drm/i915: Move updating color management to before vblank evasion
Subsystems (2)
PATCH 1/2 drm/i915: Move updating color management to before vblank evasion
This cannot be done reliably during vblank evasasion since the color management registers are not double buffered.
The original commit that moved it always during vblank evasion was wrong, so revert it to before vblank evasion again.
--- drivers/gpu/drm/i915/intel_display.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/intel_display.c +7 -5
@@ -13301,17 +13301,19 @@ static void intel_begin_crtc_commit(struct drm_crtc *crtc,to_intel_atomic_state (old_crtc_state->state);bool modeset =needs_modeset (crtc->state); +if (!modeset && + (intel_cstate->base.color_mgmt_changed || + intel_cstate->update_pipe)) { +intel_color_set_csc (crtc->state); +intel_color_load_luts (crtc->state); + } +/* Perform vblank evasion around commit operation */ intel_pipe_update_start (intel_crtc);if (modeset) goto out; -if (crtc->state->color_mgmt_changed ||to_intel_crtc_state (crtc->state)->update_pipe) { -intel_color_set_csc (crtc->state); -intel_color_load_luts (crtc->state); - } -if (intel_cstate->update_pipe)intel_update_pipe_config (intel_crtc, old_intel_cstate);else if (INTEL_GEN (dev_priv) >=9 )
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 2/2 drm/i915: Complain if we take too long under vblank evasion.
--- drivers/gpu/drm/i915/intel_sprite.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_sprite.c +10 -2
@@ -65,6 +65,8 @@ int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode, 1000 * adjusted_mode->crtc_htotal); } +#define VBLANK_EVASION_TIME_US 100 +/** * intel_pipe_update_start() - start update of a set of display registers * @crtc: the crtc of which the registers are going to be updated
@@ -92,7 +94,8 @@ void intel_pipe_update_start(struct intel_crtc *crtc) vblank_start =DIV_ROUND_UP (vblank_start,2 );/* FIXME needs to be calibrated sensibly */ - min = vblank_start -intel_usecs_to_scanlines (adjusted_mode,100 ); + min = vblank_start -intel_usecs_to_scanlines (adjusted_mode, +VBLANK_EVASION_TIME_US ); max = vblank_start -1 ;local_irq_disable ();
@@ -191,7 +194,12 @@ void intel_pipe_update_end(struct intel_crtc *crtc, struct intel_flip_work *workktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time), crtc->debug.min_vbl, crtc->debug.max_vbl, crtc->debug.scanline_start, scanline_end); - } + }else if (ktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time) > +VBLANK_EVASION_TIME_US ) +DRM_WARN ("Atomic update on pipe (%c) took %lld us, max time under evasion is %u us\n" , +pipe_name (pipe), +ktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time), +VBLANK_EVASION_TIME_US ); } static void
Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/2] drm/i915: Complain if we take too long under vblank evasion.
Quoted 29 lines · Maarten Lankhorst
Signed-off-by Maarten Lankhorst <[email protected]>--- drivers/gpu/drm/i915/intel_sprite.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)drivers/gpu/drm/i915/intel_sprite.c +4
-1@@ -65,6 +65,8 @@ int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode, 1000 * adjusted_mode->crtc_htotal); } +#define VBLANK_EVASION_TIME_US 100 +/** * intel_pipe_update_start() - start update of a set of display registers * @crtc: the crtc of which the registers are going to be updated @@ -92,7 +94,8 @@ void intel_pipe_update_start(struct intel_crtc *crtc) vblank_start =DIV_ROUND_UP (vblank_start,2 );/* FIXME needs to be calibrated sensibly */ - min = vblank_start -intel_usecs_to_scanlines (adjusted_mode,100 ); + min = vblank_start -intel_usecs_to_scanlines (adjusted_mode, +VBLANK_EVASION_TIME_US ); max = vblank_start -1 ;local_irq_disable ();
drivers/gpu/drm/i915/intel_sprite.c void intel_pipe_update_end(struct intel_crtc *crtc, struct intel_flip_work *work · line 202
Quoted 10 lines
drivers/gpu/drm/i915/intel_sprite.c +6
-1@@ -191,7 +194,12 @@ void intel_pipe_update_end(struct intel_crtc *crtc, struct intel_flip_work *workktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time), crtc->debug.min_vbl, crtc->debug.max_vbl, crtc->debug.scanline_start, scanline_end); - } + }else if (ktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time) > +VBLANK_EVASION_TIME_US ) +DRM_WARN ("Atomic update on pipe (%c) took %lld us, max time under evasion is %u us\n" , +pipe_name (pipe), +ktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time), +VBLANK_EVASION_TIME_US );
Hmm. Yeah this should make it very likely that we'll catch any potential fails even when they don't happen to hit the actual critical section. Hopefully it won't result in massive amounts of noise. One way to find out ;)
Quoted 10 lines
drivers/gpu/drm/i915/intel_sprite.c +0
-0} static voidSignature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [PATCH 2/2] drm/i915: Complain if we take too long under vblank evasion.
Op 01-03-17 om 16:40 schreef Ville Syrjälä:
Quoted 45 lines · Maarten Lankhorst
Quoted 29 lines
Signed-off-by Maarten Lankhorst <[email protected]>--- drivers/gpu/drm/i915/intel_sprite.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)drivers/gpu/drm/i915/intel_sprite.c +4
-1@@ -65,6 +65,8 @@ int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode, 1000 * adjusted_mode->crtc_htotal); } +#define VBLANK_EVASION_TIME_US 100 +/** * intel_pipe_update_start() - start update of a set of display registers * @crtc: the crtc of which the registers are going to be updated @@ -92,7 +94,8 @@ void intel_pipe_update_start(struct intel_crtc *crtc) vblank_start =DIV_ROUND_UP (vblank_start,2 );/* FIXME needs to be calibrated sensibly */ - min = vblank_start -intel_usecs_to_scanlines (adjusted_mode,100 ); + min = vblank_start -intel_usecs_to_scanlines (adjusted_mode, +VBLANK_EVASION_TIME_US ); max = vblank_start -1 ;local_irq_disable ();drivers/gpu/drm/i915/intel_sprite.c void intel_pipe_update_end(struct intel_crtc *crtc, struct intel_flip_work *work · line 202
Quoted 10 lines
drivers/gpu/drm/i915/intel_sprite.c +6
-1@@ -191,7 +194,12 @@ void intel_pipe_update_end(struct intel_crtc *crtc, struct intel_flip_work *workktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time), crtc->debug.min_vbl, crtc->debug.max_vbl, crtc->debug.scanline_start, scanline_end); - } + }else if (ktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time) > +VBLANK_EVASION_TIME_US ) +DRM_WARN ("Atomic update on pipe (%c) took %lld us, max time under evasion is %u us\n" , +pipe_name (pipe), +ktime_us_delta (end_vbl_time, crtc->debug.start_vbl_time), +VBLANK_EVASION_TIME_US );Hmm. Yeah this should make it very likely that we'll catch any potential fails even when they don't happen to hit the actual critical section. Hopefully it won't result in massive amounts of noise. One way to find out ;)
Quoted 1 line
Reviewed-by Ville Syrjälä <[email protected]>
Thanks, pushed. Also added a commit message to this commit.
_______________________________________________
Intel-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 9 lines · Maarten Lankhorst
This cannot be done reliably during vblank evasasion since the color management registers are not double buffered.
The original commit that moved it always during vblank evasion was wrong, so revert it to before vblank evasion again.
Signed-off-by Maarten Lankhorst <[email protected]>Fixes 20a34e78f0d7 ("drm/i915: Update color management during vblank evasion.")Cc [email protected] # v4.7+
Wasn't there a bugzilla + tested-by ?
Quoted 40 lines
--- drivers/gpu/drm/i915/intel_display.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-)drivers/gpu/drm/i915/intel_display.c +7
-5@@ -13301,17 +13301,19 @@ static void intel_begin_crtc_commit(struct drm_crtc *crtc,to_intel_atomic_state (old_crtc_state->state);bool modeset =needs_modeset (crtc->state); +if (!modeset && + (intel_cstate->base.color_mgmt_changed || + intel_cstate->update_pipe)) { +intel_color_set_csc (crtc->state); +intel_color_load_luts (crtc->state); + } +/* Perform vblank evasion around commit operation */ intel_pipe_update_start (intel_crtc);if (modeset) goto out; -if (crtc->state->color_mgmt_changed ||to_intel_crtc_state (crtc->state)->update_pipe) { -intel_color_set_csc (crtc->state); -intel_color_load_luts (crtc->state); - } -if (intel_cstate->update_pipe)intel_update_pipe_config (intel_crtc, old_intel_cstate);else if (INTEL_GEN (dev_priv) >=9 )Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Ville Syrjälä
Reviewed-byQuoted 4 lines · Ville Syrjälä
Quoted 3 lines · Maarten Lankhorst
This cannot be done reliably during vblank evasasion since the color management registers are not double buffered.
Quoted 8 lines
Quoted 6 lines
The original commit that moved it always during vblank evasion was wrong, so revert it to before vblank evasion again.
Signed-off-by Maarten Lankhorst <[email protected]>Fixes 20a34e78f0d7 ("drm/i915: Update color management during vblank evasion.")Cc [email protected] # v4.7+Wasn't there a bugzilla + tested-by ?
Seems we have two bugzillas now:
https://bugs.freedesktop.org/show_bug.cgi?id=91883
https://bugs.freedesktop.org/show_bug.cgi?id=99991
First one is about the atomic update fails, the second about the assert_dsi_pll() mutex vs. irq_disabled() warning.
No actual t-bs for this specific variant of the patch (yet) from what I can see.
Anyways, this is
Quoted 49 lines
Quoted 41 lines
--- drivers/gpu/drm/i915/intel_display.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-)drivers/gpu/drm/i915/intel_display.c +7
-5@@ -13301,17 +13301,19 @@ static void intel_begin_crtc_commit(struct drm_crtc *crtc,to_intel_atomic_state (old_crtc_state->state);bool modeset =needs_modeset (crtc->state); +if (!modeset && + (intel_cstate->base.color_mgmt_changed || + intel_cstate->update_pipe)) { +intel_color_set_csc (crtc->state); +intel_color_load_luts (crtc->state); + } +/* Perform vblank evasion around commit operation */ intel_pipe_update_start (intel_crtc);if (modeset) goto out; -if (crtc->state->color_mgmt_changed ||to_intel_crtc_state (crtc->state)->update_pipe) { -intel_color_set_csc (crtc->state); -intel_color_load_luts (crtc->state); - } -if (intel_cstate->update_pipe)intel_update_pipe_config (intel_crtc, old_intel_cstate);else if (INTEL_GEN (dev_priv) >=9 )Signature · 6 lines
2.7.4 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfxSignature · 6 lines
Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✓ Fi.CI.BAT: success for series starting with [1/2] drm/i915: Move updating color management to before vblank evasion
In reply to Maarten Lankhorst== Series Details ==
== Series Details ==
Series: series starting with [1/2] drm/i915: Move updating color management to before vblank evasion
URL : https://patchwork.freedesktop.org/series/20399/
State : success
== Summary ==
Series 20399v1 Series without cover letter
https://patchwork.freedesktop.org/api/1.0/series/20399/revisions/1/mbox/
Test gem_exec_flush:
Subgroup basic-batch-kernel-default-uc:
pass -> FAIL (fi-snb-2600) fdo#100007
Test kms_pipe_crc_basic:
Subgroup suspend-read-crc-pipe-b:
skip -> INCOMPLETE (fi-bsw-n3050) fdo#100004fdo#100007 https://bugs.freedesktop.org/show_bug.cgi?id=100007
fdo#100004 https://bugs.freedesktop.org/show_bug.cgi?id=100004
fi-bdw-5557u total:278 pass:267 dwarn:0 dfail:0 fail:0 skip:11 fi-bsw-n3050 total:236 pass:201 dwarn:0 dfail:0 fail:0 skip:34 fi-bxt-j4205 total:278 pass:259 dwarn:0 dfail:0 fail:0 skip:19 fi-bxt-t5700 total:108 pass:95 dwarn:0 dfail:0 fail:0 skip:12 fi-byt-n2820 total:278 pass:247 dwarn:0 dfail:0 fail:0 skip:31 fi-hsw-4770 total:278 pass:262 dwarn:0 dfail:0 fail:0 skip:16 fi-hsw-4770r total:278 pass:262 dwarn:0 dfail:0 fail:0 skip:16 fi-ilk-650 total:278 pass:228 dwarn:0 dfail:0 fail:0 skip:50 fi-ivb-3520m total:278 pass:260 dwarn:0 dfail:0 fail:0 skip:18 fi-ivb-3770 total:278 pass:260 dwarn:0 dfail:0 fail:0 skip:18 fi-kbl-7500u total:278 pass:259 dwarn:1 dfail:0 fail:0 skip:18 fi-skl-6260u total:278 pass:268 dwarn:0 dfail:0 fail:0 skip:10 fi-skl-6700hq total:278 pass:261 dwarn:0 dfail:0 fail:0 skip:17 fi-skl-6700k total:278 pass:256 dwarn:4 dfail:0 fail:0 skip:18 fi-skl-6770hq total:278 pass:268 dwarn:0 dfail:0 fail:0 skip:10 fi-snb-2520m total:278 pass:250 dwarn:0 dfail:0 fail:0 skip:28 fi-snb-2600 total:278 pass:248 dwarn:0 dfail:0 fail:1 skip:29
fi-byt-j1900 failed to collect. IGT log at Patchwork_4006/fi-byt-j1900/igt.log
5d37006b578e38562382215e8782cfced9c992ce drm-tip: 2017y-02m-28d-16h-27m-13s UTC integration manifest 5bced1e drm/i915: Complain if we take too long under vblank evasion.
628d35b drm/i915: Move updating color management to before vblank evasion
== Logs ==
For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_4006/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx