[PATCH 00/15] drm/vc4: Misc fixes and improvements for the HVS and CRTCs

[PATCH 00/15] drm/vc4: Misc fixes and improvements for the HVS and CRTCs

[Maxime Ripard]

Hi,

Here's a collection of patches for the HVS and CRTCs that fix a few
issues (with DMA accesses and timings, mostly), and add a few minor
features.

Let me know what you think,
Maxime

To: Emma Anholt <emma@anholt.net>
To: Maxime Ripard <mripard@kernel.org>
To: David Airlie <airlied@gmail.com>
To: Daniel Vetter <daniel@ffwll.ch>
To: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Cc: linux-kernel@vger.kernel.org
Cc: Dave Stevenson <dave.stevenson@raspberrypi.com>
Cc: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>

---
Dave Stevenson (12):
      drm/vc4: hvs: Configure the HVS COB allocations
      drm/vc4: hvs: Set AXI panic modes
      drm/vc4: hvs: SCALER_DISPBKGND_AUTOHS is only valid on HVS4
      drm/vc4: hvs: Correct interrupt masking bit assignment for HVS5
      drm/vc4: hvs: Support zpos on all planes
      drm/vc4: hvs: Fix colour order for xRGB1555 on HVS5
      drm/vc4: hvs: Add DRM 210101010 RGB formats
      drm/vc4: plane: Allow using 0 as a pixel order value
      drm/vc4: plane: Omit pixel_order from the hvs_format for hvs5 only formats
      drm/vc4: plane: Add 3:3:2 and 4:4:4:4 RGB/RGBX/RGBA formats
      drm/vc4: Add comments for which HVS_PIXEL_ORDER_xxx defines apply
      drm/vc4: hdmi: Correct interlaced timings again

Mateusz Kwiatkowski (2):
      drm/vc4: crtc: Fix timings for VEC modes
      drm/vc4: vec: Support progressive modes

Maxime Ripard (1):
      drm/vc4: hvs: Ignore atomic_flush if we're disabled

 drivers/gpu/drm/vc4/vc4_crtc.c  |  71 +++++++++-------
 drivers/gpu/drm/vc4/vc4_hdmi.c  |   5 +-
 drivers/gpu/drm/vc4/vc4_hvs.c   | 175 +++++++++++++++++++++++++++++++---------
 drivers/gpu/drm/vc4/vc4_kms.c   |   1 +
 drivers/gpu/drm/vc4/vc4_plane.c | 139 ++++++++++++++++++++++++++++---
 drivers/gpu/drm/vc4/vc4_regs.h  |  20 ++++-
 drivers/gpu/drm/vc4/vc4_vec.c   |   4 +-
 7 files changed, 337 insertions(+), 78 deletions(-)
---
base-commit: 99e2d98adc738597abcc5d38b03d0e9858db5c00
change-id: 20221207-rpi-hvs-crtc-misc-d97190d6a622

Best regards,
-- 
Maxime Ripard <maxime@cerno.tech>

[PATCH 01/15] drm/vc4: hvs: Configure the HVS COB allocations

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

The HVS Composite Output Buffer (COB) is the memory used to
generate the output pixel data.
Until now the vc4 driver has been relying on the firmware to
have set these to sensible values.

In testing triple screen support it has been noted that only
1 line was being assigned to HVS channel 2. Whilst that is fine
for the transposer (TXP), and indeed needed as only some pixels
have an alpha channel, it is insufficient to run a live display.

Split the COB more evenly between the 3 HVS channels.

Fixes: c54619b0bfb3 ("drm/vc4: Add support for the BCM2711 HVS5")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hvs.c | 56 ++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 55 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index c4453a5ae163..d615ba7db920 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -776,7 +776,7 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 	struct vc4_hvs *hvs = NULL;
 	int ret;
 	u32 dispctrl;
-	u32 reg;
+	u32 reg, top;
 
 	hvs = drmm_kzalloc(drm, sizeof(*hvs), GFP_KERNEL);
 	if (!hvs)
@@ -912,6 +912,60 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 
 	HVS_WRITE(SCALER_DISPCTRL, dispctrl);
 
+	/* Recompute Composite Output Buffer (COB) allocations for the displays
+	 */
+	if (!vc4->is_vc5) {
+		/* The COB is 20736 pixels, or just over 10 lines at 2048 wide.
+		 * The bottom 2048 pixels are full 32bpp RGBA (intended for the
+		 * TXP composing RGBA to memory), whilst the remainder are only
+		 * 24bpp RGB.
+		 *
+		 * Assign 3 lines to channels 1 & 2, and just over 4 lines to
+		 * channel 0.
+		 */
+		#define VC4_COB_SIZE		20736
+		#define VC4_COB_LINE_WIDTH	2048
+		#define VC4_COB_NUM_LINES	3
+		reg = 0;
+		top = VC4_COB_LINE_WIDTH * VC4_COB_NUM_LINES;
+		reg |= (top - 1) << 16;
+		HVS_WRITE(SCALER_DISPBASE2, reg);
+		reg = top;
+		top += VC4_COB_LINE_WIDTH * VC4_COB_NUM_LINES;
+		reg |= (top - 1) << 16;
+		HVS_WRITE(SCALER_DISPBASE1, reg);
+		reg = top;
+		top = VC4_COB_SIZE;
+		reg |= (top - 1) << 16;
+		HVS_WRITE(SCALER_DISPBASE0, reg);
+	} else {
+		/* The COB is 44416 pixels, or 10.8 lines at 4096 wide.
+		 * The bottom 4096 pixels are full RGBA (intended for the TXP
+		 * composing RGBA to memory), whilst the remainder are only
+		 * RGB. Addressing is always pixel wide.
+		 *
+		 * Assign 3 lines of 4096 to channels 1 & 2, and just over 4
+		 * lines. to channel 0.
+		 */
+		#define VC5_COB_SIZE		44416
+		#define VC5_COB_LINE_WIDTH	4096
+		#define VC5_COB_NUM_LINES	3
+		reg = 0;
+		top = VC5_COB_LINE_WIDTH * VC5_COB_NUM_LINES;
+		reg |= top << 16;
+		HVS_WRITE(SCALER_DISPBASE2, reg);
+		top += 16;
+		reg = top;
+		top += VC5_COB_LINE_WIDTH * VC5_COB_NUM_LINES;
+		reg |= top << 16;
+		HVS_WRITE(SCALER_DISPBASE1, reg);
+		top += 16;
+		reg = top;
+		top = VC5_COB_SIZE;
+		reg |= top << 16;
+		HVS_WRITE(SCALER_DISPBASE0, reg);
+	}
+
 	ret = devm_request_irq(dev, platform_get_irq(pdev, 0),
 			       vc4_hvs_irq_handler, 0, "vc4 hvs", drm);
 	if (ret)

-- 
2.38.1

[PATCH 02/15] drm/vc4: hvs: Set AXI panic modes

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

The HVS can change AXI request mode based on how full the COB
FIFOs are.
Until now the vc4 driver has been relying on the firmware to
have set these to sensible values.

With HVS channel 2 now being used for live video, change the
panic mode for all channels to be explicitly set by the driver,
and the same for all channels.

Fixes: c54619b0bfb3 ("drm/vc4: Add support for the BCM2711 HVS5")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hvs.c  | 11 +++++++++++
 drivers/gpu/drm/vc4/vc4_regs.h |  6 ++++++
 2 files changed, 17 insertions(+)

diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index d615ba7db920..b335815eac6a 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -910,6 +910,17 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 		      SCALER_DISPCTRL_DSPEISLUR(2) |
 		      SCALER_DISPCTRL_SCLEIRQ);
 
+	/* Set AXI panic mode.
+	 * VC4 panics when < 2 lines in FIFO.
+	 * VC5 panics when less than 1 line in the FIFO.
+	 */
+	dispctrl &= ~(SCALER_DISPCTRL_PANIC0_MASK |
+		      SCALER_DISPCTRL_PANIC1_MASK |
+		      SCALER_DISPCTRL_PANIC2_MASK);
+	dispctrl |= VC4_SET_FIELD(2, SCALER_DISPCTRL_PANIC0);
+	dispctrl |= VC4_SET_FIELD(2, SCALER_DISPCTRL_PANIC1);
+	dispctrl |= VC4_SET_FIELD(2, SCALER_DISPCTRL_PANIC2);
+
 	HVS_WRITE(SCALER_DISPCTRL, dispctrl);
 
 	/* Recompute Composite Output Buffer (COB) allocations for the displays
diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index f0290fad991d..f121905c404d 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -220,6 +220,12 @@
 #define SCALER_DISPCTRL                         0x00000000
 /* Global register for clock gating the HVS */
 # define SCALER_DISPCTRL_ENABLE			BIT(31)
+# define SCALER_DISPCTRL_PANIC0_MASK		VC4_MASK(25, 24)
+# define SCALER_DISPCTRL_PANIC0_SHIFT		24
+# define SCALER_DISPCTRL_PANIC1_MASK		VC4_MASK(27, 26)
+# define SCALER_DISPCTRL_PANIC1_SHIFT		26
+# define SCALER_DISPCTRL_PANIC2_MASK		VC4_MASK(29, 28)
+# define SCALER_DISPCTRL_PANIC2_SHIFT		28
 # define SCALER_DISPCTRL_DSP3_MUX_MASK		VC4_MASK(19, 18)
 # define SCALER_DISPCTRL_DSP3_MUX_SHIFT		18
 

-- 
2.38.1

[PATCH 03/15] drm/vc4: hvs: SCALER_DISPBKGND_AUTOHS is only valid on HVS4

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

The bit used for SCALER_DISPBKGND_AUTOHS in SCALER_DISPBKGNDX
has been repurposed on HVS5 to configure whether a display can
win back-to-back arbitration wins for the COB.

This is not desirable, therefore only select this bit on HVS4,
and explicitly clear it on HVS5.

Fixes: c54619b0bfb3 ("drm/vc4: Add support for the BCM2711 HVS5")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hvs.c  | 10 ++++++----
 drivers/gpu/drm/vc4/vc4_regs.h |  1 +
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index b335815eac6a..57d99e7199ee 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -370,28 +370,30 @@ static int vc4_hvs_init_channel(struct vc4_hvs *hvs, struct drm_crtc *crtc,
 	 * mode.
 	 */
 	dispctrl = SCALER_DISPCTRLX_ENABLE;
+	dispbkgndx = HVS_READ(SCALER_DISPBKGNDX(chan));
 
-	if (!vc4->is_vc5)
+	if (!vc4->is_vc5) {
 		dispctrl |= VC4_SET_FIELD(mode->hdisplay,
 					  SCALER_DISPCTRLX_WIDTH) |
 			    VC4_SET_FIELD(mode->vdisplay,
 					  SCALER_DISPCTRLX_HEIGHT) |
 			    (oneshot ? SCALER_DISPCTRLX_ONESHOT : 0);
-	else
+		dispbkgndx |= SCALER_DISPBKGND_AUTOHS;
+	} else {
 		dispctrl |= VC4_SET_FIELD(mode->hdisplay,
 					  SCALER5_DISPCTRLX_WIDTH) |
 			    VC4_SET_FIELD(mode->vdisplay,
 					  SCALER5_DISPCTRLX_HEIGHT) |
 			    (oneshot ? SCALER5_DISPCTRLX_ONESHOT : 0);
+		dispbkgndx &= ~SCALER5_DISPBKGND_BCK2BCK;
+	}
 
 	HVS_WRITE(SCALER_DISPCTRLX(chan), dispctrl);
 
-	dispbkgndx = HVS_READ(SCALER_DISPBKGNDX(chan));
 	dispbkgndx &= ~SCALER_DISPBKGND_GAMMA;
 	dispbkgndx &= ~SCALER_DISPBKGND_INTERLACE;
 
 	HVS_WRITE(SCALER_DISPBKGNDX(chan), dispbkgndx |
-		  SCALER_DISPBKGND_AUTOHS |
 		  ((!vc4->is_vc5) ? SCALER_DISPBKGND_GAMMA : 0) |
 		  (interlace ? SCALER_DISPBKGND_INTERLACE : 0));
 
diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index f121905c404d..95deacdc31e7 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -366,6 +366,7 @@
 
 #define SCALER_DISPBKGND0                       0x00000044
 # define SCALER_DISPBKGND_AUTOHS		BIT(31)
+# define SCALER5_DISPBKGND_BCK2BCK		BIT(31)
 # define SCALER_DISPBKGND_INTERLACE		BIT(30)
 # define SCALER_DISPBKGND_GAMMA			BIT(29)
 # define SCALER_DISPBKGND_TESTMODE_MASK		VC4_MASK(28, 25)

-- 
2.38.1

[PATCH 04/15] drm/vc4: hvs: Correct interrupt masking bit assignment for HVS5

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

HVS5 has moved the interrupt enable bits around within the
DISPCTRL register, therefore the configuration has to be updated
to account for this.

Fixes: c54619b0bfb3 ("drm/vc4: Add support for the BCM2711 HVS5")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hvs.c  | 52 +++++++++++++++++++++++++++++-------------
 drivers/gpu/drm/vc4/vc4_regs.h | 10 ++++++--
 2 files changed, 44 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index 57d99e7199ee..d9fc0d03023b 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -660,7 +660,8 @@ void vc4_hvs_mask_underrun(struct vc4_hvs *hvs, int channel)
 		return;
 
 	dispctrl = HVS_READ(SCALER_DISPCTRL);
-	dispctrl &= ~SCALER_DISPCTRL_DSPEISLUR(channel);
+	dispctrl &= ~(hvs->vc4->is_vc5 ? SCALER5_DISPCTRL_DSPEISLUR(channel) :
+					 SCALER_DISPCTRL_DSPEISLUR(channel));
 
 	HVS_WRITE(SCALER_DISPCTRL, dispctrl);
 
@@ -677,7 +678,8 @@ void vc4_hvs_unmask_underrun(struct vc4_hvs *hvs, int channel)
 		return;
 
 	dispctrl = HVS_READ(SCALER_DISPCTRL);
-	dispctrl |= SCALER_DISPCTRL_DSPEISLUR(channel);
+	dispctrl |= (hvs->vc4->is_vc5 ? SCALER5_DISPCTRL_DSPEISLUR(channel) :
+					SCALER_DISPCTRL_DSPEISLUR(channel));
 
 	HVS_WRITE(SCALER_DISPSTAT,
 		  SCALER_DISPSTAT_EUFLOW(channel));
@@ -703,6 +705,7 @@ static irqreturn_t vc4_hvs_irq_handler(int irq, void *data)
 	int channel;
 	u32 control;
 	u32 status;
+	u32 dspeislur;
 
 	/*
 	 * NOTE: We don't need to protect the register access using
@@ -719,9 +722,11 @@ static irqreturn_t vc4_hvs_irq_handler(int irq, void *data)
 	control = HVS_READ(SCALER_DISPCTRL);
 
 	for (channel = 0; channel < SCALER_CHANNELS_COUNT; channel++) {
+		dspeislur = vc4->is_vc5 ? SCALER5_DISPCTRL_DSPEISLUR(channel) :
+					  SCALER_DISPCTRL_DSPEISLUR(channel);
 		/* Interrupt masking is not always honored, so check it here. */
 		if (status & SCALER_DISPSTAT_EUFLOW(channel) &&
-		    control & SCALER_DISPCTRL_DSPEISLUR(channel)) {
+		    control & dspeislur) {
 			vc4_hvs_mask_underrun(hvs, channel);
 			vc4_hvs_report_underrun(dev);
 
@@ -898,19 +903,34 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
 		    SCALER_DISPCTRL_DISPEIRQ(1) |
 		    SCALER_DISPCTRL_DISPEIRQ(2);
 
-	dispctrl &= ~(SCALER_DISPCTRL_DMAEIRQ |
-		      SCALER_DISPCTRL_SLVWREIRQ |
-		      SCALER_DISPCTRL_SLVRDEIRQ |
-		      SCALER_DISPCTRL_DSPEIEOF(0) |
-		      SCALER_DISPCTRL_DSPEIEOF(1) |
-		      SCALER_DISPCTRL_DSPEIEOF(2) |
-		      SCALER_DISPCTRL_DSPEIEOLN(0) |
-		      SCALER_DISPCTRL_DSPEIEOLN(1) |
-		      SCALER_DISPCTRL_DSPEIEOLN(2) |
-		      SCALER_DISPCTRL_DSPEISLUR(0) |
-		      SCALER_DISPCTRL_DSPEISLUR(1) |
-		      SCALER_DISPCTRL_DSPEISLUR(2) |
-		      SCALER_DISPCTRL_SCLEIRQ);
+	if (!vc4->is_vc5)
+		dispctrl &= ~(SCALER_DISPCTRL_DMAEIRQ |
+			      SCALER_DISPCTRL_SLVWREIRQ |
+			      SCALER_DISPCTRL_SLVRDEIRQ |
+			      SCALER_DISPCTRL_DSPEIEOF(0) |
+			      SCALER_DISPCTRL_DSPEIEOF(1) |
+			      SCALER_DISPCTRL_DSPEIEOF(2) |
+			      SCALER_DISPCTRL_DSPEIEOLN(0) |
+			      SCALER_DISPCTRL_DSPEIEOLN(1) |
+			      SCALER_DISPCTRL_DSPEIEOLN(2) |
+			      SCALER_DISPCTRL_DSPEISLUR(0) |
+			      SCALER_DISPCTRL_DSPEISLUR(1) |
+			      SCALER_DISPCTRL_DSPEISLUR(2) |
+			      SCALER_DISPCTRL_SCLEIRQ);
+	else
+		dispctrl &= ~(SCALER_DISPCTRL_DMAEIRQ |
+			      SCALER5_DISPCTRL_SLVEIRQ |
+			      SCALER5_DISPCTRL_DSPEIEOF(0) |
+			      SCALER5_DISPCTRL_DSPEIEOF(1) |
+			      SCALER5_DISPCTRL_DSPEIEOF(2) |
+			      SCALER5_DISPCTRL_DSPEIEOLN(0) |
+			      SCALER5_DISPCTRL_DSPEIEOLN(1) |
+			      SCALER5_DISPCTRL_DSPEIEOLN(2) |
+			      SCALER5_DISPCTRL_DSPEISLUR(0) |
+			      SCALER5_DISPCTRL_DSPEISLUR(1) |
+			      SCALER5_DISPCTRL_DSPEISLUR(2) |
+			      SCALER_DISPCTRL_SCLEIRQ);
+
 
 	/* Set AXI panic mode.
 	 * VC4 panics when < 2 lines in FIFO.
diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index 95deacdc31e7..1256f0877ff6 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -234,15 +234,21 @@
  * always enabled.
  */
 # define SCALER_DISPCTRL_DSPEISLUR(x)		BIT(13 + (x))
+# define SCALER5_DISPCTRL_DSPEISLUR(x)		BIT(9 + ((x) * 4))
 /* Enables Display 0 end-of-line-N contribution to
  * SCALER_DISPSTAT_IRQDISP0
  */
 # define SCALER_DISPCTRL_DSPEIEOLN(x)		BIT(8 + ((x) * 2))
+# define SCALER5_DISPCTRL_DSPEIEOLN(x)		BIT(8 + ((x) * 4))
 /* Enables Display 0 EOF contribution to SCALER_DISPSTAT_IRQDISP0 */
 # define SCALER_DISPCTRL_DSPEIEOF(x)		BIT(7 + ((x) * 2))
+# define SCALER5_DISPCTRL_DSPEIEOF(x)		BIT(7 + ((x) * 4))
 
-# define SCALER_DISPCTRL_SLVRDEIRQ		BIT(6)
-# define SCALER_DISPCTRL_SLVWREIRQ		BIT(5)
+# define SCALER5_DISPCTRL_DSPEIVST(x)		BIT(6 + ((x) * 4))
+
+# define SCALER_DISPCTRL_SLVRDEIRQ		BIT(6)	/* HVS4 only */
+# define SCALER_DISPCTRL_SLVWREIRQ		BIT(5)	/* HVS4 only */
+# define SCALER5_DISPCTRL_SLVEIRQ		BIT(5)
 # define SCALER_DISPCTRL_DMAEIRQ		BIT(4)
 /* Enables interrupt generation on the enabled EOF/EOLN/EISLUR
  * bits and short frames..

-- 
2.38.1

[PATCH 05/15] drm/vc4: hvs: Support zpos on all planes

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

Adds the zpos property to all planes, and creates the dlist
by placing the fragments in the correct order based on zpos.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hvs.c   | 43 +++++++++++++++++++++++++++--------------
 drivers/gpu/drm/vc4/vc4_kms.c   |  1 +
 drivers/gpu/drm/vc4/vc4_plane.c | 22 ++++++++++++++++++---
 3 files changed, 48 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index d9fc0d03023b..3582ede1a0aa 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -570,6 +570,8 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 	bool enable_bg_fill = false;
 	u32 __iomem *dlist_start = vc4->hvs->dlist + vc4_state->mm.start;
 	u32 __iomem *dlist_next = dlist_start;
+	unsigned int zpos = 0;
+	bool found = false;
 	int idx;
 
 	if (!drm_dev_enter(dev, &idx)) {
@@ -583,23 +585,34 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 	}
 
 	/* Copy all the active planes' dlist contents to the hardware dlist. */
-	drm_atomic_crtc_for_each_plane(plane, crtc) {
-		/* Is this the first active plane? */
-		if (dlist_next == dlist_start) {
-			/* We need to enable background fill when a plane
-			 * could be alpha blending from the background, i.e.
-			 * where no other plane is underneath. It suffices to
-			 * consider the first active plane here since we set
-			 * needs_bg_fill such that either the first plane
-			 * already needs it or all planes on top blend from
-			 * the first or a lower plane.
-			 */
-			vc4_plane_state = to_vc4_plane_state(plane->state);
-			enable_bg_fill = vc4_plane_state->needs_bg_fill;
+	do {
+		found = false;
+
+		drm_atomic_crtc_for_each_plane(plane, crtc) {
+			if (plane->state->normalized_zpos != zpos)
+				continue;
+
+			/* Is this the first active plane? */
+			if (dlist_next == dlist_start) {
+				/* We need to enable background fill when a plane
+				 * could be alpha blending from the background, i.e.
+				 * where no other plane is underneath. It suffices to
+				 * consider the first active plane here since we set
+				 * needs_bg_fill such that either the first plane
+				 * already needs it or all planes on top blend from
+				 * the first or a lower plane.
+				 */
+				vc4_plane_state = to_vc4_plane_state(plane->state);
+				enable_bg_fill = vc4_plane_state->needs_bg_fill;
+			}
+
+			dlist_next += vc4_plane_write_dlist(plane, dlist_next);
+
+			found = true;
 		}
 
-		dlist_next += vc4_plane_write_dlist(plane, dlist_next);
-	}
+		zpos++;
+	} while (found);
 
 	writel(SCALER_CTL0_END, dlist_next);
 	dlist_next++;
diff --git a/drivers/gpu/drm/vc4/vc4_kms.c b/drivers/gpu/drm/vc4/vc4_kms.c
index 7282545c54a1..44deee666032 100644
--- a/drivers/gpu/drm/vc4/vc4_kms.c
+++ b/drivers/gpu/drm/vc4/vc4_kms.c
@@ -1093,6 +1093,7 @@ int vc4_kms_load(struct drm_device *dev)
 	dev->mode_config.helper_private = &vc4_mode_config_helpers;
 	dev->mode_config.preferred_depth = 24;
 	dev->mode_config.async_page_flip = true;
+	dev->mode_config.normalize_zpos = true;
 
 	ret = vc4_ctm_obj_init(vc4);
 	if (ret)
diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index 8b92a45a3c89..c212f8c10388 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -1568,9 +1568,14 @@ struct drm_plane *vc4_plane_init(struct drm_device *dev,
 					  DRM_COLOR_YCBCR_BT709,
 					  DRM_COLOR_YCBCR_LIMITED_RANGE);
 
+	if (type == DRM_PLANE_TYPE_PRIMARY)
+		drm_plane_create_zpos_immutable_property(plane, 0);
+
 	return plane;
 }
 
+#define VC4_NUM_OVERLAY_PLANES	16
+
 int vc4_plane_create_additional_planes(struct drm_device *drm)
 {
 	struct drm_plane *cursor_plane;
@@ -1586,24 +1591,35 @@ int vc4_plane_create_additional_planes(struct drm_device *drm)
 	 * modest number of planes to expose, that should hopefully
 	 * still cover any sane usecase.
 	 */
-	for (i = 0; i < 16; i++) {
+	for (i = 0; i < VC4_NUM_OVERLAY_PLANES; i++) {
 		struct drm_plane *plane =
 			vc4_plane_init(drm, DRM_PLANE_TYPE_OVERLAY,
 				       GENMASK(drm->mode_config.num_crtc - 1, 0));
 
 		if (IS_ERR(plane))
 			continue;
+
+		/* Create zpos property. Max of all the overlays + 1 primary +
+		 * 1 cursor plane on a crtc.
+		 */
+		drm_plane_create_zpos_property(plane, i + 1, 1,
+					       VC4_NUM_OVERLAY_PLANES + 1);
 	}
 
 	drm_for_each_crtc(crtc, drm) {
 		/* Set up the legacy cursor after overlay initialization,
-		 * since we overlay planes on the CRTC in the order they were
-		 * initialized.
+		 * since the zpos fallback is that planes are rendered by plane
+		 * ID order, and that then puts the cursor on top.
 		 */
 		cursor_plane = vc4_plane_init(drm, DRM_PLANE_TYPE_CURSOR,
 					      drm_crtc_mask(crtc));
 		if (!IS_ERR(cursor_plane)) {
 			crtc->cursor = cursor_plane;
+
+			drm_plane_create_zpos_property(cursor_plane,
+						       VC4_NUM_OVERLAY_PLANES + 1,
+						       1,
+						       VC4_NUM_OVERLAY_PLANES + 1);
 		}
 	}
 

-- 
2.38.1

[PATCH 06/15] drm/vc4: hvs: Fix colour order for xRGB1555 on HVS5

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

Same as the xRGB8888 formats, HVS5 has managed to swap the colour
channels for the xRGB1555 formats as well. Add the relevant
config for pixel_order_hvs5.

Fixes: c54619b0bfb3 ("drm/vc4: Add support for the BCM2711 HVS5")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_plane.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index c212f8c10388..e60d6f1a7de1 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -75,11 +75,13 @@ static const struct hvs_format {
 		.drm = DRM_FORMAT_ARGB1555,
 		.hvs = HVS_PIXEL_FORMAT_RGBA5551,
 		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
 	},
 	{
 		.drm = DRM_FORMAT_XRGB1555,
 		.hvs = HVS_PIXEL_FORMAT_RGBA5551,
 		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
 	},
 	{
 		.drm = DRM_FORMAT_RGB888,

-- 
2.38.1

[PATCH 07/15] drm/vc4: hvs: Add DRM 210101010 RGB formats

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

HVS5 supports the 210101010 RGB[A|X] formats, but they were
missing from the DRM to HVS mapping list, so weren't available.
Add them in.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_plane.c | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index e60d6f1a7de1..eb0ac2167937 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -139,6 +139,34 @@ static const struct hvs_format {
 		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
 		.hvs5_only = true,
 	},
+	{
+		.drm = DRM_FORMAT_XRGB2101010,
+		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
+		.hvs5_only = true,
+	},
+	{
+		.drm = DRM_FORMAT_ARGB2101010,
+		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
+		.hvs5_only = true,
+	},
+	{
+		.drm = DRM_FORMAT_ABGR2101010,
+		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
+		.hvs5_only = true,
+	},
+	{
+		.drm = DRM_FORMAT_XBGR2101010,
+		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
+		.hvs5_only = true,
+	},
 };
 
 static const struct hvs_format *vc4_get_hvs_format(u32 drm_format)

-- 
2.38.1

[PATCH 08/15] drm/vc4: hvs: Ignore atomic_flush if we're disabled

[Maxime Ripard]

atomic_flush will be called for each CRTC even if they aren't enabled.

The whole code we have there will thus run without a properly affected
channel, which can then result in all sorts of weird behaviour.

Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hvs.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index 3582ede1a0aa..5fadbf219542 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -579,6 +579,9 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
 		return;
 	}
 
+	if (vc4_state->assigned_channel == VC4_HVS_CHANNEL_DISABLED)
+		return;
+
 	if (debug_dump_regs) {
 		DRM_INFO("CRTC %d HVS before:\n", drm_crtc_index(crtc));
 		vc4_hvs_dump_state(hvs);

-- 
2.38.1

[PATCH 09/15] drm/vc4: plane: Allow using 0 as a pixel order value

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

vc4_plane_mode_set for HVS5 was using pixel_order unless pixel_order_hvs5
was non-zero, except 0 is a valid value for the pixel_order.

Specify pixel_order_hvs5 for all formats and remove the conditional.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_plane.c | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index eb0ac2167937..8b4805c937f0 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -65,11 +65,13 @@ static const struct hvs_format {
 		.drm = DRM_FORMAT_RGB565,
 		.hvs = HVS_PIXEL_FORMAT_RGB565,
 		.pixel_order = HVS_PIXEL_ORDER_XRGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XRGB,
 	},
 	{
 		.drm = DRM_FORMAT_BGR565,
 		.hvs = HVS_PIXEL_FORMAT_RGB565,
 		.pixel_order = HVS_PIXEL_ORDER_XBGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XBGR,
 	},
 	{
 		.drm = DRM_FORMAT_ARGB1555,
@@ -87,56 +89,67 @@ static const struct hvs_format {
 		.drm = DRM_FORMAT_RGB888,
 		.hvs = HVS_PIXEL_FORMAT_RGB888,
 		.pixel_order = HVS_PIXEL_ORDER_XRGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XRGB,
 	},
 	{
 		.drm = DRM_FORMAT_BGR888,
 		.hvs = HVS_PIXEL_FORMAT_RGB888,
 		.pixel_order = HVS_PIXEL_ORDER_XBGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XBGR,
 	},
 	{
 		.drm = DRM_FORMAT_YUV422,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCBCR,
 	},
 	{
 		.drm = DRM_FORMAT_YVU422,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCRCB,
 	},
 	{
 		.drm = DRM_FORMAT_YUV420,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCBCR,
 	},
 	{
 		.drm = DRM_FORMAT_YVU420,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCRCB,
 	},
 	{
 		.drm = DRM_FORMAT_NV12,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCBCR,
 	},
 	{
 		.drm = DRM_FORMAT_NV21,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCRCB,
 	},
 	{
 		.drm = DRM_FORMAT_NV16,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCBCR,
 	},
 	{
 		.drm = DRM_FORMAT_NV61,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE,
 		.pixel_order = HVS_PIXEL_ORDER_XYCRCB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCRCB,
 	},
 	{
 		.drm = DRM_FORMAT_P030,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_10BIT,
 		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCBCR,
 		.hvs5_only = true,
 	},
 	{
@@ -1031,15 +1044,10 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
 		vc4_dlist_write(vc4_state, 0xc0c0c0c0);
 
 	} else {
-		u32 hvs_pixel_order = format->pixel_order;
-
-		if (format->pixel_order_hvs5)
-			hvs_pixel_order = format->pixel_order_hvs5;
-
 		/* Control word */
 		vc4_dlist_write(vc4_state,
 				SCALER_CTL0_VALID |
-				(hvs_pixel_order << SCALER_CTL0_ORDER_SHIFT) |
+				(format->pixel_order_hvs5 << SCALER_CTL0_ORDER_SHIFT) |
 				(hvs_format << SCALER_CTL0_PIXEL_FORMAT_SHIFT) |
 				VC4_SET_FIELD(tiling, SCALER_CTL0_TILING) |
 				(vc4_state->is_unity ?

-- 
2.38.1

[PATCH 10/15] drm/vc4: plane: Omit pixel_order from the hvs_format for hvs5 only formats

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

pixel_order is used for the earlier versions of the HVS, so is
redundant on the 10:10:10:2 and 10bit YUV formats that are only
supported on HVS5.
Remove the assignment from the table to avoid confusion.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_plane.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index 8b4805c937f0..7b7bbe94d47a 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -148,35 +148,30 @@ static const struct hvs_format {
 	{
 		.drm = DRM_FORMAT_P030,
 		.hvs = HVS_PIXEL_FORMAT_YCBCR_10BIT,
-		.pixel_order = HVS_PIXEL_ORDER_XYCBCR,
 		.pixel_order_hvs5 = HVS_PIXEL_ORDER_XYCBCR,
 		.hvs5_only = true,
 	},
 	{
 		.drm = DRM_FORMAT_XRGB2101010,
 		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
-		.pixel_order = HVS_PIXEL_ORDER_ABGR,
 		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
 		.hvs5_only = true,
 	},
 	{
 		.drm = DRM_FORMAT_ARGB2101010,
 		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
-		.pixel_order = HVS_PIXEL_ORDER_ABGR,
 		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
 		.hvs5_only = true,
 	},
 	{
 		.drm = DRM_FORMAT_ABGR2101010,
 		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
-		.pixel_order = HVS_PIXEL_ORDER_ARGB,
 		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
 		.hvs5_only = true,
 	},
 	{
 		.drm = DRM_FORMAT_XBGR2101010,
 		.hvs = HVS_PIXEL_FORMAT_RGBA1010102,
-		.pixel_order = HVS_PIXEL_ORDER_ARGB,
 		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
 		.hvs5_only = true,
 	},

-- 
2.38.1

[PATCH 11/15] drm/vc4: plane: Add 3:3:2 and 4:4:4:4 RGB/RGBX/RGBA formats

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

The hardware supports the 332 8bpp and 4:4:4:4 16bpp formats,
but the table of supported formats didn't include them.
Add them in.

In theory they are supported for T-format as well as linear,
but without a way to test them just add them as linear for now.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_plane.c | 70 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 70 insertions(+)

diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index 7b7bbe94d47a..dee525bacd4b 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -175,6 +175,66 @@ static const struct hvs_format {
 		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
 		.hvs5_only = true,
 	},
+	{
+		.drm = DRM_FORMAT_RGB332,
+		.hvs = HVS_PIXEL_FORMAT_RGB332,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
+	},
+	{
+		.drm = DRM_FORMAT_BGR233,
+		.hvs = HVS_PIXEL_FORMAT_RGB332,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
+	},
+	{
+		.drm = DRM_FORMAT_XRGB4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
+	},
+	{
+		.drm = DRM_FORMAT_ARGB4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_ABGR,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ARGB,
+	},
+	{
+		.drm = DRM_FORMAT_XBGR4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
+	},
+	{
+		.drm = DRM_FORMAT_ABGR4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_ARGB,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_ABGR,
+	},
+	{
+		.drm = DRM_FORMAT_BGRX4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_RGBA,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_BGRA,
+	},
+	{
+		.drm = DRM_FORMAT_BGRA4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_RGBA,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_BGRA,
+	},
+	{
+		.drm = DRM_FORMAT_RGBX4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_BGRA,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_RGBA,
+	},
+	{
+		.drm = DRM_FORMAT_RGBA4444,
+		.hvs = HVS_PIXEL_FORMAT_RGBA4444,
+		.pixel_order = HVS_PIXEL_ORDER_BGRA,
+		.pixel_order_hvs5 = HVS_PIXEL_ORDER_RGBA,
+	},
 };
 
 static const struct hvs_format *vc4_get_hvs_format(u32 drm_format)
@@ -1521,6 +1581,16 @@ static bool vc4_format_mod_supported(struct drm_plane *plane,
 	case DRM_FORMAT_BGRX1010102:
 	case DRM_FORMAT_RGBA1010102:
 	case DRM_FORMAT_BGRA1010102:
+	case DRM_FORMAT_XRGB4444:
+	case DRM_FORMAT_ARGB4444:
+	case DRM_FORMAT_XBGR4444:
+	case DRM_FORMAT_ABGR4444:
+	case DRM_FORMAT_RGBX4444:
+	case DRM_FORMAT_RGBA4444:
+	case DRM_FORMAT_BGRX4444:
+	case DRM_FORMAT_BGRA4444:
+	case DRM_FORMAT_RGB332:
+	case DRM_FORMAT_BGR233:
 	case DRM_FORMAT_YUV422:
 	case DRM_FORMAT_YVU422:
 	case DRM_FORMAT_YUV420:

-- 
2.38.1

[PATCH 12/15] drm/vc4: Add comments for which HVS_PIXEL_ORDER_xxx defines apply

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

The HVS_PIXEL_ORDER_xxx defines apply to specific HVS_PIXEL_FORMAT_xxx
modes, so add comments to make this obvious.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_regs.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index 1256f0877ff6..f3763bd600f6 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -848,16 +848,19 @@ enum hvs_pixel_format {
 /* Note: the LSB is the rightmost character shown.  Only valid for
  * HVS_PIXEL_FORMAT_RGB8888, not RGB888.
  */
+/* For modes 332, 4444, 555, 5551, 6666, 8888, 10:10:10:2 */
 #define HVS_PIXEL_ORDER_RGBA			0
 #define HVS_PIXEL_ORDER_BGRA			1
 #define HVS_PIXEL_ORDER_ARGB			2
 #define HVS_PIXEL_ORDER_ABGR			3
 
+/* For modes 666 and 888 (4 & 5) */
 #define HVS_PIXEL_ORDER_XBRG			0
 #define HVS_PIXEL_ORDER_XRBG			1
 #define HVS_PIXEL_ORDER_XRGB			2
 #define HVS_PIXEL_ORDER_XBGR			3
 
+/* For YCbCr modes (8-12, and 17) */
 #define HVS_PIXEL_ORDER_XYCBCR			0
 #define HVS_PIXEL_ORDER_XYCRCB			1
 #define HVS_PIXEL_ORDER_YXCBCR			2

-- 
2.38.1

[PATCH 13/15] drm/vc4: crtc: Fix timings for VEC modes

[Maxime Ripard]

From: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>

This commit fixes vertical timings of the VEC (composite output) modes
to accurately represent the 525-line ("NTSC") and 625-line ("PAL") ITU-R
standards.

Previous timings were actually defined as 502 and 601 lines, resulting
in non-standard 62.69 Hz and 52 Hz signals being generated,
respectively.

Changes to vc4_crtc.c have also been made, to make the PixelValve
vertical timings accurately correspond to the DRM modeline in interlaced
modes. The resulting VERTA/VERTB register values have been verified
against the reference values set by the Raspberry Pi firmware.

Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_crtc.c | 71 ++++++++++++++++++++++++++----------------
 1 file changed, 44 insertions(+), 27 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c
index 333529ed3a0d..91181cac01e4 100644
--- a/drivers/gpu/drm/vc4/vc4_crtc.c
+++ b/drivers/gpu/drm/vc4/vc4_crtc.c
@@ -326,8 +326,14 @@ static void vc4_crtc_config_pv(struct drm_crtc *crtc, struct drm_encoder *encode
 	bool is_dsi = (vc4_encoder->type == VC4_ENCODER_TYPE_DSI0 ||
 		       vc4_encoder->type == VC4_ENCODER_TYPE_DSI1);
 	bool is_dsi1 = vc4_encoder->type == VC4_ENCODER_TYPE_DSI1;
+	bool is_vec = vc4_encoder->type == VC4_ENCODER_TYPE_VEC;
 	u32 format = is_dsi1 ? PV_CONTROL_FORMAT_DSIV_24 : PV_CONTROL_FORMAT_24;
 	u8 ppc = pv_data->pixels_per_clock;
+
+	u16 vert_bp = mode->crtc_vtotal - mode->crtc_vsync_end;
+	u16 vert_sync = mode->crtc_vsync_end - mode->crtc_vsync_start;
+	u16 vert_fp = mode->crtc_vsync_start - mode->crtc_vdisplay;
+
 	bool debug_dump_regs = false;
 	int idx;
 
@@ -355,49 +361,60 @@ static void vc4_crtc_config_pv(struct drm_crtc *crtc, struct drm_encoder *encode
 		   VC4_SET_FIELD(mode->hdisplay * pixel_rep / ppc,
 				 PV_HORZB_HACTIVE));
 
-	CRTC_WRITE(PV_VERTA,
-		   VC4_SET_FIELD(mode->crtc_vtotal - mode->crtc_vsync_end +
-				 interlace,
-				 PV_VERTA_VBP) |
-		   VC4_SET_FIELD(mode->crtc_vsync_end - mode->crtc_vsync_start,
-				 PV_VERTA_VSYNC));
-	CRTC_WRITE(PV_VERTB,
-		   VC4_SET_FIELD(mode->crtc_vsync_start - mode->crtc_vdisplay,
-				 PV_VERTB_VFP) |
-		   VC4_SET_FIELD(mode->crtc_vdisplay, PV_VERTB_VACTIVE));
-
 	if (interlace) {
+		bool odd_field_first = false;
+		u32 field_delay = mode->htotal * pixel_rep / (2 * ppc);
+		u16 vert_bp_even = vert_bp;
+		u16 vert_fp_even = vert_fp;
+
+		if (is_vec) {
+			/* VEC (composite output) */
+			++field_delay;
+			if (mode->htotal == 858) {
+				/* 525-line mode (NTSC or PAL-M) */
+				odd_field_first = true;
+			}
+		}
+
+		if (odd_field_first)
+			++vert_fp_even;
+		else
+			++vert_bp;
+
 		CRTC_WRITE(PV_VERTA_EVEN,
-			   VC4_SET_FIELD(mode->crtc_vtotal -
-					 mode->crtc_vsync_end,
-					 PV_VERTA_VBP) |
-			   VC4_SET_FIELD(mode->crtc_vsync_end -
-					 mode->crtc_vsync_start,
-					 PV_VERTA_VSYNC));
+			   VC4_SET_FIELD(vert_bp_even, PV_VERTA_VBP) |
+			   VC4_SET_FIELD(vert_sync, PV_VERTA_VSYNC));
 		CRTC_WRITE(PV_VERTB_EVEN,
-			   VC4_SET_FIELD(mode->crtc_vsync_start -
-					 mode->crtc_vdisplay,
-					 PV_VERTB_VFP) |
+			   VC4_SET_FIELD(vert_fp_even, PV_VERTB_VFP) |
 			   VC4_SET_FIELD(mode->crtc_vdisplay, PV_VERTB_VACTIVE));
 
-		/* We set up first field even mode for HDMI.  VEC's
-		 * NTSC mode would want first field odd instead, once
-		 * we support it (to do so, set ODD_FIRST and put the
-		 * delay in VSYNCD_EVEN instead).
+		/* We set up first field even mode for HDMI and VEC's PAL.
+		 * For NTSC, we need first field odd.
 		 */
 		CRTC_WRITE(PV_V_CONTROL,
 			   PV_VCONTROL_CONTINUOUS |
 			   (is_dsi ? PV_VCONTROL_DSI : 0) |
 			   PV_VCONTROL_INTERLACE |
-			   VC4_SET_FIELD(mode->htotal * pixel_rep / (2 * ppc),
-					 PV_VCONTROL_ODD_DELAY));
-		CRTC_WRITE(PV_VSYNCD_EVEN, 0);
+			   (odd_field_first
+				   ? PV_VCONTROL_ODD_FIRST
+				   : VC4_SET_FIELD(field_delay,
+						   PV_VCONTROL_ODD_DELAY)));
+		CRTC_WRITE(PV_VSYNCD_EVEN,
+			   (odd_field_first ? field_delay : 0));
 	} else {
 		CRTC_WRITE(PV_V_CONTROL,
 			   PV_VCONTROL_CONTINUOUS |
 			   (is_dsi ? PV_VCONTROL_DSI : 0));
+		CRTC_WRITE(PV_VSYNCD_EVEN, 0);
 	}
 
+	CRTC_WRITE(PV_VERTA,
+		   VC4_SET_FIELD(vert_bp, PV_VERTA_VBP) |
+		   VC4_SET_FIELD(vert_sync, PV_VERTA_VSYNC));
+	CRTC_WRITE(PV_VERTB,
+		   VC4_SET_FIELD(vert_fp, PV_VERTB_VFP) |
+		   VC4_SET_FIELD(mode->crtc_vdisplay, PV_VERTB_VACTIVE));
+
 	if (is_dsi)
 		CRTC_WRITE(PV_HACT_ACT, mode->hdisplay * pixel_rep);
 

-- 
2.38.1

[PATCH 14/15] drm/vc4: hdmi: Correct interlaced timings again

[Maxime Ripard]

From: Dave Stevenson <dave.stevenson@raspberrypi.com>

The back porch timings were correct, only the sync offset was wrong.
Correct timing is now reported for 1080i and 576i, but the h offset is
incorrect for 480i for non-obvious reasons.

Fixes: c0ba150ce468 ("drm/vc4: Correct HDMI timing registers for interlaced modes")
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_hdmi.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c
index 12a00d644b61..e027381b7331 100644
--- a/drivers/gpu/drm/vc4/vc4_hdmi.c
+++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
@@ -1298,11 +1298,12 @@ static void vc5_hdmi_set_timings(struct vc4_hdmi *vc4_hdmi,
 		     VC4_SET_FIELD(mode->crtc_vdisplay, VC5_HDMI_VERTA_VAL));
 	u32 vertb = (VC4_SET_FIELD(mode->htotal >> (2 - pixel_rep),
 				   VC5_HDMI_VERTB_VSPO) |
-		     VC4_SET_FIELD(mode->crtc_vtotal - mode->crtc_vsync_end,
+		     VC4_SET_FIELD(mode->crtc_vtotal - mode->crtc_vsync_end +
+				   interlaced,
 				   VC4_HDMI_VERTB_VBP));
 	u32 vertb_even = (VC4_SET_FIELD(0, VC5_HDMI_VERTB_VSPO) |
 			  VC4_SET_FIELD(mode->crtc_vtotal -
-					mode->crtc_vsync_end - interlaced,
+					mode->crtc_vsync_end,
 					VC4_HDMI_VERTB_VBP));
 	unsigned long flags;
 	unsigned char gcp;

-- 
2.38.1

[PATCH 15/15] drm/vc4: vec: Support progressive modes

[Maxime Ripard]

From: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>

The VEC is able to output progressive analog modes, but the driver has
never set the proper bit to do so.

Signed-off-by: Mateusz Kwiatkowski <kfyatek+publicgit@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
---
 drivers/gpu/drm/vc4/vc4_vec.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/vc4/vc4_vec.c b/drivers/gpu/drm/vc4/vc4_vec.c
index e270a4099be3..66359d0debbc 100644
--- a/drivers/gpu/drm/vc4/vc4_vec.c
+++ b/drivers/gpu/drm/vc4/vc4_vec.c
@@ -604,7 +604,9 @@ static void vc4_vec_encoder_enable(struct drm_encoder *encoder,
 	VEC_WRITE(VEC_CLMP0_START, 0xac);
 	VEC_WRITE(VEC_CLMP0_END, 0xec);
 	VEC_WRITE(VEC_CONFIG2,
-		  VEC_CONFIG2_UV_DIG_DIS | VEC_CONFIG2_RGB_DIG_DIS);
+		  VEC_CONFIG2_UV_DIG_DIS |
+		  VEC_CONFIG2_RGB_DIG_DIS |
+		  ((adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) ? 0 : VEC_CONFIG2_PROG_SCAN));
 	VEC_WRITE(VEC_CONFIG3, VEC_CONFIG3_HORIZ_LEN_STD);
 	VEC_WRITE(VEC_DAC_CONFIG, vec->variant->dac_config);
 

-- 
2.38.1

Re: [PATCH 08/15] drm/vc4: hvs: Ignore atomic_flush if we're disabled

[Dave Stevenson]

Hi Maxime

On Wed, 7 Dec 2022 at 11:55, Maxime Ripard <maxime@cerno.tech> wrote:
>
> atomic_flush will be called for each CRTC even if they aren't enabled.
>
> The whole code we have there will thus run without a properly affected
> channel, which can then result in all sorts of weird behaviour.
>
> Signed-off-by: Maxime Ripard <maxime@cerno.tech>

Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>

> ---
>  drivers/gpu/drm/vc4/vc4_hvs.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
> index 3582ede1a0aa..5fadbf219542 100644
> --- a/drivers/gpu/drm/vc4/vc4_hvs.c
> +++ b/drivers/gpu/drm/vc4/vc4_hvs.c
> @@ -579,6 +579,9 @@ void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
>                 return;
>         }
>
> +       if (vc4_state->assigned_channel == VC4_HVS_CHANNEL_DISABLED)
> +               return;
> +
>         if (debug_dump_regs) {
>                 DRM_INFO("CRTC %d HVS before:\n", drm_crtc_index(crtc));
>                 vc4_hvs_dump_state(hvs);
>
> --
> 2.38.1

Re: [PATCH 00/15] drm/vc4: Misc fixes and improvements for the HVS and CRTCs

[Maxime Ripard]

[-- Attachment #1: Type: text/plain, Size: 253 bytes --]

Hi,

On Wed, Dec 07, 2022 at 12:53:11PM +0100, Maxime Ripard wrote:
> Here's a collection of patches for the HVS and CRTCs that fix a few
> issues (with DMA accesses and timings, mostly), and add a few minor
> features.

Applied to drm-misc-next

Maxime

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]