summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/vc4
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/vc4')
-rw-r--r--drivers/gpu/drm/vc4/Kconfig2
-rw-r--r--drivers/gpu/drm/vc4/Makefile2
-rw-r--r--drivers/gpu/drm/vc4/vc4_crtc.c113
-rw-r--r--drivers/gpu/drm/vc4/vc4_debugfs.c8
-rw-r--r--drivers/gpu/drm/vc4/vc4_drv.c20
-rw-r--r--drivers/gpu/drm/vc4/vc4_drv.h18
-rw-r--r--drivers/gpu/drm/vc4/vc4_dsi.c1725
-rw-r--r--drivers/gpu/drm/vc4/vc4_gem.c20
-rw-r--r--drivers/gpu/drm/vc4/vc4_hdmi.c18
-rw-r--r--drivers/gpu/drm/vc4/vc4_hvs.c17
-rw-r--r--drivers/gpu/drm/vc4/vc4_kms.c37
-rw-r--r--drivers/gpu/drm/vc4/vc4_plane.c16
-rw-r--r--drivers/gpu/drm/vc4/vc4_regs.h8
-rw-r--r--drivers/gpu/drm/vc4/vc4_render_cl.c2
-rw-r--r--drivers/gpu/drm/vc4/vc4_v3d.c2
-rw-r--r--drivers/gpu/drm/vc4/vc4_validate.c24
-rw-r--r--drivers/gpu/drm/vc4/vc4_validate_shaders.c82
-rw-r--r--drivers/gpu/drm/vc4/vc4_vec.c657
18 files changed, 2659 insertions, 112 deletions
diff --git a/drivers/gpu/drm/vc4/Kconfig b/drivers/gpu/drm/vc4/Kconfig
index e53df59cb139..e1517d07cb7d 100644
--- a/drivers/gpu/drm/vc4/Kconfig
+++ b/drivers/gpu/drm/vc4/Kconfig
@@ -2,10 +2,12 @@ config DRM_VC4
tristate "Broadcom VC4 Graphics"
depends on ARCH_BCM2835 || COMPILE_TEST
depends on DRM
+ depends on COMMON_CLK
select DRM_KMS_HELPER
select DRM_KMS_CMA_HELPER
select DRM_GEM_CMA_HELPER
select DRM_PANEL
+ select DRM_MIPI_DSI
help
Choose this option if you have a system that has a Broadcom
VC4 GPU, such as the Raspberry Pi or other BCM2708/BCM2835.
diff --git a/drivers/gpu/drm/vc4/Makefile b/drivers/gpu/drm/vc4/Makefile
index fb77db755e0a..61f45d122bd0 100644
--- a/drivers/gpu/drm/vc4/Makefile
+++ b/drivers/gpu/drm/vc4/Makefile
@@ -8,9 +8,11 @@ vc4-y := \
vc4_crtc.o \
vc4_drv.o \
vc4_dpi.o \
+ vc4_dsi.o \
vc4_kms.o \
vc4_gem.o \
vc4_hdmi.o \
+ vc4_vec.o \
vc4_hvs.o \
vc4_irq.o \
vc4_plane.o \
diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c
index 7f08d681a74b..9fcf05ca492b 100644
--- a/drivers/gpu/drm/vc4/vc4_crtc.c
+++ b/drivers/gpu/drm/vc4/vc4_crtc.c
@@ -83,8 +83,7 @@ struct vc4_crtc_data {
/* Which channel of the HVS this pixelvalve sources from. */
int hvs_channel;
- enum vc4_encoder_type encoder0_type;
- enum vc4_encoder_type encoder1_type;
+ enum vc4_encoder_type encoder_types[4];
};
#define CRTC_WRITE(offset, val) writel(val, vc4_crtc->regs + (offset))
@@ -157,7 +156,8 @@ int vc4_crtc_get_scanoutpos(struct drm_device *dev, unsigned int crtc_id,
const struct drm_display_mode *mode)
{
struct vc4_dev *vc4 = to_vc4_dev(dev);
- struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id];
+ struct drm_crtc *crtc = drm_crtc_from_index(dev, crtc_id);
+ struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
u32 val;
int fifo_lines;
int vblank_lines;
@@ -273,9 +273,7 @@ int vc4_crtc_get_vblank_timestamp(struct drm_device *dev, unsigned int crtc_id,
int *max_error, struct timeval *vblank_time,
unsigned flags)
{
- struct vc4_dev *vc4 = to_vc4_dev(dev);
- struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id];
- struct drm_crtc *crtc = &vc4_crtc->base;
+ struct drm_crtc *crtc = drm_crtc_from_index(dev, crtc_id);
struct drm_crtc_state *state = crtc->state;
/* Helper routine in DRM core does all the work: */
@@ -350,38 +348,40 @@ static u32 vc4_get_fifo_full_level(u32 format)
}
/*
- * Returns the clock select bit for the connector attached to the
- * CRTC.
+ * Returns the encoder attached to the CRTC.
+ *
+ * VC4 can only scan out to one encoder at a time, while the DRM core
+ * allows drivers to push pixels to more than one encoder from the
+ * same CRTC.
*/
-static int vc4_get_clock_select(struct drm_crtc *crtc)
+static struct drm_encoder *vc4_get_crtc_encoder(struct drm_crtc *crtc)
{
struct drm_connector *connector;
drm_for_each_connector(connector, crtc->dev) {
if (connector->state->crtc == crtc) {
- struct drm_encoder *encoder = connector->encoder;
- struct vc4_encoder *vc4_encoder =
- to_vc4_encoder(encoder);
-
- return vc4_encoder->clock_select;
+ return connector->encoder;
}
}
- return -1;
+ return NULL;
}
static void vc4_crtc_mode_set_nofb(struct drm_crtc *crtc)
{
struct drm_device *dev = crtc->dev;
struct vc4_dev *vc4 = to_vc4_dev(dev);
+ struct drm_encoder *encoder = vc4_get_crtc_encoder(crtc);
+ struct vc4_encoder *vc4_encoder = to_vc4_encoder(encoder);
struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
struct drm_crtc_state *state = crtc->state;
struct drm_display_mode *mode = &state->adjusted_mode;
bool interlace = mode->flags & DRM_MODE_FLAG_INTERLACE;
u32 pixel_rep = (mode->flags & DRM_MODE_FLAG_DBLCLK) ? 2 : 1;
- u32 format = PV_CONTROL_FORMAT_24;
+ bool is_dsi = (vc4_encoder->type == VC4_ENCODER_TYPE_DSI0 ||
+ vc4_encoder->type == VC4_ENCODER_TYPE_DSI1);
+ u32 format = is_dsi ? PV_CONTROL_FORMAT_DSIV_24 : PV_CONTROL_FORMAT_24;
bool debug_dump_regs = false;
- int clock_select = vc4_get_clock_select(crtc);
if (debug_dump_regs) {
DRM_INFO("CRTC %d regs before:\n", drm_crtc_index(crtc));
@@ -437,17 +437,19 @@ static void vc4_crtc_mode_set_nofb(struct drm_crtc *crtc)
*/
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,
PV_VCONTROL_ODD_DELAY));
CRTC_WRITE(PV_VSYNCD_EVEN, 0);
} else {
- CRTC_WRITE(PV_V_CONTROL, PV_VCONTROL_CONTINUOUS);
+ CRTC_WRITE(PV_V_CONTROL,
+ PV_VCONTROL_CONTINUOUS |
+ (is_dsi ? PV_VCONTROL_DSI : 0));
}
CRTC_WRITE(PV_HACT_ACT, mode->hdisplay * pixel_rep);
-
CRTC_WRITE(PV_CONTROL,
VC4_SET_FIELD(format, PV_CONTROL_FORMAT) |
VC4_SET_FIELD(vc4_get_fifo_full_level(format),
@@ -456,7 +458,8 @@ static void vc4_crtc_mode_set_nofb(struct drm_crtc *crtc)
PV_CONTROL_CLR_AT_START |
PV_CONTROL_TRIGGER_UNDERFLOW |
PV_CONTROL_WAIT_HSTART |
- VC4_SET_FIELD(clock_select, PV_CONTROL_CLK_SELECT) |
+ VC4_SET_FIELD(vc4_encoder->clock_select,
+ PV_CONTROL_CLK_SELECT) |
PV_CONTROL_FIFO_CLR |
PV_CONTROL_EN);
@@ -590,7 +593,7 @@ static int vc4_crtc_atomic_check(struct drm_crtc *crtc,
spin_lock_irqsave(&vc4->hvs->mm_lock, flags);
ret = drm_mm_insert_node(&vc4->hvs->dlist_mm, &vc4_state->mm,
- dlist_count, 1, 0);
+ dlist_count);
spin_unlock_irqrestore(&vc4->hvs->mm_lock, flags);
if (ret)
return ret;
@@ -653,8 +656,8 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc,
int vc4_enable_vblank(struct drm_device *dev, unsigned int crtc_id)
{
- struct vc4_dev *vc4 = to_vc4_dev(dev);
- struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id];
+ struct drm_crtc *crtc = drm_crtc_from_index(dev, crtc_id);
+ struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
CRTC_WRITE(PV_INTEN, PV_INT_VFP_START);
@@ -663,12 +666,20 @@ int vc4_enable_vblank(struct drm_device *dev, unsigned int crtc_id)
void vc4_disable_vblank(struct drm_device *dev, unsigned int crtc_id)
{
- struct vc4_dev *vc4 = to_vc4_dev(dev);
- struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id];
+ struct drm_crtc *crtc = drm_crtc_from_index(dev, crtc_id);
+ struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
CRTC_WRITE(PV_INTEN, 0);
}
+/* Must be called with the event lock held */
+bool vc4_event_pending(struct drm_crtc *crtc)
+{
+ struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+
+ return !!vc4_crtc->event;
+}
+
static void vc4_crtc_handle_page_flip(struct vc4_crtc *vc4_crtc)
{
struct drm_crtc *crtc = &vc4_crtc->base;
@@ -832,7 +843,18 @@ static void vc4_crtc_destroy_state(struct drm_crtc *crtc,
}
- __drm_atomic_helper_crtc_destroy_state(state);
+ drm_atomic_helper_crtc_destroy_state(crtc, state);
+}
+
+static void
+vc4_crtc_reset(struct drm_crtc *crtc)
+{
+ if (crtc->state)
+ __drm_atomic_helper_crtc_destroy_state(crtc->state);
+
+ crtc->state = kzalloc(sizeof(struct vc4_crtc_state), GFP_KERNEL);
+ if (crtc->state)
+ crtc->state->crtc = crtc;
}
static const struct drm_crtc_funcs vc4_crtc_funcs = {
@@ -842,7 +864,7 @@ static const struct drm_crtc_funcs vc4_crtc_funcs = {
.set_property = NULL,
.cursor_set = NULL, /* handled by drm_mode_cursor_universal */
.cursor_move = NULL, /* handled by drm_mode_cursor_universal */
- .reset = drm_atomic_helper_crtc_reset,
+ .reset = vc4_crtc_reset,
.atomic_duplicate_state = vc4_crtc_duplicate_state,
.atomic_destroy_state = vc4_crtc_destroy_state,
.gamma_set = vc4_crtc_gamma_set,
@@ -859,20 +881,26 @@ static const struct drm_crtc_helper_funcs vc4_crtc_helper_funcs = {
static const struct vc4_crtc_data pv0_data = {
.hvs_channel = 0,
- .encoder0_type = VC4_ENCODER_TYPE_DSI0,
- .encoder1_type = VC4_ENCODER_TYPE_DPI,
+ .encoder_types = {
+ [PV_CONTROL_CLK_SELECT_DSI] = VC4_ENCODER_TYPE_DSI0,
+ [PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_DPI,
+ },
};
static const struct vc4_crtc_data pv1_data = {
.hvs_channel = 2,
- .encoder0_type = VC4_ENCODER_TYPE_DSI1,
- .encoder1_type = VC4_ENCODER_TYPE_SMI,
+ .encoder_types = {
+ [PV_CONTROL_CLK_SELECT_DSI] = VC4_ENCODER_TYPE_DSI1,
+ [PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_SMI,
+ },
};
static const struct vc4_crtc_data pv2_data = {
.hvs_channel = 1,
- .encoder0_type = VC4_ENCODER_TYPE_VEC,
- .encoder1_type = VC4_ENCODER_TYPE_HDMI,
+ .encoder_types = {
+ [PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI] = VC4_ENCODER_TYPE_HDMI,
+ [PV_CONTROL_CLK_SELECT_VEC] = VC4_ENCODER_TYPE_VEC,
+ },
};
static const struct of_device_id vc4_crtc_dt_match[] = {
@@ -886,17 +914,20 @@ static void vc4_set_crtc_possible_masks(struct drm_device *drm,
struct drm_crtc *crtc)
{
struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+ const struct vc4_crtc_data *crtc_data = vc4_crtc->data;
+ const enum vc4_encoder_type *encoder_types = crtc_data->encoder_types;
struct drm_encoder *encoder;
drm_for_each_encoder(encoder, drm) {
struct vc4_encoder *vc4_encoder = to_vc4_encoder(encoder);
-
- if (vc4_encoder->type == vc4_crtc->data->encoder0_type) {
- vc4_encoder->clock_select = 0;
- encoder->possible_crtcs |= drm_crtc_mask(crtc);
- } else if (vc4_encoder->type == vc4_crtc->data->encoder1_type) {
- vc4_encoder->clock_select = 1;
- encoder->possible_crtcs |= drm_crtc_mask(crtc);
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(crtc_data->encoder_types); i++) {
+ if (vc4_encoder->type == encoder_types[i]) {
+ vc4_encoder->clock_select = i;
+ encoder->possible_crtcs |= drm_crtc_mask(crtc);
+ break;
+ }
}
}
}
@@ -921,7 +952,6 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data)
{
struct platform_device *pdev = to_platform_device(dev);
struct drm_device *drm = dev_get_drvdata(master);
- struct vc4_dev *vc4 = to_vc4_dev(drm);
struct vc4_crtc *vc4_crtc;
struct drm_crtc *crtc;
struct drm_plane *primary_plane, *cursor_plane, *destroy_plane, *temp;
@@ -959,7 +989,6 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data)
&vc4_crtc_funcs, NULL);
drm_crtc_helper_add(crtc, &vc4_crtc_helper_funcs);
primary_plane->crtc = crtc;
- vc4->crtc[drm_crtc_index(crtc)] = vc4_crtc;
vc4_crtc->channel = vc4_crtc->data->hvs_channel;
drm_mode_crtc_set_gamma_size(crtc, ARRAY_SIZE(vc4_crtc->lut_r));
diff --git a/drivers/gpu/drm/vc4/vc4_debugfs.c b/drivers/gpu/drm/vc4/vc4_debugfs.c
index 245115d49c46..5db06bdb5f27 100644
--- a/drivers/gpu/drm/vc4/vc4_debugfs.c
+++ b/drivers/gpu/drm/vc4/vc4_debugfs.c
@@ -18,7 +18,9 @@
static const struct drm_info_list vc4_debugfs_list[] = {
{"bo_stats", vc4_bo_stats_debugfs, 0},
{"dpi_regs", vc4_dpi_debugfs_regs, 0},
+ {"dsi1_regs", vc4_dsi_debugfs_regs, 0, (void *)(uintptr_t)1},
{"hdmi_regs", vc4_hdmi_debugfs_regs, 0},
+ {"vec_regs", vc4_vec_debugfs_regs, 0},
{"hvs_regs", vc4_hvs_debugfs_regs, 0},
{"crtc0_regs", vc4_crtc_debugfs_regs, 0, (void *)(uintptr_t)0},
{"crtc1_regs", vc4_crtc_debugfs_regs, 0, (void *)(uintptr_t)1},
@@ -35,9 +37,3 @@ vc4_debugfs_init(struct drm_minor *minor)
return drm_debugfs_create_files(vc4_debugfs_list, VC4_DEBUGFS_ENTRIES,
minor->debugfs_root, minor);
}
-
-void
-vc4_debugfs_cleanup(struct drm_minor *minor)
-{
- drm_debugfs_remove_files(vc4_debugfs_list, VC4_DEBUGFS_ENTRIES, minor);
-}
diff --git a/drivers/gpu/drm/vc4/vc4_drv.c b/drivers/gpu/drm/vc4/vc4_drv.c
index 8703f56b7947..a459745e96f7 100644
--- a/drivers/gpu/drm/vc4/vc4_drv.c
+++ b/drivers/gpu/drm/vc4/vc4_drv.c
@@ -61,23 +61,28 @@ static int vc4_get_param_ioctl(struct drm_device *dev, void *data,
if (ret < 0)
return ret;
args->value = V3D_READ(V3D_IDENT0);
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
break;
case DRM_VC4_PARAM_V3D_IDENT1:
ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
if (ret < 0)
return ret;
args->value = V3D_READ(V3D_IDENT1);
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
break;
case DRM_VC4_PARAM_V3D_IDENT2:
ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev);
if (ret < 0)
return ret;
args->value = V3D_READ(V3D_IDENT2);
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
break;
case DRM_VC4_PARAM_SUPPORTS_BRANCHES:
+ case DRM_VC4_PARAM_SUPPORTS_ETC1:
+ case DRM_VC4_PARAM_SUPPORTS_THREADED_FS:
args->value = true;
break;
default:
@@ -103,9 +108,7 @@ static const struct file_operations vc4_drm_fops = {
.mmap = vc4_mmap,
.poll = drm_poll,
.read = drm_read,
-#ifdef CONFIG_COMPAT
.compat_ioctl = drm_compat_ioctl,
-#endif
.llseek = noop_llseek,
};
@@ -142,7 +145,6 @@ static struct drm_driver vc4_drm_driver = {
#if defined(CONFIG_DEBUG_FS)
.debugfs_init = vc4_debugfs_init,
- .debugfs_cleanup = vc4_debugfs_cleanup,
#endif
.gem_create_object = vc4_create_object,
@@ -274,12 +276,14 @@ static void vc4_drm_unbind(struct device *dev)
struct drm_device *drm = platform_get_drvdata(pdev);
struct vc4_dev *vc4 = to_vc4_dev(drm);
+ drm_dev_unregister(drm);
+
if (vc4->fbdev)
drm_fbdev_cma_fini(vc4->fbdev);
drm_mode_config_cleanup(drm);
- drm_put_dev(drm);
+ drm_dev_unref(drm);
}
static const struct component_master_ops vc4_drm_ops = {
@@ -289,7 +293,9 @@ static const struct component_master_ops vc4_drm_ops = {
static struct platform_driver *const component_drivers[] = {
&vc4_hdmi_driver,
+ &vc4_vec_driver,
&vc4_dpi_driver,
+ &vc4_dsi_driver,
&vc4_hvs_driver,
&vc4_crtc_driver,
&vc4_v3d_driver,
diff --git a/drivers/gpu/drm/vc4/vc4_drv.h b/drivers/gpu/drm/vc4/vc4_drv.h
index 7c1e4d97486f..0e59f3ee1b83 100644
--- a/drivers/gpu/drm/vc4/vc4_drv.h
+++ b/drivers/gpu/drm/vc4/vc4_drv.h
@@ -9,14 +9,17 @@
#include "drmP.h"
#include "drm_gem_cma_helper.h"
+#include <drm/drm_encoder.h>
+
struct vc4_dev {
struct drm_device *dev;
struct vc4_hdmi *hdmi;
struct vc4_hvs *hvs;
- struct vc4_crtc *crtc[3];
struct vc4_v3d *v3d;
struct vc4_dpi *dpi;
+ struct vc4_dsi *dsi1;
+ struct vc4_vec *vec;
struct drm_fbdev_cma *fbdev;
@@ -194,6 +197,7 @@ to_vc4_plane(struct drm_plane *plane)
}
enum vc4_encoder_type {
+ VC4_ENCODER_TYPE_NONE,
VC4_ENCODER_TYPE_HDMI,
VC4_ENCODER_TYPE_VEC,
VC4_ENCODER_TYPE_DSI0,
@@ -381,6 +385,8 @@ struct vc4_validated_shader_info {
uint32_t num_uniform_addr_offsets;
uint32_t *uniform_addr_offsets;
+
+ bool is_threaded;
};
/**
@@ -440,6 +446,7 @@ int vc4_bo_stats_debugfs(struct seq_file *m, void *arg);
extern struct platform_driver vc4_crtc_driver;
int vc4_enable_vblank(struct drm_device *dev, unsigned int crtc_id);
void vc4_disable_vblank(struct drm_device *dev, unsigned int crtc_id);
+bool vc4_event_pending(struct drm_crtc *crtc);
int vc4_crtc_debugfs_regs(struct seq_file *m, void *arg);
int vc4_crtc_get_scanoutpos(struct drm_device *dev, unsigned int crtc_id,
unsigned int flags, int *vpos, int *hpos,
@@ -451,7 +458,6 @@ int vc4_crtc_get_vblank_timestamp(struct drm_device *dev, unsigned int crtc_id,
/* vc4_debugfs.c */
int vc4_debugfs_init(struct drm_minor *minor);
-void vc4_debugfs_cleanup(struct drm_minor *minor);
/* vc4_drv.c */
void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index);
@@ -460,6 +466,10 @@ void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index);
extern struct platform_driver vc4_dpi_driver;
int vc4_dpi_debugfs_regs(struct seq_file *m, void *unused);
+/* vc4_dsi.c */
+extern struct platform_driver vc4_dsi_driver;
+int vc4_dsi_debugfs_regs(struct seq_file *m, void *unused);
+
/* vc4_gem.c */
void vc4_gem_init(struct drm_device *dev);
void vc4_gem_destroy(struct drm_device *dev);
@@ -483,6 +493,10 @@ int vc4_queue_seqno_cb(struct drm_device *dev,
extern struct platform_driver vc4_hdmi_driver;
int vc4_hdmi_debugfs_regs(struct seq_file *m, void *unused);
+/* vc4_hdmi.c */
+extern struct platform_driver vc4_vec_driver;
+int vc4_vec_debugfs_regs(struct seq_file *m, void *unused);
+
/* vc4_irq.c */
irqreturn_t vc4_irq(int irq, void *arg);
void vc4_irq_preinstall(struct drm_device *dev);
diff --git a/drivers/gpu/drm/vc4/vc4_dsi.c b/drivers/gpu/drm/vc4/vc4_dsi.c
new file mode 100644
index 000000000000..2736b0331beb
--- /dev/null
+++ b/drivers/gpu/drm/vc4/vc4_dsi.c
@@ -0,0 +1,1725 @@
+/*
+ * Copyright (C) 2016 Broadcom
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+/**
+ * DOC: VC4 DSI0/DSI1 module
+ *
+ * BCM2835 contains two DSI modules, DSI0 and DSI1. DSI0 is a
+ * single-lane DSI controller, while DSI1 is a more modern 4-lane DSI
+ * controller.
+ *
+ * Most Raspberry Pi boards expose DSI1 as their "DISPLAY" connector,
+ * while the compute module brings both DSI0 and DSI1 out.
+ *
+ * This driver has been tested for DSI1 video-mode display only
+ * currently, with most of the information necessary for DSI0
+ * hopefully present.
+ */
+
+#include "drm_atomic_helper.h"
+#include "drm_crtc_helper.h"
+#include "drm_edid.h"
+#include "drm_mipi_dsi.h"
+#include "drm_panel.h"
+#include "linux/clk.h"
+#include "linux/clk-provider.h"
+#include "linux/completion.h"
+#include "linux/component.h"
+#include "linux/dmaengine.h"
+#include "linux/i2c.h"
+#include "linux/of_address.h"
+#include "linux/of_platform.h"
+#include "linux/pm_runtime.h"
+#include "vc4_drv.h"
+#include "vc4_regs.h"
+
+#define DSI_CMD_FIFO_DEPTH 16
+#define DSI_PIX_FIFO_DEPTH 256
+#define DSI_PIX_FIFO_WIDTH 4
+
+#define DSI0_CTRL 0x00
+
+/* Command packet control. */
+#define DSI0_TXPKT1C 0x04 /* AKA PKTC */
+#define DSI1_TXPKT1C 0x04
+# define DSI_TXPKT1C_TRIG_CMD_MASK VC4_MASK(31, 24)
+# define DSI_TXPKT1C_TRIG_CMD_SHIFT 24
+# define DSI_TXPKT1C_CMD_REPEAT_MASK VC4_MASK(23, 10)
+# define DSI_TXPKT1C_CMD_REPEAT_SHIFT 10
+
+# define DSI_TXPKT1C_DISPLAY_NO_MASK VC4_MASK(9, 8)
+# define DSI_TXPKT1C_DISPLAY_NO_SHIFT 8
+/* Short, trigger, BTA, or a long packet that fits all in CMDFIFO. */
+# define DSI_TXPKT1C_DISPLAY_NO_SHORT 0
+/* Primary display where cmdfifo provides part of the payload and
+ * pixelvalve the rest.
+ */
+# define DSI_TXPKT1C_DISPLAY_NO_PRIMARY 1
+/* Secondary display where cmdfifo provides part of the payload and
+ * pixfifo the rest.
+ */
+# define DSI_TXPKT1C_DISPLAY_NO_SECONDARY 2
+
+# define DSI_TXPKT1C_CMD_TX_TIME_MASK VC4_MASK(7, 6)
+# define DSI_TXPKT1C_CMD_TX_TIME_SHIFT 6
+
+# define DSI_TXPKT1C_CMD_CTRL_MASK VC4_MASK(5, 4)
+# define DSI_TXPKT1C_CMD_CTRL_SHIFT 4
+/* Command only. Uses TXPKT1H and DISPLAY_NO */
+# define DSI_TXPKT1C_CMD_CTRL_TX 0
+/* Command with BTA for either ack or read data. */
+# define DSI_TXPKT1C_CMD_CTRL_RX 1
+/* Trigger according to TRIG_CMD */
+# define DSI_TXPKT1C_CMD_CTRL_TRIG 2
+/* BTA alone for getting error status after a command, or a TE trigger
+ * without a previous command.
+ */
+# define DSI_TXPKT1C_CMD_CTRL_BTA 3
+
+# define DSI_TXPKT1C_CMD_MODE_LP BIT(3)
+# define DSI_TXPKT1C_CMD_TYPE_LONG BIT(2)
+# define DSI_TXPKT1C_CMD_TE_EN BIT(1)
+# define DSI_TXPKT1C_CMD_EN BIT(0)
+
+/* Command packet header. */
+#define DSI0_TXPKT1H 0x08 /* AKA PKTH */
+#define DSI1_TXPKT1H 0x08
+# define DSI_TXPKT1H_BC_CMDFIFO_MASK VC4_MASK(31, 24)
+# define DSI_TXPKT1H_BC_CMDFIFO_SHIFT 24
+# define DSI_TXPKT1H_BC_PARAM_MASK VC4_MASK(23, 8)
+# define DSI_TXPKT1H_BC_PARAM_SHIFT 8
+# define DSI_TXPKT1H_BC_DT_MASK VC4_MASK(7, 0)
+# define DSI_TXPKT1H_BC_DT_SHIFT 0
+
+#define DSI0_RXPKT1H 0x0c /* AKA RX1_PKTH */
+#define DSI1_RXPKT1H 0x14
+# define DSI_RXPKT1H_CRC_ERR BIT(31)
+# define DSI_RXPKT1H_DET_ERR BIT(30)
+# define DSI_RXPKT1H_ECC_ERR BIT(29)
+# define DSI_RXPKT1H_COR_ERR BIT(28)
+# define DSI_RXPKT1H_INCOMP_PKT BIT(25)
+# define DSI_RXPKT1H_PKT_TYPE_LONG BIT(24)
+/* Byte count if DSI_RXPKT1H_PKT_TYPE_LONG */
+# define DSI_RXPKT1H_BC_PARAM_MASK VC4_MASK(23, 8)
+# define DSI_RXPKT1H_BC_PARAM_SHIFT 8
+/* Short return bytes if !DSI_RXPKT1H_PKT_TYPE_LONG */
+# define DSI_RXPKT1H_SHORT_1_MASK VC4_MASK(23, 16)
+# define DSI_RXPKT1H_SHORT_1_SHIFT 16
+# define DSI_RXPKT1H_SHORT_0_MASK VC4_MASK(15, 8)
+# define DSI_RXPKT1H_SHORT_0_SHIFT 8
+# define DSI_RXPKT1H_DT_LP_CMD_MASK VC4_MASK(7, 0)
+# define DSI_RXPKT1H_DT_LP_CMD_SHIFT 0
+
+#define DSI0_RXPKT2H 0x10 /* AKA RX2_PKTH */
+#define DSI1_RXPKT2H 0x18
+# define DSI_RXPKT1H_DET_ERR BIT(30)
+# define DSI_RXPKT1H_ECC_ERR BIT(29)
+# define DSI_RXPKT1H_COR_ERR BIT(28)
+# define DSI_RXPKT1H_INCOMP_PKT BIT(25)
+# define DSI_RXPKT1H_BC_PARAM_MASK VC4_MASK(23, 8)
+# define DSI_RXPKT1H_BC_PARAM_SHIFT 8
+# define DSI_RXPKT1H_DT_MASK VC4_MASK(7, 0)
+# define DSI_RXPKT1H_DT_SHIFT 0
+
+#define DSI0_TXPKT_CMD_FIFO 0x14 /* AKA CMD_DATAF */
+#define DSI1_TXPKT_CMD_FIFO 0x1c
+
+#define DSI0_DISP0_CTRL 0x18
+# define DSI_DISP0_PIX_CLK_DIV_MASK VC4_MASK(21, 13)
+# define DSI_DISP0_PIX_CLK_DIV_SHIFT 13
+# define DSI_DISP0_LP_STOP_CTRL_MASK VC4_MASK(12, 11)
+# define DSI_DISP0_LP_STOP_CTRL_SHIFT 11
+# define DSI_DISP0_LP_STOP_DISABLE 0
+# define DSI_DISP0_LP_STOP_PERLINE 1
+# define DSI_DISP0_LP_STOP_PERFRAME 2
+
+/* Transmit RGB pixels and null packets only during HACTIVE, instead
+ * of going to LP-STOP.
+ */
+# define DSI_DISP_HACTIVE_NULL BIT(10)
+/* Transmit blanking packet only during vblank, instead of allowing LP-STOP. */
+# define DSI_DISP_VBLP_CTRL BIT(9)
+/* Transmit blanking packet only during HFP, instead of allowing LP-STOP. */
+# define DSI_DISP_HFP_CTRL BIT(8)
+/* Transmit blanking packet only during HBP, instead of allowing LP-STOP. */
+# define DSI_DISP_HBP_CTRL BIT(7)
+# define DSI_DISP0_CHANNEL_MASK VC4_MASK(6, 5)
+# define DSI_DISP0_CHANNEL_SHIFT 5
+/* Enables end events for HSYNC/VSYNC, not just start events. */
+# define DSI_DISP0_ST_END BIT(4)
+# define DSI_DISP0_PFORMAT_MASK VC4_MASK(3, 2)
+# define DSI_DISP0_PFORMAT_SHIFT 2
+# define DSI_PFORMAT_RGB565 0
+# define DSI_PFORMAT_RGB666_PACKED 1
+# define DSI_PFORMAT_RGB666 2
+# define DSI_PFORMAT_RGB888 3
+/* Default is VIDEO mode. */
+# define DSI_DISP0_COMMAND_MODE BIT(1)
+# define DSI_DISP0_ENABLE BIT(0)
+
+#define DSI0_DISP1_CTRL 0x1c
+#define DSI1_DISP1_CTRL 0x2c
+/* Format of the data written to TXPKT_PIX_FIFO. */
+# define DSI_DISP1_PFORMAT_MASK VC4_MASK(2, 1)
+# define DSI_DISP1_PFORMAT_SHIFT 1
+# define DSI_DISP1_PFORMAT_16BIT 0
+# define DSI_DISP1_PFORMAT_24BIT 1
+# define DSI_DISP1_PFORMAT_32BIT_LE 2
+# define DSI_DISP1_PFORMAT_32BIT_BE 3
+
+/* DISP1 is always command mode. */
+# define DSI_DISP1_ENABLE BIT(0)
+
+#define DSI0_TXPKT_PIX_FIFO 0x20 /* AKA PIX_FIFO */
+
+#define DSI0_INT_STAT 0x24
+#define DSI0_INT_EN 0x28
+# define DSI1_INT_PHY_D3_ULPS BIT(30)
+# define DSI1_INT_PHY_D3_STOP BIT(29)
+# define DSI1_INT_PHY_D2_ULPS BIT(28)
+# define DSI1_INT_PHY_D2_STOP BIT(27)
+# define DSI1_INT_PHY_D1_ULPS BIT(26)
+# define DSI1_INT_PHY_D1_STOP BIT(25)
+# define DSI1_INT_PHY_D0_ULPS BIT(24)
+# define DSI1_INT_PHY_D0_STOP BIT(23)
+# define DSI1_INT_FIFO_ERR BIT(22)
+# define DSI1_INT_PHY_DIR_RTF BIT(21)
+# define DSI1_INT_PHY_RXLPDT BIT(20)
+# define DSI1_INT_PHY_RXTRIG BIT(19)
+# define DSI1_INT_PHY_D0_LPDT BIT(18)
+# define DSI1_INT_PHY_DIR_FTR BIT(17)
+
+/* Signaled when the clock lane enters the given state. */
+# define DSI1_INT_PHY_CLOCK_ULPS BIT(16)
+# define DSI1_INT_PHY_CLOCK_HS BIT(15)
+# define DSI1_INT_PHY_CLOCK_STOP BIT(14)
+
+/* Signaled on timeouts */
+# define DSI1_INT_PR_TO BIT(13)
+# define DSI1_INT_TA_TO BIT(12)
+# define DSI1_INT_LPRX_TO BIT(11)
+# define DSI1_INT_HSTX_TO BIT(10)
+
+/* Contention on a line when trying to drive the line low */
+# define DSI1_INT_ERR_CONT_LP1 BIT(9)
+# define DSI1_INT_ERR_CONT_LP0 BIT(8)
+
+/* Control error: incorrect line state sequence on data lane 0. */
+# define DSI1_INT_ERR_CONTROL BIT(7)
+/* LPDT synchronization error (bits received not a multiple of 8. */
+
+# define DSI1_INT_ERR_SYNC_ESC BIT(6)
+/* Signaled after receiving an error packet from the display in
+ * response to a read.
+ */
+# define DSI1_INT_RXPKT2 BIT(5)
+/* Signaled after receiving a packet. The header and optional short
+ * response will be in RXPKT1H, and a long response will be in the
+ * RXPKT_FIFO.
+ */
+# define DSI1_INT_RXPKT1 BIT(4)
+# define DSI1_INT_TXPKT2_DONE BIT(3)
+# define DSI1_INT_TXPKT2_END BIT(2)
+/* Signaled after all repeats of TXPKT1 are transferred. */
+# define DSI1_INT_TXPKT1_DONE BIT(1)
+/* Signaled after each TXPKT1 repeat is scheduled. */
+# define DSI1_INT_TXPKT1_END BIT(0)
+
+#define DSI1_INTERRUPTS_ALWAYS_ENABLED (DSI1_INT_ERR_SYNC_ESC | \
+ DSI1_INT_ERR_CONTROL | \
+ DSI1_INT_ERR_CONT_LP0 | \
+ DSI1_INT_ERR_CONT_LP1 | \
+ DSI1_INT_HSTX_TO | \
+ DSI1_INT_LPRX_TO | \
+ DSI1_INT_TA_TO | \
+ DSI1_INT_PR_TO)
+
+#define DSI0_STAT 0x2c
+#define DSI0_HSTX_TO_CNT 0x30
+#define DSI0_LPRX_TO_CNT 0x34
+#define DSI0_TA_TO_CNT 0x38
+#define DSI0_PR_TO_CNT 0x3c
+#define DSI0_PHYC 0x40
+# define DSI1_PHYC_ESC_CLK_LPDT_MASK VC4_MASK(25, 20)
+# define DSI1_PHYC_ESC_CLK_LPDT_SHIFT 20
+# define DSI1_PHYC_HS_CLK_CONTINUOUS BIT(18)
+# define DSI0_PHYC_ESC_CLK_LPDT_MASK VC4_MASK(17, 12)
+# define DSI0_PHYC_ESC_CLK_LPDT_SHIFT 12
+# define DSI1_PHYC_CLANE_ULPS BIT(17)
+# define DSI1_PHYC_CLANE_ENABLE BIT(16)
+# define DSI_PHYC_DLANE3_ULPS BIT(13)
+# define DSI_PHYC_DLANE3_ENABLE BIT(12)
+# define DSI0_PHYC_HS_CLK_CONTINUOUS BIT(10)
+# define DSI0_PHYC_CLANE_ULPS BIT(9)
+# define DSI_PHYC_DLANE2_ULPS BIT(9)
+# define DSI0_PHYC_CLANE_ENABLE BIT(8)
+# define DSI_PHYC_DLANE2_ENABLE BIT(8)
+# define DSI_PHYC_DLANE1_ULPS BIT(5)
+# define DSI_PHYC_DLANE1_ENABLE BIT(4)
+# define DSI_PHYC_DLANE0_FORCE_STOP BIT(2)
+# define DSI_PHYC_DLANE0_ULPS BIT(1)
+# define DSI_PHYC_DLANE0_ENABLE BIT(0)
+
+#define DSI0_HS_CLT0 0x44
+#define DSI0_HS_CLT1 0x48
+#define DSI0_HS_CLT2 0x4c
+#define DSI0_HS_DLT3 0x50
+#define DSI0_HS_DLT4 0x54
+#define DSI0_HS_DLT5 0x58
+#define DSI0_HS_DLT6 0x5c
+#define DSI0_HS_DLT7 0x60
+
+#define DSI0_PHY_AFEC0 0x64
+# define DSI0_PHY_AFEC0_DDR2CLK_EN BIT(26)
+# define DSI0_PHY_AFEC0_DDRCLK_EN BIT(25)
+# define DSI0_PHY_AFEC0_LATCH_ULPS BIT(24)
+# define DSI1_PHY_AFEC0_IDR_DLANE3_MASK VC4_MASK(31, 29)
+# define DSI1_PHY_AFEC0_IDR_DLANE3_SHIFT 29
+# define DSI1_PHY_AFEC0_IDR_DLANE2_MASK VC4_MASK(28, 26)
+# define DSI1_PHY_AFEC0_IDR_DLANE2_SHIFT 26
+# define DSI1_PHY_AFEC0_IDR_DLANE1_MASK VC4_MASK(27, 23)
+# define DSI1_PHY_AFEC0_IDR_DLANE1_SHIFT 23
+# define DSI1_PHY_AFEC0_IDR_DLANE0_MASK VC4_MASK(22, 20)
+# define DSI1_PHY_AFEC0_IDR_DLANE0_SHIFT 20
+# define DSI1_PHY_AFEC0_IDR_CLANE_MASK VC4_MASK(19, 17)
+# define DSI1_PHY_AFEC0_IDR_CLANE_SHIFT 17
+# define DSI0_PHY_AFEC0_ACTRL_DLANE1_MASK VC4_MASK(23, 20)
+# define DSI0_PHY_AFEC0_ACTRL_DLANE1_SHIFT 20
+# define DSI0_PHY_AFEC0_ACTRL_DLANE0_MASK VC4_MASK(19, 16)
+# define DSI0_PHY_AFEC0_ACTRL_DLANE0_SHIFT 16
+# define DSI0_PHY_AFEC0_ACTRL_CLANE_MASK VC4_MASK(15, 12)
+# define DSI0_PHY_AFEC0_ACTRL_CLANE_SHIFT 12
+# define DSI1_PHY_AFEC0_DDR2CLK_EN BIT(16)
+# define DSI1_PHY_AFEC0_DDRCLK_EN BIT(15)
+# define DSI1_PHY_AFEC0_LATCH_ULPS BIT(14)
+# define DSI1_PHY_AFEC0_RESET BIT(13)
+# define DSI1_PHY_AFEC0_PD BIT(12)
+# define DSI0_PHY_AFEC0_RESET BIT(11)
+# define DSI1_PHY_AFEC0_PD_BG BIT(11)
+# define DSI0_PHY_AFEC0_PD BIT(10)
+# define DSI1_PHY_AFEC0_PD_DLANE3 BIT(10)
+# define DSI0_PHY_AFEC0_PD_BG BIT(9)
+# define DSI1_PHY_AFEC0_PD_DLANE2 BIT(9)
+# define DSI0_PHY_AFEC0_PD_DLANE1 BIT(8)
+# define DSI1_PHY_AFEC0_PD_DLANE1 BIT(8)
+# define DSI_PHY_AFEC0_PTATADJ_MASK VC4_MASK(7, 4)
+# define DSI_PHY_AFEC0_PTATADJ_SHIFT 4
+# define DSI_PHY_AFEC0_CTATADJ_MASK VC4_MASK(3, 0)
+# define DSI_PHY_AFEC0_CTATADJ_SHIFT 0
+
+#define DSI0_PHY_AFEC1 0x68
+# define DSI0_PHY_AFEC1_IDR_DLANE1_MASK VC4_MASK(10, 8)
+# define DSI0_PHY_AFEC1_IDR_DLANE1_SHIFT 8
+# define DSI0_PHY_AFEC1_IDR_DLANE0_MASK VC4_MASK(6, 4)
+# define DSI0_PHY_AFEC1_IDR_DLANE0_SHIFT 4
+# define DSI0_PHY_AFEC1_IDR_CLANE_MASK VC4_MASK(2, 0)
+# define DSI0_PHY_AFEC1_IDR_CLANE_SHIFT 0
+
+#define DSI0_TST_SEL 0x6c
+#define DSI0_TST_MON 0x70
+#define DSI0_ID 0x74
+# define DSI_ID_VALUE 0x00647369
+
+#define DSI1_CTRL 0x00
+# define DSI_CTRL_HS_CLKC_MASK VC4_MASK(15, 14)
+# define DSI_CTRL_HS_CLKC_SHIFT 14
+# define DSI_CTRL_HS_CLKC_BYTE 0
+# define DSI_CTRL_HS_CLKC_DDR2 1
+# define DSI_CTRL_HS_CLKC_DDR 2
+
+# define DSI_CTRL_RX_LPDT_EOT_DISABLE BIT(13)
+# define DSI_CTRL_LPDT_EOT_DISABLE BIT(12)
+# define DSI_CTRL_HSDT_EOT_DISABLE BIT(11)
+# define DSI_CTRL_SOFT_RESET_CFG BIT(10)
+# define DSI_CTRL_CAL_BYTE BIT(9)
+# define DSI_CTRL_INV_BYTE BIT(8)
+# define DSI_CTRL_CLR_LDF BIT(7)
+# define DSI0_CTRL_CLR_PBCF BIT(6)
+# define DSI1_CTRL_CLR_RXF BIT(6)
+# define DSI0_CTRL_CLR_CPBCF BIT(5)
+# define DSI1_CTRL_CLR_PDF BIT(5)
+# define DSI0_CTRL_CLR_PDF BIT(4)
+# define DSI1_CTRL_CLR_CDF BIT(4)
+# define DSI0_CTRL_CLR_CDF BIT(3)
+# define DSI0_CTRL_CTRL2 BIT(2)
+# define DSI1_CTRL_DISABLE_DISP_CRCC BIT(2)
+# define DSI0_CTRL_CTRL1 BIT(1)
+# define DSI1_CTRL_DISABLE_DISP_ECCC BIT(1)
+# define DSI0_CTRL_CTRL0 BIT(0)
+# define DSI1_CTRL_EN BIT(0)
+# define DSI0_CTRL_RESET_FIFOS (DSI_CTRL_CLR_LDF | \
+ DSI0_CTRL_CLR_PBCF | \
+ DSI0_CTRL_CLR_CPBCF | \
+ DSI0_CTRL_CLR_PDF | \
+ DSI0_CTRL_CLR_CDF)
+# define DSI1_CTRL_RESET_FIFOS (DSI_CTRL_CLR_LDF | \
+ DSI1_CTRL_CLR_RXF | \
+ DSI1_CTRL_CLR_PDF | \
+ DSI1_CTRL_CLR_CDF)
+
+#define DSI1_TXPKT2C 0x0c
+#define DSI1_TXPKT2H 0x10
+#define DSI1_TXPKT_PIX_FIFO 0x20
+#define DSI1_RXPKT_FIFO 0x24
+#define DSI1_DISP0_CTRL 0x28
+#define DSI1_INT_STAT 0x30
+#define DSI1_INT_EN 0x34
+/* State reporting bits. These mostly behave like INT_STAT, where
+ * writing a 1 clears the bit.
+ */
+#define DSI1_STAT 0x38
+# define DSI1_STAT_PHY_D3_ULPS BIT(31)
+# define DSI1_STAT_PHY_D3_STOP BIT(30)
+# define DSI1_STAT_PHY_D2_ULPS BIT(29)
+# define DSI1_STAT_PHY_D2_STOP BIT(28)
+# define DSI1_STAT_PHY_D1_ULPS BIT(27)
+# define DSI1_STAT_PHY_D1_STOP BIT(26)
+# define DSI1_STAT_PHY_D0_ULPS BIT(25)
+# define DSI1_STAT_PHY_D0_STOP BIT(24)
+# define DSI1_STAT_FIFO_ERR BIT(23)
+# define DSI1_STAT_PHY_RXLPDT BIT(22)
+# define DSI1_STAT_PHY_RXTRIG BIT(21)
+# define DSI1_STAT_PHY_D0_LPDT BIT(20)
+/* Set when in forward direction */
+# define DSI1_STAT_PHY_DIR BIT(19)
+# define DSI1_STAT_PHY_CLOCK_ULPS BIT(18)
+# define DSI1_STAT_PHY_CLOCK_HS BIT(17)
+# define DSI1_STAT_PHY_CLOCK_STOP BIT(16)
+# define DSI1_STAT_PR_TO BIT(15)
+# define DSI1_STAT_TA_TO BIT(14)
+# define DSI1_STAT_LPRX_TO BIT(13)
+# define DSI1_STAT_HSTX_TO BIT(12)
+# define DSI1_STAT_ERR_CONT_LP1 BIT(11)
+# define DSI1_STAT_ERR_CONT_LP0 BIT(10)
+# define DSI1_STAT_ERR_CONTROL BIT(9)
+# define DSI1_STAT_ERR_SYNC_ESC BIT(8)
+# define DSI1_STAT_RXPKT2 BIT(7)
+# define DSI1_STAT_RXPKT1 BIT(6)
+# define DSI1_STAT_TXPKT2_BUSY BIT(5)
+# define DSI1_STAT_TXPKT2_DONE BIT(4)
+# define DSI1_STAT_TXPKT2_END BIT(3)
+# define DSI1_STAT_TXPKT1_BUSY BIT(2)
+# define DSI1_STAT_TXPKT1_DONE BIT(1)
+# define DSI1_STAT_TXPKT1_END BIT(0)
+
+#define DSI1_HSTX_TO_CNT 0x3c
+#define DSI1_LPRX_TO_CNT 0x40
+#define DSI1_TA_TO_CNT 0x44
+#define DSI1_PR_TO_CNT 0x48
+#define DSI1_PHYC 0x4c
+
+#define DSI1_HS_CLT0 0x50
+# define DSI_HS_CLT0_CZERO_MASK VC4_MASK(26, 18)
+# define DSI_HS_CLT0_CZERO_SHIFT 18
+# define DSI_HS_CLT0_CPRE_MASK VC4_MASK(17, 9)
+# define DSI_HS_CLT0_CPRE_SHIFT 9
+# define DSI_HS_CLT0_CPREP_MASK VC4_MASK(8, 0)
+# define DSI_HS_CLT0_CPREP_SHIFT 0
+
+#define DSI1_HS_CLT1 0x54
+# define DSI_HS_CLT1_CTRAIL_MASK VC4_MASK(17, 9)
+# define DSI_HS_CLT1_CTRAIL_SHIFT 9
+# define DSI_HS_CLT1_CPOST_MASK VC4_MASK(8, 0)
+# define DSI_HS_CLT1_CPOST_SHIFT 0
+
+#define DSI1_HS_CLT2 0x58
+# define DSI_HS_CLT2_WUP_MASK VC4_MASK(23, 0)
+# define DSI_HS_CLT2_WUP_SHIFT 0
+
+#define DSI1_HS_DLT3 0x5c
+# define DSI_HS_DLT3_EXIT_MASK VC4_MASK(26, 18)
+# define DSI_HS_DLT3_EXIT_SHIFT 18
+# define DSI_HS_DLT3_ZERO_MASK VC4_MASK(17, 9)
+# define DSI_HS_DLT3_ZERO_SHIFT 9
+# define DSI_HS_DLT3_PRE_MASK VC4_MASK(8, 0)
+# define DSI_HS_DLT3_PRE_SHIFT 0
+
+#define DSI1_HS_DLT4 0x60
+# define DSI_HS_DLT4_ANLAT_MASK VC4_MASK(22, 18)
+# define DSI_HS_DLT4_ANLAT_SHIFT 18
+# define DSI_HS_DLT4_TRAIL_MASK VC4_MASK(17, 9)
+# define DSI_HS_DLT4_TRAIL_SHIFT 9
+# define DSI_HS_DLT4_LPX_MASK VC4_MASK(8, 0)
+# define DSI_HS_DLT4_LPX_SHIFT 0
+
+#define DSI1_HS_DLT5 0x64
+# define DSI_HS_DLT5_INIT_MASK VC4_MASK(23, 0)
+# define DSI_HS_DLT5_INIT_SHIFT 0
+
+#define DSI1_HS_DLT6 0x68
+# define DSI_HS_DLT6_TA_GET_MASK VC4_MASK(31, 24)
+# define DSI_HS_DLT6_TA_GET_SHIFT 24
+# define DSI_HS_DLT6_TA_SURE_MASK VC4_MASK(23, 16)
+# define DSI_HS_DLT6_TA_SURE_SHIFT 16
+# define DSI_HS_DLT6_TA_GO_MASK VC4_MASK(15, 8)
+# define DSI_HS_DLT6_TA_GO_SHIFT 8
+# define DSI_HS_DLT6_LP_LPX_MASK VC4_MASK(7, 0)
+# define DSI_HS_DLT6_LP_LPX_SHIFT 0
+
+#define DSI1_HS_DLT7 0x6c
+# define DSI_HS_DLT7_LP_WUP_MASK VC4_MASK(23, 0)
+# define DSI_HS_DLT7_LP_WUP_SHIFT 0
+
+#define DSI1_PHY_AFEC0 0x70
+
+#define DSI1_PHY_AFEC1 0x74
+# define DSI1_PHY_AFEC1_ACTRL_DLANE3_MASK VC4_MASK(19, 16)
+# define DSI1_PHY_AFEC1_ACTRL_DLANE3_SHIFT 16
+# define DSI1_PHY_AFEC1_ACTRL_DLANE2_MASK VC4_MASK(15, 12)
+# define DSI1_PHY_AFEC1_ACTRL_DLANE2_SHIFT 12
+# define DSI1_PHY_AFEC1_ACTRL_DLANE1_MASK VC4_MASK(11, 8)
+# define DSI1_PHY_AFEC1_ACTRL_DLANE1_SHIFT 8
+# define DSI1_PHY_AFEC1_ACTRL_DLANE0_MASK VC4_MASK(7, 4)
+# define DSI1_PHY_AFEC1_ACTRL_DLANE0_SHIFT 4
+# define DSI1_PHY_AFEC1_ACTRL_CLANE_MASK VC4_MASK(3, 0)
+# define DSI1_PHY_AFEC1_ACTRL_CLANE_SHIFT 0
+
+#define DSI1_TST_SEL 0x78
+#define DSI1_TST_MON 0x7c
+#define DSI1_PHY_TST1 0x80
+#define DSI1_PHY_TST2 0x84
+#define DSI1_PHY_FIFO_STAT 0x88
+/* Actually, all registers in the range that aren't otherwise claimed
+ * will return the ID.
+ */
+#define DSI1_ID 0x8c
+
+/* General DSI hardware state. */
+struct vc4_dsi {
+ struct platform_device *pdev;
+
+ struct mipi_dsi_host dsi_host;
+ struct drm_encoder *encoder;
+ struct drm_connector *connector;
+ struct drm_panel *panel;
+
+ void __iomem *regs;
+
+ struct dma_chan *reg_dma_chan;
+ dma_addr_t reg_dma_paddr;
+ u32 *reg_dma_mem;
+ dma_addr_t reg_paddr;
+
+ /* Whether we're on bcm2835's DSI0 or DSI1. */
+ int port;
+
+ /* DSI channel for the panel we're connected to. */
+ u32 channel;
+ u32 lanes;
+ enum mipi_dsi_pixel_format format;
+ u32 mode_flags;
+
+ /* Input clock from CPRMAN to the digital PHY, for the DSI
+ * escape clock.
+ */
+ struct clk *escape_clock;
+
+ /* Input clock to the analog PHY, used to generate the DSI bit
+ * clock.
+ */
+ struct clk *pll_phy_clock;
+
+ /* HS Clocks generated within the DSI analog PHY. */
+ struct clk_fixed_factor phy_clocks[3];
+
+ struct clk_hw_onecell_data *clk_onecell;
+
+ /* Pixel clock output to the pixelvalve, generated from the HS
+ * clock.
+ */
+ struct clk *pixel_clock;
+
+ struct completion xfer_completion;
+ int xfer_result;
+};
+
+#define host_to_dsi(host) container_of(host, struct vc4_dsi, dsi_host)
+
+static inline void
+dsi_dma_workaround_write(struct vc4_dsi *dsi, u32 offset, u32 val)
+{
+ struct dma_chan *chan = dsi->reg_dma_chan;
+ struct dma_async_tx_descriptor *tx;
+ dma_cookie_t cookie;
+ int ret;
+
+ /* DSI0 should be able to write normally. */
+ if (!chan) {
+ writel(val, dsi->regs + offset);
+ return;
+ }
+
+ *dsi->reg_dma_mem = val;
+
+ tx = chan->device->device_prep_dma_memcpy(chan,
+ dsi->reg_paddr + offset,
+ dsi->reg_dma_paddr,
+ 4, 0);
+ if (!tx) {
+ DRM_ERROR("Failed to set up DMA register write\n");
+ return;
+ }
+
+ cookie = tx->tx_submit(tx);
+ ret = dma_submit_error(cookie);
+ if (ret) {
+ DRM_ERROR("Failed to submit DMA: %d\n", ret);
+ return;
+ }
+ ret = dma_sync_wait(chan, cookie);
+ if (ret)
+ DRM_ERROR("Failed to wait for DMA: %d\n", ret);
+}
+
+#define DSI_READ(offset) readl(dsi->regs + (offset))
+#define DSI_WRITE(offset, val) dsi_dma_workaround_write(dsi, offset, val)
+#define DSI_PORT_READ(offset) \
+ DSI_READ(dsi->port ? DSI1_##offset : DSI0_##offset)
+#define DSI_PORT_WRITE(offset, val) \
+ DSI_WRITE(dsi->port ? DSI1_##offset : DSI0_##offset, val)
+#define DSI_PORT_BIT(bit) (dsi->port ? DSI1_##bit : DSI0_##bit)
+
+/* VC4 DSI encoder KMS struct */
+struct vc4_dsi_encoder {
+ struct vc4_encoder base;
+ struct vc4_dsi *dsi;
+};
+
+static inline struct vc4_dsi_encoder *
+to_vc4_dsi_encoder(struct drm_encoder *encoder)
+{
+ return container_of(encoder, struct vc4_dsi_encoder, base.base);
+}
+
+/* VC4 DSI connector KMS struct */
+struct vc4_dsi_connector {
+ struct drm_connector base;
+ struct vc4_dsi *dsi;
+};
+
+static inline struct vc4_dsi_connector *
+to_vc4_dsi_connector(struct drm_connector *connector)
+{
+ return container_of(connector, struct vc4_dsi_connector, base);
+}
+
+#define DSI_REG(reg) { reg, #reg }
+static const struct {
+ u32 reg;
+ const char *name;
+} dsi0_regs[] = {
+ DSI_REG(DSI0_CTRL),
+ DSI_REG(DSI0_STAT),
+ DSI_REG(DSI0_HSTX_TO_CNT),
+ DSI_REG(DSI0_LPRX_TO_CNT),
+ DSI_REG(DSI0_TA_TO_CNT),
+ DSI_REG(DSI0_PR_TO_CNT),
+ DSI_REG(DSI0_DISP0_CTRL),
+ DSI_REG(DSI0_DISP1_CTRL),
+ DSI_REG(DSI0_INT_STAT),
+ DSI_REG(DSI0_INT_EN),
+ DSI_REG(DSI0_PHYC),
+ DSI_REG(DSI0_HS_CLT0),
+ DSI_REG(DSI0_HS_CLT1),
+ DSI_REG(DSI0_HS_CLT2),
+ DSI_REG(DSI0_HS_DLT3),
+ DSI_REG(DSI0_HS_DLT4),
+ DSI_REG(DSI0_HS_DLT5),
+ DSI_REG(DSI0_HS_DLT6),
+ DSI_REG(DSI0_HS_DLT7),
+ DSI_REG(DSI0_PHY_AFEC0),
+ DSI_REG(DSI0_PHY_AFEC1),
+ DSI_REG(DSI0_ID),
+};
+
+static const struct {
+ u32 reg;
+ const char *name;
+} dsi1_regs[] = {
+ DSI_REG(DSI1_CTRL),
+ DSI_REG(DSI1_STAT),
+ DSI_REG(DSI1_HSTX_TO_CNT),
+ DSI_REG(DSI1_LPRX_TO_CNT),
+ DSI_REG(DSI1_TA_TO_CNT),
+ DSI_REG(DSI1_PR_TO_CNT),
+ DSI_REG(DSI1_DISP0_CTRL),
+ DSI_REG(DSI1_DISP1_CTRL),
+ DSI_REG(DSI1_INT_STAT),
+ DSI_REG(DSI1_INT_EN),
+ DSI_REG(DSI1_PHYC),
+ DSI_REG(DSI1_HS_CLT0),
+ DSI_REG(DSI1_HS_CLT1),
+ DSI_REG(DSI1_HS_CLT2),
+ DSI_REG(DSI1_HS_DLT3),
+ DSI_REG(DSI1_HS_DLT4),
+ DSI_REG(DSI1_HS_DLT5),
+ DSI_REG(DSI1_HS_DLT6),
+ DSI_REG(DSI1_HS_DLT7),
+ DSI_REG(DSI1_PHY_AFEC0),
+ DSI_REG(DSI1_PHY_AFEC1),
+ DSI_REG(DSI1_ID),
+};
+
+static void vc4_dsi_dump_regs(struct vc4_dsi *dsi)
+{
+ int i;
+
+ if (dsi->port == 0) {
+ for (i = 0; i < ARRAY_SIZE(dsi0_regs); i++) {
+ DRM_INFO("0x%04x (%s): 0x%08x\n",
+ dsi0_regs[i].reg, dsi0_regs[i].name,
+ DSI_READ(dsi0_regs[i].reg));
+ }
+ } else {
+ for (i = 0; i < ARRAY_SIZE(dsi1_regs); i++) {
+ DRM_INFO("0x%04x (%s): 0x%08x\n",
+ dsi1_regs[i].reg, dsi1_regs[i].name,
+ DSI_READ(dsi1_regs[i].reg));
+ }
+ }
+}
+
+#ifdef CONFIG_DEBUG_FS
+int vc4_dsi_debugfs_regs(struct seq_file *m, void *unused)
+{
+ struct drm_info_node *node = (struct drm_info_node *)m->private;
+ struct drm_device *drm = node->minor->dev;
+ struct vc4_dev *vc4 = to_vc4_dev(drm);
+ int dsi_index = (uintptr_t)node->info_ent->data;
+ struct vc4_dsi *dsi = (dsi_index == 1 ? vc4->dsi1 : NULL);
+ int i;
+
+ if (!dsi)
+ return 0;
+
+ if (dsi->port == 0) {
+ for (i = 0; i < ARRAY_SIZE(dsi0_regs); i++) {
+ seq_printf(m, "0x%04x (%s): 0x%08x\n",
+ dsi0_regs[i].reg, dsi0_regs[i].name,
+ DSI_READ(dsi0_regs[i].reg));
+ }
+ } else {
+ for (i = 0; i < ARRAY_SIZE(dsi1_regs); i++) {
+ seq_printf(m, "0x%04x (%s): 0x%08x\n",
+ dsi1_regs[i].reg, dsi1_regs[i].name,
+ DSI_READ(dsi1_regs[i].reg));
+ }
+ }
+
+ return 0;
+}
+#endif
+
+static enum drm_connector_status
+vc4_dsi_connector_detect(struct drm_connector *connector, bool force)
+{
+ struct vc4_dsi_connector *vc4_connector =
+ to_vc4_dsi_connector(connector);
+ struct vc4_dsi *dsi = vc4_connector->dsi;
+
+ if (dsi->panel)
+ return connector_status_connected;
+ else
+ return connector_status_disconnected;
+}
+
+static void vc4_dsi_connector_destroy(struct drm_connector *connector)
+{
+ drm_connector_unregister(connector);
+ drm_connector_cleanup(connector);
+}
+
+static int vc4_dsi_connector_get_modes(struct drm_connector *connector)
+{
+ struct vc4_dsi_connector *vc4_connector =
+ to_vc4_dsi_connector(connector);
+ struct vc4_dsi *dsi = vc4_connector->dsi;
+
+ if (dsi->panel)
+ return drm_panel_get_modes(dsi->panel);
+
+ return 0;
+}
+
+static const struct drm_connector_funcs vc4_dsi_connector_funcs = {
+ .dpms = drm_atomic_helper_connector_dpms,
+ .detect = vc4_dsi_connector_detect,
+ .fill_modes = drm_helper_probe_single_connector_modes,
+ .destroy = vc4_dsi_connector_destroy,
+ .reset = drm_atomic_helper_connector_reset,
+ .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+ .atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+};
+
+static const struct drm_connector_helper_funcs vc4_dsi_connector_helper_funcs = {
+ .get_modes = vc4_dsi_connector_get_modes,
+};
+
+static struct drm_connector *vc4_dsi_connector_init(struct drm_device *dev,
+ struct vc4_dsi *dsi)
+{
+ struct drm_connector *connector = NULL;
+ struct vc4_dsi_connector *dsi_connector;
+ int ret = 0;
+
+ dsi_connector = devm_kzalloc(dev->dev, sizeof(*dsi_connector),
+ GFP_KERNEL);
+ if (!dsi_connector) {
+ ret = -ENOMEM;
+ goto fail;
+ }
+ connector = &dsi_connector->base;
+
+ dsi_connector->dsi = dsi;
+
+ drm_connector_init(dev, connector, &vc4_dsi_connector_funcs,
+ DRM_MODE_CONNECTOR_DSI);
+ drm_connector_helper_add(connector, &vc4_dsi_connector_helper_funcs);
+
+ connector->polled = 0;
+ connector->interlace_allowed = 0;
+ connector->doublescan_allowed = 0;
+
+ drm_mode_connector_attach_encoder(connector, dsi->encoder);
+
+ return connector;
+
+fail:
+ if (connector)
+ vc4_dsi_connector_destroy(connector);
+
+ return ERR_PTR(ret);
+}
+
+static void vc4_dsi_encoder_destroy(struct drm_encoder *encoder)
+{
+ drm_encoder_cleanup(encoder);
+}
+
+static const struct drm_encoder_funcs vc4_dsi_encoder_funcs = {
+ .destroy = vc4_dsi_encoder_destroy,
+};
+
+static void vc4_dsi_latch_ulps(struct vc4_dsi *dsi, bool latch)
+{
+ u32 afec0 = DSI_PORT_READ(PHY_AFEC0);
+
+ if (latch)
+ afec0 |= DSI_PORT_BIT(PHY_AFEC0_LATCH_ULPS);
+ else
+ afec0 &= ~DSI_PORT_BIT(PHY_AFEC0_LATCH_ULPS);
+
+ DSI_PORT_WRITE(PHY_AFEC0, afec0);
+}
+
+/* Enters or exits Ultra Low Power State. */
+static void vc4_dsi_ulps(struct vc4_dsi *dsi, bool ulps)
+{
+ bool continuous = dsi->mode_flags & MIPI_DSI_CLOCK_NON_CONTINUOUS;
+ u32 phyc_ulps = ((continuous ? DSI_PORT_BIT(PHYC_CLANE_ULPS) : 0) |
+ DSI_PHYC_DLANE0_ULPS |
+ (dsi->lanes > 1 ? DSI_PHYC_DLANE1_ULPS : 0) |
+ (dsi->lanes > 2 ? DSI_PHYC_DLANE2_ULPS : 0) |
+ (dsi->lanes > 3 ? DSI_PHYC_DLANE3_ULPS : 0));
+ u32 stat_ulps = ((continuous ? DSI1_STAT_PHY_CLOCK_ULPS : 0) |
+ DSI1_STAT_PHY_D0_ULPS |
+ (dsi->lanes > 1 ? DSI1_STAT_PHY_D1_ULPS : 0) |
+ (dsi->lanes > 2 ? DSI1_STAT_PHY_D2_ULPS : 0) |
+ (dsi->lanes > 3 ? DSI1_STAT_PHY_D3_ULPS : 0));
+ u32 stat_stop = ((continuous ? DSI1_STAT_PHY_CLOCK_STOP : 0) |
+ DSI1_STAT_PHY_D0_STOP |
+ (dsi->lanes > 1 ? DSI1_STAT_PHY_D1_STOP : 0) |
+ (dsi->lanes > 2 ? DSI1_STAT_PHY_D2_STOP : 0) |
+ (dsi->lanes > 3 ? DSI1_STAT_PHY_D3_STOP : 0));
+ int ret;
+
+ DSI_PORT_WRITE(STAT, stat_ulps);
+ DSI_PORT_WRITE(PHYC, DSI_PORT_READ(PHYC) | phyc_ulps);
+ ret = wait_for((DSI_PORT_READ(STAT) & stat_ulps) == stat_ulps, 200);
+ if (ret) {
+ dev_warn(&dsi->pdev->dev,
+ "Timeout waiting for DSI ULPS entry: STAT 0x%08x",
+ DSI_PORT_READ(STAT));
+ DSI_PORT_WRITE(PHYC, DSI_PORT_READ(PHYC) & ~phyc_ulps);
+ vc4_dsi_latch_ulps(dsi, false);
+ return;
+ }
+
+ /* The DSI module can't be disabled while the module is
+ * generating ULPS state. So, to be able to disable the
+ * module, we have the AFE latch the ULPS state and continue
+ * on to having the module enter STOP.
+ */
+ vc4_dsi_latch_ulps(dsi, ulps);
+
+ DSI_PORT_WRITE(STAT, stat_stop);
+ DSI_PORT_WRITE(PHYC, DSI_PORT_READ(PHYC) & ~phyc_ulps);
+ ret = wait_for((DSI_PORT_READ(STAT) & stat_stop) == stat_stop, 200);
+ if (ret) {
+ dev_warn(&dsi->pdev->dev,
+ "Timeout waiting for DSI STOP entry: STAT 0x%08x",
+ DSI_PORT_READ(STAT));
+ DSI_PORT_WRITE(PHYC, DSI_PORT_READ(PHYC) & ~phyc_ulps);
+ return;
+ }
+}
+
+static u32
+dsi_hs_timing(u32 ui_ns, u32 ns, u32 ui)
+{
+ /* The HS timings have to be rounded up to a multiple of 8
+ * because we're using the byte clock.
+ */
+ return roundup(ui + DIV_ROUND_UP(ns, ui_ns), 8);
+}
+
+/* ESC always runs at 100Mhz. */
+#define ESC_TIME_NS 10
+
+static u32
+dsi_esc_timing(u32 ns)
+{
+ return DIV_ROUND_UP(ns, ESC_TIME_NS);
+}
+
+static void vc4_dsi_encoder_disable(struct drm_encoder *encoder)
+{
+ struct vc4_dsi_encoder *vc4_encoder = to_vc4_dsi_encoder(encoder);
+ struct vc4_dsi *dsi = vc4_encoder->dsi;
+ struct device *dev = &dsi->pdev->dev;
+
+ drm_panel_disable(dsi->panel);
+
+ vc4_dsi_ulps(dsi, true);
+
+ drm_panel_unprepare(dsi->panel);
+
+ clk_disable_unprepare(dsi->pll_phy_clock);
+ clk_disable_unprepare(dsi->escape_clock);
+ clk_disable_unprepare(dsi->pixel_clock);
+
+ pm_runtime_put(dev);
+}
+
+static void vc4_dsi_encoder_enable(struct drm_encoder *encoder)
+{
+ struct drm_display_mode *mode = &encoder->crtc->mode;
+ struct vc4_dsi_encoder *vc4_encoder = to_vc4_dsi_encoder(encoder);
+ struct vc4_dsi *dsi = vc4_encoder->dsi;
+ struct device *dev = &dsi->pdev->dev;
+ u32 format = 0, divider = 0;
+ bool debug_dump_regs = false;
+ unsigned long hs_clock;
+ u32 ui_ns;
+ /* Minimum LP state duration in escape clock cycles. */
+ u32 lpx = dsi_esc_timing(60);
+ unsigned long pixel_clock_hz = mode->clock * 1000;
+ unsigned long dsip_clock;
+ unsigned long phy_clock;
+ int ret;
+
+ ret = pm_runtime_get_sync(dev);
+ if (ret) {
+ DRM_ERROR("Failed to runtime PM enable on DSI%d\n", dsi->port);
+ return;
+ }
+
+ ret = drm_panel_prepare(dsi->panel);
+ if (ret) {
+ DRM_ERROR("Panel failed to prepare\n");
+ return;
+ }
+
+ if (debug_dump_regs) {
+ DRM_INFO("DSI regs before:\n");
+ vc4_dsi_dump_regs(dsi);
+ }
+
+ switch (dsi->format) {
+ case MIPI_DSI_FMT_RGB888:
+ format = DSI_PFORMAT_RGB888;
+ divider = 24 / dsi->lanes;
+ break;
+ case MIPI_DSI_FMT_RGB666:
+ format = DSI_PFORMAT_RGB666;
+ divider = 24 / dsi->lanes;
+ break;
+ case MIPI_DSI_FMT_RGB666_PACKED:
+ format = DSI_PFORMAT_RGB666_PACKED;
+ divider = 18 / dsi->lanes;
+ break;
+ case MIPI_DSI_FMT_RGB565:
+ format = DSI_PFORMAT_RGB565;
+ divider = 16 / dsi->lanes;
+ break;
+ }
+
+ phy_clock = pixel_clock_hz * divider;
+ ret = clk_set_rate(dsi->pll_phy_clock, phy_clock);
+ if (ret) {
+ dev_err(&dsi->pdev->dev,
+ "Failed to set phy clock to %ld: %d\n", phy_clock, ret);
+ }
+
+ /* Reset the DSI and all its fifos. */
+ DSI_PORT_WRITE(CTRL,
+ DSI_CTRL_SOFT_RESET_CFG |
+ DSI_PORT_BIT(CTRL_RESET_FIFOS));
+
+ DSI_PORT_WRITE(CTRL,
+ DSI_CTRL_HSDT_EOT_DISABLE |
+ DSI_CTRL_RX_LPDT_EOT_DISABLE);
+
+ /* Clear all stat bits so we see what has happened during enable. */
+ DSI_PORT_WRITE(STAT, DSI_PORT_READ(STAT));
+
+ /* Set AFE CTR00/CTR1 to release powerdown of analog. */
+ if (dsi->port == 0) {
+ u32 afec0 = (VC4_SET_FIELD(7, DSI_PHY_AFEC0_PTATADJ) |
+ VC4_SET_FIELD(7, DSI_PHY_AFEC0_CTATADJ));
+
+ if (dsi->lanes < 2)
+ afec0 |= DSI0_PHY_AFEC0_PD_DLANE1;
+
+ if (!(dsi->mode_flags & MIPI_DSI_MODE_VIDEO))
+ afec0 |= DSI0_PHY_AFEC0_RESET;
+
+ DSI_PORT_WRITE(PHY_AFEC0, afec0);
+
+ DSI_PORT_WRITE(PHY_AFEC1,
+ VC4_SET_FIELD(6, DSI0_PHY_AFEC1_IDR_DLANE1) |
+ VC4_SET_FIELD(6, DSI0_PHY_AFEC1_IDR_DLANE0) |
+ VC4_SET_FIELD(6, DSI0_PHY_AFEC1_IDR_CLANE));
+ } else {
+ u32 afec0 = (VC4_SET_FIELD(7, DSI_PHY_AFEC0_PTATADJ) |
+ VC4_SET_FIELD(7, DSI_PHY_AFEC0_CTATADJ) |
+ VC4_SET_FIELD(6, DSI1_PHY_AFEC0_IDR_CLANE) |
+ VC4_SET_FIELD(6, DSI1_PHY_AFEC0_IDR_DLANE0) |
+ VC4_SET_FIELD(6, DSI1_PHY_AFEC0_IDR_DLANE1) |
+ VC4_SET_FIELD(6, DSI1_PHY_AFEC0_IDR_DLANE2) |
+ VC4_SET_FIELD(6, DSI1_PHY_AFEC0_IDR_DLANE3));
+
+ if (dsi->lanes < 4)
+ afec0 |= DSI1_PHY_AFEC0_PD_DLANE3;
+ if (dsi->lanes < 3)
+ afec0 |= DSI1_PHY_AFEC0_PD_DLANE2;
+ if (dsi->lanes < 2)
+ afec0 |= DSI1_PHY_AFEC0_PD_DLANE1;
+
+ afec0 |= DSI1_PHY_AFEC0_RESET;
+
+ DSI_PORT_WRITE(PHY_AFEC0, afec0);
+
+ DSI_PORT_WRITE(PHY_AFEC1, 0);
+
+ /* AFEC reset hold time */
+ mdelay(1);
+ }
+
+ ret = clk_prepare_enable(dsi->escape_clock);
+ if (ret) {
+ DRM_ERROR("Failed to turn on DSI escape clock: %d\n", ret);
+ return;
+ }
+
+ ret = clk_prepare_enable(dsi->pll_phy_clock);
+ if (ret) {
+ DRM_ERROR("Failed to turn on DSI PLL: %d\n", ret);
+ return;
+ }
+
+ hs_clock = clk_get_rate(dsi->pll_phy_clock);
+
+ /* Yes, we set the DSI0P/DSI1P pixel clock to the byte rate,
+ * not the pixel clock rate. DSIxP take from the APHY's byte,
+ * DDR2, or DDR4 clock (we use byte) and feed into the PV at
+ * that rate. Separately, a value derived from PIX_CLK_DIV
+ * and HS_CLKC is fed into the PV to divide down to the actual
+ * pixel clock for pushing pixels into DSI.
+ */
+ dsip_clock = phy_clock / 8;
+ ret = clk_set_rate(dsi->pixel_clock, dsip_clock);
+ if (ret) {
+ dev_err(dev, "Failed to set pixel clock to %ldHz: %d\n",
+ dsip_clock, ret);
+ }
+
+ ret = clk_prepare_enable(dsi->pixel_clock);
+ if (ret) {
+ DRM_ERROR("Failed to turn on DSI pixel clock: %d\n", ret);
+ return;
+ }
+
+ /* How many ns one DSI unit interval is. Note that the clock
+ * is DDR, so there's an extra divide by 2.
+ */
+ ui_ns = DIV_ROUND_UP(500000000, hs_clock);
+
+ DSI_PORT_WRITE(HS_CLT0,
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 262, 0),
+ DSI_HS_CLT0_CZERO) |
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 0, 8),
+ DSI_HS_CLT0_CPRE) |
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 38, 0),
+ DSI_HS_CLT0_CPREP));
+
+ DSI_PORT_WRITE(HS_CLT1,
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 60, 0),
+ DSI_HS_CLT1_CTRAIL) |
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 60, 52),
+ DSI_HS_CLT1_CPOST));
+
+ DSI_PORT_WRITE(HS_CLT2,
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 1000000, 0),
+ DSI_HS_CLT2_WUP));
+
+ DSI_PORT_WRITE(HS_DLT3,
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 100, 0),
+ DSI_HS_DLT3_EXIT) |
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 105, 6),
+ DSI_HS_DLT3_ZERO) |
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, 40, 4),
+ DSI_HS_DLT3_PRE));
+
+ DSI_PORT_WRITE(HS_DLT4,
+ VC4_SET_FIELD(dsi_hs_timing(ui_ns, lpx * ESC_TIME_NS, 0),
+ DSI_HS_DLT4_LPX) |
+ VC4_SET_FIELD(max(dsi_hs_timing(ui_ns, 0, 8),
+ dsi_hs_timing(ui_ns, 60, 4)),
+ DSI_HS_DLT4_TRAIL) |
+ VC4_SET_FIELD(0, DSI_HS_DLT4_ANLAT));
+
+ DSI_PORT_WRITE(HS_DLT5, VC4_SET_FIELD(dsi_hs_timing(ui_ns, 1000, 5000),
+ DSI_HS_DLT5_INIT));
+
+ DSI_PORT_WRITE(HS_DLT6,
+ VC4_SET_FIELD(lpx * 5, DSI_HS_DLT6_TA_GET) |
+ VC4_SET_FIELD(lpx, DSI_HS_DLT6_TA_SURE) |
+ VC4_SET_FIELD(lpx * 4, DSI_HS_DLT6_TA_GO) |
+ VC4_SET_FIELD(lpx, DSI_HS_DLT6_LP_LPX));
+
+ DSI_PORT_WRITE(HS_DLT7,
+ VC4_SET_FIELD(dsi_esc_timing(1000000),
+ DSI_HS_DLT7_LP_WUP));
+
+ DSI_PORT_WRITE(PHYC,
+ DSI_PHYC_DLANE0_ENABLE |
+ (dsi->lanes >= 2 ? DSI_PHYC_DLANE1_ENABLE : 0) |
+ (dsi->lanes >= 3 ? DSI_PHYC_DLANE2_ENABLE : 0) |
+ (dsi->lanes >= 4 ? DSI_PHYC_DLANE3_ENABLE : 0) |
+ DSI_PORT_BIT(PHYC_CLANE_ENABLE) |
+ ((dsi->mode_flags & MIPI_DSI_CLOCK_NON_CONTINUOUS) ?
+ 0 : DSI_PORT_BIT(PHYC_HS_CLK_CONTINUOUS)) |
+ (dsi->port == 0 ?
+ VC4_SET_FIELD(lpx - 1, DSI0_PHYC_ESC_CLK_LPDT) :
+ VC4_SET_FIELD(lpx - 1, DSI1_PHYC_ESC_CLK_LPDT)));
+
+ DSI_PORT_WRITE(CTRL,
+ DSI_PORT_READ(CTRL) |
+ DSI_CTRL_CAL_BYTE);
+
+ /* HS timeout in HS clock cycles: disabled. */
+ DSI_PORT_WRITE(HSTX_TO_CNT, 0);
+ /* LP receive timeout in HS clocks. */
+ DSI_PORT_WRITE(LPRX_TO_CNT, 0xffffff);
+ /* Bus turnaround timeout */
+ DSI_PORT_WRITE(TA_TO_CNT, 100000);
+ /* Display reset sequence timeout */
+ DSI_PORT_WRITE(PR_TO_CNT, 100000);
+
+ if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO) {
+ DSI_PORT_WRITE(DISP0_CTRL,
+ VC4_SET_FIELD(divider, DSI_DISP0_PIX_CLK_DIV) |
+ VC4_SET_FIELD(format, DSI_DISP0_PFORMAT) |
+ VC4_SET_FIELD(DSI_DISP0_LP_STOP_PERFRAME,
+ DSI_DISP0_LP_STOP_CTRL) |
+ DSI_DISP0_ST_END |
+ DSI_DISP0_ENABLE);
+ } else {
+ DSI_PORT_WRITE(DISP0_CTRL,
+ DSI_DISP0_COMMAND_MODE |
+ DSI_DISP0_ENABLE);
+ }
+
+ /* Set up DISP1 for transferring long command payloads through
+ * the pixfifo.
+ */
+ DSI_PORT_WRITE(DISP1_CTRL,
+ VC4_SET_FIELD(DSI_DISP1_PFORMAT_32BIT_LE,
+ DSI_DISP1_PFORMAT) |
+ DSI_DISP1_ENABLE);
+
+ /* Ungate the block. */
+ if (dsi->port == 0)
+ DSI_PORT_WRITE(CTRL, DSI_PORT_READ(CTRL) | DSI0_CTRL_CTRL0);
+ else
+ DSI_PORT_WRITE(CTRL, DSI_PORT_READ(CTRL) | DSI1_CTRL_EN);
+
+ /* Bring AFE out of reset. */
+ if (dsi->port == 0) {
+ } else {
+ DSI_PORT_WRITE(PHY_AFEC0,
+ DSI_PORT_READ(PHY_AFEC0) &
+ ~DSI1_PHY_AFEC0_RESET);
+ }
+
+ vc4_dsi_ulps(dsi, false);
+
+ if (debug_dump_regs) {
+ DRM_INFO("DSI regs after:\n");
+ vc4_dsi_dump_regs(dsi);
+ }
+
+ ret = drm_panel_enable(dsi->panel);
+ if (ret) {
+ DRM_ERROR("Panel failed to enable\n");
+ drm_panel_unprepare(dsi->panel);
+ return;
+ }
+}
+
+static ssize_t vc4_dsi_host_transfer(struct mipi_dsi_host *host,
+ const struct mipi_dsi_msg *msg)
+{
+ struct vc4_dsi *dsi = host_to_dsi(host);
+ struct mipi_dsi_packet packet;
+ u32 pkth = 0, pktc = 0;
+ int i, ret;
+ bool is_long = mipi_dsi_packet_format_is_long(msg->type);
+ u32 cmd_fifo_len = 0, pix_fifo_len = 0;
+
+ mipi_dsi_create_packet(&packet, msg);
+
+ pkth |= VC4_SET_FIELD(packet.header[0], DSI_TXPKT1H_BC_DT);
+ pkth |= VC4_SET_FIELD(packet.header[1] |
+ (packet.header[2] << 8),
+ DSI_TXPKT1H_BC_PARAM);
+ if (is_long) {
+ /* Divide data across the various FIFOs we have available.
+ * The command FIFO takes byte-oriented data, but is of
+ * limited size. The pixel FIFO (never actually used for
+ * pixel data in reality) is word oriented, and substantially
+ * larger. So, we use the pixel FIFO for most of the data,
+ * sending the residual bytes in the command FIFO at the start.
+ *
+ * With this arrangement, the command FIFO will never get full.
+ */
+ if (packet.payload_length <= 16) {
+ cmd_fifo_len = packet.payload_length;
+ pix_fifo_len = 0;
+ } else {
+ cmd_fifo_len = (packet.payload_length %
+ DSI_PIX_FIFO_WIDTH);
+ pix_fifo_len = ((packet.payload_length - cmd_fifo_len) /
+ DSI_PIX_FIFO_WIDTH);
+ }
+
+ WARN_ON_ONCE(pix_fifo_len >= DSI_PIX_FIFO_DEPTH);
+
+ pkth |= VC4_SET_FIELD(cmd_fifo_len, DSI_TXPKT1H_BC_CMDFIFO);
+ }
+
+ if (msg->rx_len) {
+ pktc |= VC4_SET_FIELD(DSI_TXPKT1C_CMD_CTRL_RX,
+ DSI_TXPKT1C_CMD_CTRL);
+ } else {
+ pktc |= VC4_SET_FIELD(DSI_TXPKT1C_CMD_CTRL_TX,
+ DSI_TXPKT1C_CMD_CTRL);
+ }
+
+ for (i = 0; i < cmd_fifo_len; i++)
+ DSI_PORT_WRITE(TXPKT_CMD_FIFO, packet.payload[i]);
+ for (i = 0; i < pix_fifo_len; i++) {
+ const u8 *pix = packet.payload + cmd_fifo_len + i * 4;
+
+ DSI_PORT_WRITE(TXPKT_PIX_FIFO,
+ pix[0] |
+ pix[1] << 8 |
+ pix[2] << 16 |
+ pix[3] << 24);
+ }
+
+ if (msg->flags & MIPI_DSI_MSG_USE_LPM)
+ pktc |= DSI_TXPKT1C_CMD_MODE_LP;
+ if (is_long)
+ pktc |= DSI_TXPKT1C_CMD_TYPE_LONG;
+
+ /* Send one copy of the packet. Larger repeats are used for pixel
+ * data in command mode.
+ */
+ pktc |= VC4_SET_FIELD(1, DSI_TXPKT1C_CMD_REPEAT);
+
+ pktc |= DSI_TXPKT1C_CMD_EN;
+ if (pix_fifo_len) {
+ pktc |= VC4_SET_FIELD(DSI_TXPKT1C_DISPLAY_NO_SECONDARY,
+ DSI_TXPKT1C_DISPLAY_NO);
+ } else {
+ pktc |= VC4_SET_FIELD(DSI_TXPKT1C_DISPLAY_NO_SHORT,
+ DSI_TXPKT1C_DISPLAY_NO);
+ }
+
+ /* Enable the appropriate interrupt for the transfer completion. */
+ dsi->xfer_result = 0;
+ reinit_completion(&dsi->xfer_completion);
+ DSI_PORT_WRITE(INT_STAT, DSI1_INT_TXPKT1_DONE | DSI1_INT_PHY_DIR_RTF);
+ if (msg->rx_len) {
+ DSI_PORT_WRITE(INT_EN, (DSI1_INTERRUPTS_ALWAYS_ENABLED |
+ DSI1_INT_PHY_DIR_RTF));
+ } else {
+ DSI_PORT_WRITE(INT_EN, (DSI1_INTERRUPTS_ALWAYS_ENABLED |
+ DSI1_INT_TXPKT1_DONE));
+ }
+
+ /* Send the packet. */
+ DSI_PORT_WRITE(TXPKT1H, pkth);
+ DSI_PORT_WRITE(TXPKT1C, pktc);
+
+ if (!wait_for_completion_timeout(&dsi->xfer_completion,
+ msecs_to_jiffies(1000))) {
+ dev_err(&dsi->pdev->dev, "transfer interrupt wait timeout");
+ dev_err(&dsi->pdev->dev, "instat: 0x%08x\n",
+ DSI_PORT_READ(INT_STAT));
+ ret = -ETIMEDOUT;
+ } else {
+ ret = dsi->xfer_result;
+ }
+
+ DSI_PORT_WRITE(INT_EN, DSI1_INTERRUPTS_ALWAYS_ENABLED);
+
+ if (ret)
+ goto reset_fifo_and_return;
+
+ if (ret == 0 && msg->rx_len) {
+ u32 rxpkt1h = DSI_PORT_READ(RXPKT1H);
+ u8 *msg_rx = msg->rx_buf;
+
+ if (rxpkt1h & DSI_RXPKT1H_PKT_TYPE_LONG) {
+ u32 rxlen = VC4_GET_FIELD(rxpkt1h,
+ DSI_RXPKT1H_BC_PARAM);
+
+ if (rxlen != msg->rx_len) {
+ DRM_ERROR("DSI returned %db, expecting %db\n",
+ rxlen, (int)msg->rx_len);
+ ret = -ENXIO;
+ goto reset_fifo_and_return;
+ }
+
+ for (i = 0; i < msg->rx_len; i++)
+ msg_rx[i] = DSI_READ(DSI1_RXPKT_FIFO);
+ } else {
+ /* FINISHME: Handle AWER */
+
+ msg_rx[0] = VC4_GET_FIELD(rxpkt1h,
+ DSI_RXPKT1H_SHORT_0);
+ if (msg->rx_len > 1) {
+ msg_rx[1] = VC4_GET_FIELD(rxpkt1h,
+ DSI_RXPKT1H_SHORT_1);
+ }
+ }
+ }
+
+ return ret;
+
+reset_fifo_and_return:
+ DRM_ERROR("DSI transfer failed, resetting: %d\n", ret);
+
+ DSI_PORT_WRITE(TXPKT1C, DSI_PORT_READ(TXPKT1C) & ~DSI_TXPKT1C_CMD_EN);
+ udelay(1);
+ DSI_PORT_WRITE(CTRL,
+ DSI_PORT_READ(CTRL) |
+ DSI_PORT_BIT(CTRL_RESET_FIFOS));
+
+ DSI_PORT_WRITE(TXPKT1C, 0);
+ DSI_PORT_WRITE(INT_EN, DSI1_INTERRUPTS_ALWAYS_ENABLED);
+ return ret;
+}
+
+static int vc4_dsi_host_attach(struct mipi_dsi_host *host,
+ struct mipi_dsi_device *device)
+{
+ struct vc4_dsi *dsi = host_to_dsi(host);
+ int ret = 0;
+
+ dsi->lanes = device->lanes;
+ dsi->channel = device->channel;
+ dsi->format = device->format;
+ dsi->mode_flags = device->mode_flags;
+
+ if (!(dsi->mode_flags & MIPI_DSI_MODE_VIDEO)) {
+ dev_err(&dsi->pdev->dev,
+ "Only VIDEO mode panels supported currently.\n");
+ return 0;
+ }
+
+ dsi->panel = of_drm_find_panel(device->dev.of_node);
+ if (!dsi->panel)
+ return 0;
+
+ ret = drm_panel_attach(dsi->panel, dsi->connector);
+ if (ret != 0)
+ return ret;
+
+ drm_helper_hpd_irq_event(dsi->connector->dev);
+
+ return 0;
+}
+
+static int vc4_dsi_host_detach(struct mipi_dsi_host *host,
+ struct mipi_dsi_device *device)
+{
+ struct vc4_dsi *dsi = host_to_dsi(host);
+
+ if (dsi->panel) {
+ int ret = drm_panel_detach(dsi->panel);
+
+ if (ret)
+ return ret;
+
+ dsi->panel = NULL;
+
+ drm_helper_hpd_irq_event(dsi->connector->dev);
+ }
+
+ return 0;
+}
+
+static const struct mipi_dsi_host_ops vc4_dsi_host_ops = {
+ .attach = vc4_dsi_host_attach,
+ .detach = vc4_dsi_host_detach,
+ .transfer = vc4_dsi_host_transfer,
+};
+
+static const struct drm_encoder_helper_funcs vc4_dsi_encoder_helper_funcs = {
+ .disable = vc4_dsi_encoder_disable,
+ .enable = vc4_dsi_encoder_enable,
+};
+
+static const struct of_device_id vc4_dsi_dt_match[] = {
+ { .compatible = "brcm,bcm2835-dsi1", (void *)(uintptr_t)1 },
+ {}
+};
+
+static void dsi_handle_error(struct vc4_dsi *dsi,
+ irqreturn_t *ret, u32 stat, u32 bit,
+ const char *type)
+{
+ if (!(stat & bit))
+ return;
+
+ DRM_ERROR("DSI%d: %s error\n", dsi->port, type);
+ *ret = IRQ_HANDLED;
+}
+
+static irqreturn_t vc4_dsi_irq_handler(int irq, void *data)
+{
+ struct vc4_dsi *dsi = data;
+ u32 stat = DSI_PORT_READ(INT_STAT);
+ irqreturn_t ret = IRQ_NONE;
+
+ DSI_PORT_WRITE(INT_STAT, stat);
+
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_ERR_SYNC_ESC, "LPDT sync");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_ERR_CONTROL, "data lane 0 sequence");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_ERR_CONT_LP0, "LP0 contention");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_ERR_CONT_LP1, "LP1 contention");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_HSTX_TO, "HSTX timeout");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_LPRX_TO, "LPRX timeout");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_TA_TO, "turnaround timeout");
+ dsi_handle_error(dsi, &ret, stat,
+ DSI1_INT_PR_TO, "peripheral reset timeout");
+
+ if (stat & (DSI1_INT_TXPKT1_DONE | DSI1_INT_PHY_DIR_RTF)) {
+ complete(&dsi->xfer_completion);
+ ret = IRQ_HANDLED;
+ } else if (stat & DSI1_INT_HSTX_TO) {
+ complete(&dsi->xfer_completion);
+ dsi->xfer_result = -ETIMEDOUT;
+ ret = IRQ_HANDLED;
+ }
+
+ return ret;
+}
+
+/**
+ * Exposes clocks generated by the analog PHY that are consumed by
+ * CPRMAN (clk-bcm2835.c).
+ */
+static int
+vc4_dsi_init_phy_clocks(struct vc4_dsi *dsi)
+{
+ struct device *dev = &dsi->pdev->dev;
+ const char *parent_name = __clk_get_name(dsi->pll_phy_clock);
+ static const struct {
+ const char *dsi0_name, *dsi1_name;
+ int div;
+ } phy_clocks[] = {
+ { "dsi0_byte", "dsi1_byte", 8 },
+ { "dsi0_ddr2", "dsi1_ddr2", 4 },
+ { "dsi0_ddr", "dsi1_ddr", 2 },
+ };
+ int i;
+
+ dsi->clk_onecell = devm_kzalloc(dev,
+ sizeof(*dsi->clk_onecell) +
+ ARRAY_SIZE(phy_clocks) *
+ sizeof(struct clk_hw *),
+ GFP_KERNEL);
+ if (!dsi->clk_onecell)
+ return -ENOMEM;
+ dsi->clk_onecell->num = ARRAY_SIZE(phy_clocks);
+
+ for (i = 0; i < ARRAY_SIZE(phy_clocks); i++) {
+ struct clk_fixed_factor *fix = &dsi->phy_clocks[i];
+ struct clk_init_data init;
+ int ret;
+
+ /* We just use core fixed factor clock ops for the PHY
+ * clocks. The clocks are actually gated by the
+ * PHY_AFEC0_DDRCLK_EN bits, which we should be
+ * setting if we use the DDR/DDR2 clocks. However,
+ * vc4_dsi_encoder_enable() is setting up both AFEC0,
+ * setting both our parent DSI PLL's rate and this
+ * clock's rate, so it knows if DDR/DDR2 are going to
+ * be used and could enable the gates itself.
+ */
+ fix->mult = 1;
+ fix->div = phy_clocks[i].div;
+ fix->hw.init = &init;
+
+ memset(&init, 0, sizeof(init));
+ init.parent_names = &parent_name;
+ init.num_parents = 1;
+ if (dsi->port == 1)
+ init.name = phy_clocks[i].dsi1_name;
+ else
+ init.name = phy_clocks[i].dsi0_name;
+ init.ops = &clk_fixed_factor_ops;
+
+ ret = devm_clk_hw_register(dev, &fix->hw);
+ if (ret)
+ return ret;
+
+ dsi->clk_onecell->hws[i] = &fix->hw;
+ }
+
+ return of_clk_add_hw_provider(dev->of_node,
+ of_clk_hw_onecell_get,
+ dsi->clk_onecell);
+}
+
+static int vc4_dsi_bind(struct device *dev, struct device *master, void *data)
+{
+ struct platform_device *pdev = to_platform_device(dev);
+ struct drm_device *drm = dev_get_drvdata(master);
+ struct vc4_dev *vc4 = to_vc4_dev(drm);
+ struct vc4_dsi *dsi;
+ struct vc4_dsi_encoder *vc4_dsi_encoder;
+ const struct of_device_id *match;
+ dma_cap_mask_t dma_mask;
+ int ret;
+
+ dsi = devm_kzalloc(dev, sizeof(*dsi), GFP_KERNEL);
+ if (!dsi)
+ return -ENOMEM;
+
+ match = of_match_device(vc4_dsi_dt_match, dev);
+ if (!match)
+ return -ENODEV;
+
+ dsi->port = (uintptr_t)match->data;
+
+ vc4_dsi_encoder = devm_kzalloc(dev, sizeof(*vc4_dsi_encoder),
+ GFP_KERNEL);
+ if (!vc4_dsi_encoder)
+ return -ENOMEM;
+ vc4_dsi_encoder->base.type = VC4_ENCODER_TYPE_DSI1;
+ vc4_dsi_encoder->dsi = dsi;
+ dsi->encoder = &vc4_dsi_encoder->base.base;
+
+ dsi->pdev = pdev;
+ dsi->regs = vc4_ioremap_regs(pdev, 0);
+ if (IS_ERR(dsi->regs))
+ return PTR_ERR(dsi->regs);
+
+ if (DSI_PORT_READ(ID) != DSI_ID_VALUE) {
+ dev_err(dev, "Port returned 0x%08x for ID instead of 0x%08x\n",
+ DSI_PORT_READ(ID), DSI_ID_VALUE);
+ return -ENODEV;
+ }
+
+ /* DSI1 has a broken AXI slave that doesn't respond to writes
+ * from the ARM. It does handle writes from the DMA engine,
+ * so set up a channel for talking to it.
+ */
+ if (dsi->port == 1) {
+ dsi->reg_dma_mem = dma_alloc_coherent(dev, 4,
+ &dsi->reg_dma_paddr,
+ GFP_KERNEL);
+ if (!dsi->reg_dma_mem) {
+ DRM_ERROR("Failed to get DMA memory\n");
+ return -ENOMEM;
+ }
+
+ dma_cap_zero(dma_mask);
+ dma_cap_set(DMA_MEMCPY, dma_mask);
+ dsi->reg_dma_chan = dma_request_chan_by_mask(&dma_mask);
+ if (IS_ERR(dsi->reg_dma_chan)) {
+ ret = PTR_ERR(dsi->reg_dma_chan);
+ if (ret != -EPROBE_DEFER)
+ DRM_ERROR("Failed to get DMA channel: %d\n",
+ ret);
+ return ret;
+ }
+
+ /* Get the physical address of the device's registers. The
+ * struct resource for the regs gives us the bus address
+ * instead.
+ */
+ dsi->reg_paddr = be32_to_cpup(of_get_address(dev->of_node,
+ 0, NULL, NULL));
+ }
+
+ init_completion(&dsi->xfer_completion);
+ /* At startup enable error-reporting interrupts and nothing else. */
+ DSI_PORT_WRITE(INT_EN, DSI1_INTERRUPTS_ALWAYS_ENABLED);
+ /* Clear any existing interrupt state. */
+ DSI_PORT_WRITE(INT_STAT, DSI_PORT_READ(INT_STAT));
+
+ ret = devm_request_irq(dev, platform_get_irq(pdev, 0),
+ vc4_dsi_irq_handler, 0, "vc4 dsi", dsi);
+ if (ret) {
+ if (ret != -EPROBE_DEFER)
+ dev_err(dev, "Failed to get interrupt: %d\n", ret);
+ return ret;
+ }
+
+ dsi->escape_clock = devm_clk_get(dev, "escape");
+ if (IS_ERR(dsi->escape_clock)) {
+ ret = PTR_ERR(dsi->escape_clock);
+ if (ret != -EPROBE_DEFER)
+ dev_err(dev, "Failed to get escape clock: %d\n", ret);
+ return ret;
+ }
+
+ dsi->pll_phy_clock = devm_clk_get(dev, "phy");
+ if (IS_ERR(dsi->pll_phy_clock)) {
+ ret = PTR_ERR(dsi->pll_phy_clock);
+ if (ret != -EPROBE_DEFER)
+ dev_err(dev, "Failed to get phy clock: %d\n", ret);
+ return ret;
+ }
+
+ dsi->pixel_clock = devm_clk_get(dev, "pixel");
+ if (IS_ERR(dsi->pixel_clock)) {
+ ret = PTR_ERR(dsi->pixel_clock);
+ if (ret != -EPROBE_DEFER)
+ dev_err(dev, "Failed to get pixel clock: %d\n", ret);
+ return ret;
+ }
+
+ /* The esc clock rate is supposed to always be 100Mhz. */
+ ret = clk_set_rate(dsi->escape_clock, 100 * 1000000);
+ if (ret) {
+ dev_err(dev, "Failed to set esc clock: %d\n", ret);
+ return ret;
+ }
+
+ ret = vc4_dsi_init_phy_clocks(dsi);
+ if (ret)
+ return ret;
+
+ if (dsi->port == 1)
+ vc4->dsi1 = dsi;
+
+ drm_encoder_init(drm, dsi->encoder, &vc4_dsi_encoder_funcs,
+ DRM_MODE_ENCODER_DSI, NULL);
+ drm_encoder_helper_add(dsi->encoder, &vc4_dsi_encoder_helper_funcs);
+
+ dsi->connector = vc4_dsi_connector_init(drm, dsi);
+ if (IS_ERR(dsi->connector)) {
+ ret = PTR_ERR(dsi->connector);
+ goto err_destroy_encoder;
+ }
+
+ dsi->dsi_host.ops = &vc4_dsi_host_ops;
+ dsi->dsi_host.dev = dev;
+
+ mipi_dsi_host_register(&dsi->dsi_host);
+
+ dev_set_drvdata(dev, dsi);
+
+ pm_runtime_enable(dev);
+
+ return 0;
+
+err_destroy_encoder:
+ vc4_dsi_encoder_destroy(dsi->encoder);
+
+ return ret;
+}
+
+static void vc4_dsi_unbind(struct device *dev, struct device *master,
+ void *data)
+{
+ struct drm_device *drm = dev_get_drvdata(master);
+ struct vc4_dev *vc4 = to_vc4_dev(drm);
+ struct vc4_dsi *dsi = dev_get_drvdata(dev);
+
+ pm_runtime_disable(dev);
+
+ vc4_dsi_connector_destroy(dsi->connector);
+ vc4_dsi_encoder_destroy(dsi->encoder);
+
+ mipi_dsi_host_unregister(&dsi->dsi_host);
+
+ clk_disable_unprepare(dsi->pll_phy_clock);
+ clk_disable_unprepare(dsi->escape_clock);
+
+ if (dsi->port == 1)
+ vc4->dsi1 = NULL;
+}
+
+static const struct component_ops vc4_dsi_ops = {
+ .bind = vc4_dsi_bind,
+ .unbind = vc4_dsi_unbind,
+};
+
+static int vc4_dsi_dev_probe(struct platform_device *pdev)
+{
+ return component_add(&pdev->dev, &vc4_dsi_ops);
+}
+
+static int vc4_dsi_dev_remove(struct platform_device *pdev)
+{
+ component_del(&pdev->dev, &vc4_dsi_ops);
+ return 0;
+}
+
+struct platform_driver vc4_dsi_driver = {
+ .probe = vc4_dsi_dev_probe,
+ .remove = vc4_dsi_dev_remove,
+ .driver = {
+ .name = "vc4_dsi",
+ .of_match_table = vc4_dsi_dt_match,
+ },
+};
diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c
index 47a095f392f8..1eef98c3331d 100644
--- a/drivers/gpu/drm/vc4/vc4_gem.c
+++ b/drivers/gpu/drm/vc4/vc4_gem.c
@@ -26,6 +26,7 @@
#include <linux/pm_runtime.h>
#include <linux/device.h>
#include <linux/io.h>
+#include <linux/sched/signal.h>
#include "uapi/drm/vc4_drm.h"
#include "vc4_drv.h"
@@ -544,14 +545,15 @@ vc4_cl_lookup_bos(struct drm_device *dev,
handles = drm_malloc_ab(exec->bo_count, sizeof(uint32_t));
if (!handles) {
+ ret = -ENOMEM;
DRM_ERROR("Failed to allocate incoming GEM handles\n");
goto fail;
}
- ret = copy_from_user(handles,
- (void __user *)(uintptr_t)args->bo_handles,
- exec->bo_count * sizeof(uint32_t));
- if (ret) {
+ if (copy_from_user(handles,
+ (void __user *)(uintptr_t)args->bo_handles,
+ exec->bo_count * sizeof(uint32_t))) {
+ ret = -EFAULT;
DRM_ERROR("Failed to copy in GEM handles\n");
goto fail;
}
@@ -593,12 +595,14 @@ vc4_get_bcl(struct drm_device *dev, struct vc4_exec_info *exec)
args->shader_rec_count);
struct vc4_bo *bo;
- if (uniforms_offset < shader_rec_offset ||
+ if (shader_rec_offset < args->bin_cl_size ||
+ uniforms_offset < shader_rec_offset ||
exec_size < uniforms_offset ||
args->shader_rec_count >= (UINT_MAX /
sizeof(struct vc4_shader_state)) ||
temp_size < exec_size) {
DRM_ERROR("overflow in exec arguments\n");
+ ret = -EINVAL;
goto fail;
}
@@ -708,8 +712,10 @@ vc4_complete_exec(struct drm_device *dev, struct vc4_exec_info *exec)
}
mutex_lock(&vc4->power_lock);
- if (--vc4->power_refcount == 0)
- pm_runtime_put(&vc4->v3d->pdev->dev);
+ if (--vc4->power_refcount == 0) {
+ pm_runtime_mark_last_busy(&vc4->v3d->pdev->dev);
+ pm_runtime_put_autosuspend(&vc4->v3d->pdev->dev);
+ }
mutex_unlock(&vc4->power_lock);
kfree(exec);
diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c
index c4cb2e26de32..93d5994f3a04 100644
--- a/drivers/gpu/drm/vc4/vc4_hdmi.c
+++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
@@ -356,15 +356,11 @@ static void vc4_hdmi_set_avi_infoframe(struct drm_encoder *encoder)
return;
}
- if (vc4_encoder->rgb_range_selectable) {
- if (vc4_encoder->limited_rgb_range) {
- frame.avi.quantization_range =
- HDMI_QUANTIZATION_RANGE_LIMITED;
- } else {
- frame.avi.quantization_range =
- HDMI_QUANTIZATION_RANGE_FULL;
- }
- }
+ drm_hdmi_avi_infoframe_quant_range(&frame.avi, mode,
+ vc4_encoder->limited_rgb_range ?
+ HDMI_QUANTIZATION_RANGE_LIMITED :
+ HDMI_QUANTIZATION_RANGE_FULL,
+ vc4_encoder->rgb_range_selectable);
vc4_hdmi_write_infoframe(encoder, &frame);
}
@@ -463,7 +459,9 @@ static void vc4_hdmi_encoder_mode_set(struct drm_encoder *encoder,
csc_ctl = VC4_SET_FIELD(VC4_HD_CSC_CTL_ORDER_BGR,
VC4_HD_CSC_CTL_ORDER);
- if (vc4_encoder->hdmi_monitor && drm_match_cea_mode(mode) > 1) {
+ if (vc4_encoder->hdmi_monitor &&
+ drm_default_rgb_quant_range(mode) ==
+ HDMI_QUANTIZATION_RANGE_LIMITED) {
/* CEA VICs other than #1 requre limited range RGB
* output unless overridden by an AVI infoframe.
* Apply a colorspace conversion to squash 0-255 down
diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c
index 6fbab1c82cb1..f7f7677f6d8d 100644
--- a/drivers/gpu/drm/vc4/vc4_hvs.c
+++ b/drivers/gpu/drm/vc4/vc4_hvs.c
@@ -141,8 +141,7 @@ static int vc4_hvs_upload_linear_kernel(struct vc4_hvs *hvs,
int ret, i;
u32 __iomem *dst_kernel;
- ret = drm_mm_insert_node(&hvs->dlist_mm, space, VC4_KERNEL_DWORDS, 1,
- 0);
+ ret = drm_mm_insert_node(&hvs->dlist_mm, space, VC4_KERNEL_DWORDS);
if (ret) {
DRM_ERROR("Failed to allocate space for filter kernel: %d\n",
ret);
@@ -170,6 +169,7 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
struct vc4_dev *vc4 = drm->dev_private;
struct vc4_hvs *hvs = NULL;
int ret;
+ u32 dispctrl;
hvs = devm_kzalloc(&pdev->dev, sizeof(*hvs), GFP_KERNEL);
if (!hvs)
@@ -211,6 +211,19 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data)
return ret;
vc4->hvs = hvs;
+
+ dispctrl = HVS_READ(SCALER_DISPCTRL);
+
+ dispctrl |= SCALER_DISPCTRL_ENABLE;
+
+ /* Set DSP3 (PV1) to use HVS channel 2, which would otherwise
+ * be unused.
+ */
+ dispctrl &= ~SCALER_DISPCTRL_DSP3_MUX_MASK;
+ dispctrl |= VC4_SET_FIELD(2, SCALER_DISPCTRL_DSP3_MUX);
+
+ HVS_WRITE(SCALER_DISPCTRL, dispctrl);
+
return 0;
}
diff --git a/drivers/gpu/drm/vc4/vc4_kms.c b/drivers/gpu/drm/vc4/vc4_kms.c
index c1f65c6c8e60..ad7925a9e0ea 100644
--- a/drivers/gpu/drm/vc4/vc4_kms.c
+++ b/drivers/gpu/drm/vc4/vc4_kms.c
@@ -61,7 +61,7 @@ vc4_atomic_complete_commit(struct vc4_commit *c)
drm_atomic_helper_cleanup_planes(dev, state);
- drm_atomic_state_free(state);
+ drm_atomic_state_put(state);
up(&vc4->async_modeset);
@@ -119,17 +119,34 @@ static int vc4_atomic_commit(struct drm_device *dev,
/* Make sure that any outstanding modesets have finished. */
if (nonblock) {
- ret = down_trylock(&vc4->async_modeset);
- if (ret) {
+ struct drm_crtc *crtc;
+ struct drm_crtc_state *crtc_state;
+ unsigned long flags;
+ bool busy = false;
+
+ /*
+ * If there's an undispatched event to send then we're
+ * obviously still busy. If there isn't, then we can
+ * unconditionally wait for the semaphore because it
+ * shouldn't be contended (for long).
+ *
+ * This is to prevent a race where queuing a new flip
+ * from userspace immediately on receipt of an event
+ * beats our clean-up and returns EBUSY.
+ */
+ spin_lock_irqsave(&dev->event_lock, flags);
+ for_each_crtc_in_state(state, crtc, crtc_state, i)
+ busy |= vc4_event_pending(crtc);
+ spin_unlock_irqrestore(&dev->event_lock, flags);
+ if (busy) {
kfree(c);
return -EBUSY;
}
- } else {
- ret = down_interruptible(&vc4->async_modeset);
- if (ret) {
- kfree(c);
- return ret;
- }
+ }
+ ret = down_interruptible(&vc4->async_modeset);
+ if (ret) {
+ kfree(c);
+ return ret;
}
ret = drm_atomic_helper_prepare_planes(dev, state);
@@ -173,6 +190,7 @@ static int vc4_atomic_commit(struct drm_device *dev,
* current layout.
*/
+ drm_atomic_state_get(state);
if (nonblock) {
vc4_queue_seqno_cb(dev, &c->cb, wait_seqno,
vc4_atomic_complete_commit_seqno_cb);
@@ -213,7 +231,6 @@ int vc4_kms_load(struct drm_device *dev)
drm_mode_config_reset(dev);
vc4->fbdev = drm_fbdev_cma_init(dev, 32,
- dev->mode_config.num_crtc,
dev->mode_config.num_connector);
if (IS_ERR(vc4->fbdev))
vc4->fbdev = NULL;
diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c
index 881bf489478b..f7a229df572d 100644
--- a/drivers/gpu/drm/vc4/vc4_plane.c
+++ b/drivers/gpu/drm/vc4/vc4_plane.c
@@ -295,8 +295,8 @@ static int vc4_plane_setup_clipping_and_scaling(struct drm_plane_state *state)
struct drm_framebuffer *fb = state->fb;
struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0);
u32 subpixel_src_mask = (1 << 16) - 1;
- u32 format = fb->pixel_format;
- int num_planes = drm_format_num_planes(format);
+ u32 format = fb->format->format;
+ int num_planes = fb->format->num_planes;
u32 h_subsample = 1;
u32 v_subsample = 1;
int i;
@@ -369,7 +369,7 @@ static int vc4_plane_setup_clipping_and_scaling(struct drm_plane_state *state)
*/
if (vc4_state->crtc_x < 0) {
for (i = 0; i < num_planes; i++) {
- u32 cpp = drm_format_plane_cpp(fb->pixel_format, i);
+ u32 cpp = fb->format->cpp[i];
u32 subs = ((i == 0) ? 1 : h_subsample);
vc4_state->offsets[i] += (cpp *
@@ -496,7 +496,7 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
struct vc4_plane_state *vc4_state = to_vc4_plane_state(state);
struct drm_framebuffer *fb = state->fb;
u32 ctl0_offset = vc4_state->dlist_count;
- const struct hvs_format *format = vc4_get_hvs_format(fb->pixel_format);
+ const struct hvs_format *format = vc4_get_hvs_format(fb->format->format);
int num_planes = drm_format_num_planes(format->drm);
u32 scl0, scl1;
u32 lbm_size;
@@ -514,9 +514,9 @@ static int vc4_plane_mode_set(struct drm_plane *plane,
if (lbm_size) {
if (!vc4_state->lbm.allocated) {
spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags);
- ret = drm_mm_insert_node(&vc4->hvs->lbm_mm,
- &vc4_state->lbm,
- lbm_size, 32, 0);
+ ret = drm_mm_insert_node_generic(&vc4->hvs->lbm_mm,
+ &vc4_state->lbm,
+ lbm_size, 32, 0, 0);
spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags);
} else {
WARN_ON_ONCE(lbm_size != vc4_state->lbm.size);
@@ -858,7 +858,7 @@ struct drm_plane *vc4_plane_init(struct drm_device *dev,
}
}
plane = &vc4_plane->base;
- ret = drm_universal_plane_init(dev, plane, 0xff,
+ ret = drm_universal_plane_init(dev, plane, 0,
&vc4_plane_funcs,
formats, num_formats,
type, NULL);
diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h
index 1aa44c2db556..385405a2df05 100644
--- a/drivers/gpu/drm/vc4/vc4_regs.h
+++ b/drivers/gpu/drm/vc4/vc4_regs.h
@@ -177,8 +177,9 @@
# define PV_CONTROL_WAIT_HSTART BIT(12)
# define PV_CONTROL_PIXEL_REP_MASK VC4_MASK(5, 4)
# define PV_CONTROL_PIXEL_REP_SHIFT 4
-# define PV_CONTROL_CLK_SELECT_DSI_VEC 0
+# define PV_CONTROL_CLK_SELECT_DSI 0
# define PV_CONTROL_CLK_SELECT_DPI_SMI_HDMI 1
+# define PV_CONTROL_CLK_SELECT_VEC 2
# define PV_CONTROL_CLK_SELECT_MASK VC4_MASK(3, 2)
# define PV_CONTROL_CLK_SELECT_SHIFT 2
# define PV_CONTROL_FIFO_CLR BIT(1)
@@ -189,6 +190,8 @@
# define PV_VCONTROL_ODD_DELAY_SHIFT 6
# define PV_VCONTROL_ODD_FIRST BIT(5)
# define PV_VCONTROL_INTERLACE BIT(4)
+# define PV_VCONTROL_DSI BIT(3)
+# define PV_VCONTROL_COMMAND BIT(2)
# define PV_VCONTROL_CONTINUOUS BIT(1)
# define PV_VCONTROL_VIDEN BIT(0)
@@ -243,6 +246,9 @@
# define SCALER_DISPCTRL_ENABLE BIT(31)
# define SCALER_DISPCTRL_DSP2EISLUR BIT(15)
# define SCALER_DISPCTRL_DSP1EISLUR BIT(14)
+# define SCALER_DISPCTRL_DSP3_MUX_MASK VC4_MASK(19, 18)
+# define SCALER_DISPCTRL_DSP3_MUX_SHIFT 18
+
/* Enables Display 0 short line and underrun contribution to
* SCALER_DISPSTAT_IRQDISP0. Note that short frame contributions are
* always enabled.
diff --git a/drivers/gpu/drm/vc4/vc4_render_cl.c b/drivers/gpu/drm/vc4/vc4_render_cl.c
index 08886a309757..5cdd003605f5 100644
--- a/drivers/gpu/drm/vc4/vc4_render_cl.c
+++ b/drivers/gpu/drm/vc4/vc4_render_cl.c
@@ -461,7 +461,7 @@ static int vc4_rcl_surface_setup(struct vc4_exec_info *exec,
}
ret = vc4_full_res_bounds_check(exec, *obj, surf);
- if (!ret)
+ if (ret)
return ret;
return 0;
diff --git a/drivers/gpu/drm/vc4/vc4_v3d.c b/drivers/gpu/drm/vc4/vc4_v3d.c
index e6d3c6028341..7cc346ad9b0b 100644
--- a/drivers/gpu/drm/vc4/vc4_v3d.c
+++ b/drivers/gpu/drm/vc4/vc4_v3d.c
@@ -222,6 +222,8 @@ static int vc4_v3d_bind(struct device *dev, struct device *master, void *data)
return ret;
}
+ pm_runtime_use_autosuspend(dev);
+ pm_runtime_set_autosuspend_delay(dev, 40); /* a little over 2 frames. */
pm_runtime_enable(dev);
return 0;
diff --git a/drivers/gpu/drm/vc4/vc4_validate.c b/drivers/gpu/drm/vc4/vc4_validate.c
index 26503e307438..9fd171c361c2 100644
--- a/drivers/gpu/drm/vc4/vc4_validate.c
+++ b/drivers/gpu/drm/vc4/vc4_validate.c
@@ -644,6 +644,13 @@ reloc_tex(struct vc4_exec_info *exec,
cpp = 1;
break;
case VC4_TEXTURE_TYPE_ETC1:
+ /* ETC1 is arranged as 64-bit blocks, where each block is 4x4
+ * pixels.
+ */
+ cpp = 8;
+ width = (width + 3) >> 2;
+ height = (height + 3) >> 2;
+ break;
case VC4_TEXTURE_TYPE_BW1:
case VC4_TEXTURE_TYPE_A4:
case VC4_TEXTURE_TYPE_A1:
@@ -782,11 +789,6 @@ validate_gl_shader_rec(struct drm_device *dev,
exec->shader_rec_v += roundup(packet_size, 16);
exec->shader_rec_size -= packet_size;
- if (!(*(uint16_t *)pkt_u & VC4_SHADER_FLAG_FS_SINGLE_THREAD)) {
- DRM_ERROR("Multi-threaded fragment shaders not supported.\n");
- return -EINVAL;
- }
-
for (i = 0; i < shader_reloc_count; i++) {
if (src_handles[i] > exec->bo_count) {
DRM_ERROR("Shader handle %d too big\n", src_handles[i]);
@@ -803,6 +805,18 @@ validate_gl_shader_rec(struct drm_device *dev,
return -EINVAL;
}
+ if (((*(uint16_t *)pkt_u & VC4_SHADER_FLAG_FS_SINGLE_THREAD) == 0) !=
+ to_vc4_bo(&bo[0]->base)->validated_shader->is_threaded) {
+ DRM_ERROR("Thread mode of CL and FS do not match\n");
+ return -EINVAL;
+ }
+
+ if (to_vc4_bo(&bo[1]->base)->validated_shader->is_threaded ||
+ to_vc4_bo(&bo[2]->base)->validated_shader->is_threaded) {
+ DRM_ERROR("cs and vs cannot be threaded\n");
+ return -EINVAL;
+ }
+
for (i = 0; i < shader_reloc_count; i++) {
struct vc4_validated_shader_info *validated_shader;
uint32_t o = shader_reloc_offsets[i];
diff --git a/drivers/gpu/drm/vc4/vc4_validate_shaders.c b/drivers/gpu/drm/vc4/vc4_validate_shaders.c
index 2543cf5b8b51..5dba13dd1e9b 100644
--- a/drivers/gpu/drm/vc4/vc4_validate_shaders.c
+++ b/drivers/gpu/drm/vc4/vc4_validate_shaders.c
@@ -83,6 +83,13 @@ struct vc4_shader_validation_state {
* basic blocks.
