Skip to contentSkip to conversation

drm/i915/skl: Increase ddb blocks to support large cursor sizes

PATCH In review Radhakrishna Sripada 5 participants 8 messages Dec 18, 2015 – Dec 22, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 drm/i915/skl: Increase ddb blocks to support large cursor sizes

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

---
 drivers/gpu/drm/i915/intel_pm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +2 -2
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
 static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
 {
 	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
-	return 8;
+	return 16;
 }
 
 static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw
Quoted 10 lines · Radhakrishna Sripada

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

Quoted 8 lines

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

Can you try my patch "drm/i915/skl: Use proper plane dimensions for DDB and WM calculations" first and see if it solves your problems? It's possible that your problems were caused by my fumble that I'm fixing with that patch; if that patch fixes the problem, then we may not need to alter the cursor's fixed allocation after all.

Of course as we noted, there are a handful of new workarounds in the bspec (SAGV disabling, system memory bandwidth checking, etc.) that we still don't have in our driver yet, so if any of those are the true root cause, then your patch here is probably a reasonable short-term fix until we can take action on the new workarounds. In that case, you can consider this patch

Unfortunately I only have a BXT which doesn't exhibit the problems you're seeing, so I can't actually test this out myself.

Matt

Quoted 23 lines
---
 drivers/gpu/drm/i915/intel_pm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +2 -2
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
 static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
 {
 	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
-	return 8;
+	return 16;
 }
 
 static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
Signature · 1 line
1.9.1
Signature · 9 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
Quoted 14 lines · Matt Roper
Quoted 13 lines · Radhakrishna Sripada

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

Quoted 8 lines
Quoted 5 lines

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

Can you try my patch "drm/i915/skl: Use proper plane dimensions for DDB and WM calculations" first and see if it solves your problems? It's

Where can I find this patch ? Couldn't find in drm-intel branches or the mailing list.

Regards
Shobhit

Quoted 34 lines

possible that your problems were caused by my fumble that I'm fixing with that patch; if that patch fixes the problem, then we may not need to alter the cursor's fixed allocation after all.

Of course as we noted, there are a handful of new workarounds in the bspec (SAGV disabling, system memory bandwidth checking, etc.) that we still don't have in our driver yet, so if any of those are the true root cause, then your patch here is probably a reasonable short-term fix until we can take action on the new workarounds. In that case, you can consider this patch

Unfortunately I only have a BXT which doesn't exhibit the problems you're seeing, so I can't actually test this out myself.

Matt

Quoted 15 lines
---
  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +1 -1
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
  static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
  {
  	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
Quoted 9 lines
Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +1 -2
-	return 8;
+	return 16;
  }
 
  static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
--

1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw
Quoted 18 lines · Shobhit
Quoted 17 lines · Matt Roper
Quoted 16 lines · Radhakrishna Sripada

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Can you try my patch "drm/i915/skl: Use proper plane dimensions for DDB and WM calculations" first and see if it solves your problems? It's

Where can I find this patch ? Couldn't find in drm-intel branches or the mailing list.

Yeah, looks like I mistyped the mailing list address for git-send-email; just resent with the proper address...you can find it here:

        http://patchwork.freedesktop.org/patch/68786/

Matt

Quoted 47 lines

Regards
Shobhit

Quoted 43 lines

possible that your problems were caused by my fumble that I'm fixing with that patch; if that patch fixes the problem, then we may not need to alter the cursor's fixed allocation after all.

Of course as we noted, there are a handful of new workarounds in the bspec (SAGV disabling, system memory bandwidth checking, etc.) that we still don't have in our driver yet, so if any of those are the true root cause, then your patch here is probably a reasonable short-term fix until we can take action on the new workarounds. In that case, you can consider this patch

Unfortunately I only have a BXT which doesn't exhibit the problems you're seeing, so I can't actually test this out myself.

Matt

Quoted 23 lines
---
 drivers/gpu/drm/i915/intel_pm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +2 -3
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
 static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
 {
 	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
-	return 8;
+	return 16;
 }
 
 static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
--

1.9.1

Signature · 9 lines
Matt Roper
Graphics Software Engineer
IoTG Platform Enabling & Development
Intel Corporation
(916) 356-2795
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw
Quoted 25 lines · Matt Roper
Quoted 24 lines · Shobhit
Quoted 22 lines · Matt Roper
Quoted 10 lines · Radhakrishna Sripada

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

Quoted 8 lines

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

Can you try my patch "drm/i915/skl: Use proper plane dimensions for DDB and WM calculations" first and see if it solves your problems? It's

Quoted 8 lines
Quoted 2 lines

Where can I find this patch ? Couldn't find in drm-intel branches or the mailing list.

Yeah, looks like I mistyped the mailing list address for git-send-email; just resent with the proper address...you can find it here:

         http://patchwork.freedesktop.org/patch/68786/

Thanks Matt. I tried this patch against this issue reverting the cursor DDB bump up from Sripada and I could still reproduce the blanking/unblanking on external display on cursor move to the external display. Tested with a 4k DP monitor.

Didn't have a high res eDP panel handy right away to test single display with hardware cursor.

Regards
Shobhit

Quoted 51 lines

Matt

Quoted 47 lines

Regards
Shobhit

Quoted 43 lines

possible that your problems were caused by my fumble that I'm fixing with that patch; if that patch fixes the problem, then we may not need to alter the cursor's fixed allocation after all.

Of course as we noted, there are a handful of new workarounds in the bspec (SAGV disabling, system memory bandwidth checking, etc.) that we still don't have in our driver yet, so if any of those are the true root cause, then your patch here is probably a reasonable short-term fix until we can take action on the new workarounds. In that case, you can consider this patch

Unfortunately I only have a BXT which doesn't exhibit the problems you're seeing, so I can't actually test this out myself.

Matt

Quoted 23 lines
---
  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +2 -3
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
  static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
  {
  	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
-	return 8;
+	return 16;
  }
 
  static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
--

1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#5 lore raw

✗ warning: Fi.CI.BAT

In reply to Radhakrishna Sripada
== Summary ==

== Summary ==

Built on 7cdc548e77f503593b83a1c5d58f4dcc862c17e2 drm-intel-nightly: 2015y-12m-18d-19h-26m-21s UTC integration manifest

Test gem_storedw_loop:

        Subgroup basic-render:
                pass       -> DMESG-WARN (skl-i5k-2)
Test kms_flip:
        Subgroup basic-flip-vs-modeset:
                dmesg-warn -> PASS       (hsw-brixbox)
                pass       -> DMESG-WARN (skl-i7k-2)
                dmesg-warn -> PASS       (byt-nuc)
Test kms_pipe_crc_basic:
        Subgroup read-crc-pipe-a-frame-sequence:
                dmesg-warn -> PASS       (byt-nuc)
        Subgroup read-crc-pipe-b-frame-sequence:
                pass       -> DMESG-WARN (byt-nuc)
        Subgroup read-crc-pipe-c:
                dmesg-warn -> PASS       (skl-i7k-2)
Test kms_psr_sink_crc:
        Subgroup psr_basic:
                pass       -> DMESG-WARN (bdw-ultra)
Test kms_setmode:
        Subgroup basic-clone-single-crtc:
                dmesg-warn -> PASS       (snb-dellxps)
Test pm_rpm:
        Subgroup basic-pci-d3-state:
                dmesg-warn -> PASS       (bdw-ultra)
bdw-nuci7        total:135  pass:125  dwarn:1   dfail:0   fail:0   skip:9  
bdw-ultra        total:132  pass:124  dwarn:2   dfail:0   fail:0   skip:6  
byt-nuc          total:135  pass:120  dwarn:2   dfail:0   fail:0   skip:13 
hsw-brixbox      total:135  pass:127  dwarn:1   dfail:0   fail:0   skip:7  
hsw-gt2          total:135  pass:130  dwarn:1   dfail:0   fail:0   skip:4  
ilk-hp8440p      total:135  pass:100  dwarn:0   dfail:0   fail:0   skip:35 
ivb-t430s        total:135  pass:127  dwarn:2   dfail:0   fail:0   skip:6  
skl-i5k-2        total:135  pass:121  dwarn:6   dfail:0   fail:0   skip:8  
skl-i7k-2        total:135  pass:122  dwarn:5   dfail:0   fail:0   skip:8  
snb-dellxps      total:135  pass:122  dwarn:1   dfail:0   fail:0   skip:12 
snb-x220t        total:135  pass:121  dwarn:2   dfail:0   fail:1   skip:11 

Results at /archive/results/CI_IGT_test/Patchwork_728/

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#6 lore raw
In reply to Radhakrishna Sripada
Quoted 8 lines · Radhakrishna Sripada

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700
Quoted 7 lines

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

This does fix the blanking/unblanking in case of external 4k@60 DP monitor. Will do some test on the patch suggested by Matt in the other mail.

Regards
Shobhit

Quoted 14 lines
---
  drivers/gpu/drm/i915/intel_pm.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
  static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)

drivers/gpu/drm/i915/intel_pm.c skl_ddb_get_pipe_allocation_limits(struct drm_device *dev, · line 2810

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +2 -2
  {
  	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
-	return 8;
+	return 16;
  }
 
  static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

maintainer
#7 lore raw
In reply to Radhakrishna Sripada
Quoted 8 lines · Radhakrishna Sripada

Original value of 32 blocks is not sufficient when using cursor size of 256x256 causing FIFO underruns when the reworked wm caluclations in

commit 024c9045221fe45482863c47c4b4c47d37f97cbf

Author: Matt Roper <[email protected]>
Date:   Thu Sep 24 15:53:11 2015 -0700
Quoted 8 lines

drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v4)

are used. Increasing the number of blocks to 52 to make cursor plane tolerate SAGV block time for the maximum possible cursor size.

v2: Included Matt's suggestion, bumping up the no of blocks in multi-pipe case to 16.

Sorry the documentation isn't there yet (working on it), but please always addd

in the commit message when you're sending a patch there.

BR,
Jani.

Quoted 22 lines
---
 drivers/gpu/drm/i915/intel_pm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +2 -2
index d385d99..c5ba4e5 100644
@@ -2802,9 +2802,9 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
 static unsigned int skl_cursor_allocation(const struct intel_wm_config *config)
 {
 	if (config->num_pipes_active == 1)
-		return 32;
+		return 52;
 
-	return 8;
+	return 16;
 }
 
 static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx