From: "Zanoni, Paulo R" <paulo.r.zanoni@intel.com>
To: "intel-gfx@lists.freedesktop.org"
<intel-gfx@lists.freedesktop.org>,
"maarten.lankhorst@linux.intel.com"
<maarten.lankhorst@linux.intel.com>
Subject: Re: [PATCH v4.1 6/8] drm/i915: Nuke fbc members from intel_crtc->atomic, v3.
Date: Mon, 29 Feb 2016 21:06:27 +0000 [thread overview]
Message-ID: <1456779986.2408.23.camel@intel.com> (raw)
In-Reply-To: <56D4164F.4080205@linux.intel.com>
Em Seg, 2016-02-29 às 10:58 +0100, Maarten Lankhorst escreveu:
> Whenever there's an update to the primary plane,
> fbc_pre_update and fbc_post_update are called. Kill off
> intel_crtc->atomic.update_fbc and now that intel_crtc->atomic
> is empty, kill it off too.
>
> Changes since v1:
> - Add a intel_fbc_supports_rotation helper.
> Changes sinec v2:
> - Remove intel_fbc_supports_rotation_helper.
> - Remove unrelated changes.
>
> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> ---
> This version doesn't change intel_fbc_enable any more!
Ok, so I decided to test this.
First of all, patch 05 needs a rebase too. I did my own local rebase
and I suppose I got things right.
Second, after I did actual testing, I discovered that, contrary to my
previous belief, "old_pri_state" will also be valid when we're only
changing the sprite or cursor planes. So if we apply this patch, we'll
be calling pre_update+enable+post_update for every plane/cursor
enable/disable operation, which is not what we want since these are
unnecessary calls, as FBC only deals with the primary plane. I think we
need to fix this.
The tests I've been using to check this were:
- kms_frontbuffer_tracking/fbc-1p-primscrn-cur-indfb-onoff
- kms_frontbuffer_tracking/fbc-1p-primscrn-spr-indfb-onoff
>
> diff --git a/drivers/gpu/drm/i915/intel_display.c
> b/drivers/gpu/drm/i915/intel_display.c
> index 78f72bef25aa..c8404092aee9 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -4789,11 +4789,9 @@ static void intel_post_plane_update(struct
> intel_crtc_state *old_crtc_state)
> {
> struct intel_crtc *crtc = to_intel_crtc(old_crtc_state-
> >base.crtc);
> struct drm_atomic_state *old_state = old_crtc_state-
> >base.state;
> - struct intel_crtc_atomic_commit *atomic = &crtc->atomic;
> struct intel_crtc_state *pipe_config =
> to_intel_crtc_state(crtc->base.state);
> struct drm_device *dev = crtc->base.dev;
> - struct drm_i915_private *dev_priv = dev->dev_private;
> struct drm_plane *primary = crtc->base.primary;
> struct drm_plane_state *old_pri_state =
> drm_atomic_get_existing_plane_state(old_state,
> primary);
> @@ -4805,22 +4803,19 @@ static void intel_post_plane_update(struct
> intel_crtc_state *old_crtc_state)
> if (pipe_config->wm_changed && pipe_config->base.active)
> intel_update_watermarks(&crtc->base);
>
> - if (atomic->update_fbc)
> - intel_fbc_post_update(crtc);
> -
> if (old_pri_state) {
> struct intel_plane_state *primary_state =
> to_intel_plane_state(primary->state);
> struct intel_plane_state *old_primary_state =
> to_intel_plane_state(old_pri_state);
>
> + intel_fbc_post_update(crtc);
> +
> if (primary_state->visible &&
> (needs_modeset(&pipe_config->base) ||
> !old_primary_state->visible))
> intel_post_enable_primary(&crtc->base);
> }
> -
> - memset(atomic, 0, sizeof(*atomic));
> }
>
> static void intel_pre_plane_update(struct intel_crtc_state
> *old_crtc_state)
> @@ -4828,7 +4823,6 @@ static void intel_pre_plane_update(struct
> intel_crtc_state *old_crtc_state)
> struct intel_crtc *crtc = to_intel_crtc(old_crtc_state-
> >base.crtc);
> struct drm_device *dev = crtc->base.dev;
> struct drm_i915_private *dev_priv = dev->dev_private;
> - struct intel_crtc_atomic_commit *atomic = &crtc->atomic;
> struct intel_crtc_state *pipe_config =
> to_intel_crtc_state(crtc->base.state);
> struct drm_atomic_state *old_state = old_crtc_state-
> >base.state;
> @@ -4837,15 +4831,14 @@ static void intel_pre_plane_update(struct
> intel_crtc_state *old_crtc_state)
> drm_atomic_get_existing_plane_state(old_state,
> primary);
> bool modeset = needs_modeset(&pipe_config->base);
>
> - if (atomic->update_fbc)
> - intel_fbc_pre_update(crtc);
> -
> if (old_pri_state) {
> struct intel_plane_state *primary_state =
> to_intel_plane_state(primary->state);
> struct intel_plane_state *old_primary_state =
> to_intel_plane_state(old_pri_state);
>
> + intel_fbc_pre_update(crtc);
> +
> if (old_primary_state->visible &&
> (modeset || !primary_state->visible))
> intel_pre_disable_primary(&crtc->base);
> @@ -11877,27 +11870,17 @@ int intel_plane_atomic_calc_changes(struct
> drm_crtc_state *crtc_state,
> if (visible || was_visible)
> pipe_config->fb_bits |= to_intel_plane(plane)-
> >frontbuffer_bit;
>
> - switch (plane->type) {
> - case DRM_PLANE_TYPE_PRIMARY:
> - intel_crtc->atomic.update_fbc = true;
> -
> - break;
> - case DRM_PLANE_TYPE_CURSOR:
> - break;
> - case DRM_PLANE_TYPE_OVERLAY:
> - /*
> - * WaCxSRDisabledForSpriteScaling:ivb
> - *
> - * cstate->update_wm was already set above, so this
> flag will
> - * take effect when we commit and program
> watermarks.
> - */
> - if (IS_IVYBRIDGE(dev) &&
> - needs_scaling(to_intel_plane_state(plane_state))
> &&
> - !needs_scaling(old_plane_state))
> - pipe_config->disable_lp_wm = true;
> + /*
> + * WaCxSRDisabledForSpriteScaling:ivb
> + *
> + * cstate->update_wm was already set above, so this flag
> will
> + * take effect when we commit and program watermarks.
> + */
> + if (plane->type == DRM_PLANE_TYPE_OVERLAY &&
> IS_IVYBRIDGE(dev) &&
> + needs_scaling(to_intel_plane_state(plane_state)) &&
> + !needs_scaling(old_plane_state))
> + pipe_config->disable_lp_wm = true;
>
> - break;
> - }
> return 0;
> }
>
> @@ -13296,9 +13279,6 @@ static int intel_atomic_check(struct
> drm_device *dev,
> struct intel_crtc_state *pipe_config =
> to_intel_crtc_state(crtc_state);
>
> - memset(&to_intel_crtc(crtc)->atomic, 0,
> - sizeof(struct intel_crtc_atomic_commit));
> -
> /* Catch I915_MODE_FLAG_INHERITED */
> if (crtc_state->mode.private_flags != crtc->state-
> >mode.private_flags)
> crtc_state->mode_changed = true;
> @@ -13606,7 +13586,8 @@ static int intel_atomic_commit(struct
> drm_device *dev,
> if (!modeset)
> intel_pre_plane_update(to_intel_crtc_state(c
> rtc_state));
>
> - if (crtc->state->active && intel_crtc-
> >atomic.update_fbc)
> + if (crtc->state->active &&
> + drm_atomic_get_existing_plane_state(state, crtc-
> >primary))
> intel_fbc_enable(intel_crtc);
>
> if (crtc->state->active &&
> diff --git a/drivers/gpu/drm/i915/intel_drv.h
> b/drivers/gpu/drm/i915/intel_drv.h
> index 058bb10e53ae..5e5b64fff30d 100644
> --- a/drivers/gpu/drm/i915/intel_drv.h
> +++ b/drivers/gpu/drm/i915/intel_drv.h
> @@ -538,19 +538,6 @@ struct intel_mmio_flip {
> unsigned int rotation;
> };
>
> -/*
> - * Tracking of operations that need to be performed at the
> beginning/end of an
> - * atomic commit, outside the atomic section where interrupts are
> disabled.
> - * These are generally operations that grab mutexes or might
> otherwise sleep
> - * and thus can't be run with interrupts disabled.
> - */
> -struct intel_crtc_atomic_commit {
> - /* Sleepable operations to perform before commit */
> -
> - /* Sleepable operations to perform after commit */
> - bool update_fbc;
> -};
> -
> struct intel_crtc {
> struct drm_crtc base;
> enum pipe pipe;
> @@ -610,8 +597,6 @@ struct intel_crtc {
> int scanline_start;
> } debug;
>
> - struct intel_crtc_atomic_commit atomic;
> -
> /* scalers available on this crtc */
> int num_scalers;
>
>
> _______________________________________________
> 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
next prev parent reply other threads:[~2016-02-29 21:06 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-10 12:49 [PATCH v4 0/8] Kill off intel_crtc->atomic! Maarten Lankhorst
2016-02-10 12:49 ` [PATCH v4 1/8] drm/i915: Pass crtc state to modeset_get_crtc_power_domains Maarten Lankhorst
2016-02-17 17:54 ` Zanoni, Paulo R
2016-02-18 9:51 ` Maarten Lankhorst
2016-02-18 13:08 ` Zanoni, Paulo R
2016-02-18 13:21 ` Maarten Lankhorst
2016-02-10 12:49 ` [PATCH v4 2/8] drm/i915: Unify power domain handling Maarten Lankhorst
2016-02-17 19:54 ` Zanoni, Paulo R
2016-02-10 12:49 ` [PATCH v4 3/8] drm/i915: Kill off intel_crtc->atomic.wait_vblank, v4 Maarten Lankhorst
2016-02-17 21:20 ` Zanoni, Paulo R
2016-02-18 13:22 ` Maarten Lankhorst
2016-02-18 14:14 ` Zanoni, Paulo R
2016-02-18 14:46 ` Maarten Lankhorst
2016-02-18 17:02 ` Zanoni, Paulo R
2016-02-24 10:24 ` [PATCH v6 3/8] drm/i915: Kill off intel_crtc->atomic.wait_vblank, v6 Maarten Lankhorst
2016-02-24 14:50 ` Zanoni, Paulo R
2016-02-10 12:49 ` [PATCH v4 4/8] drm/i915: Remove update_sprite_watermarks Maarten Lankhorst
2016-02-10 12:49 ` [PATCH v4 5/8] drm/i915: Remove some post-commit members from intel_crtc->atomic, v2 Maarten Lankhorst
2016-02-10 12:49 ` [PATCH v4 6/8] drm/i915: Nuke fbc " Maarten Lankhorst
2016-02-12 13:56 ` Zanoni, Paulo R
2016-02-15 14:31 ` Maarten Lankhorst
2016-02-18 17:17 ` Zanoni, Paulo R
2016-02-29 9:58 ` [PATCH v4.1 6/8] drm/i915: Nuke fbc members from intel_crtc->atomic, v3 Maarten Lankhorst
2016-02-29 21:06 ` Zanoni, Paulo R [this message]
2016-03-01 8:27 ` Maarten Lankhorst
2016-02-10 12:49 ` [PATCH v4 7/8] drm/i915: Do not compute watermarks on a noop Maarten Lankhorst
2016-02-18 20:51 ` Zanoni, Paulo R
2016-03-01 10:11 ` Maarten Lankhorst
2016-02-10 12:49 ` [PATCH v4 8/8] drm/i915: Remove vblank wait from hsw_enable_ips Maarten Lankhorst
2016-02-12 12:06 ` Zanoni, Paulo R
2016-02-16 10:32 ` Maarten Lankhorst
2016-03-23 13:33 ` [PATCH v5 8/8] drm/i915: Remove vblank wait from hsw_enable_ips, v2 Maarten Lankhorst
2016-03-23 14:43 ` Zanoni, Paulo R
2016-02-15 13:35 ` ✗ Fi.CI.BAT: failure for Kill off intel_crtc->atomic! (rev8) 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=1456779986.2408.23.camel@intel.com \
--to=paulo.r.zanoni@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=maarten.lankhorst@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.