public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
To: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 20/21] drm/i915: Make GuC GGTT reservation work on ggtt
Date: Thu, 6 Jun 2019 06:44:59 -0700	[thread overview]
Message-ID: <20190606134459.GD3452@intel.com> (raw)
In-Reply-To: <20190606093639.9372-21-tvrtko.ursulin@linux.intel.com>

On Thu, Jun 06, 2019 at 10:36:38AM +0100, Tvrtko Ursulin wrote:
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> These functions operate on ggtt so make them take that directly as
> parameter.

This patch makes me wonder where we really want and need to go.

We need to move out of dev_priv and global i915...
but do we need to go and reduce to all minimal stuff used like
uncore and ggtt or could we find a middle solution where
each group has its own "class"?

like this guc stuff would keep the intel_guc, but the i915_gem
stuff or intel_gt stuff would have their own structs where
we have everything needed for that group?

> 
> At the same time move the USES_GUC conditional down to
> intel_guc_reserve_ggtt_top for symmetry with
> intel_guc_reserved_gtt_size.
> 
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> ---
>  drivers/gpu/drm/i915/i915_gem_gtt.c | 14 ++++++--------
>  drivers/gpu/drm/i915/intel_guc.c    | 18 ++++++++----------
>  drivers/gpu/drm/i915/intel_guc.h    |  6 +++---
>  3 files changed, 17 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
> index d3b3676d10f3..d967a4e9ceb0 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.c
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
> @@ -2912,7 +2912,7 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
>  	 * why.
>  	 */
>  	ggtt->pin_bias = max_t(u32, I915_GTT_PAGE_SIZE,
> -			       intel_guc_reserved_gtt_size(&dev_priv->guc));
> +			       intel_guc_reserved_gtt_size(ggtt));
>  
>  	ret = intel_vgt_balloon(ggtt);
>  	if (ret)
> @@ -2926,11 +2926,9 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
>  	if (ret)
>  		return ret;
>  
> -	if (USES_GUC(dev_priv)) {
> -		ret = intel_guc_reserve_ggtt_top(&dev_priv->guc);
> -		if (ret)
> -			goto err_reserve;
> -	}
> +	ret = intel_guc_reserve_ggtt_top(ggtt);
> +	if (ret)
> +		goto err_reserve;
>  
>  	/* Clear any non-preallocated blocks */
>  	drm_mm_for_each_hole(entry, &ggtt->vm.mm, hole_start, hole_end) {
> @@ -2952,7 +2950,7 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
>  	return 0;
>  
>  err_appgtt:
> -	intel_guc_release_ggtt_top(&dev_priv->guc);
> +	intel_guc_release_ggtt_top(ggtt);
>  err_reserve:
>  	drm_mm_remove_node(&ggtt->error_capture);
>  	return ret;
> @@ -2979,7 +2977,7 @@ void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv)
>  	if (drm_mm_node_allocated(&ggtt->error_capture))
>  		drm_mm_remove_node(&ggtt->error_capture);
>  
> -	intel_guc_release_ggtt_top(&dev_priv->guc);
> +	intel_guc_release_ggtt_top(ggtt);
>  
>  	if (drm_mm_initialized(&ggtt->vm.mm)) {
>  		intel_vgt_deballoon(ggtt);
> diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c
> index b88c349c4fa6..633248b7da25 100644
> --- a/drivers/gpu/drm/i915/intel_guc.c
> +++ b/drivers/gpu/drm/i915/intel_guc.c
> @@ -719,7 +719,7 @@ struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size)
>  
>  /**
>   * intel_guc_reserved_gtt_size()
> - * @guc:	intel_guc structure
> + * @ggtt:	Pointer to struct i915_ggtt
>   *
>   * The GuC WOPCM mapping shadows the lower part of the GGTT, so if we are using
>   * GuC we can't have any objects pinned in that region. This function returns
> @@ -729,18 +729,19 @@ struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size)
>   * 0 if GuC is not present or not in use.
>   * Otherwise, the GuC WOPCM size.
>   */
> -u32 intel_guc_reserved_gtt_size(struct intel_guc *guc)
> +u32 intel_guc_reserved_gtt_size(struct i915_ggtt *ggtt)
>  {
> -	return guc_to_i915(guc)->wopcm.guc.size;
> +	return ggtt->vm.i915->wopcm.guc.size;
>  }
>  
> -int intel_guc_reserve_ggtt_top(struct intel_guc *guc)
> +int intel_guc_reserve_ggtt_top(struct i915_ggtt *ggtt)
>  {
> -	struct drm_i915_private *i915 = guc_to_i915(guc);
> -	struct i915_ggtt *ggtt = &i915->ggtt;
>  	u64 size;
>  	int ret;
>  
> +	if (!USES_GUC(ggtt->vm.i915))
> +		return 0;
> +
>  	size = ggtt->vm.total - GUC_GGTT_TOP;
>  
>  	ret = i915_gem_gtt_reserve(&ggtt->vm, &ggtt->uc_fw, size,
> @@ -752,11 +753,8 @@ int intel_guc_reserve_ggtt_top(struct intel_guc *guc)
>  	return ret;
>  }
>  
> -void intel_guc_release_ggtt_top(struct intel_guc *guc)
> +void intel_guc_release_ggtt_top(struct i915_ggtt *ggtt)
>  {
> -	struct drm_i915_private *i915 = guc_to_i915(guc);
> -	struct i915_ggtt *ggtt = &i915->ggtt;
> -
>  	if (drm_mm_node_allocated(&ggtt->uc_fw))
>  		drm_mm_remove_node(&ggtt->uc_fw);
>  }
> diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h
> index cbfed7a77c8b..55ea14176c5e 100644
> --- a/drivers/gpu/drm/i915/intel_guc.h
> +++ b/drivers/gpu/drm/i915/intel_guc.h
> @@ -173,9 +173,9 @@ int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset);
>  int intel_guc_suspend(struct intel_guc *guc);
>  int intel_guc_resume(struct intel_guc *guc);
>  struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size);
> -u32 intel_guc_reserved_gtt_size(struct intel_guc *guc);
> -int intel_guc_reserve_ggtt_top(struct intel_guc *guc);
> -void intel_guc_release_ggtt_top(struct intel_guc *guc);
> +u32 intel_guc_reserved_gtt_size(struct i915_ggtt *ggtt);
> +int intel_guc_reserve_ggtt_top(struct i915_ggtt *ggtt);
> +void intel_guc_release_ggtt_top(struct i915_ggtt *ggtt);
>  
>  static inline bool intel_guc_is_loaded(struct intel_guc *guc)
>  {
> -- 
> 2.20.1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2019-06-06 13:44 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-06-06  9:36 [PATCH 00/21] Implicit dev_priv removal Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 01/21] drm/i915: Reset only affected engines when handling error capture Tvrtko Ursulin
2019-06-06  9:44   ` Chris Wilson
2019-06-06  9:49     ` Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 02/21] drm/i915: Tidy engine mask types in hangcheck Tvrtko Ursulin
2019-06-06  9:45   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 03/21] drm/i915: Make Gen6/7 RING_FAULT_REG access engine centric Tvrtko Ursulin
2019-06-06  9:47   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 04/21] drm/i915: Extract engine fault reset to a helper Tvrtko Ursulin
2019-06-06  9:48   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 05/21] drm/i915: Make i915_clear_error_registers take uncore Tvrtko Ursulin
2019-06-06  9:50   ` Chris Wilson
2019-06-06  9:51     ` Chris Wilson
2019-06-06 10:29     ` Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 06/21] drm/i915: Convert some more bits to use engine mmio accessors Tvrtko Ursulin
2019-06-06  9:53   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 07/21] drm/i915: Make read_subslice_reg take uncore Tvrtko Ursulin
2019-06-06  9:54   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 08/21] drm/i915: Tidy intel_execlists_submission_init Tvrtko Ursulin
2019-06-06  9:55   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 09/21] drm/i915: Make i915_check_and_clear_faults take uncore Tvrtko Ursulin
2019-06-06  9:57   ` Chris Wilson
2019-06-06 10:31     ` Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 10/21] drm/i915: Move scheduler caps init to i915_gem_init Tvrtko Ursulin
2019-06-06  9:59   ` Chris Wilson
2019-06-06  9:36 ` [PATCH 11/21] drm/i915: Remove impossible path from i915_gem_init_swizzling Tvrtko Ursulin
2019-06-06 10:01   ` Chris Wilson
2019-06-06 10:23     ` Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 12/21] drm/i915: Convert i915_gem_init_swizzling to uncore Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 13/21] drm/i915: Convert init_unused_rings " Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 14/21] drm/i915: Convert gt workarounds " Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 15/21] drm/i915: Convert intel_mocs_init_l3cc_table " Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 16/21] drm/i915: Convert i915_ppgtt_init_hw " Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 17/21] drm/i915: Consolidate some open coded mmio rmw Tvrtko Ursulin
2019-06-06 13:46   ` Rodrigo Vivi
2019-06-06  9:36 ` [PATCH 18/21] drm/i915: Convert i915_gem_init_hw to uncore Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 19/21] drm/i915: Convert intel_vgt_(de)balloon " Tvrtko Ursulin
2019-06-06  9:36 ` [PATCH 20/21] drm/i915: Make GuC GGTT reservation work on ggtt Tvrtko Ursulin
2019-06-06 11:58   ` Michal Wajdeczko
2019-06-06 12:23     ` Tvrtko Ursulin
2019-06-06 13:21       ` Michal Wajdeczko
2019-06-06 13:44   ` Rodrigo Vivi [this message]
2019-06-06  9:36 ` [PATCH 21/21] drm/i915: Unexport i915_gem_init/fini_aliasing_ppgtt Tvrtko Ursulin
2019-06-06 13:39   ` Rodrigo Vivi
2019-06-06 10:05 ` [PATCH 00/21] Implicit dev_priv removal Chris Wilson
2019-06-06 10:35   ` Tvrtko Ursulin
2019-06-06 10:10 ` ✗ Fi.CI.CHECKPATCH: warning for " Patchwork
2019-06-06 10:19 ` ✗ Fi.CI.SPARSE: " Patchwork
2019-06-06 12:42 ` ✓ Fi.CI.BAT: success " Patchwork
2019-06-08 15:11 ` ✗ Fi.CI.IGT: failure " 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=20190606134459.GD3452@intel.com \
    --to=rodrigo.vivi@intel.com \
    --cc=Intel-gfx@lists.freedesktop.org \
    --cc=tvrtko.ursulin@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