2018-11-06 09:40:47

by Neil Armstrong

[permalink] [raw]
Subject: [PATCH 0/3] drm/meson: Add support for Overlay and OSD Plane scaling

This serie adds support for :
- Overlay Plane blended with the primary plane, we can describe as "behind"
- Primary Plane up-scaling up to 5x factor to support the OSD plane being
upscaled from 1920x1080 when the display mode is set as 3840x2160.

These patches depends on the "drm/meson: Allow using optional canvas provider"
patchset from Maxime Jourdan at [1]

[1] https://patchwork.kernel.org/project/linux-amlogic/list/?series=38355

Neil Armstrong (3):
drm/meson: Support Overlay plane for video rendering
drm/meson: move OSD scaler management into plane atomic update
drm/meson: Add primary plane scaling

drivers/gpu/drm/meson/Makefile | 2 +-
drivers/gpu/drm/meson/meson_canvas.c | 7 +-
drivers/gpu/drm/meson/meson_canvas.h | 11 +-
drivers/gpu/drm/meson/meson_crtc.c | 251 +++++++++++++-
drivers/gpu/drm/meson/meson_drv.c | 29 +-
drivers/gpu/drm/meson/meson_drv.h | 62 ++++
drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
drivers/gpu/drm/meson/meson_overlay.h | 14 +
drivers/gpu/drm/meson/meson_plane.c | 177 ++++++++--
drivers/gpu/drm/meson/meson_registers.h | 3 +
drivers/gpu/drm/meson/meson_viu.c | 15 +
drivers/gpu/drm/meson/meson_vpp.c | 90 +++--
12 files changed, 1156 insertions(+), 91 deletions(-)
create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
create mode 100644 drivers/gpu/drm/meson/meson_overlay.h

--
2.7.4



2018-11-06 09:40:58

by Neil Armstrong

[permalink] [raw]
Subject: [PATCH 3/3] drm/meson: Add primary plane scaling

This patch adds support for the Primary Plane scaling.

On the Amlogic GX SoCs, the primary plane is used as On-Screen-Display
layer on top of video, and it's needed to keep the OSD layer to a lower
size as the physical display size to :
- lower the memory bandwidth
- lower the OSD rendering
- lower the memory usage

This use-case is used when setting the display mode to 3840x2160 and the
OSD layer is rendered using the GPU. In this case, the GXBB & GXL cannot
work on more than 2000x2000 buffer, thus needing the OSD layer to be kept
at 1920x1080 and upscaled to 3840x2160 in hardware.

The primary plane atomic check still allow 1:1 scaling, allowing native
3840x2160 if needed by user-space applications.

Signed-off-by: Neil Armstrong <[email protected]>
---
drivers/gpu/drm/meson/meson_plane.c | 186 +++++++++++++++++++++++++++---------
1 file changed, 141 insertions(+), 45 deletions(-)

diff --git a/drivers/gpu/drm/meson/meson_plane.c b/drivers/gpu/drm/meson/meson_plane.c
index f915a79..12a47b4 100644
--- a/drivers/gpu/drm/meson/meson_plane.c
+++ b/drivers/gpu/drm/meson/meson_plane.c
@@ -24,6 +24,7 @@
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/mutex.h>
+#include <linux/bitfield.h>
#include <linux/platform_device.h>
#include <drm/drmP.h>
#include <drm/drm_atomic.h>
@@ -39,12 +40,50 @@
#include "meson_canvas.h"
#include "meson_registers.h"

+/* OSD_SCI_WH_M1 */
+#define SCI_WH_M1_W(w) FIELD_PREP(GENMASK(28, 16), w)
+#define SCI_WH_M1_H(h) FIELD_PREP(GENMASK(12, 0), h)
+
+/* OSD_SCO_H_START_END */
+/* OSD_SCO_V_START_END */
+#define SCO_HV_START(start) FIELD_PREP(GENMASK(27, 16), start)
+#define SCO_HV_END(end) FIELD_PREP(GENMASK(11, 0), end)
+
+/* OSD_SC_CTRL0 */
+#define SC_CTRL0_PATH_EN BIT(3)
+#define SC_CTRL0_SEL_OSD1 BIT(2)
+
+/* OSD_VSC_CTRL0 */
+#define VSC_BANK_LEN(value) FIELD_PREP(GENMASK(2, 0), value)
+#define VSC_TOP_INI_RCV_NUM(value) FIELD_PREP(GENMASK(6, 3), value)
+#define VSC_TOP_RPT_L0_NUM(value) FIELD_PREP(GENMASK(9, 8), value)
+#define VSC_BOT_INI_RCV_NUM(value) FIELD_PREP(GENMASK(14, 11), value)
+#define VSC_BOT_RPT_L0_NUM(value) FIELD_PREP(GENMASK(17, 16), value)
+#define VSC_PROG_INTERLACE BIT(23)
+#define VSC_VERTICAL_SCALER_EN BIT(24)
+
+/* OSD_VSC_INI_PHASE */
+#define VSC_INI_PHASE_BOT(bottom) FIELD_PREP(GENMASK(31, 16), bottom)
+#define VSC_INI_PHASE_TOP(top) FIELD_PREP(GENMASK(15, 0), top)
+
+/* OSD_HSC_CTRL0 */
+#define HSC_BANK_LENGTH(value) FIELD_PREP(GENMASK(2, 0), value)
+#define HSC_INI_RCV_NUM0(value) FIELD_PREP(GENMASK(6, 3), value)
+#define HSC_RPT_P0_NUM0(value) FIELD_PREP(GENMASK(9, 8), value)
+#define HSC_HORIZ_SCALER_EN BIT(22)
+
+/* VPP_OSD_VSC_PHASE_STEP */
+/* VPP_OSD_HSC_PHASE_STEP */
+#define SC_PHASE_STEP(value) FIELD_PREP(GENMASK(27, 0), value)
+
struct meson_plane {
struct drm_plane base;
struct meson_drm *priv;
};
#define to_meson_plane(x) container_of(x, struct meson_plane, base)

+#define FRAC_16_16(mult, div) (((mult) << 16) / (div))
+
static int meson_plane_atomic_check(struct drm_plane *plane,
struct drm_plane_state *state)
{
@@ -57,10 +96,15 @@ static int meson_plane_atomic_check(struct drm_plane *plane,
if (IS_ERR(crtc_state))
return PTR_ERR(crtc_state);

+ /*
+ * Only allow :
+ * - Upscaling up to 5x, vertical and horizontal
+ * - Final coordinates must match crtc size
+ */
return drm_atomic_helper_check_plane_state(state, crtc_state,
+ FRAC_16_16(1, 5),
DRM_PLANE_HELPER_NO_SCALING,
- DRM_PLANE_HELPER_NO_SCALING,
- true, true);
+ false, true);
}

/* Takes a fixed 16.16 number and converts it to integer. */
@@ -74,22 +118,19 @@ static void meson_plane_atomic_update(struct drm_plane *plane,
{
struct meson_plane *meson_plane = to_meson_plane(plane);
struct drm_plane_state *state = plane->state;
- struct drm_framebuffer *fb = state->fb;
+ struct drm_rect dest = drm_plane_state_dest(state);
struct meson_drm *priv = meson_plane->priv;
+ struct drm_framebuffer *fb = state->fb;
struct drm_gem_cma_object *gem;
- struct drm_rect src = {
- .x1 = (state->src_x),
- .y1 = (state->src_y),
- .x2 = (state->src_x + state->src_w),
- .y2 = (state->src_y + state->src_h),
- };
- struct drm_rect dest = {
- .x1 = state->crtc_x,
- .y1 = state->crtc_y,
- .x2 = state->crtc_x + state->crtc_w,
- .y2 = state->crtc_y + state->crtc_h,
- };
unsigned long flags;
+ int vsc_ini_rcv_num, vsc_ini_rpt_p0_num;
+ int vsc_bot_rcv_num, vsc_bot_rpt_p0_num;
+ int hsc_ini_rcv_num, hsc_ini_rpt_p0_num;
+ int hf_phase_step, vf_phase_step;
+ int src_w, src_h, dst_w, dst_h;
+ int bot_ini_phase;
+ int hf_bank_len;
+ int vf_bank_len;
u8 canvas_id_osd1;

/*
@@ -143,6 +184,27 @@ static void meson_plane_atomic_update(struct drm_plane *plane,
break;
};

+ /* Default scaler parameters */
+ vsc_bot_rcv_num = 0;
+ vsc_bot_rpt_p0_num = 0;
+ hf_bank_len = 4;
+ vf_bank_len = 4;
+
+ if (state->crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) {
+ vsc_bot_rcv_num = 6;
+ vsc_bot_rpt_p0_num = 2;
+ }
+
+ hsc_ini_rcv_num = hf_bank_len;
+ vsc_ini_rcv_num = vf_bank_len;
+ hsc_ini_rpt_p0_num = (hf_bank_len / 2) - 1;
+ vsc_ini_rpt_p0_num = (vf_bank_len / 2) - 1;
+
+ src_w = fixed16_to_int(state->src_w);
+ src_h = fixed16_to_int(state->src_h);
+ dst_w = state->crtc_w;
+ dst_h = state->crtc_h;
+
/*
* When the output is interlaced, the OSD must switch between
* each field using the INTERLACE_SEL_ODD (0) of VIU_OSD1_BLK0_CFG_W0
@@ -151,41 +213,73 @@ static void meson_plane_atomic_update(struct drm_plane *plane,
* is configured for 2:1 scaling with interlace options enabled.
*/
if (state->crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) {
- priv->viu.osd1_interlace = true;
-
dest.y1 /= 2;
dest.y2 /= 2;
+ dst_h /= 2;
+ }

- priv->viu.osd_sc_ctrl0 = BIT(3)| /* Enable scaler */
- BIT(2); /* Select OSD1 */
+ hf_phase_step = ((src_w << 18) / dst_w) << 6;
+ vf_phase_step = (src_h << 20) / dst_h;

- /* 2:1 scaling */
- priv->viu.osd_sc_i_wh_m1 = ((drm_rect_width(&dest) - 1) << 16) |
- (drm_rect_height(&dest) - 1);
- priv->viu.osd_sc_o_h_start_end = (dest.x1 << 16) | dest.x2;
- priv->viu.osd_sc_o_v_start_end = (dest.y1 << 16) | dest.y2;
+ if (state->crtc->mode.flags & DRM_MODE_FLAG_INTERLACE)
+ bot_ini_phase = ((vf_phase_step / 2) >> 4);
+ else
+ bot_ini_phase = 0;
+
+ vf_phase_step = (vf_phase_step << 4);
+
+ /* In interlaced mode, scaler is always active */
+ if (src_h != dst_h || src_w != dst_w) {
+ priv->viu.osd_sc_i_wh_m1 = SCI_WH_M1_W(src_w - 1) |
+ SCI_WH_M1_H(src_h - 1);
+ priv->viu.osd_sc_o_h_start_end = SCO_HV_START(dest.x1) |
+ SCO_HV_END(dest.x2 - 1);
+ priv->viu.osd_sc_o_v_start_end = SCO_HV_START(dest.y1) |
+ SCO_HV_END(dest.y2 - 1);
+ /* Enable OSD Scaler */
+ priv->viu.osd_sc_ctrl0 = SC_CTRL0_PATH_EN | SC_CTRL0_SEL_OSD1;
+ } else {
+ priv->viu.osd_sc_i_wh_m1 = 0;
+ priv->viu.osd_sc_o_h_start_end = 0;
+ priv->viu.osd_sc_o_v_start_end = 0;
+ priv->viu.osd_sc_ctrl0 = 0;
+ }

- /* 2:1 vertical scaling values */
- priv->viu.osd_sc_v_ini_phase = BIT(16);
- priv->viu.osd_sc_v_phase_step = BIT(25);
+ /* In interlaced mode, vertical scaler is always active */
+ if (src_h != dst_h) {
priv->viu.osd_sc_v_ctrl0 =
- (4 << 0) | /* osd_vsc_bank_length */
- (4 << 3) | /* osd_vsc_top_ini_rcv_num0 */
- (1 << 8) | /* osd_vsc_top_rpt_p0_num0 */
- (6 << 11) | /* osd_vsc_bot_ini_rcv_num0 */
- (2 << 16) | /* osd_vsc_bot_rpt_p0_num0 */
- BIT(23) | /* osd_prog_interlace */
- BIT(24); /* Enable vertical scaler */
-
- /* No horizontal scaling */
+ VSC_BANK_LEN(vf_bank_len) |
+ VSC_TOP_INI_RCV_NUM(vsc_ini_rcv_num) |
+ VSC_TOP_RPT_L0_NUM(vsc_ini_rpt_p0_num) |
+ VSC_VERTICAL_SCALER_EN;
+
+ if (state->crtc->mode.flags & DRM_MODE_FLAG_INTERLACE)
+ priv->viu.osd_sc_v_ctrl0 |=
+ VSC_BOT_INI_RCV_NUM(vsc_bot_rcv_num) |
+ VSC_BOT_RPT_L0_NUM(vsc_bot_rpt_p0_num) |
+ VSC_PROG_INTERLACE;
+
+ priv->viu.osd_sc_v_phase_step = SC_PHASE_STEP(vf_phase_step);
+ priv->viu.osd_sc_v_ini_phase = VSC_INI_PHASE_BOT(bot_ini_phase);
+ } else {
+ priv->viu.osd_sc_v_ctrl0 = 0;
+ priv->viu.osd_sc_v_phase_step = 0;
+ priv->viu.osd_sc_v_ini_phase = 0;
+ }
+
+ /* Horizontal scaler is only used if width does not match */
+ if (src_w != dst_w) {
+ priv->viu.osd_sc_h_ctrl0 =
+ HSC_BANK_LENGTH(hf_bank_len) |
+ HSC_INI_RCV_NUM0(hsc_ini_rcv_num) |
+ HSC_RPT_P0_NUM0(hsc_ini_rpt_p0_num) |
+ HSC_HORIZ_SCALER_EN;
+ priv->viu.osd_sc_h_phase_step = SC_PHASE_STEP(hf_phase_step);
priv->viu.osd_sc_h_ini_phase = 0;
- priv->viu.osd_sc_h_phase_step = 0;
- priv->viu.osd_sc_h_ctrl0 = 0;
} else {
- priv->viu.osd1_interlace = false;
- priv->viu.osd_sc_ctrl0 = 0;
priv->viu.osd_sc_h_ctrl0 = 0;
- priv->viu.osd_sc_v_ctrl0 = 0;
+ priv->viu.osd_sc_h_phase_step = 0;
+ priv->viu.osd_sc_h_ini_phase = 0;
}

/*
@@ -193,10 +287,12 @@ static void meson_plane_atomic_update(struct drm_plane *plane,
* where x2 is exclusive.
* e.g. +30x1920 would be (1919 << 16) | 30
*/
- priv->viu.osd1_blk0_cfg[1] = ((fixed16_to_int(src.x2) - 1) << 16) |
- fixed16_to_int(src.x1);
- priv->viu.osd1_blk0_cfg[2] = ((fixed16_to_int(src.y2) - 1) << 16) |
- fixed16_to_int(src.y1);
+ priv->viu.osd1_blk0_cfg[1] =
+ ((fixed16_to_int(state->src.x2) - 1) << 16) |
+ fixed16_to_int(state->src.x1);
+ priv->viu.osd1_blk0_cfg[2] =
+ ((fixed16_to_int(state->src.y2) - 1) << 16) |
+ fixed16_to_int(state->src.y1);
priv->viu.osd1_blk0_cfg[3] = ((dest.x2 - 1) << 16) | dest.x1;
priv->viu.osd1_blk0_cfg[4] = ((dest.y2 - 1) << 16) | dest.y1;

--
2.7.4


2018-11-06 09:42:06

by Neil Armstrong

[permalink] [raw]
Subject: [PATCH 2/3] drm/meson: move OSD scaler management into plane atomic update

In preparation to support the Primary Plane scaling, move the basic
OSD Interlace-Only scaler setup code into the primary plane atomic
update callback and handle the vsync scaler update like the overlay
plane scaling registers update.

Signed-off-by: Neil Armstrong <[email protected]>
---
drivers/gpu/drm/meson/meson_crtc.c | 35 ++++++++++++++++------------
drivers/gpu/drm/meson/meson_drv.h | 10 ++++++++
drivers/gpu/drm/meson/meson_plane.c | 39 ++++++++++++++++++++++++++++++-
drivers/gpu/drm/meson/meson_vpp.c | 46 -------------------------------------
4 files changed, 68 insertions(+), 62 deletions(-)

diff --git a/drivers/gpu/drm/meson/meson_crtc.c b/drivers/gpu/drm/meson/meson_crtc.c
index 38686c9..d78168f 100644
--- a/drivers/gpu/drm/meson/meson_crtc.c
+++ b/drivers/gpu/drm/meson/meson_crtc.c
@@ -189,21 +189,26 @@ void meson_crtc_irq(struct meson_drm *priv)
priv->io_base + _REG(VIU_OSD1_BLK0_CFG_W3));
writel_relaxed(priv->viu.osd1_blk0_cfg[4],
priv->io_base + _REG(VIU_OSD1_BLK0_CFG_W4));
-
- /* If output is interlace, make use of the Scaler */
- if (priv->viu.osd1_interlace) {
- struct drm_plane *plane = priv->primary_plane;
- struct drm_plane_state *state = plane->state;
- struct drm_rect dest = {
- .x1 = state->crtc_x,
- .y1 = state->crtc_y,
- .x2 = state->crtc_x + state->crtc_w,
- .y2 = state->crtc_y + state->crtc_h,
- };
-
- meson_vpp_setup_interlace_vscaler_osd1(priv, &dest);
- } else
- meson_vpp_disable_interlace_vscaler_osd1(priv);
+ writel_relaxed(priv->viu.osd_sc_ctrl0,
+ priv->io_base + _REG(VPP_OSD_SC_CTRL0));
+ writel_relaxed(priv->viu.osd_sc_i_wh_m1,
+ priv->io_base + _REG(VPP_OSD_SCI_WH_M1));
+ writel_relaxed(priv->viu.osd_sc_o_h_start_end,
+ priv->io_base + _REG(VPP_OSD_SCO_H_START_END));
+ writel_relaxed(priv->viu.osd_sc_o_v_start_end,
+ priv->io_base + _REG(VPP_OSD_SCO_V_START_END));
+ writel_relaxed(priv->viu.osd_sc_v_ini_phase,
+ priv->io_base + _REG(VPP_OSD_VSC_INI_PHASE));
+ writel_relaxed(priv->viu.osd_sc_v_phase_step,
+ priv->io_base + _REG(VPP_OSD_VSC_PHASE_STEP));
+ writel_relaxed(priv->viu.osd_sc_h_ini_phase,
+ priv->io_base + _REG(VPP_OSD_HSC_INI_PHASE));
+ writel_relaxed(priv->viu.osd_sc_h_phase_step,
+ priv->io_base + _REG(VPP_OSD_HSC_PHASE_STEP));
+ writel_relaxed(priv->viu.osd_sc_h_ctrl0,
+ priv->io_base + _REG(VPP_OSD_HSC_CTRL0));
+ writel_relaxed(priv->viu.osd_sc_v_ctrl0,
+ priv->io_base + _REG(VPP_OSD_VSC_CTRL0));

if (priv->canvas)
meson_canvas_config(priv->canvas, priv->canvas_id_osd1,
diff --git a/drivers/gpu/drm/meson/meson_drv.h b/drivers/gpu/drm/meson/meson_drv.h
index c971557..a955354 100644
--- a/drivers/gpu/drm/meson/meson_drv.h
+++ b/drivers/gpu/drm/meson/meson_drv.h
@@ -54,6 +54,16 @@ struct meson_drm {
uint32_t osd1_addr;
uint32_t osd1_stride;
uint32_t osd1_height;
+ uint32_t osd_sc_ctrl0;
+ uint32_t osd_sc_i_wh_m1;
+ uint32_t osd_sc_o_h_start_end;
+ uint32_t osd_sc_o_v_start_end;
+ uint32_t osd_sc_v_ini_phase;
+ uint32_t osd_sc_v_phase_step;
+ uint32_t osd_sc_h_ini_phase;
+ uint32_t osd_sc_h_phase_step;
+ uint32_t osd_sc_h_ctrl0;
+ uint32_t osd_sc_v_ctrl0;

bool vd1_enabled;
bool vd1_commit;
diff --git a/drivers/gpu/drm/meson/meson_plane.c b/drivers/gpu/drm/meson/meson_plane.c
index 51bec8e..f915a79 100644
--- a/drivers/gpu/drm/meson/meson_plane.c
+++ b/drivers/gpu/drm/meson/meson_plane.c
@@ -143,13 +143,50 @@ static void meson_plane_atomic_update(struct drm_plane *plane,
break;
};

+ /*
+ * When the output is interlaced, the OSD must switch between
+ * each field using the INTERLACE_SEL_ODD (0) of VIU_OSD1_BLK0_CFG_W0
+ * at each vsync.
+ * But the vertical scaler can provide such funtionnality if
+ * is configured for 2:1 scaling with interlace options enabled.
+ */
if (state->crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) {
priv->viu.osd1_interlace = true;

dest.y1 /= 2;
dest.y2 /= 2;
- } else
+
+ priv->viu.osd_sc_ctrl0 = BIT(3) | /* Enable scaler */
+ BIT(2); /* Select OSD1 */
+
+ /* 2:1 scaling */
+ priv->viu.osd_sc_i_wh_m1 = ((drm_rect_width(&dest) - 1) << 16) |
+ (drm_rect_height(&dest) - 1);
+ priv->viu.osd_sc_o_h_start_end = (dest.x1 << 16) | dest.x2;
+ priv->viu.osd_sc_o_v_start_end = (dest.y1 << 16) | dest.y2;
+
+ /* 2:1 vertical scaling values */
+ priv->viu.osd_sc_v_ini_phase = BIT(16);
+ priv->viu.osd_sc_v_phase_step = BIT(25);
+ priv->viu.osd_sc_v_ctrl0 =
+ (4 << 0) | /* osd_vsc_bank_length */
+ (4 << 3) | /* osd_vsc_top_ini_rcv_num0 */
+ (1 << 8) | /* osd_vsc_top_rpt_p0_num0 */
+ (6 << 11) | /* osd_vsc_bot_ini_rcv_num0 */
+ (2 << 16) | /* osd_vsc_bot_rpt_p0_num0 */
+ BIT(23) | /* osd_prog_interlace */
+ BIT(24); /* Enable vertical scaler */
+
+ /* No horizontal scaling */
+ priv->viu.osd_sc_h_ini_phase = 0;
+ priv->viu.osd_sc_h_phase_step = 0;
+ priv->viu.osd_sc_h_ctrl0 = 0;
+ } else {
priv->viu.osd1_interlace = false;
+ priv->viu.osd_sc_ctrl0 = 0;
+ priv->viu.osd_sc_h_ctrl0 = 0;
+ priv->viu.osd_sc_v_ctrl0 = 0;
+ }

/*
* The format of these registers is (x2 << 16 | x1),
diff --git a/drivers/gpu/drm/meson/meson_vpp.c b/drivers/gpu/drm/meson/meson_vpp.c
index 5dc24a9..f9efb43 100644
--- a/drivers/gpu/drm/meson/meson_vpp.c
+++ b/drivers/gpu/drm/meson/meson_vpp.c
@@ -51,52 +51,6 @@ void meson_vpp_setup_mux(struct meson_drm *priv, unsigned int mux)
writel(mux, priv->io_base + _REG(VPU_VIU_VENC_MUX_CTRL));
}

-/*
- * When the output is interlaced, the OSD must switch between
- * each field using the INTERLACE_SEL_ODD (0) of VIU_OSD1_BLK0_CFG_W0
- * at each vsync.
- * But the vertical scaler can provide such funtionnality if
- * is configured for 2:1 scaling with interlace options enabled.
- */
-void meson_vpp_setup_interlace_vscaler_osd1(struct meson_drm *priv,
- struct drm_rect *input)
-{
- writel_relaxed(BIT(3) /* Enable scaler */ |
- BIT(2), /* Select OSD1 */
- priv->io_base + _REG(VPP_OSD_SC_CTRL0));
-
- writel_relaxed(((drm_rect_width(input) - 1) << 16) |
- (drm_rect_height(input) - 1),
- priv->io_base + _REG(VPP_OSD_SCI_WH_M1));
- /* 2:1 scaling */
- writel_relaxed(((input->x1) << 16) | (input->x2),
- priv->io_base + _REG(VPP_OSD_SCO_H_START_END));
- writel_relaxed(((input->y1 >> 1) << 16) | (input->y2 >> 1),
- priv->io_base + _REG(VPP_OSD_SCO_V_START_END));
-
- /* 2:1 scaling values */
- writel_relaxed(BIT(16), priv->io_base + _REG(VPP_OSD_VSC_INI_PHASE));
- writel_relaxed(BIT(25), priv->io_base + _REG(VPP_OSD_VSC_PHASE_STEP));
-
- writel_relaxed(0, priv->io_base + _REG(VPP_OSD_HSC_CTRL0));
-
- writel_relaxed((4 << 0) /* osd_vsc_bank_length */ |
- (4 << 3) /* osd_vsc_top_ini_rcv_num0 */ |
- (1 << 8) /* osd_vsc_top_rpt_p0_num0 */ |
- (6 << 11) /* osd_vsc_bot_ini_rcv_num0 */ |
- (2 << 16) /* osd_vsc_bot_rpt_p0_num0 */ |
- BIT(23) /* osd_prog_interlace */ |
- BIT(24), /* Enable vertical scaler */
- priv->io_base + _REG(VPP_OSD_VSC_CTRL0));
-}
-
-void meson_vpp_disable_interlace_vscaler_osd1(struct meson_drm *priv)
-{
- writel_relaxed(0, priv->io_base + _REG(VPP_OSD_SC_CTRL0));
- writel_relaxed(0, priv->io_base + _REG(VPP_OSD_VSC_CTRL0));
- writel_relaxed(0, priv->io_base + _REG(VPP_OSD_HSC_CTRL0));
-}
-
static unsigned int vpp_filter_coefs_4point_bspline[] = {
0x15561500, 0x14561600, 0x13561700, 0x12561800,
0x11551a00, 0x11541b00, 0x10541c00, 0x0f541d00,
--
2.7.4


2018-11-06 09:52:33

by Neil Armstrong

[permalink] [raw]
Subject: [PATCH 1/3] drm/meson: Support Overlay plane for video rendering

The Amlogic Meson GX SoCs support an Overlay plane behind the primary
plane for video rendering.

This Overlay plane support various YUV layouts :
- YUYV
- NV12 / NV21
- YUV444 / 422 / 420 / 411 / 410

The scaler supports a wide range of scaling ratios, but for simplicity,
plane atomic check limits the scaling from x5 to /5 in vertical and
horizontal scaling.

The z-order is fixed and always behind the primary plane and cannot be changed.

The scaling parameter algorithm was taken from the Amlogic vendor kernel
code and rewritten to match the atomic universal plane requirements.

The video rendering using this overlay plane support has been tested using
the new Kodi DRM-KMS Prime rendering path along the in-review V4L2 Mem2Mem
Hardware Video Decoder up to 3840x2160 NV12 frames on various display modes.

Signed-off-by: Neil Armstrong <[email protected]>
---
drivers/gpu/drm/meson/Makefile | 2 +-
drivers/gpu/drm/meson/meson_canvas.c | 7 +-
drivers/gpu/drm/meson/meson_canvas.h | 11 +-
drivers/gpu/drm/meson/meson_crtc.c | 216 +++++++++++-
drivers/gpu/drm/meson/meson_drv.c | 29 +-
drivers/gpu/drm/meson/meson_drv.h | 52 +++
drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
drivers/gpu/drm/meson/meson_overlay.h | 14 +
drivers/gpu/drm/meson/meson_registers.h | 3 +
drivers/gpu/drm/meson/meson_viu.c | 15 +
drivers/gpu/drm/meson/meson_vpp.c | 44 ++-
11 files changed, 971 insertions(+), 8 deletions(-)
create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
create mode 100644 drivers/gpu/drm/meson/meson_overlay.h

diff --git a/drivers/gpu/drm/meson/Makefile b/drivers/gpu/drm/meson/Makefile
index c5c4cc3..7709f2f 100644
--- a/drivers/gpu/drm/meson/Makefile
+++ b/drivers/gpu/drm/meson/Makefile
@@ -1,5 +1,5 @@
meson-drm-y := meson_drv.o meson_plane.o meson_crtc.o meson_venc_cvbs.o
-meson-drm-y += meson_viu.o meson_vpp.o meson_venc.o meson_vclk.o meson_canvas.o
+meson-drm-y += meson_viu.o meson_vpp.o meson_venc.o meson_vclk.o meson_canvas.o meson_overlay.o

obj-$(CONFIG_DRM_MESON) += meson-drm.o
obj-$(CONFIG_DRM_MESON_DW_HDMI) += meson_dw_hdmi.o
diff --git a/drivers/gpu/drm/meson/meson_canvas.c b/drivers/gpu/drm/meson/meson_canvas.c
index 08f6073..5de11aa 100644
--- a/drivers/gpu/drm/meson/meson_canvas.c
+++ b/drivers/gpu/drm/meson/meson_canvas.c
@@ -39,6 +39,7 @@
#define CANVAS_WIDTH_HBIT 0
#define CANVAS_HEIGHT_BIT 9
#define CANVAS_BLKMODE_BIT 24
+#define CANVAS_ENDIAN_BIT 26
#define DMC_CAV_LUT_ADDR 0x50 /* 0x14 offset in data sheet */
#define CANVAS_LUT_WR_EN (0x2 << 8)
#define CANVAS_LUT_RD_EN (0x1 << 8)
@@ -47,7 +48,8 @@ void meson_canvas_setup(struct meson_drm *priv,
uint32_t canvas_index, uint32_t addr,
uint32_t stride, uint32_t height,
unsigned int wrap,
- unsigned int blkmode)
+ unsigned int blkmode,
+ unsigned int endian)
{
unsigned int val;

@@ -60,7 +62,8 @@ void meson_canvas_setup(struct meson_drm *priv,
CANVAS_WIDTH_HBIT) |
(height << CANVAS_HEIGHT_BIT) |
(wrap << 22) |
- (blkmode << CANVAS_BLKMODE_BIT));
+ (blkmode << CANVAS_BLKMODE_BIT) |
+ (endian << CANVAS_ENDIAN_BIT));

regmap_write(priv->dmc, DMC_CAV_LUT_ADDR,
CANVAS_LUT_WR_EN | canvas_index);
diff --git a/drivers/gpu/drm/meson/meson_canvas.h b/drivers/gpu/drm/meson/meson_canvas.h
index af1759d..85dbf26 100644
--- a/drivers/gpu/drm/meson/meson_canvas.h
+++ b/drivers/gpu/drm/meson/meson_canvas.h
@@ -23,6 +23,9 @@
#define __MESON_CANVAS_H

#define MESON_CANVAS_ID_OSD1 0x4e
+#define MESON_CANVAS_ID_VD1_0 0x60
+#define MESON_CANVAS_ID_VD1_1 0x61
+#define MESON_CANVAS_ID_VD1_2 0x62

/* Canvas configuration. */
#define MESON_CANVAS_WRAP_NONE 0x00
@@ -33,10 +36,16 @@
#define MESON_CANVAS_BLKMODE_32x32 0x01
#define MESON_CANVAS_BLKMODE_64x64 0x02

+#define MESON_CANVAS_ENDIAN_SWAP16 0x1
+#define MESON_CANVAS_ENDIAN_SWAP32 0x3
+#define MESON_CANVAS_ENDIAN_SWAP64 0x7
+#define MESON_CANVAS_ENDIAN_SWAP128 0xf
+
void meson_canvas_setup(struct meson_drm *priv,
uint32_t canvas_index, uint32_t addr,
uint32_t stride, uint32_t height,
unsigned int wrap,
- unsigned int blkmode);
+ unsigned int blkmode,
+ unsigned int endian);

#endif /* __MESON_CANVAS_H */
diff --git a/drivers/gpu/drm/meson/meson_crtc.c b/drivers/gpu/drm/meson/meson_crtc.c
index b3bc0b0..38686c9 100644
--- a/drivers/gpu/drm/meson/meson_crtc.c
+++ b/drivers/gpu/drm/meson/meson_crtc.c
@@ -25,6 +25,7 @@
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/platform_device.h>
+#include <linux/bitfield.h>
#include <drm/drmP.h>
#include <drm/drm_atomic.h>
#include <drm/drm_atomic_helper.h>
@@ -98,6 +99,10 @@ static void meson_crtc_atomic_enable(struct drm_crtc *crtc,
writel(crtc_state->mode.hdisplay,
priv->io_base + _REG(VPP_POSTBLEND_H_SIZE));

+ /* VD1 Preblend vertical start/end */
+ writel(FIELD_PREP(GENMASK(11, 0), 2303),
+ priv->io_base + _REG(VPP_PREBLEND_VD1_V_START_END));
+
writel_bits_relaxed(VPP_POSTBLEND_ENABLE, VPP_POSTBLEND_ENABLE,
priv->io_base + _REG(VPP_MISC));

@@ -110,11 +115,17 @@ static void meson_crtc_atomic_disable(struct drm_crtc *crtc,
struct meson_crtc *meson_crtc = to_meson_crtc(crtc);
struct meson_drm *priv = meson_crtc->priv;

+ DRM_DEBUG_DRIVER("\n");
+
priv->viu.osd1_enabled = false;
priv->viu.osd1_commit = false;

+ priv->viu.vd1_enabled = false;
+ priv->viu.vd1_commit = false;
+
/* Disable VPP Postblend */
- writel_bits_relaxed(VPP_POSTBLEND_ENABLE, 0,
+ writel_bits_relaxed(VPP_OSD1_POSTBLEND | VPP_VD1_POSTBLEND |
+ VPP_VD1_PREBLEND | VPP_POSTBLEND_ENABLE, 0,
priv->io_base + _REG(VPP_MISC));

if (crtc->state->event && !crtc->state->active) {
@@ -149,6 +160,7 @@ static void meson_crtc_atomic_flush(struct drm_crtc *crtc,
struct meson_drm *priv = meson_crtc->priv;

priv->viu.osd1_commit = true;
+ priv->viu.vd1_commit = true;
}

static const struct drm_crtc_helper_funcs meson_crtc_helper_funcs = {
@@ -202,7 +214,7 @@ void meson_crtc_irq(struct meson_drm *priv)
meson_canvas_setup(priv, MESON_CANVAS_ID_OSD1,
priv->viu.osd1_addr, priv->viu.osd1_stride,
priv->viu.osd1_height, MESON_CANVAS_WRAP_NONE,
- MESON_CANVAS_BLKMODE_LINEAR);
+ MESON_CANVAS_BLKMODE_LINEAR, 0);

/* Enable OSD1 */
writel_bits_relaxed(VPP_OSD1_POSTBLEND, VPP_OSD1_POSTBLEND,
@@ -211,6 +223,206 @@ void meson_crtc_irq(struct meson_drm *priv)
priv->viu.osd1_commit = false;
}

+ /* Update the VD1 registers */
+ if (priv->viu.vd1_enabled && priv->viu.vd1_commit) {
+
+ switch (priv->viu.vd1_planes) {
+ case 3:
+ if (priv->canvas)
+ meson_canvas_config(priv->canvas,
+ priv->canvas_id_vd1_2,
+ priv->viu.vd1_addr2,
+ priv->viu.vd1_stride2,
+ priv->viu.vd1_height2,
+ MESON_CANVAS_WRAP_NONE,
+ MESON_CANVAS_BLKMODE_LINEAR,
+ MESON_CANVAS_ENDIAN_SWAP64);
+ else
+ meson_canvas_setup(priv, MESON_CANVAS_ID_VD1_2,
+ priv->viu.vd1_addr2,
+ priv->viu.vd1_stride2,
+ priv->viu.vd1_height2,
+ MESON_CANVAS_WRAP_NONE,
+ MESON_CANVAS_BLKMODE_LINEAR,
+ MESON_CANVAS_ENDIAN_SWAP64);
+ /* fallthrough */
+ case 2:
+ if (priv->canvas)
+ meson_canvas_config(priv->canvas,
+ priv->canvas_id_vd1_1,
+ priv->viu.vd1_addr1,
+ priv->viu.vd1_stride1,
+ priv->viu.vd1_height1,
+ MESON_CANVAS_WRAP_NONE,
+ MESON_CANVAS_BLKMODE_LINEAR,
+ MESON_CANVAS_ENDIAN_SWAP64);
+ else
+ meson_canvas_setup(priv, MESON_CANVAS_ID_VD1_1,
+ priv->viu.vd1_addr2,
+ priv->viu.vd1_stride2,
+ priv->viu.vd1_height2,
+ MESON_CANVAS_WRAP_NONE,
+ MESON_CANVAS_BLKMODE_LINEAR,
+ MESON_CANVAS_ENDIAN_SWAP64);
+ /* fallthrough */
+ case 1:
+ if (priv->canvas)
+ meson_canvas_config(priv->canvas,
+ priv->canvas_id_vd1_0,
+ priv->viu.vd1_addr0,
+ priv->viu.vd1_stride0,
+ priv->viu.vd1_height0,
+ MESON_CANVAS_WRAP_NONE,
+ MESON_CANVAS_BLKMODE_LINEAR,
+ MESON_CANVAS_ENDIAN_SWAP64);
+ else
+ meson_canvas_setup(priv, MESON_CANVAS_ID_VD1_0,
+ priv->viu.vd1_addr2,
+ priv->viu.vd1_stride2,
+ priv->viu.vd1_height2,
+ MESON_CANVAS_WRAP_NONE,
+ MESON_CANVAS_BLKMODE_LINEAR,
+ MESON_CANVAS_ENDIAN_SWAP64);
+ };
+
+ writel_relaxed(priv->viu.vd1_if0_gen_reg,
+ priv->io_base + _REG(VD1_IF0_GEN_REG));
+ writel_relaxed(priv->viu.vd1_if0_gen_reg,
+ priv->io_base + _REG(VD2_IF0_GEN_REG));
+ writel_relaxed(priv->viu.vd1_if0_gen_reg2,
+ priv->io_base + _REG(VD1_IF0_GEN_REG2));
+ writel_relaxed(priv->viu.viu_vd1_fmt_ctrl,
+ priv->io_base + _REG(VIU_VD1_FMT_CTRL));
+ writel_relaxed(priv->viu.viu_vd1_fmt_ctrl,
+ priv->io_base + _REG(VIU_VD2_FMT_CTRL));
+ writel_relaxed(priv->viu.viu_vd1_fmt_w,
+ priv->io_base + _REG(VIU_VD1_FMT_W));
+ writel_relaxed(priv->viu.viu_vd1_fmt_w,
+ priv->io_base + _REG(VIU_VD2_FMT_W));
+ writel_relaxed(priv->viu.vd1_if0_canvas0,
+ priv->io_base + _REG(VD1_IF0_CANVAS0));
+ writel_relaxed(priv->viu.vd1_if0_canvas0,
+ priv->io_base + _REG(VD1_IF0_CANVAS1));
+ writel_relaxed(priv->viu.vd1_if0_canvas0,
+ priv->io_base + _REG(VD2_IF0_CANVAS0));
+ writel_relaxed(priv->viu.vd1_if0_canvas0,
+ priv->io_base + _REG(VD2_IF0_CANVAS1));
+ writel_relaxed(priv->viu.vd1_if0_luma_x0,
+ priv->io_base + _REG(VD1_IF0_LUMA_X0));
+ writel_relaxed(priv->viu.vd1_if0_luma_x0,
+ priv->io_base + _REG(VD1_IF0_LUMA_X1));
+ writel_relaxed(priv->viu.vd1_if0_luma_x0,
+ priv->io_base + _REG(VD2_IF0_LUMA_X0));
+ writel_relaxed(priv->viu.vd1_if0_luma_x0,
+ priv->io_base + _REG(VD2_IF0_LUMA_X1));
+ writel_relaxed(priv->viu.vd1_if0_luma_y0,
+ priv->io_base + _REG(VD1_IF0_LUMA_Y0));
+ writel_relaxed(priv->viu.vd1_if0_luma_y0,
+ priv->io_base + _REG(VD1_IF0_LUMA_Y1));
+ writel_relaxed(priv->viu.vd1_if0_luma_y0,
+ priv->io_base + _REG(VD2_IF0_LUMA_Y0));
+ writel_relaxed(priv->viu.vd1_if0_luma_y0,
+ priv->io_base + _REG(VD2_IF0_LUMA_Y1));
+ writel_relaxed(priv->viu.vd1_if0_chroma_x0,
+ priv->io_base + _REG(VD1_IF0_CHROMA_X0));
+ writel_relaxed(priv->viu.vd1_if0_chroma_x0,
+ priv->io_base + _REG(VD1_IF0_CHROMA_X1));
+ writel_relaxed(priv->viu.vd1_if0_chroma_x0,
+ priv->io_base + _REG(VD2_IF0_CHROMA_X0));
+ writel_relaxed(priv->viu.vd1_if0_chroma_x0,
+ priv->io_base + _REG(VD2_IF0_CHROMA_X1));
+ writel_relaxed(priv->viu.vd1_if0_chroma_y0,
+ priv->io_base + _REG(VD1_IF0_CHROMA_Y0));
+ writel_relaxed(priv->viu.vd1_if0_chroma_y0,
+ priv->io_base + _REG(VD1_IF0_CHROMA_Y1));
+ writel_relaxed(priv->viu.vd1_if0_chroma_y0,
+ priv->io_base + _REG(VD2_IF0_CHROMA_Y0));
+ writel_relaxed(priv->viu.vd1_if0_chroma_y0,
+ priv->io_base + _REG(VD2_IF0_CHROMA_Y1));
+ writel_relaxed(priv->viu.vd1_if0_repeat_loop,
+ priv->io_base + _REG(VD1_IF0_RPT_LOOP));
+ writel_relaxed(priv->viu.vd1_if0_repeat_loop,
+ priv->io_base + _REG(VD2_IF0_RPT_LOOP));
+ writel_relaxed(priv->viu.vd1_if0_luma0_rpt_pat,
+ priv->io_base + _REG(VD1_IF0_LUMA0_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_luma0_rpt_pat,
+ priv->io_base + _REG(VD2_IF0_LUMA0_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_luma0_rpt_pat,
+ priv->io_base + _REG(VD1_IF0_LUMA1_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_luma0_rpt_pat,
+ priv->io_base + _REG(VD2_IF0_LUMA1_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_chroma0_rpt_pat,
+ priv->io_base + _REG(VD1_IF0_CHROMA0_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_chroma0_rpt_pat,
+ priv->io_base + _REG(VD2_IF0_CHROMA0_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_chroma0_rpt_pat,
+ priv->io_base + _REG(VD1_IF0_CHROMA1_RPT_PAT));
+ writel_relaxed(priv->viu.vd1_if0_chroma0_rpt_pat,
+ priv->io_base + _REG(VD2_IF0_CHROMA1_RPT_PAT));
+ writel_relaxed(0, priv->io_base + _REG(VD1_IF0_LUMA_PSEL));
+ writel_relaxed(0, priv->io_base + _REG(VD1_IF0_CHROMA_PSEL));
+ writel_relaxed(0, priv->io_base + _REG(VD2_IF0_LUMA_PSEL));
+ writel_relaxed(0, priv->io_base + _REG(VD2_IF0_CHROMA_PSEL));
+ writel_relaxed(priv->viu.vd1_range_map_y,
+ priv->io_base + _REG(VD1_IF0_RANGE_MAP_Y));
+ writel_relaxed(priv->viu.vd1_range_map_cb,
+ priv->io_base + _REG(VD1_IF0_RANGE_MAP_CB));
+ writel_relaxed(priv->viu.vd1_range_map_cr,
+ priv->io_base + _REG(VD1_IF0_RANGE_MAP_CR));
+ writel_relaxed(0x78404,
+ priv->io_base + _REG(VPP_SC_MISC));
+ writel_relaxed(priv->viu.vpp_pic_in_height,
+ priv->io_base + _REG(VPP_PIC_IN_HEIGHT));
+ writel_relaxed(priv->viu.vpp_postblend_vd1_h_start_end,
+ priv->io_base + _REG(VPP_POSTBLEND_VD1_H_START_END));
+ writel_relaxed(priv->viu.vpp_blend_vd2_h_start_end,
+ priv->io_base + _REG(VPP_BLEND_VD2_H_START_END));
+ writel_relaxed(priv->viu.vpp_postblend_vd1_v_start_end,
+ priv->io_base + _REG(VPP_POSTBLEND_VD1_V_START_END));
+ writel_relaxed(priv->viu.vpp_blend_vd2_v_start_end,
+ priv->io_base + _REG(VPP_BLEND_VD2_V_START_END));
+ writel_relaxed(priv->viu.vpp_hsc_region12_startp,
+ priv->io_base + _REG(VPP_HSC_REGION12_STARTP));
+ writel_relaxed(priv->viu.vpp_hsc_region34_startp,
+ priv->io_base + _REG(VPP_HSC_REGION34_STARTP));
+ writel_relaxed(priv->viu.vpp_hsc_region4_endp,
+ priv->io_base + _REG(VPP_HSC_REGION4_ENDP));
+ writel_relaxed(priv->viu.vpp_hsc_start_phase_step,
+ priv->io_base + _REG(VPP_HSC_START_PHASE_STEP));
+ writel_relaxed(priv->viu.vpp_hsc_region1_phase_slope,
+ priv->io_base + _REG(VPP_HSC_REGION1_PHASE_SLOPE));
+ writel_relaxed(priv->viu.vpp_hsc_region3_phase_slope,
+ priv->io_base + _REG(VPP_HSC_REGION3_PHASE_SLOPE));
+ writel_relaxed(priv->viu.vpp_line_in_length,
+ priv->io_base + _REG(VPP_LINE_IN_LENGTH));
+ writel_relaxed(priv->viu.vpp_preblend_h_size,
+ priv->io_base + _REG(VPP_PREBLEND_H_SIZE));
+ writel_relaxed(priv->viu.vpp_vsc_region12_startp,
+ priv->io_base + _REG(VPP_VSC_REGION12_STARTP));
+ writel_relaxed(priv->viu.vpp_vsc_region34_startp,
+ priv->io_base + _REG(VPP_VSC_REGION34_STARTP));
+ writel_relaxed(priv->viu.vpp_vsc_region4_endp,
+ priv->io_base + _REG(VPP_VSC_REGION4_ENDP));
+ writel_relaxed(priv->viu.vpp_vsc_start_phase_step,
+ priv->io_base + _REG(VPP_VSC_START_PHASE_STEP));
+ writel_relaxed(priv->viu.vpp_vsc_ini_phase,
+ priv->io_base + _REG(VPP_VSC_INI_PHASE));
+ writel_relaxed(priv->viu.vpp_vsc_phase_ctrl,
+ priv->io_base + _REG(VPP_VSC_PHASE_CTRL));
+ writel_relaxed(priv->viu.vpp_hsc_phase_ctrl,
+ priv->io_base + _REG(VPP_HSC_PHASE_CTRL));
+ writel_relaxed(0x42, priv->io_base + _REG(VPP_SCALE_COEF_IDX));
+
+ /* Enable VD1 */
+ writel_bits_relaxed(VPP_VD1_PREBLEND | VPP_VD1_POSTBLEND |
+ VPP_COLOR_MNG_ENABLE,
+ VPP_VD1_PREBLEND | VPP_VD1_POSTBLEND |
+ VPP_COLOR_MNG_ENABLE,
+ priv->io_base + _REG(VPP_MISC));
+
+ priv->viu.vd1_commit = false;
+ }
+
drm_crtc_handle_vblank(priv->crtc);

spin_lock_irqsave(&priv->drm->event_lock, flags);
diff --git a/drivers/gpu/drm/meson/meson_drv.c b/drivers/gpu/drm/meson/meson_drv.c
index b39c38c..b55e03d 100644
--- a/drivers/gpu/drm/meson/meson_drv.c
+++ b/drivers/gpu/drm/meson/meson_drv.c
@@ -41,6 +41,7 @@

#include "meson_drv.h"
#include "meson_plane.h"
+#include "meson_overlay.h"
#include "meson_crtc.h"
#include "meson_venc_cvbs.h"

@@ -221,6 +222,24 @@ static int meson_drv_bind_master(struct device *dev, bool has_components)
ret = meson_canvas_alloc(priv->canvas, &priv->canvas_id_osd1);
if (ret)
goto free_drm;
+ ret = meson_canvas_alloc(priv->canvas, &priv->canvas_id_vd1_0);
+ if (ret) {
+ meson_canvas_free(priv->canvas, priv->canvas_id_osd1);
+ goto free_drm;
+ }
+ ret = meson_canvas_alloc(priv->canvas, &priv->canvas_id_vd1_1);
+ if (ret) {
+ meson_canvas_free(priv->canvas, priv->canvas_id_osd1);
+ meson_canvas_free(priv->canvas, priv->canvas_id_vd1_0);
+ goto free_drm;
+ }
+ ret = meson_canvas_alloc(priv->canvas, &priv->canvas_id_vd1_2);
+ if (ret) {
+ meson_canvas_free(priv->canvas, priv->canvas_id_osd1);
+ meson_canvas_free(priv->canvas, priv->canvas_id_vd1_0);
+ meson_canvas_free(priv->canvas, priv->canvas_id_vd1_1);
+ goto free_drm;
+ }
} else {
priv->canvas = NULL;

@@ -281,6 +300,10 @@ static int meson_drv_bind_master(struct device *dev, bool has_components)
if (ret)
goto free_drm;

+ ret = meson_overlay_create(priv);
+ if (ret)
+ goto free_drm;
+
ret = meson_crtc_create(priv);
if (ret)
goto free_drm;
@@ -324,8 +347,12 @@ static void meson_drv_unbind(struct device *dev)
struct drm_device *drm = dev_get_drvdata(dev);
struct meson_drm *priv = drm->dev_private;

- if (priv->canvas)
+ if (priv->canvas) {
meson_canvas_free(priv->canvas, priv->canvas_id_osd1);
+ meson_canvas_free(priv->canvas, priv->canvas_id_vd1_0);
+ meson_canvas_free(priv->canvas, priv->canvas_id_vd1_1);
+ meson_canvas_free(priv->canvas, priv->canvas_id_vd1_2);
+ }

drm_dev_unregister(drm);
drm_kms_helper_poll_fini(drm);
diff --git a/drivers/gpu/drm/meson/meson_drv.h b/drivers/gpu/drm/meson/meson_drv.h
index 728d0ca..c971557 100644
--- a/drivers/gpu/drm/meson/meson_drv.h
+++ b/drivers/gpu/drm/meson/meson_drv.h
@@ -34,11 +34,15 @@ struct meson_drm {

struct meson_canvas *canvas;
u8 canvas_id_osd1;
+ u8 canvas_id_vd1_0;
+ u8 canvas_id_vd1_1;
+ u8 canvas_id_vd1_2;

struct drm_device *drm;
struct drm_crtc *crtc;
struct drm_fbdev_cma *fbdev;
struct drm_plane *primary_plane;
+ struct drm_plane *overlay_plane;

/* Components Data */
struct {
@@ -50,6 +54,54 @@ struct meson_drm {
uint32_t osd1_addr;
uint32_t osd1_stride;
uint32_t osd1_height;
+
+ bool vd1_enabled;
+ bool vd1_commit;
+ unsigned int vd1_planes;
+ uint32_t vd1_if0_gen_reg;
+ uint32_t vd1_if0_luma_x0;
+ uint32_t vd1_if0_luma_y0;
+ uint32_t vd1_if0_chroma_x0;
+ uint32_t vd1_if0_chroma_y0;
+ uint32_t vd1_if0_repeat_loop;
+ uint32_t vd1_if0_luma0_rpt_pat;
+ uint32_t vd1_if0_chroma0_rpt_pat;
+ uint32_t vd1_range_map_y;
+ uint32_t vd1_range_map_cb;
+ uint32_t vd1_range_map_cr;
+ uint32_t viu_vd1_fmt_w;
+ uint32_t vd1_if0_canvas0;
+ uint32_t vd1_if0_gen_reg2;
+ uint32_t viu_vd1_fmt_ctrl;
+ uint32_t vd1_addr0;
+ uint32_t vd1_addr1;
+ uint32_t vd1_addr2;
+ uint32_t vd1_stride0;
+ uint32_t vd1_stride1;
+ uint32_t vd1_stride2;
+ uint32_t vd1_height0;
+ uint32_t vd1_height1;
+ uint32_t vd1_height2;
+ uint32_t vpp_pic_in_height;
+ uint32_t vpp_postblend_vd1_h_start_end;
+ uint32_t vpp_postblend_vd1_v_start_end;
+ uint32_t vpp_hsc_region12_startp;
+ uint32_t vpp_hsc_region34_startp;
+ uint32_t vpp_hsc_region4_endp;
+ uint32_t vpp_hsc_start_phase_step;
+ uint32_t vpp_hsc_region1_phase_slope;
+ uint32_t vpp_hsc_region3_phase_slope;
+ uint32_t vpp_line_in_length;
+ uint32_t vpp_preblend_h_size;
+ uint32_t vpp_vsc_region12_startp;
+ uint32_t vpp_vsc_region34_startp;
+ uint32_t vpp_vsc_region4_endp;
+ uint32_t vpp_vsc_start_phase_step;
+ uint32_t vpp_vsc_ini_phase;
+ uint32_t vpp_vsc_phase_ctrl;
+ uint32_t vpp_hsc_phase_ctrl;
+ uint32_t vpp_blend_vd2_h_start_end;
+ uint32_t vpp_blend_vd2_v_start_end;
} viu;

struct {
diff --git a/drivers/gpu/drm/meson/meson_overlay.c b/drivers/gpu/drm/meson/meson_overlay.c
new file mode 100644
index 0000000..9aebc5e
--- /dev/null
+++ b/drivers/gpu/drm/meson/meson_overlay.c
@@ -0,0 +1,586 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2018 BayLibre, SAS
+ * Author: Neil Armstrong <[email protected]>
+ * Copyright (C) 2015 Amlogic, Inc. All rights reserved.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/bitfield.h>
+#include <linux/platform_device.h>
+#include <drm/drmP.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_plane_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_rect.h>
+
+#include "meson_overlay.h"
+#include "meson_vpp.h"
+#include "meson_viu.h"
+#include "meson_canvas.h"
+#include "meson_registers.h"
+
+/* VD1_IF0_GEN_REG */
+#define VD_URGENT_CHROMA BIT(28)
+#define VD_URGENT_LUMA BIT(27)
+#define VD_HOLD_LINES(lines) FIELD_PREP(GENMASK(24, 19), lines)
+#define VD_DEMUX_MODE_RGB BIT(16)
+#define VD_BYTES_PER_PIXEL(val) FIELD_PREP(GENMASK(15, 14), val)
+#define VD_CHRO_RPT_LASTL_CTRL BIT(6)
+#define VD_LITTLE_ENDIAN BIT(4)
+#define VD_SEPARATE_EN BIT(1)
+#define VD_ENABLE BIT(0)
+
+/* VD1_IF0_CANVAS0 */
+#define CANVAS_ADDR2(addr) FIELD_PREP(GENMASK(23, 16), addr)
+#define CANVAS_ADDR1(addr) FIELD_PREP(GENMASK(15, 8), addr)
+#define CANVAS_ADDR0(addr) FIELD_PREP(GENMASK(7, 0), addr)
+
+/* VD1_IF0_LUMA_X0 VD1_IF0_CHROMA_X0 */
+#define VD_X_START(value) FIELD_PREP(GENMASK(14, 0), value)
+#define VD_X_END(value) FIELD_PREP(GENMASK(30, 16), value)
+
+/* VD1_IF0_LUMA_Y0 VD1_IF0_CHROMA_Y0 */
+#define VD_Y_START(value) FIELD_PREP(GENMASK(12, 0), value)
+#define VD_Y_END(value) FIELD_PREP(GENMASK(28, 16), value)
+
+/* VD1_IF0_GEN_REG2 */
+#define VD_COLOR_MAP(value) FIELD_PREP(GENMASK(1, 0), value)
+
+/* VIU_VD1_FMT_CTRL */
+#define VD_HORZ_Y_C_RATIO(value) FIELD_PREP(GENMASK(22, 21), value)
+#define VD_HORZ_FMT_EN BIT(20)
+#define VD_VERT_RPT_LINE0 BIT(16)
+#define VD_VERT_INITIAL_PHASE(value) FIELD_PREP(GENMASK(11, 8), value)
+#define VD_VERT_PHASE_STEP(value) FIELD_PREP(GENMASK(7, 1), value)
+#define VD_VERT_FMT_EN BIT(0)
+
+/* VPP_POSTBLEND_VD1_H_START_END */
+#define VD_H_END(value) FIELD_PREP(GENMASK(11, 0), value)
+#define VD_H_START(value) FIELD_PREP(GENMASK(27, 16), value)
+
+/* VPP_POSTBLEND_VD1_V_START_END */
+#define VD_V_END(value) FIELD_PREP(GENMASK(11, 0), value)
+#define VD_V_START(value) FIELD_PREP(GENMASK(27, 16), value)
+
+/* VPP_BLEND_VD2_V_START_END */
+#define VD2_V_END(value) FIELD_PREP(GENMASK(11, 0), value)
+#define VD2_V_START(value) FIELD_PREP(GENMASK(27, 16), value)
+
+/* VIU_VD1_FMT_W */
+#define VD_V_WIDTH(value) FIELD_PREP(GENMASK(11, 0), value)
+#define VD_H_WIDTH(value) FIELD_PREP(GENMASK(27, 16), value)
+
+/* VPP_HSC_REGION12_STARTP VPP_HSC_REGION34_STARTP */
+#define VD_REGION24_START(value) FIELD_PREP(GENMASK(11, 0), value)
+#define VD_REGION13_END(value) FIELD_PREP(GENMASK(27, 16), value)
+
+struct meson_overlay {
+ struct drm_plane base;
+ struct meson_drm *priv;
+};
+#define to_meson_overlay(x) container_of(x, struct meson_overlay, base)
+
+#define FRAC_16_16(mult, div) (((mult) << 16) / (div))
+
+static int meson_overlay_atomic_check(struct drm_plane *plane,
+ struct drm_plane_state *state)
+{
+ struct drm_crtc_state *crtc_state;
+
+ if (!state->crtc)
+ return 0;
+
+ crtc_state = drm_atomic_get_crtc_state(state->state, state->crtc);
+ if (IS_ERR(crtc_state))
+ return PTR_ERR(crtc_state);
+
+ return drm_atomic_helper_check_plane_state(state, crtc_state,
+ FRAC_16_16(1, 5),
+ FRAC_16_16(5, 1),
+ true, true);
+}
+
+/* Takes a fixed 16.16 number and converts it to integer. */
+static inline int64_t fixed16_to_int(int64_t value)
+{
+ return value >> 16;
+}
+
+static const uint8_t skip_tab[6] = {
+ 0x24, 0x04, 0x68, 0x48, 0x28, 0x08,
+};
+
+static void meson_overlay_get_vertical_phase(unsigned int ratio_y, int *phase,
+ int *repeat, bool interlace)
+{
+ int offset_in = 0;
+ int offset_out = 0;
+ int repeat_skip = 0;
+
+ if (!interlace && ratio_y > (1 << 18))
+ offset_out = (1 * ratio_y) >> 10;
+
+ while ((offset_in + (4 << 8)) <= offset_out) {
+ repeat_skip++;
+ offset_in += 4 << 8;
+ }
+
+ *phase = (offset_out - offset_in) >> 2;
+
+ if (*phase > 0x100)
+ repeat_skip++;
+
+ *phase = *phase & 0xff;
+
+ if (repeat_skip > 5)
+ repeat_skip = 5;
+
+ *repeat = skip_tab[repeat_skip];
+}
+
+static void meson_overlay_setup_scaler_params(struct meson_drm *priv,
+ struct drm_plane *plane,
+ bool interlace_mode)
+{
+ struct drm_crtc_state *crtc_state = priv->crtc->state;
+ int video_top, video_left, video_width, video_height;
+ struct drm_plane_state *state = plane->state;
+ unsigned int vd_start_lines, vd_end_lines;
+ unsigned int hd_start_lines, hd_end_lines;
+ unsigned int crtc_height, crtc_width;
+ unsigned int vsc_startp, vsc_endp;
+ unsigned int hsc_startp, hsc_endp;
+ unsigned int crop_top, crop_left;
+ int vphase, vphase_repeat_skip;
+ unsigned int ratio_x, ratio_y;
+ int temp_height, temp_width;
+ unsigned int w_in, h_in;
+ int temp, start, end;
+
+ if (!crtc_state) {
+ DRM_ERROR("Invalid crtc_state\n");
+ return;
+ }
+
+ crtc_height = crtc_state->mode.vdisplay;
+ crtc_width = crtc_state->mode.hdisplay;
+
+ w_in = fixed16_to_int(state->src_w);
+ h_in = fixed16_to_int(state->src_h);
+ crop_top = fixed16_to_int(state->src_x);
+ crop_left = fixed16_to_int(state->src_x);
+
+ video_top = state->crtc_y;
+ video_left = state->crtc_x;
+ video_width = state->crtc_w;
+ video_height = state->crtc_h;
+
+ DRM_DEBUG("crtc_width %d crtc_height %d interlace %d\n",
+ crtc_width, crtc_height, interlace_mode);
+ DRM_DEBUG("w_in %d h_in %d crop_top %d crop_left %d\n",
+ w_in, h_in, crop_top, crop_left);
+ DRM_DEBUG("video top %d left %d width %d height %d\n",
+ video_top, video_left, video_width, video_height);
+
+ ratio_x = (w_in << 18) / video_width;
+ ratio_y = (h_in << 18) / video_height;
+
+ if (ratio_x * video_width < (w_in << 18))
+ ratio_x++;
+
+ DRM_DEBUG("ratio x 0x%x y 0x%x\n", ratio_x, ratio_y);
+
+ meson_overlay_get_vertical_phase(ratio_y, &vphase, &vphase_repeat_skip,
+ interlace_mode);
+
+ DRM_DEBUG("vphase 0x%x skip %d\n", vphase, vphase_repeat_skip);
+
+ /* Vertical */
+
+ start = video_top + video_height / 2 - ((h_in << 17) / ratio_y);
+ end = (h_in << 18) / ratio_y + start - 1;
+
+ if (video_top < 0 && start < 0)
+ vd_start_lines = (-(start) * ratio_y) >> 18;
+ else if (start < video_top)
+ vd_start_lines = ((video_top - start) * ratio_y) >> 18;
+ else
+ vd_start_lines = 0;
+
+ if (video_top < 0)
+ temp_height = min_t(unsigned int,
+ video_top + video_height - 1,
+ crtc_height - 1);
+ else
+ temp_height = min_t(unsigned int,
+ video_top + video_height - 1,
+ crtc_height - 1) - video_top + 1;
+
+ temp = vd_start_lines + (temp_height * ratio_y >> 18);
+ vd_end_lines = (temp <= (h_in - 1)) ? temp : (h_in - 1);
+
+ vd_start_lines += crop_left;
+ vd_end_lines += crop_left;
+
+ /*
+ * TOFIX: Input frames are handled and scaled like progressive frames,
+ * proper handling of interlaced field input frames need to be figured
+ * out using the proper framebuffer flags set by userspace.
+ */
+ if (interlace_mode) {
+ start >>= 1;
+ end >>= 1;
+ }
+
+ vsc_startp = max_t(int, start,
+ max_t(int, 0, video_top));
+ vsc_endp = min_t(int, end,
+ min_t(int, crtc_height - 1,
+ video_top + video_height - 1));
+
+ DRM_DEBUG("vsc startp %d endp %d start_lines %d end_lines %d\n",
+ vsc_startp, vsc_endp, vd_start_lines, vd_end_lines);
+
+ /* Horizontal */
+
+ start = video_left + video_width / 2 - ((w_in << 17) / ratio_x);
+ end = (w_in << 18) / ratio_x + start - 1;
+
+ if (video_left < 0 && start < 0)
+ hd_start_lines = (-(start) * ratio_x) >> 18;
+ else if (start < video_left)
+ hd_start_lines = ((video_left - start) * ratio_x) >> 18;
+ else
+ hd_start_lines = 0;
+
+ if (video_left < 0)
+ temp_width = min_t(unsigned int,
+ video_left + video_width - 1,
+ crtc_width - 1);
+ else
+ temp_width = min_t(unsigned int,
+ video_left + video_width - 1,
+ crtc_width - 1) - video_left + 1;
+
+ temp = hd_start_lines + (temp_width * ratio_x >> 18);
+ hd_end_lines = (temp <= (w_in - 1)) ? temp : (w_in - 1);
+
+ priv->viu.vpp_line_in_length = hd_end_lines - hd_start_lines + 1;
+ hsc_startp = max_t(int, start, max_t(int, 0, video_left));
+ hsc_endp = min_t(int, end, min_t(int, crtc_width - 1,
+ video_left + video_width - 1));
+
+ hd_start_lines += crop_top;
+ hd_end_lines += crop_top;
+
+ DRM_DEBUG("hsc startp %d endp %d start_lines %d end_lines %d\n",
+ hsc_startp, hsc_endp, hd_start_lines, hd_end_lines);
+
+ priv->viu.vpp_vsc_start_phase_step = ratio_y << 6;
+
+ priv->viu.vpp_vsc_ini_phase = vphase << 8;
+ priv->viu.vpp_vsc_phase_ctrl = (1 << 13) | (4 << 8) |
+ vphase_repeat_skip;
+
+ priv->viu.vd1_if0_luma_x0 = VD_X_START(hd_start_lines) |
+ VD_X_END(hd_end_lines);
+ priv->viu.vd1_if0_chroma_x0 = VD_X_START(hd_start_lines >> 1) |
+ VD_X_END(hd_end_lines >> 1);
+
+ priv->viu.viu_vd1_fmt_w =
+ VD_H_WIDTH(hd_end_lines - hd_start_lines + 1) |
+ VD_V_WIDTH(hd_end_lines/2 - hd_start_lines/2 + 1);
+
+ priv->viu.vd1_if0_luma_y0 = VD_Y_START(vd_start_lines) |
+ VD_Y_END(vd_end_lines);
+
+ priv->viu.vd1_if0_chroma_y0 = VD_Y_START(vd_start_lines >> 1) |
+ VD_Y_END(vd_end_lines >> 1);
+
+ priv->viu.vpp_pic_in_height = h_in;
+
+ priv->viu.vpp_postblend_vd1_h_start_end = VD_H_START(hsc_startp) |
+ VD_H_END(hsc_endp);
+ priv->viu.vpp_blend_vd2_h_start_end = VD_H_START(hd_start_lines) |
+ VD_H_END(hd_end_lines);
+ priv->viu.vpp_hsc_region12_startp = VD_REGION13_END(0) |
+ VD_REGION24_START(hsc_startp);
+ priv->viu.vpp_hsc_region34_startp =
+ VD_REGION13_END(hsc_startp) |
+ VD_REGION24_START(hsc_endp - hsc_startp);
+ priv->viu.vpp_hsc_region4_endp = hsc_endp - hsc_startp;
+ priv->viu.vpp_hsc_start_phase_step = ratio_x << 6;
+ priv->viu.vpp_hsc_region1_phase_slope = 0;
+ priv->viu.vpp_hsc_region3_phase_slope = 0;
+ priv->viu.vpp_hsc_phase_ctrl = (1 << 21) | (4 << 16);
+
+ priv->viu.vpp_line_in_length = hd_end_lines - hd_start_lines + 1;
+ priv->viu.vpp_preblend_h_size = hd_end_lines - hd_start_lines + 1;
+
+ priv->viu.vpp_postblend_vd1_v_start_end = VD_V_START(vsc_startp) |
+ VD_V_END(vsc_endp);
+ priv->viu.vpp_blend_vd2_v_start_end =
+ VD2_V_START((vd_end_lines + 1) >> 1) |
+ VD2_V_END(vd_end_lines);
+
+ priv->viu.vpp_vsc_region12_startp = 0;
+ priv->viu.vpp_vsc_region34_startp =
+ VD_REGION13_END(vsc_endp - vsc_startp) |
+ VD_REGION24_START(vsc_endp - vsc_startp);
+ priv->viu.vpp_vsc_region4_endp = vsc_endp - vsc_startp;
+ priv->viu.vpp_vsc_start_phase_step = ratio_y << 6;
+}
+
+static void meson_overlay_atomic_update(struct drm_plane *plane,
+ struct drm_plane_state *old_state)
+{
+ struct meson_overlay *meson_overlay = to_meson_overlay(plane);
+ struct drm_plane_state *state = plane->state;
+ struct drm_framebuffer *fb = state->fb;
+ struct meson_drm *priv = meson_overlay->priv;
+ struct drm_gem_cma_object *gem;
+ unsigned long flags;
+ bool interlace_mode;
+
+ DRM_DEBUG_DRIVER("\n");
+
+ /* Fallback is canvas provider is not available */
+ if (!priv->canvas) {
+ priv->canvas_id_vd1_0 = MESON_CANVAS_ID_VD1_0;
+ priv->canvas_id_vd1_1 = MESON_CANVAS_ID_VD1_1;
+ priv->canvas_id_vd1_2 = MESON_CANVAS_ID_VD1_2;
+ }
+
+ interlace_mode = state->crtc->mode.flags & DRM_MODE_FLAG_INTERLACE;
+
+ spin_lock_irqsave(&priv->drm->event_lock, flags);
+
+ priv->viu.vd1_if0_gen_reg = VD_URGENT_CHROMA |
+ VD_URGENT_LUMA |
+ VD_HOLD_LINES(9) |
+ VD_CHRO_RPT_LASTL_CTRL |
+ VD_ENABLE;
+
+ /* Setup scaler params */
+ meson_overlay_setup_scaler_params(priv, plane, interlace_mode);
+
+ priv->viu.vd1_if0_repeat_loop = 0;
+ priv->viu.vd1_if0_luma0_rpt_pat = interlace_mode ? 8 : 0;
+ priv->viu.vd1_if0_chroma0_rpt_pat = interlace_mode ? 8 : 0;
+ priv->viu.vd1_range_map_y = 0;
+ priv->viu.vd1_range_map_cb = 0;
+ priv->viu.vd1_range_map_cr = 0;
+
+ /* Default values for RGB888/YUV444 */
+ priv->viu.vd1_if0_gen_reg2 = 0;
+ priv->viu.viu_vd1_fmt_ctrl = 0;
+
+ switch (fb->format->format) {
+ /* TOFIX DRM_FORMAT_RGB888 should be supported */
+ case DRM_FORMAT_YUYV:
+ priv->viu.vd1_if0_gen_reg |= VD_BYTES_PER_PIXEL(1);
+ priv->viu.vd1_if0_canvas0 =
+ CANVAS_ADDR2(priv->canvas_id_vd1_0) |
+ CANVAS_ADDR1(priv->canvas_id_vd1_0) |
+ CANVAS_ADDR0(priv->canvas_id_vd1_0);
+ priv->viu.viu_vd1_fmt_ctrl = VD_HORZ_Y_C_RATIO(1) | /* /2 */
+ VD_HORZ_FMT_EN |
+ VD_VERT_RPT_LINE0 |
+ VD_VERT_INITIAL_PHASE(12) |
+ VD_VERT_PHASE_STEP(16) | /* /2 */
+ VD_VERT_FMT_EN;
+ break;
+ case DRM_FORMAT_NV12:
+ case DRM_FORMAT_NV21:
+ priv->viu.vd1_if0_gen_reg |= VD_SEPARATE_EN;
+ priv->viu.vd1_if0_canvas0 =
+ CANVAS_ADDR2(priv->canvas_id_vd1_1) |
+ CANVAS_ADDR1(priv->canvas_id_vd1_1) |
+ CANVAS_ADDR0(priv->canvas_id_vd1_0);
+ if (fb->format->format == DRM_FORMAT_NV12)
+ priv->viu.vd1_if0_gen_reg2 = VD_COLOR_MAP(1);
+ else
+ priv->viu.vd1_if0_gen_reg2 = VD_COLOR_MAP(2);
+ priv->viu.viu_vd1_fmt_ctrl = VD_HORZ_Y_C_RATIO(1) | /* /2 */
+ VD_HORZ_FMT_EN |
+ VD_VERT_RPT_LINE0 |
+ VD_VERT_INITIAL_PHASE(12) |
+ VD_VERT_PHASE_STEP(8) | /* /4 */
+ VD_VERT_FMT_EN;
+ break;
+ case DRM_FORMAT_YUV444:
+ case DRM_FORMAT_YUV422:
+ case DRM_FORMAT_YUV420:
+ case DRM_FORMAT_YUV411:
+ case DRM_FORMAT_YUV410:
+ priv->viu.vd1_if0_gen_reg |= VD_SEPARATE_EN;
+ priv->viu.vd1_if0_canvas0 =
+ CANVAS_ADDR2(priv->canvas_id_vd1_2) |
+ CANVAS_ADDR1(priv->canvas_id_vd1_1) |
+ CANVAS_ADDR0(priv->canvas_id_vd1_0);
+ switch (fb->format->format) {
+ case DRM_FORMAT_YUV422:
+ priv->viu.viu_vd1_fmt_ctrl =
+ VD_HORZ_Y_C_RATIO(1) | /* /2 */
+ VD_HORZ_FMT_EN |
+ VD_VERT_RPT_LINE0 |
+ VD_VERT_INITIAL_PHASE(12) |
+ VD_VERT_PHASE_STEP(16) | /* /2 */
+ VD_VERT_FMT_EN;
+ break;
+ case DRM_FORMAT_YUV420:
+ priv->viu.viu_vd1_fmt_ctrl =
+ VD_HORZ_Y_C_RATIO(1) | /* /2 */
+ VD_HORZ_FMT_EN |
+ VD_VERT_RPT_LINE0 |
+ VD_VERT_INITIAL_PHASE(12) |
+ VD_VERT_PHASE_STEP(8) | /* /4 */
+ VD_VERT_FMT_EN;
+ break;
+ case DRM_FORMAT_YUV411:
+ priv->viu.viu_vd1_fmt_ctrl =
+ VD_HORZ_Y_C_RATIO(2) | /* /4 */
+ VD_HORZ_FMT_EN |
+ VD_VERT_RPT_LINE0 |
+ VD_VERT_INITIAL_PHASE(12) |
+ VD_VERT_PHASE_STEP(16) | /* /2 */
+ VD_VERT_FMT_EN;
+ break;
+ case DRM_FORMAT_YUV410:
+ priv->viu.viu_vd1_fmt_ctrl =
+ VD_HORZ_Y_C_RATIO(2) | /* /4 */
+ VD_HORZ_FMT_EN |
+ VD_VERT_RPT_LINE0 |
+ VD_VERT_INITIAL_PHASE(12) |
+ VD_VERT_PHASE_STEP(8) | /* /4 */
+ VD_VERT_FMT_EN;
+ break;
+ }
+ break;
+ }
+
+ /* Update Canvas with buffer address */
+ priv->viu.vd1_planes = drm_format_num_planes(fb->format->format);
+
+ switch (priv->viu.vd1_planes) {
+ case 3:
+ gem = drm_fb_cma_get_gem_obj(fb, 2);
+ priv->viu.vd1_addr2 = gem->paddr + fb->offsets[2];
+ priv->viu.vd1_stride2 = fb->pitches[2];
+ priv->viu.vd1_height2 =
+ drm_format_plane_height(fb->height,
+ fb->format->format, 2);
+ DRM_DEBUG("plane 2 addr 0x%x stride %d height %d\n",
+ priv->viu.vd1_addr2,
+ priv->viu.vd1_stride2,
+ priv->viu.vd1_height2);
+ /* fallthrough */
+ case 2:
+ gem = drm_fb_cma_get_gem_obj(fb, 1);
+ priv->viu.vd1_addr1 = gem->paddr + fb->offsets[1];
+ priv->viu.vd1_stride1 = fb->pitches[1];
+ priv->viu.vd1_height1 =
+ drm_format_plane_height(fb->height,
+ fb->format->format, 1);
+ DRM_DEBUG("plane 1 addr 0x%x stride %d height %d\n",
+ priv->viu.vd1_addr1,
+ priv->viu.vd1_stride1,
+ priv->viu.vd1_height1);
+ /* fallthrough */
+ case 1:
+ gem = drm_fb_cma_get_gem_obj(fb, 0);
+ priv->viu.vd1_addr0 = gem->paddr + fb->offsets[0];
+ priv->viu.vd1_stride0 = fb->pitches[0];
+ priv->viu.vd1_height0 =
+ drm_format_plane_height(fb->height,
+ fb->format->format, 0);
+ DRM_DEBUG("plane 0 addr 0x%x stride %d height %d\n",
+ priv->viu.vd1_addr0,
+ priv->viu.vd1_stride0,
+ priv->viu.vd1_height0);
+ }
+
+ priv->viu.vd1_enabled = true;
+
+ spin_unlock_irqrestore(&priv->drm->event_lock, flags);
+
+ DRM_DEBUG_DRIVER("\n");
+}
+
+static void meson_overlay_atomic_disable(struct drm_plane *plane,
+ struct drm_plane_state *old_state)
+{
+ struct meson_overlay *meson_overlay = to_meson_overlay(plane);
+ struct meson_drm *priv = meson_overlay->priv;
+
+ DRM_DEBUG_DRIVER("\n");
+
+ priv->viu.vd1_enabled = false;
+
+ /* Disable VD1 */
+ writel_bits_relaxed(VPP_VD1_POSTBLEND | VPP_VD1_PREBLEND, 0,
+ priv->io_base + _REG(VPP_MISC));
+
+}
+
+static const struct drm_plane_helper_funcs meson_overlay_helper_funcs = {
+ .atomic_check = meson_overlay_atomic_check,
+ .atomic_disable = meson_overlay_atomic_disable,
+ .atomic_update = meson_overlay_atomic_update,
+};
+
+static const struct drm_plane_funcs meson_overlay_funcs = {
+ .update_plane = drm_atomic_helper_update_plane,
+ .disable_plane = drm_atomic_helper_disable_plane,
+ .destroy = drm_plane_cleanup,
+ .reset = drm_atomic_helper_plane_reset,
+ .atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state,
+ .atomic_destroy_state = drm_atomic_helper_plane_destroy_state,
+};
+
+static const uint32_t supported_drm_formats[] = {
+ DRM_FORMAT_YUYV,
+ DRM_FORMAT_NV12,
+ DRM_FORMAT_NV21,
+ DRM_FORMAT_YUV444,
+ DRM_FORMAT_YUV422,
+ DRM_FORMAT_YUV420,
+ DRM_FORMAT_YUV411,
+ DRM_FORMAT_YUV410,
+};
+
+int meson_overlay_create(struct meson_drm *priv)
+{
+ struct meson_overlay *meson_overlay;
+ struct drm_plane *plane;
+
+ DRM_DEBUG_DRIVER("\n");
+
+ meson_overlay = devm_kzalloc(priv->drm->dev, sizeof(*meson_overlay),
+ GFP_KERNEL);
+ if (!meson_overlay)
+ return -ENOMEM;
+
+ meson_overlay->priv = priv;
+ plane = &meson_overlay->base;
+
+ drm_universal_plane_init(priv->drm, plane, 0xFF,
+ &meson_overlay_funcs,
+ supported_drm_formats,
+ ARRAY_SIZE(supported_drm_formats),
+ NULL,
+ DRM_PLANE_TYPE_OVERLAY, "meson_overlay_plane");
+
+ drm_plane_helper_add(plane, &meson_overlay_helper_funcs);
+
+ priv->overlay_plane = plane;
+
+ DRM_DEBUG_DRIVER("\n");
+
+ return 0;
+}
diff --git a/drivers/gpu/drm/meson/meson_overlay.h b/drivers/gpu/drm/meson/meson_overlay.h
new file mode 100644
index 0000000..dae24f5
--- /dev/null
+++ b/drivers/gpu/drm/meson/meson_overlay.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright (C) 2018 BayLibre, SAS
+ * Author: Neil Armstrong <[email protected]>
+ */
+
+#ifndef __MESON_OVERLAY_H
+#define __MESON_OVERLAY_H
+
+#include "meson_drv.h"
+
+int meson_overlay_create(struct meson_drm *priv);
+
+#endif /* __MESON_OVERLAY_H */
diff --git a/drivers/gpu/drm/meson/meson_registers.h b/drivers/gpu/drm/meson/meson_registers.h
index bca8714..5c7e02c 100644
--- a/drivers/gpu/drm/meson/meson_registers.h
+++ b/drivers/gpu/drm/meson/meson_registers.h
@@ -286,6 +286,7 @@
#define VIU_OSD1_MATRIX_COEF22_30 0x1a9d
#define VIU_OSD1_MATRIX_COEF31_32 0x1a9e
#define VIU_OSD1_MATRIX_COEF40_41 0x1a9f
+#define VD1_IF0_GEN_REG3 0x1aa7
#define VIU_OSD1_EOTF_CTL 0x1ad4
#define VIU_OSD1_EOTF_COEF00_01 0x1ad5
#define VIU_OSD1_EOTF_COEF02_10 0x1ad6
@@ -297,6 +298,7 @@
#define VIU_OSD1_OETF_CTL 0x1adc
#define VIU_OSD1_OETF_LUT_ADDR_PORT 0x1add
#define VIU_OSD1_OETF_LUT_DATA_PORT 0x1ade
+#define AFBC_ENABLE 0x1ae0

/* vpp */
#define VPP_DUMMY_DATA 0x1d00
@@ -349,6 +351,7 @@
#define VPP_VD2_PREBLEND BIT(15)
#define VPP_OSD1_PREBLEND BIT(16)
#define VPP_OSD2_PREBLEND BIT(17)
+#define VPP_COLOR_MNG_ENABLE BIT(28)
#define VPP_OFIFO_SIZE 0x1d27
#define VPP_FIFO_STATUS 0x1d28
#define VPP_SMOKE_CTRL 0x1d29
diff --git a/drivers/gpu/drm/meson/meson_viu.c b/drivers/gpu/drm/meson/meson_viu.c
index 6bcfa52..2dffb98 100644
--- a/drivers/gpu/drm/meson/meson_viu.c
+++ b/drivers/gpu/drm/meson/meson_viu.c
@@ -329,6 +329,21 @@ void meson_viu_init(struct meson_drm *priv)
0xff << OSD_REPLACE_SHIFT,
priv->io_base + _REG(VIU_OSD2_CTRL_STAT2));

+ /* Disable VD1 AFBC */
+ /* di_mif0_en=0 mif0_to_vpp_en=0 di_mad_en=0 */
+ writel_bits_relaxed(0x7 << 16, 0,
+ priv->io_base + _REG(VIU_MISC_CTRL0));
+ /* afbc vd1 set=0 */
+ writel_bits_relaxed(BIT(20), 0,
+ priv->io_base + _REG(VIU_MISC_CTRL0));
+ writel_relaxed(0, priv->io_base + _REG(AFBC_ENABLE));
+
+ writel_relaxed(0x00FF00C0,
+ priv->io_base + _REG(VD1_IF0_LUMA_FIFO_SIZE));
+ writel_relaxed(0x00FF00C0,
+ priv->io_base + _REG(VD2_IF0_LUMA_FIFO_SIZE));
+
+
priv->viu.osd1_enabled = false;
priv->viu.osd1_commit = false;
priv->viu.osd1_interlace = false;
diff --git a/drivers/gpu/drm/meson/meson_vpp.c b/drivers/gpu/drm/meson/meson_vpp.c
index 27356f8..5dc24a9 100644
--- a/drivers/gpu/drm/meson/meson_vpp.c
+++ b/drivers/gpu/drm/meson/meson_vpp.c
@@ -122,6 +122,31 @@ static void meson_vpp_write_scaling_filter_coefs(struct meson_drm *priv,
priv->io_base + _REG(VPP_OSD_SCALE_COEF));
}

+static const uint32_t vpp_filter_coefs_bicubic[] = {
+ 0x00800000, 0x007f0100, 0xff7f0200, 0xfe7f0300,
+ 0xfd7e0500, 0xfc7e0600, 0xfb7d0800, 0xfb7c0900,
+ 0xfa7b0b00, 0xfa7a0dff, 0xf9790fff, 0xf97711ff,
+ 0xf87613ff, 0xf87416fe, 0xf87218fe, 0xf8701afe,
+ 0xf76f1dfd, 0xf76d1ffd, 0xf76b21fd, 0xf76824fd,
+ 0xf76627fc, 0xf76429fc, 0xf7612cfc, 0xf75f2ffb,
+ 0xf75d31fb, 0xf75a34fb, 0xf75837fa, 0xf7553afa,
+ 0xf8523cfa, 0xf8503ff9, 0xf84d42f9, 0xf84a45f9,
+ 0xf84848f8
+};
+
+static void meson_vpp_write_vd_scaling_filter_coefs(struct meson_drm *priv,
+ const unsigned int *coefs,
+ bool is_horizontal)
+{
+ int i;
+
+ writel_relaxed(is_horizontal ? BIT(8) : 0,
+ priv->io_base + _REG(VPP_SCALE_COEF_IDX));
+ for (i = 0; i < 33; i++)
+ writel_relaxed(coefs[i],
+ priv->io_base + _REG(VPP_SCALE_COEF));
+}
+
void meson_vpp_init(struct meson_drm *priv)
{
/* set dummy data default YUV black */
@@ -150,17 +175,34 @@ void meson_vpp_init(struct meson_drm *priv)

/* Force all planes off */
writel_bits_relaxed(VPP_OSD1_POSTBLEND | VPP_OSD2_POSTBLEND |
- VPP_VD1_POSTBLEND | VPP_VD2_POSTBLEND, 0,
+ VPP_VD1_POSTBLEND | VPP_VD2_POSTBLEND |
+ VPP_VD1_PREBLEND | VPP_VD2_PREBLEND, 0,
priv->io_base + _REG(VPP_MISC));

+ /* Setup default VD settings */
+ writel_relaxed(4096,
+ priv->io_base + _REG(VPP_PREBLEND_VD1_H_START_END));
+ writel_relaxed(4096,
+ priv->io_base + _REG(VPP_BLEND_VD2_H_START_END));
+
/* Disable Scalers */
writel_relaxed(0, priv->io_base + _REG(VPP_OSD_SC_CTRL0));
writel_relaxed(0, priv->io_base + _REG(VPP_OSD_VSC_CTRL0));
writel_relaxed(0, priv->io_base + _REG(VPP_OSD_HSC_CTRL0));
+ writel_relaxed(4 | (4 << 8) | BIT(15),
+ priv->io_base + _REG(VPP_SC_MISC));
+
+ writel_relaxed(1, priv->io_base + _REG(VPP_VADJ_CTRL));

/* Write in the proper filter coefficients. */
meson_vpp_write_scaling_filter_coefs(priv,
vpp_filter_coefs_4point_bspline, false);
meson_vpp_write_scaling_filter_coefs(priv,
vpp_filter_coefs_4point_bspline, true);
+
+ /* Write the VD proper filter coefficients. */
+ meson_vpp_write_vd_scaling_filter_coefs(priv, vpp_filter_coefs_bicubic,
+ false);
+ meson_vpp_write_vd_scaling_filter_coefs(priv, vpp_filter_coefs_bicubic,
+ true);
}
--
2.7.4


2018-11-06 11:13:53

by Daniel Vetter

[permalink] [raw]
Subject: Re: [PATCH 0/3] drm/meson: Add support for Overlay and OSD Plane scaling

On Tue, Nov 06, 2018 at 10:39:59AM +0100, Neil Armstrong wrote:
> This serie adds support for :
> - Overlay Plane blended with the primary plane, we can describe as "behind"
> - Primary Plane up-scaling up to 5x factor to support the OSD plane being
> upscaled from 1920x1080 when the display mode is set as 3840x2160.
>
> These patches depends on the "drm/meson: Allow using optional canvas provider"
> patchset from Maxime Jourdan at [1]
>
> [1] https://patchwork.kernel.org/project/linux-amlogic/list/?series=38355
>
> Neil Armstrong (3):
> drm/meson: Support Overlay plane for video rendering
> drm/meson: move OSD scaler management into plane atomic update
> drm/meson: Add primary plane scaling

Got bored, read a bit your patch series. Two things I've noticed:
- We need to drop the crtc_state argument from
drm_atomic_helper_check_plane_state(), some drivers (not meson) do
really funny stuff with this.

- I think your osd1_commit = true update is racy with the interrupt
handler checking it. Needs at least a few memory barriers. Since helpers
guarantee that you can't overtake a preceending update (you do the
vblank handling after the updating) that should be ok, except you arm
the vblank before setting up registers, so this might race too.

Anyway all unrelated, so feel free to smash an

Acked-by: Daniel Vetter <[email protected]>

on the entire series.

Cheers, Daniel

>
> drivers/gpu/drm/meson/Makefile | 2 +-
> drivers/gpu/drm/meson/meson_canvas.c | 7 +-
> drivers/gpu/drm/meson/meson_canvas.h | 11 +-
> drivers/gpu/drm/meson/meson_crtc.c | 251 +++++++++++++-
> drivers/gpu/drm/meson/meson_drv.c | 29 +-
> drivers/gpu/drm/meson/meson_drv.h | 62 ++++
> drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
> drivers/gpu/drm/meson/meson_overlay.h | 14 +
> drivers/gpu/drm/meson/meson_plane.c | 177 ++++++++--
> drivers/gpu/drm/meson/meson_registers.h | 3 +
> drivers/gpu/drm/meson/meson_viu.c | 15 +
> drivers/gpu/drm/meson/meson_vpp.c | 90 +++--
> 12 files changed, 1156 insertions(+), 91 deletions(-)
> create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
> create mode 100644 drivers/gpu/drm/meson/meson_overlay.h
>
> --
> 2.7.4
>
> _______________________________________________
> dri-devel mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

2018-11-06 13:38:28

by Neil Armstrong

[permalink] [raw]
Subject: Re: [PATCH 0/3] drm/meson: Add support for Overlay and OSD Plane scaling

Hi Daniel,

On 06/11/2018 12:13, Daniel Vetter wrote:
> On Tue, Nov 06, 2018 at 10:39:59AM +0100, Neil Armstrong wrote:
>> This serie adds support for :
>> - Overlay Plane blended with the primary plane, we can describe as "behind"
>> - Primary Plane up-scaling up to 5x factor to support the OSD plane being
>> upscaled from 1920x1080 when the display mode is set as 3840x2160.
>>
>> These patches depends on the "drm/meson: Allow using optional canvas provider"
>> patchset from Maxime Jourdan at [1]
>>
>> [1] https://patchwork.kernel.org/project/linux-amlogic/list/?series=38355
>>
>> Neil Armstrong (3):
>> drm/meson: Support Overlay plane for video rendering
>> drm/meson: move OSD scaler management into plane atomic update
>> drm/meson: Add primary plane scaling
>
> Got bored, read a bit your patch series. Two things I've noticed:

Great you got bored today !

> - We need to drop the crtc_state argument from
> drm_atomic_helper_check_plane_state(), some drivers (not meson) do
> really funny stuff with this.

Should drm_atomic_get_crtc_state() be integrated into drm_atomic_helper_check_plane_state() ?

>
> - I think your osd1_commit = true update is racy with the interrupt
> handler checking it. Needs at least a few memory barriers. Since helpers
> guarantee that you can't overtake a preceending update (you do the
> vblank handling after the updating) that should be ok, except you arm
> the vblank before setting up registers, so this might race too.

Yeah, I need to do a cleanup of this at some point...

>
> Anyway all unrelated, so feel free to smash an
>
> Acked-by: Daniel Vetter <[email protected]>
>
> on the entire series.

thanks !

Will push to drm-misc-next once the branch is ready.

Neil

>
> Cheers, Daniel
>
>>
>> drivers/gpu/drm/meson/Makefile | 2 +-
>> drivers/gpu/drm/meson/meson_canvas.c | 7 +-
>> drivers/gpu/drm/meson/meson_canvas.h | 11 +-
>> drivers/gpu/drm/meson/meson_crtc.c | 251 +++++++++++++-
>> drivers/gpu/drm/meson/meson_drv.c | 29 +-
>> drivers/gpu/drm/meson/meson_drv.h | 62 ++++
>> drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
>> drivers/gpu/drm/meson/meson_overlay.h | 14 +
>> drivers/gpu/drm/meson/meson_plane.c | 177 ++++++++--
>> drivers/gpu/drm/meson/meson_registers.h | 3 +
>> drivers/gpu/drm/meson/meson_viu.c | 15 +
>> drivers/gpu/drm/meson/meson_vpp.c | 90 +++--
>> 12 files changed, 1156 insertions(+), 91 deletions(-)
>> create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
>> create mode 100644 drivers/gpu/drm/meson/meson_overlay.h
>>
>> --
>> 2.7.4
>>
>> _______________________________________________
>> dri-devel mailing list
>> [email protected]
>> https://lists.freedesktop.org/mailman/listinfo/dri-devel
>


2018-11-06 14:01:27

by Daniel Vetter

[permalink] [raw]
Subject: Re: [PATCH 0/3] drm/meson: Add support for Overlay and OSD Plane scaling

On Tue, Nov 6, 2018 at 2:37 PM Neil Armstrong <[email protected]> wrote:
>
> Hi Daniel,
>
> On 06/11/2018 12:13, Daniel Vetter wrote:
> > On Tue, Nov 06, 2018 at 10:39:59AM +0100, Neil Armstrong wrote:
> >> This serie adds support for :
> >> - Overlay Plane blended with the primary plane, we can describe as "behind"
> >> - Primary Plane up-scaling up to 5x factor to support the OSD plane being
> >> upscaled from 1920x1080 when the display mode is set as 3840x2160.
> >>
> >> These patches depends on the "drm/meson: Allow using optional canvas provider"
> >> patchset from Maxime Jourdan at [1]
> >>
> >> [1] https://patchwork.kernel.org/project/linux-amlogic/list/?series=38355
> >>
> >> Neil Armstrong (3):
> >> drm/meson: Support Overlay plane for video rendering
> >> drm/meson: move OSD scaler management into plane atomic update
> >> drm/meson: Add primary plane scaling
> >
> > Got bored, read a bit your patch series. Two things I've noticed:
>
> Great you got bored today !
>
> > - We need to drop the crtc_state argument from
> > drm_atomic_helper_check_plane_state(), some drivers (not meson) do
> > really funny stuff with this.
>
> Should drm_atomic_get_crtc_state() be integrated into drm_atomic_helper_check_plane_state() ?

Yeah. But I think I'll bake that into a nice todo.rst task for the
next intern. It's kinda the perfect task, including some cleanup in
armada and a bit of wtf in hdlcd.

> > - I think your osd1_commit = true update is racy with the interrupt
> > handler checking it. Needs at least a few memory barriers. Since helpers
> > guarantee that you can't overtake a preceending update (you do the
> > vblank handling after the updating) that should be ok, except you arm
> > the vblank before setting up registers, so this might race too.
>
> Yeah, I need to do a cleanup of this at some point...

Usually we solved this by wrapping everything (i.e. from atomic_begin
to atomic_flush) under a commit spinlock. Same with the entire
interrupt handler. You can do more complicated ofc, but this tends to
get the job done and very obviously avoids all the races. Also as long
as you don't have any register writes in there it should be fast
enough. If not, prepare all the register data as a separate allocation
and then just exchange the pointers.
-Daniel

>
> >
> > Anyway all unrelated, so feel free to smash an
> >
> > Acked-by: Daniel Vetter <[email protected]>
> >
> > on the entire series.
>
> thanks !
>
> Will push to drm-misc-next once the branch is ready.
>
> Neil
>
> >
> > Cheers, Daniel
> >
> >>
> >> drivers/gpu/drm/meson/Makefile | 2 +-
> >> drivers/gpu/drm/meson/meson_canvas.c | 7 +-
> >> drivers/gpu/drm/meson/meson_canvas.h | 11 +-
> >> drivers/gpu/drm/meson/meson_crtc.c | 251 +++++++++++++-
> >> drivers/gpu/drm/meson/meson_drv.c | 29 +-
> >> drivers/gpu/drm/meson/meson_drv.h | 62 ++++
> >> drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
> >> drivers/gpu/drm/meson/meson_overlay.h | 14 +
> >> drivers/gpu/drm/meson/meson_plane.c | 177 ++++++++--
> >> drivers/gpu/drm/meson/meson_registers.h | 3 +
> >> drivers/gpu/drm/meson/meson_viu.c | 15 +
> >> drivers/gpu/drm/meson/meson_vpp.c | 90 +++--
> >> 12 files changed, 1156 insertions(+), 91 deletions(-)
> >> create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
> >> create mode 100644 drivers/gpu/drm/meson/meson_overlay.h
> >>
> >> --
> >> 2.7.4
> >>
> >> _______________________________________________
> >> dri-devel mailing list
> >> [email protected]
> >> https://lists.freedesktop.org/mailman/listinfo/dri-devel
> >
>
> _______________________________________________
> dri-devel mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/dri-devel



--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

2018-11-07 09:43:33

by Maxime Jourdan

[permalink] [raw]
Subject: Re: [PATCH 1/3] drm/meson: Support Overlay plane for video rendering

On Tue, Nov 6, 2018 at 10:40 AM Neil Armstrong <[email protected]> wrote:
>
> The Amlogic Meson GX SoCs support an Overlay plane behind the primary
> plane for video rendering.
>
> This Overlay plane support various YUV layouts :
> - YUYV
> - NV12 / NV21
> - YUV444 / 422 / 420 / 411 / 410
>
> The scaler supports a wide range of scaling ratios, but for simplicity,
> plane atomic check limits the scaling from x5 to /5 in vertical and
> horizontal scaling.
>
> The z-order is fixed and always behind the primary plane and cannot be changed.
>
> The scaling parameter algorithm was taken from the Amlogic vendor kernel
> code and rewritten to match the atomic universal plane requirements.
>
> The video rendering using this overlay plane support has been tested using
> the new Kodi DRM-KMS Prime rendering path along the in-review V4L2 Mem2Mem
> Hardware Video Decoder up to 3840x2160 NV12 frames on various display modes.
>
> Signed-off-by: Neil Armstrong <[email protected]>
> ---
> drivers/gpu/drm/meson/Makefile | 2 +-
> drivers/gpu/drm/meson/meson_canvas.c | 7 +-
> drivers/gpu/drm/meson/meson_canvas.h | 11 +-
> drivers/gpu/drm/meson/meson_crtc.c | 216 +++++++++++-
> drivers/gpu/drm/meson/meson_drv.c | 29 +-
> drivers/gpu/drm/meson/meson_drv.h | 52 +++
> drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
> drivers/gpu/drm/meson/meson_overlay.h | 14 +
> drivers/gpu/drm/meson/meson_registers.h | 3 +
> drivers/gpu/drm/meson/meson_viu.c | 15 +
> drivers/gpu/drm/meson/meson_vpp.c | 44 ++-
> 11 files changed, 971 insertions(+), 8 deletions(-)
> create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
> create mode 100644 drivers/gpu/drm/meson/meson_overlay.h
>

Tested by rendering video with:
- gstreamer 1.14.4 using kmssink
- ffmpeg-drm (https://github.com/Elyotna/ffmpeg-drm),
by applying this patch on top of the video decoder patch series.

Used different resolutions and pixel aspect ratios.

Tested-by: Maxime Jourdan <[email protected]>

[snip]

2018-11-13 13:34:36

by Neil Armstrong

[permalink] [raw]
Subject: Re: [PATCH 0/3] drm/meson: Add support for Overlay and OSD Plane scaling

On 06/11/2018 10:39, Neil Armstrong wrote:
> This serie adds support for :
> - Overlay Plane blended with the primary plane, we can describe as "behind"
> - Primary Plane up-scaling up to 5x factor to support the OSD plane being
> upscaled from 1920x1080 when the display mode is set as 3840x2160.
>
> These patches depends on the "drm/meson: Allow using optional canvas provider"
> patchset from Maxime Jourdan at [1]
>
> [1] https://patchwork.kernel.org/project/linux-amlogic/list/?series=38355
>
> Neil Armstrong (3):
> drm/meson: Support Overlay plane for video rendering
> drm/meson: move OSD scaler management into plane atomic update
> drm/meson: Add primary plane scaling
>
> drivers/gpu/drm/meson/Makefile | 2 +-
> drivers/gpu/drm/meson/meson_canvas.c | 7 +-
> drivers/gpu/drm/meson/meson_canvas.h | 11 +-
> drivers/gpu/drm/meson/meson_crtc.c | 251 +++++++++++++-
> drivers/gpu/drm/meson/meson_drv.c | 29 +-
> drivers/gpu/drm/meson/meson_drv.h | 62 ++++
> drivers/gpu/drm/meson/meson_overlay.c | 586 ++++++++++++++++++++++++++++++++
> drivers/gpu/drm/meson/meson_overlay.h | 14 +
> drivers/gpu/drm/meson/meson_plane.c | 177 ++++++++--
> drivers/gpu/drm/meson/meson_registers.h | 3 +
> drivers/gpu/drm/meson/meson_viu.c | 15 +
> drivers/gpu/drm/meson/meson_vpp.c | 90 +++--
> 12 files changed, 1156 insertions(+), 91 deletions(-)
> create mode 100644 drivers/gpu/drm/meson/meson_overlay.c
> create mode 100644 drivers/gpu/drm/meson/meson_overlay.h
>

Applied