intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: intel-gfx@lists.freedesktop.org, intel-xe@lists.freedesktop.org
Subject: Re: [PATCH 16/20] drm/i915/fbdev: Extract intel_fbdev_fb_prefer_stolen()
Date: Wed, 10 Jul 2024 14:47:56 +0300	[thread overview]
Message-ID: <Zo507NB2VMELZAxH@intel.com> (raw)
In-Reply-To: <4mw2rthck4nmipknligxj6hu2siwzg236ge2vwor3w4ae7pydu@i2hwfhwouxon>

On Tue, Jul 09, 2024 at 03:28:15PM -0500, Lucas De Marchi wrote:
> On Fri, Jul 05, 2024 at 05:52:50PM GMT, Ville Syrjälä wrote:
> >From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> >
> >Consolidate the "should we allocate fbdev fb in stolen?"
> >check into a helper function. Makes it easier to change the
> >heuristics without having to change so many places.
> >
> >Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> >---
> > drivers/gpu/drm/i915/display/intel_fbdev_fb.c | 24 ++++++++++++-------
> > drivers/gpu/drm/i915/display/intel_fbdev_fb.h |  5 +++-
> > .../drm/i915/display/intel_plane_initial.c    | 10 +++-----
> > 3 files changed, 23 insertions(+), 16 deletions(-)
> >
> >diff --git a/drivers/gpu/drm/i915/display/intel_fbdev_fb.c b/drivers/gpu/drm/i915/display/intel_fbdev_fb.c
> >index 497525ef9668..0a6445acb100 100644
> >--- a/drivers/gpu/drm/i915/display/intel_fbdev_fb.c
> >+++ b/drivers/gpu/drm/i915/display/intel_fbdev_fb.c
> >@@ -11,6 +11,19 @@
> > #include "intel_display_types.h"
> > #include "intel_fbdev_fb.h"
> >
> >+bool intel_fbdev_fb_prefer_stolen(struct intel_display *display,
> >+				  unsigned int size)
> >+{
> >+	struct drm_i915_private *i915 = to_i915(display->drm);
> >+
> >+	/*
> >+	 * If the FB is too big, just don't use it since fbdev is not very
> >+	 * important and we should probably use that space with FBC or other
> >+	 * features.
> >+	 */
> >+	return i915->dsm.usable_size >= size * 2;
> >+}
> >+
> > struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper,
> > 					       struct drm_fb_helper_surface_size *sizes)
> > {
> >@@ -42,14 +55,9 @@ struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper,
> > 						  I915_BO_ALLOC_CONTIGUOUS |
> > 						  I915_BO_ALLOC_USER);
> > 	} else {
> >-		/*
> >-		 * If the FB is too big, just don't use it since fbdev is not very
> >-		 * important and we should probably use that space with FBC or other
> >-		 * features.
> >-		 *
> >-		 * Also skip stolen on MTL as Wa_22018444074 mitigation.
> >-		 */
> >-		if (!(IS_METEORLAKE(dev_priv)) && size * 2 < dev_priv->dsm.usable_size)
> >+		/* skip stolen on MTL as Wa_22018444074 mitigation */
> >+		if (!IS_METEORLAKE(dev_priv) &&
> 
> shouldn't this be inside intel_fbdev_fb_prefer_stolen()?

That would also apply it to the BIOS fb takeover, so change the
behaviour. The correct answer is likely just removing the MTL check
entirely, but I left that out for now to avoid too many functional
changes.

> 
> And also pull the same logic on the xe side a few patches after this.
> 
> Lucas De Marchi
> 
> >+		    intel_fbdev_fb_prefer_stolen(&dev_priv->display, size))
> > 			obj = i915_gem_object_create_stolen(dev_priv, size);
> > 		if (IS_ERR(obj))
> > 			obj = i915_gem_object_create_shmem(dev_priv, size);
> >diff --git a/drivers/gpu/drm/i915/display/intel_fbdev_fb.h b/drivers/gpu/drm/i915/display/intel_fbdev_fb.h
> >index 4832fe688fbf..3b9033bd2160 100644
> >--- a/drivers/gpu/drm/i915/display/intel_fbdev_fb.h
> >+++ b/drivers/gpu/drm/i915/display/intel_fbdev_fb.h
> >@@ -6,16 +6,19 @@
> > #ifndef __INTEL_FBDEV_FB_H__
> > #define __INTEL_FBDEV_FB_H__
> >
> >+#include <linux/types.h>
> >+
> > struct drm_fb_helper;
> > struct drm_fb_helper_surface_size;
> > struct drm_i915_gem_object;
> > struct drm_i915_private;
> > struct fb_info;
> > struct i915_vma;
> >+struct intel_display;
> >
> > struct intel_framebuffer *intel_fbdev_fb_alloc(struct drm_fb_helper *helper,
> > 					       struct drm_fb_helper_surface_size *sizes);
> > int intel_fbdev_fb_fill_info(struct drm_i915_private *i915, struct fb_info *info,
> > 			     struct drm_i915_gem_object *obj, struct i915_vma *vma);
> >-
> >+bool intel_fbdev_fb_prefer_stolen(struct intel_display *display, unsigned int size);
> > #endif
> >diff --git a/drivers/gpu/drm/i915/display/intel_plane_initial.c b/drivers/gpu/drm/i915/display/intel_plane_initial.c
> >index ada1792df5b3..4622bb5f3426 100644
> >--- a/drivers/gpu/drm/i915/display/intel_plane_initial.c
> >+++ b/drivers/gpu/drm/i915/display/intel_plane_initial.c
> >@@ -11,6 +11,7 @@
> > #include "intel_display.h"
> > #include "intel_display_types.h"
> > #include "intel_fb.h"
> >+#include "intel_fbdev_fb.h"
> > #include "intel_frontbuffer.h"
> > #include "intel_plane_initial.h"
> >
> >@@ -160,15 +161,10 @@ initial_plane_vma(struct drm_i915_private *i915,
> > 			mem->min_page_size);
> > 	size -= base;
> >
> >-	/*
> >-	 * If the FB is too big, just don't use it since fbdev is not very
> >-	 * important and we should probably use that space with FBC or other
> >-	 * features.
> >-	 */
> > 	if (IS_ENABLED(CONFIG_FRAMEBUFFER_CONSOLE) &&
> > 	    mem == i915->mm.stolen_region &&
> >-	    size * 2 > i915->dsm.usable_size) {
> >-		drm_dbg_kms(&i915->drm, "Initial FB size exceeds half of stolen, discarding\n");
> >+	    !intel_fbdev_fb_prefer_stolen(&i915->display, size)) {
> >+		drm_dbg_kms(&i915->drm, "Initial FB size uses too much stolen, discarding\n");
> > 		return NULL;
> > 	}
> >
> >-- 
> >2.44.2
> >

-- 
Ville Syrjälä
Intel

  reply	other threads:[~2024-07-10 11:48 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-07-05 14:52 [PATCH 00/20] drm/{i915, xe}: FBC cleanups + tweak fbdev stolen usage Ville Syrjala
2024-07-05 14:52 ` [PATCH 01/20] drm/i915/fbc: Extract intel_fbc_has_fences() Ville Syrjala
2024-07-09 19:46   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 02/20] drm/i915/fbc: Convert to intel_display, mostly Ville Syrjala
2024-07-09 19:49   ` Rodrigo Vivi
2024-07-12 21:38     ` Ville Syrjälä
2024-07-05 14:52 ` [PATCH 03/20] drm/i915/fbc: s/_intel_fbc_cfb_stride()/intel_fbc_plane_cfb_stride()/ Ville Syrjala
2024-07-09 19:50   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 04/20] drm/i915/fbc: Extract intel_fbc_max_plane_size() Ville Syrjala
2024-07-09 19:51   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 05/20] drm/i915/fbc: Extract intel_fbc_max_surface_size() Ville Syrjala
2024-07-09 19:51   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 06/20] drm/i915/fbc: s/intel_fbc_hw_tracking_covers_screen()/intel_fbc_surface_size_ok()/ Ville Syrjala
2024-07-09 19:52   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 07/20] drm/i915/fbc: Adjust g4x+ platform checks Ville Syrjala
2024-07-09 19:54   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 08/20] drm/i915/fbc: Extract _intel_fbc_cfb_stride() Ville Syrjala
2024-07-10  8:01   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 09/20] drm/i915/fbc: s/lines/height/ Ville Syrjala
2024-07-09 20:00   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 10/20] drm/i915/fbc: Reoder CFB max height platform checks Ville Syrjala
2024-07-09 20:00   ` Rodrigo Vivi
2024-07-05 14:52 ` [PATCH 11/20] drm/i915/fbc: Extract intel_fbc_max_cfb_height() Ville Syrjala
2024-07-10  8:26   ` Shankar, Uma
2024-07-12 21:34     ` Ville Syrjälä
2024-07-05 14:52 ` [PATCH 12/20] drm/i915/fbc: Extract _intel_fbc_cfb_size() Ville Syrjala
2024-07-10  8:28   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 13/20] drm/i915/fbc: Extract intel_fbc_cfb_cpp() Ville Syrjala
2024-07-10  8:30   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 14/20] drm/i915/fbc: Introduce intel_fbc_preferred_cfb_size() Ville Syrjala
2024-07-10  8:36   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 15/20] drm/xe/fbdev: Fix BIOS FB vs.s stolen size checke Ville Syrjala
2024-07-10  8:42   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 16/20] drm/i915/fbdev: Extract intel_fbdev_fb_prefer_stolen() Ville Syrjala
2024-07-06 12:06   ` kernel test robot
2024-07-09 20:28   ` Lucas De Marchi
2024-07-10 11:47     ` Ville Syrjälä [this message]
2024-07-10  8:51   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 17/20] drm/xe/fbdev: " Ville Syrjala
2024-07-10  8:58   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 18/20] drm/xe/fbdev: Use the same logic for fbdev stolen takever and fresh allocation Ville Syrjala
2024-07-10  9:08   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 19/20] drm/i915/fbdev: Adjust fbdev stolen mem usage heuristic Ville Syrjala
2024-07-10  9:11   ` Shankar, Uma
2024-07-05 14:52 ` [PATCH 20/20] drm/xe/fbdev: " Ville Syrjala
2024-07-10  9:12   ` Shankar, Uma
2024-07-05 15:31 ` ✗ Fi.CI.CHECKPATCH: warning for drm/{i915, xe}: FBC cleanups + tweak fbdev stolen usage Patchwork
2024-07-05 15:31 ` ✗ Fi.CI.SPARSE: " Patchwork
2024-07-05 15:39 ` ✓ Fi.CI.BAT: success " Patchwork
2024-07-06 18:52 ` ✗ Fi.CI.IGT: failure " Patchwork
2024-07-12 21:39 ` [PATCH 00/20] drm/{i915,xe}: " Ville Syrjälä

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=Zo507NB2VMELZAxH@intel.com \
    --to=ville.syrjala@linux.intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=lucas.demarchi@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;
as well as URLs for NNTP newsgroup(s).