd += 4;
> +
> +		bit_group_0 |= PIPE_CONTROL0_HDC_PIPELINE_FLUSH;
> +
> +		switch (rq->engine->class) {
> +		case VIDEO_DECODE_CLASS:
> +			bit_group_1 |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
> +			bit_group_1 |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
> +			bit_group_1 |= PIPE_CONTROL_DC_FLUSH_ENABLE;
> +			bit_group_1 |= PIPE_CONTROL_FLUSH_L3;
> +			bit_group_1 |= PIPE_CONTROL_CS_STALL;
> +
> +			/*
> +			 * When required, in MTL+ platforms we need to
> +			 * set the CCS_FLUSH bit in the pipe control
> +			 */
> +			if (GRAPHICS_VER_FULL(rq->i915) >= IP_VER(12, 70))
> +				bit_group_0 |= PIPE_CONTROL_CCS_FLUSH;
> +
> +			break;
> +
> +		case VIDEO_ENHANCEMENT_CLASS:
> +		case COMPUTE_CLASS:
> +			bit_group_1 |= MI_FLUSH_DW;
> +
> +			break;
> +
> +		case COPY_ENGINE_CLASS:
> +			/*
> +			 * When required, in MTL+ platforms we need to
> +			 * set the CCS_FLUSH bit in the pipe control
> +			 */
> +			if (GRAPHICS_VER_FULL(rq->i915) >= IP_VER(12, 70))
> +				bit_group_0 |= PIPE_CONTROL_CCS_FLUSH;
> +			break;
>   		}
> +
> +		if (bit_group_1 || bit_group_0)
> +			intel_emit_pipe_control_cs(rq, bit_group_0, bit_group_1,
> +						   LRC_PPHWSP_SCRATCH_ADDR);


The pipe control is limited to render and compute engines only and

gen12_emit_flush_xcs() gets called only for other engines(BCS,VE,VD) AFAIU. So I imagine changes for this patch as:

gen12_emit_flush_rcs()
	pipe_control with CCS_FLUSH
	AUX CCS inval
gen12_emit_flush_xcs()
	MI_FLUSH_DW (with CCS flush for BCS)
	AUX CCS inval

(Note that ccs flush bit for MI_FLUSH_DW is at 16 )

Regards,
Nirmoy

>   	}
>   
>   	cs = intel_ring_begin(rq, cmd);
> diff --git a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> index 5d143e2a8db03..5df7cce23197c 100644
> --- a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> +++ b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> @@ -299,6 +299,7 @@
>   #define   PIPE_CONTROL_QW_WRITE				(1<<14)
>   #define   PIPE_CONTROL_POST_SYNC_OP_MASK                (3<<14)
>   #define   PIPE_CONTROL_DEPTH_STALL			(1<<13)
> +#define   PIPE_CONTROL_CCS_FLUSH			(1<<13) /* MTL+ */
>   #define   PIPE_CONTROL_WRITE_FLUSH			(1<<12)
>   #define   PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH	(1<<12) /* gen6+ */
>   #define   PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE	(1<<11) /* MBZ on ILK */

  reply	other threads:[~2023-07-18 14:13 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-18 13:38 [Intel-gfx] [PATCH v5 0/9] Update AUX invalidation sequence Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 1/9] drm-tip: 2023y-07m-17d-16h-04m-53s UTC integration manifest Andi Shyti
2023-07-18 13:44   ` Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 2/9] drm/i915/gt: Cleanup aux invalidation registers Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 3/9] drm/i915: Add the has_aux_ccs device property Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 4/9] drm/i915/gt: Ensure memory quiesced before invalidation Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 5/9] drm/i915/gt: Rename flags with bit_group_X according to the datasheet Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 6/9] drm/i915/gt: Refactor intel_emit_pipe_control_cs() in a single function Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 7/9] drm/i915/gt: Ensure memory quiesced before invalidation for all engines Andi Shyti
2023-07-18 14:12   ` Nirmoy Das [this message]
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 8/9] drm/i915/gt: Poll aux invalidation register bit on invalidation Andi Shyti
2023-07-18 13:38 ` [Intel-gfx] [PATCH v5 9/9] drm/i915/gt: Support aux invalidation on all engines Andi Shyti
2023-07-18 15:51 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Update AUX invalidation sequence (rev5) Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=00d7515b-8e02-e3e1-2018-67fe0073b427@intel.com \
    --to=nirmoy.das@intel.com \
    --cc=andi.shyti@linux.intel.com \
    --cc=andrzej.hajda@intel.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=jonathan.cavitt@intel.com \
    --cc=matthew.d.roper@intel.com \
    --cc=mika.kuoppala@linux.intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox