drm/vc4: crtc: Move HVS setup code to the HVS driver

The CRTC in vc4 is backed by two devices, the HVS that does the composition
and the PixelValve that does the timing generation.

The writeback is kind of a special case since it doesn't have an associated
pixelvalve but goes straight from the HVS to the TXP. Therefore, it makes
sense to move out the HVS setup code into helpers so that we can also reuse
them from the TXP driver.
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Reviewed-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/96443394e81429ee38f070cfe231701b07e56d69.1591882579.git-series.maxime@cerno.tech

drivers/gpu/drm/vc4/vc4_drv.h

@@ -857,6 +857,11 @@ void vc4_irq_reset(struct drm_device *dev);
 
 /* vc4_hvs.c */
 extern struct platform_driver vc4_hvs_driver;
+int vc4_hvs_atomic_check(struct drm_crtc *crtc, struct drm_crtc_state *state);
+void vc4_hvs_atomic_enable(struct drm_crtc *crtc, struct drm_crtc_state *old_state);
+void vc4_hvs_atomic_disable(struct drm_crtc *crtc, struct drm_crtc_state *old_state);
+void vc4_hvs_atomic_flush(struct drm_crtc *crtc, struct drm_crtc_state *state);
+void vc4_hvs_mode_set_nofb(struct drm_crtc *crtc);
 void vc4_hvs_dump_state(struct drm_device *dev);
 void vc4_hvs_unmask_underrun(struct drm_device *dev, int channel);
 void vc4_hvs_mask_underrun(struct drm_device *dev, int channel);

drivers/gpu/drm/vc4/vc4_hvs.c

@@ -23,6 +23,7 @@
 #include <linux/platform_device.h>
 
 #include <drm/drm_atomic_helper.h>
+#include <drm/drm_vblank.h>
 
 #include "vc4_drv.h"
 #include "vc4_regs.h"
@@ -154,6 +155,296 @@ static int vc4_hvs_upload_linear_kernel(struct vc4_hvs *hvs,
 	return 0;
 }
 
+static void vc4_hvs_lut_load(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	u32 i;
+
+	/* The LUT memory is laid out with each HVS channel in order,
+	 * each of which takes 256 writes for R, 256 for G, then 256
+	 * for B.
+	 */
+	HVS_WRITE(SCALER_GAMADDR,
+		  SCALER_GAMADDR_AUTOINC |
+		  (vc4_crtc->channel * 3 * crtc->gamma_size));
+
+	for (i = 0; i < crtc->gamma_size; i++)
+		HVS_WRITE(SCALER_GAMDATA, vc4_crtc->lut_r[i]);
+	for (i = 0; i < crtc->gamma_size; i++)
+		HVS_WRITE(SCALER_GAMDATA, vc4_crtc->lut_g[i]);
+	for (i = 0; i < crtc->gamma_size; i++)
+		HVS_WRITE(SCALER_GAMDATA, vc4_crtc->lut_b[i]);
+}
+
+static void vc4_hvs_update_gamma_lut(struct drm_crtc *crtc)
+{
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct drm_color_lut *lut = crtc->state->gamma_lut->data;
+	u32 length = drm_color_lut_size(crtc->state->gamma_lut);
+	u32 i;
+
+	for (i = 0; i < length; i++) {
+		vc4_crtc->lut_r[i] = drm_color_lut_extract(lut[i].red, 8);
+		vc4_crtc->lut_g[i] = drm_color_lut_extract(lut[i].green, 8);
+		vc4_crtc->lut_b[i] = drm_color_lut_extract(lut[i].blue, 8);
+	}
+
+	vc4_hvs_lut_load(crtc);
+}
+
+int vc4_hvs_atomic_check(struct drm_crtc *crtc,
+			 struct drm_crtc_state *state)
+{
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(state);
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct drm_plane *plane;
+	unsigned long flags;
+	const struct drm_plane_state *plane_state;
+	u32 dlist_count = 0;
+	int ret;
+
+	/* The pixelvalve can only feed one encoder (and encoders are
+	 * 1:1 with connectors.)
+	 */
+	if (hweight32(state->connector_mask) > 1)
+		return -EINVAL;
+
+	drm_atomic_crtc_state_for_each_plane_state(plane, plane_state, state)
+		dlist_count += vc4_plane_dlist_size(plane_state);
+
+	dlist_count++; /* Account for SCALER_CTL0_END. */
+
+	spin_lock_irqsave(&vc4->hvs->mm_lock, flags);
+	ret = drm_mm_insert_node(&vc4->hvs->dlist_mm, &vc4_state->mm,
+				 dlist_count);
+	spin_unlock_irqrestore(&vc4->hvs->mm_lock, flags);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void vc4_hvs_update_dlist(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
+
+	if (crtc->state->event) {
+		unsigned long flags;
+
+		crtc->state->event->pipe = drm_crtc_index(crtc);
+
+		WARN_ON(drm_crtc_vblank_get(crtc) != 0);
+
+		spin_lock_irqsave(&dev->event_lock, flags);
+
+		if (!vc4_state->feed_txp || vc4_state->txp_armed) {
+			vc4_crtc->event = crtc->state->event;
+			crtc->state->event = NULL;
+		}
+
+		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+			  vc4_state->mm.start);
+
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	} else {
+		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+			  vc4_state->mm.start);
+	}
+}
+
+void vc4_hvs_atomic_enable(struct drm_crtc *crtc,
+			   struct drm_crtc_state *old_state)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
+	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
+	bool oneshot = vc4_state->feed_txp;
+	u32 dispctrl;
+
+	vc4_hvs_update_dlist(crtc);
+
+	/* Turn on the scaler, which will wait for vstart to start
+	 * compositing.
+	 * When feeding the transposer, we should operate in oneshot
+	 * mode.
+	 */
+	dispctrl = SCALER_DISPCTRLX_ENABLE;
+	dispctrl |= VC4_SET_FIELD(mode->hdisplay,
+				  SCALER_DISPCTRLX_WIDTH) |
+		    VC4_SET_FIELD(mode->vdisplay,
+				  SCALER_DISPCTRLX_HEIGHT) |
+		    (oneshot ? SCALER_DISPCTRLX_ONESHOT : 0);
+
+	HVS_WRITE(SCALER_DISPCTRLX(vc4_crtc->channel), dispctrl);
+}
+
+void vc4_hvs_atomic_disable(struct drm_crtc *crtc,
+			    struct drm_crtc_state *old_state)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	u32 chan = vc4_crtc->channel;
+
+	if (HVS_READ(SCALER_DISPCTRLX(chan)) &
+	    SCALER_DISPCTRLX_ENABLE) {
+		HVS_WRITE(SCALER_DISPCTRLX(chan),
+			  SCALER_DISPCTRLX_RESET);
+
+		/* While the docs say that reset is self-clearing, it
+		 * seems it doesn't actually.
+		 */
+		HVS_WRITE(SCALER_DISPCTRLX(chan), 0);
+	}
+
+	/* Once we leave, the scaler should be disabled and its fifo empty. */
+
+	WARN_ON_ONCE(HVS_READ(SCALER_DISPCTRLX(chan)) & SCALER_DISPCTRLX_RESET);
+
+	WARN_ON_ONCE(VC4_GET_FIELD(HVS_READ(SCALER_DISPSTATX(chan)),
+				   SCALER_DISPSTATX_MODE) !=
+		     SCALER_DISPSTATX_MODE_DISABLED);
+
+	WARN_ON_ONCE((HVS_READ(SCALER_DISPSTATX(chan)) &
+		      (SCALER_DISPSTATX_FULL | SCALER_DISPSTATX_EMPTY)) !=
+		     SCALER_DISPSTATX_EMPTY);
+}
+
+void vc4_hvs_atomic_flush(struct drm_crtc *crtc,
+			  struct drm_crtc_state *old_state)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
+	struct drm_plane *plane;
+	struct vc4_plane_state *vc4_plane_state;
+	bool debug_dump_regs = false;
+	bool enable_bg_fill = false;
+	u32 __iomem *dlist_start = vc4->hvs->dlist + vc4_state->mm.start;
+	u32 __iomem *dlist_next = dlist_start;
+
+	if (debug_dump_regs) {
+		DRM_INFO("CRTC %d HVS before:\n", drm_crtc_index(crtc));
+		vc4_hvs_dump_state(dev);
+	}
+
+	/* 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;
+		}
+
+		dlist_next += vc4_plane_write_dlist(plane, dlist_next);
+	}
+
+	writel(SCALER_CTL0_END, dlist_next);
+	dlist_next++;
+
+	WARN_ON_ONCE(dlist_next - dlist_start != vc4_state->mm.size);
+
+	if (enable_bg_fill)
+		/* This sets a black background color fill, as is the case
+		 * with other DRM drivers.
+		 */
+		HVS_WRITE(SCALER_DISPBKGNDX(vc4_crtc->channel),
+			  HVS_READ(SCALER_DISPBKGNDX(vc4_crtc->channel)) |
+			  SCALER_DISPBKGND_FILL);
+
+	/* Only update DISPLIST if the CRTC was already running and is not
+	 * being disabled.
+	 * vc4_crtc_enable() takes care of updating the dlist just after
+	 * re-enabling VBLANK interrupts and before enabling the engine.
+	 * If the CRTC is being disabled, there's no point in updating this
+	 * information.
+	 */
+	if (crtc->state->active && old_state->active)
+		vc4_hvs_update_dlist(crtc);
+
+	if (crtc->state->color_mgmt_changed) {
+		u32 dispbkgndx = HVS_READ(SCALER_DISPBKGNDX(vc4_crtc->channel));
+
+		if (crtc->state->gamma_lut) {
+			vc4_hvs_update_gamma_lut(crtc);
+			dispbkgndx |= SCALER_DISPBKGND_GAMMA;
+		} else {
+			/* Unsetting DISPBKGND_GAMMA skips the gamma lut step
+			 * in hardware, which is the same as a linear lut that
+			 * DRM expects us to use in absence of a user lut.
+			 */
+			dispbkgndx &= ~SCALER_DISPBKGND_GAMMA;
+		}
+		HVS_WRITE(SCALER_DISPBKGNDX(vc4_crtc->channel), dispbkgndx);
+	}
+
+	if (debug_dump_regs) {
+		DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc));
+		vc4_hvs_dump_state(dev);
+	}
+}
+
+void vc4_hvs_mode_set_nofb(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
+	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
+	bool interlace = mode->flags & DRM_MODE_FLAG_INTERLACE;
+
+	if (vc4_crtc->data->hvs_channel == 2) {
+		u32 dispctrl;
+		u32 dsp3_mux;
+
+		/*
+		 * SCALER_DISPCTRL_DSP3 = X, where X < 2 means 'connect DSP3 to
+		 * FIFO X'.
+		 * SCALER_DISPCTRL_DSP3 = 3 means 'disable DSP 3'.
+		 *
+		 * DSP3 is connected to FIFO2 unless the transposer is
+		 * enabled. In this case, FIFO 2 is directly accessed by the
+		 * TXP IP, and we need to disable the FIFO2 -> pixelvalve1
+		 * route.
+		 */
+		if (vc4_state->feed_txp)
+			dsp3_mux = VC4_SET_FIELD(3, SCALER_DISPCTRL_DSP3_MUX);
+		else
+			dsp3_mux = VC4_SET_FIELD(2, SCALER_DISPCTRL_DSP3_MUX);
+
+		dispctrl = HVS_READ(SCALER_DISPCTRL) &
+			   ~SCALER_DISPCTRL_DSP3_MUX_MASK;
+		HVS_WRITE(SCALER_DISPCTRL, dispctrl | dsp3_mux);
+	}
+
+	HVS_WRITE(SCALER_DISPBKGNDX(vc4_crtc->channel),
+		  SCALER_DISPBKGND_AUTOHS |
+		  SCALER_DISPBKGND_GAMMA |
+		  (interlace ? SCALER_DISPBKGND_INTERLACE : 0));
+
+	/* Reload the LUT, since the SRAMs would have been disabled if
+	 * all CRTCs had SCALER_DISPBKGND_GAMMA unset at once.
+	 */
+	vc4_hvs_lut_load(crtc);
+}
+
 void vc4_hvs_mask_underrun(struct drm_device *dev, int channel)
 {
 	struct vc4_dev *vc4 = to_vc4_dev(dev);