*/
bool needs_uniform_address_for_loop;
+
+ /* Set when we find an instruction writing the top half of the
+ * register files. If we allowed writing the unusable regs in
+ * a threaded shader, then the other shader running on our
+ * QPU's clamp validation would be invalid.
+ */
+ bool all_registers_used;
};
static uint32_t
@@ -119,6 +126,13 @@ raddr_add_a_to_live_reg_index(uint64_t inst)
}
static bool
+live_reg_is_upper_half(uint32_t lri)
+{
+ return (lri >= 16 && lri < 32) ||
+ (lri >= 32 + 16 && lri < 32 + 32);
+}
+
+static bool
is_tmu_submit(uint32_t waddr)
{
return (waddr == QPU_W_TMU0_S ||
@@ -390,6 +404,9 @@ check_reg_write(struct vc4_validated_shader_info *validated_shader,
} else {
validation_state->live_immediates[lri] = ~0;
}
+
+ if (live_reg_is_upper_half(lri))
+ validation_state->all_registers_used = true;
}
switch (waddr) {
@@ -598,6 +615,11 @@ check_instruction_reads(struct vc4_validated_shader_info *validated_shader,
}
}
+ if ((raddr_a >= 16 && raddr_a < 32) ||
+ (raddr_b >= 16 && raddr_b < 32 && sig != QPU_SIG_SMALL_IMM)) {
+ validation_state->all_registers_used = true;
+ }
+
return true;
}
@@ -608,9 +630,7 @@ static bool
vc4_validate_branches(struct vc4_shader_validation_state *validation_state)
{
uint32_t max_branch_target = 0;
- bool found_shader_end = false;
int ip;
- int shader_end_ip = 0;
int last_branch = -2;
for (ip = 0; ip < validation_state->max_ip; ip++) {
@@ -621,8 +641,13 @@ vc4_validate_branches(struct vc4_shader_validation_state *validation_state)
uint32_t branch_target_ip;
if (sig == QPU_SIG_PROG_END) {
- shader_end_ip = ip;
- found_shader_end = true;
+ /* There are two delay slots after program end is
+ * signaled that are still executed, then we're
+ * finished. validation_state->max_ip is the
+ * instruction after the last valid instruction in the
+ * program.
+ */
+ validation_state->max_ip = ip + 3;
continue;
}
@@ -676,15 +701,9 @@ vc4_validate_branches(struct vc4_shader_validation_state *validation_state)
}
set_bit(after_delay_ip, validation_state->branch_targets);
max_branch_target = max(max_branch_target, after_delay_ip);
-
- /* There are two delay slots after program end is signaled
- * that are still executed, then we're finished.
- */
- if (found_shader_end && ip == shader_end_ip + 2)
- break;
}
- if (max_branch_target > shader_end_ip) {
+ if (max_branch_target > validation_state->max_ip - 3) {
DRM_ERROR("Branch landed after QPU_SIG_PROG_END");
return false;
}
@@ -756,6 +775,7 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
{
bool found_shader_end = false;
int shader_end_ip = 0;
+ uint32_t last_thread_switch_ip = -3;
uint32_t ip;
struct vc4_validated_shader_info *validated_shader = NULL;
struct vc4_shader_validation_state validation_state;
@@ -788,6 +808,17 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
if (!vc4_handle_branch_target(&validation_state))
goto fail;
+ if (ip == last_thread_switch_ip + 3) {
+ /* Reset r0-r3 live clamp data */
+ int i;
+
+ for (i = 64; i < LIVE_REG_COUNT; i++) {
+ validation_state.live_min_clamp_offsets[i] = ~0;
+ validation_state.live_max_clamp_regs[i] = false;
+ validation_state.live_immediates[i] = ~0;
+ }
+ }
+
switch (sig) {
case QPU_SIG_NONE:
case QPU_SIG_WAIT_FOR_SCOREBOARD:
@@ -797,6 +828,8 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
case QPU_SIG_LOAD_TMU1:
case QPU_SIG_PROG_END:
case QPU_SIG_SMALL_IMM:
+ case QPU_SIG_THREAD_SWITCH:
+ case QPU_SIG_LAST_THREAD_SWITCH:
if (!check_instruction_writes(validated_shader,
&validation_state)) {
DRM_ERROR("Bad write at ip %d\n", ip);
@@ -812,6 +845,18 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
shader_end_ip = ip;
}
+ if (sig == QPU_SIG_THREAD_SWITCH ||
+ sig == QPU_SIG_LAST_THREAD_SWITCH) {
+ validated_shader->is_threaded = true;
+
+ if (ip < last_thread_switch_ip + 3) {
+ DRM_ERROR("Thread switch too soon after "
+ "last switch at ip %d\n", ip);
+ goto fail;
+ }
+ last_thread_switch_ip = ip;
+ }
+
break;
case QPU_SIG_LOAD_IMM:
@@ -826,6 +871,13 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
if (!check_branch(inst, validated_shader,
&validation_state, ip))
goto fail;
+
+ if (ip < last_thread_switch_ip + 3) {
+ DRM_ERROR("Branch in thread switch at ip %d",
+ ip);
+ goto fail;
+ }
+
break;
default:
DRM_ERROR("Unsupported QPU signal %d at "
@@ -847,6 +899,14 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj)
goto fail;
}
+ /* Might corrupt other thread */
+ if (validated_shader->is_threaded &&
+ validation_state.all_registers_used) {
+ DRM_ERROR("Shader uses threading, but uses the upper "
+ "half of the registers, too\n");
+ goto fail;
+ }
+
/* If we did a backwards branch and we haven't emitted a uniforms
* reset since then, we still need the uniforms stream to have the
* uniforms address available so that the backwards branch can do its
diff --git a/drivers/gpu/drm/vc4/vc4_vec.c b/drivers/gpu/drm/vc4/vc4_vec.c
new file mode 100644
index 000000000000..32bb8ef985fb
--- /dev/null
+++ b/drivers/gpu/drm/vc4/vc4_vec.c
@@ -0,0 +1,657 @@
+/*
+ * Copyright (C) 2016 Broadcom
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+/**
+ * DOC: VC4 SDTV module
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_panel.h>
+#include <linux/clk.h>
+#include <linux/component.h>
+#include <linux/of_graph.h>
+#include <linux/of_platform.h>
+#include <linux/pm_runtime.h>
+
+#include "vc4_drv.h"
+#include "vc4_regs.h"
+
+/* WSE Registers */
+#define VEC_WSE_RESET 0xc0
+
+#define VEC_WSE_CONTROL 0xc4
+#define VEC_WSE_WSS_ENABLE BIT(7)
+
+#define VEC_WSE_WSS_DATA 0xc8
+#define VEC_WSE_VPS_DATA1 0xcc
+#define VEC_WSE_VPS_CONTROL 0xd0
+
+/* VEC Registers */
+#define VEC_REVID 0x100
+
+#define VEC_CONFIG0 0x104
+#define VEC_CONFIG0_YDEL_MASK GENMASK(28, 26)
+#define VEC_CONFIG0_YDEL(x) ((x) << 26)
+#define VEC_CONFIG0_CDEL_MASK GENMASK(25, 24)
+#define VEC_CONFIG0_CDEL(x) ((x) << 24)
+#define VEC_CONFIG0_PBPR_FIL BIT(18)
+#define VEC_CONFIG0_CHROMA_GAIN_MASK GENMASK(17, 16)
+#define VEC_CONFIG0_CHROMA_GAIN_UNITY (0 << 16)
+#define VEC_CONFIG0_CHROMA_GAIN_1_32 (1 << 16)
+#define VEC_CONFIG0_CHROMA_GAIN_1_16 (2 << 16)
+#define VEC_CONFIG0_CHROMA_GAIN_1_8 (3 << 16)
+#define VEC_CONFIG0_CBURST_GAIN_MASK GENMASK(14, 13)
+#define VEC_CONFIG0_CBURST_GAIN_UNITY (0 << 13)
+#define VEC_CONFIG0_CBURST_GAIN_1_128 (1 << 13)
+#define VEC_CONFIG0_CBURST_GAIN_1_64 (2 << 13)
+#define VEC_CONFIG0_CBURST_GAIN_1_32 (3 << 13)
+#define VEC_CONFIG0_CHRBW1 BIT(11)
+#define VEC_CONFIG0_CHRBW0 BIT(10)
+#define VEC_CONFIG0_SYNCDIS BIT(9)
+#define VEC_CONFIG0_BURDIS BIT(8)
+#define VEC_CONFIG0_CHRDIS BIT(7)
+#define VEC_CONFIG0_PDEN BIT(6)
+#define VEC_CONFIG0_YCDELAY BIT(4)
+#define VEC_CONFIG0_RAMPEN BIT(2)
+#define VEC_CONFIG0_YCDIS BIT(2)
+#define VEC_CONFIG0_STD_MASK GENMASK(1, 0)
+#define VEC_CONFIG0_NTSC_STD 0
+#define VEC_CONFIG0_PAL_BDGHI_STD 1
+#define VEC_CONFIG0_PAL_N_STD 3
+
+#define VEC_SCHPH 0x108
+#define VEC_SOFT_RESET 0x10c
+#define VEC_CLMP0_START 0x144
+#define VEC_CLMP0_END 0x148
+#define VEC_FREQ3_2 0x180
+#define VEC_FREQ1_0 0x184
+
+#define VEC_CONFIG1 0x188
+#define VEC_CONFIG_VEC_RESYNC_OFF BIT(18)
+#define VEC_CONFIG_RGB219 BIT(17)
+#define VEC_CONFIG_CBAR_EN BIT(16)
+#define VEC_CONFIG_TC_OBB BIT(15)
+#define VEC_CONFIG1_OUTPUT_MODE_MASK GENMASK(12, 10)
+#define VEC_CONFIG1_C_Y_CVBS (0 << 10)
+#define VEC_CONFIG1_CVBS_Y_C (1 << 10)
+#define VEC_CONFIG1_PR_Y_PB (2 << 10)
+#define VEC_CONFIG1_RGB (4 << 10)
+#define VEC_CONFIG1_Y_C_CVBS (5 << 10)
+#define VEC_CONFIG1_C_CVBS_Y (6 << 10)
+#define VEC_CONFIG1_C_CVBS_CVBS (7 << 10)
+#define VEC_CONFIG1_DIS_CHR BIT(9)
+#define VEC_CONFIG1_DIS_LUMA BIT(8)
+#define VEC_CONFIG1_YCBCR_IN BIT(6)
+#define VEC_CONFIG1_DITHER_TYPE_LFSR 0
+#define VEC_CONFIG1_DITHER_TYPE_COUNTER BIT(5)
+#define VEC_CONFIG1_DITHER_EN BIT(4)
+#define VEC_CONFIG1_CYDELAY BIT(3)
+#define VEC_CONFIG1_LUMADIS BIT(2)
+#define VEC_CONFIG1_COMPDIS BIT(1)
+#define VEC_CONFIG1_CUSTOM_FREQ BIT(0)
+
+#define VEC_CONFIG2 0x18c
+#define VEC_CONFIG2_PROG_SCAN BIT(15)
+#define VEC_CONFIG2_SYNC_ADJ_MASK GENMASK(14, 12)
+#define VEC_CONFIG2_SYNC_ADJ(x) (((x) / 2) << 12)
+#define VEC_CONFIG2_PBPR_EN BIT(10)
+#define VEC_CONFIG2_UV_DIG_DIS BIT(6)
+#define VEC_CONFIG2_RGB_DIG_DIS BIT(5)
+#define VEC_CONFIG2_TMUX_MASK GENMASK(3, 2)
+#define VEC_CONFIG2_TMUX_DRIVE0 (0 << 2)
+#define VEC_CONFIG2_TMUX_RG_COMP (1 << 2)
+#define VEC_CONFIG2_TMUX_UV_YC (2 << 2)
+#define VEC_CONFIG2_TMUX_SYNC_YC (3 << 2)
+
+#define VEC_INTERRUPT_CONTROL 0x190
+#define VEC_INTERRUPT_STATUS 0x194
+#define VEC_FCW_SECAM_B 0x198
+#define VEC_SECAM_GAIN_VAL 0x19c
+
+#define VEC_CONFIG3 0x1a0
+#define VEC_CONFIG3_HORIZ_LEN_STD (0 << 0)
+#define VEC_CONFIG3_HORIZ_LEN_MPEG1_SIF (1 << 0)
+#define VEC_CONFIG3_SHAPE_NON_LINEAR BIT(1)
+
+#define VEC_STATUS0 0x200
+#define VEC_MASK0 0x204
+
+#define VEC_CFG 0x208
+#define VEC_CFG_SG_MODE_MASK GENMASK(6, 5)
+#define VEC_CFG_SG_MODE(x) ((x) << 5)
+#define VEC_CFG_SG_EN BIT(4)
+#define VEC_CFG_VEC_EN BIT(3)
+#define VEC_CFG_MB_EN BIT(2)
+#define VEC_CFG_ENABLE BIT(1)
+#define VEC_CFG_TB_EN BIT(0)
+
+#define VEC_DAC_TEST 0x20c
+
+#define VEC_DAC_CONFIG 0x210
+#define VEC_DAC_CONFIG_LDO_BIAS_CTRL(x) ((x) << 24)
+#define VEC_DAC_CONFIG_DRIVER_CTRL(x) ((x) << 16)
+#define VEC_DAC_CONFIG_DAC_CTRL(x) (x)
+
+#define VEC_DAC_MISC 0x214
+#define VEC_DAC_MISC_VCD_CTRL_MASK GENMASK(31, 16)
+#define VEC_DAC_MISC_VCD_CTRL(x) ((x) << 16)
+#define VEC_DAC_MISC_VID_ACT BIT(8)
+#define VEC_DAC_MISC_VCD_PWRDN BIT(6)
+#define VEC_DAC_MISC_BIAS_PWRDN BIT(5)
+#define VEC_DAC_MISC_DAC_PWRDN BIT(2)
+#define VEC_DAC_MISC_LDO_PWRDN BIT(1)
+#define VEC_DAC_MISC_DAC_RST_N BIT(0)
+
+
+/* General VEC hardware state. */
+struct vc4_vec {
+ struct platform_device *pdev;
+
+ struct drm_encoder *encoder;
+ struct drm_connector *connector;
+
+ void __iomem *regs;
+
+ struct clk *clock;
+
+ const struct vc4_vec_tv_mode *tv_mode;
+};
+
+#define VEC_READ(offset) readl(vec->regs + (offset))
+#define VEC_WRITE(offset, val) writel(val, vec->regs + (offset))
+
+/* VC4 VEC encoder KMS struct */
+struct vc4_vec_encoder {
+ struct vc4_encoder base;
+ struct vc4_vec *vec;
+};
+
+static inline struct vc4_vec_encoder *
+to_vc4_vec_encoder(struct drm_encoder *encoder)
+{
+ return container_of(encoder, struct vc4_vec_encoder, base.base);
+}
+
+/* VC4 VEC connector KMS struct */
+struct vc4_vec_connector {
+ struct drm_connector base;
+ struct vc4_vec *vec;
+
+ /* Since the connector is attached to just the one encoder,
+ * this is the reference to it so we can do the best_encoder()
+ * hook.
+ */
+ struct drm_encoder *encoder;
+};
+
+static inline struct vc4_vec_connector *
+to_vc4_vec_connector(struct drm_connector *connector)
+{
+ return container_of(connector, struct vc4_vec_connector, base);
+}
+
+enum vc4_vec_tv_mode_id {
+ VC4_VEC_TV_MODE_NTSC,
+ VC4_VEC_TV_MODE_NTSC_J,
+ VC4_VEC_TV_MODE_PAL,
+ VC4_VEC_TV_MODE_PAL_M,
+};
+
+struct vc4_vec_tv_mode {
+ const struct drm_display_mode *mode;
+ void (*mode_set)(struct vc4_vec *vec);
+};
+
+#define VEC_REG(reg) { reg, #reg }
+static const struct {
+ u32 reg;
+ const char *name;
+} vec_regs[] = {
+ VEC_REG(VEC_WSE_CONTROL),
+ VEC_REG(VEC_WSE_WSS_DATA),
+ VEC_REG(VEC_WSE_VPS_DATA1),
+ VEC_REG(VEC_WSE_VPS_CONTROL),
+ VEC_REG(VEC_REVID),
+ VEC_REG(VEC_CONFIG0),
+ VEC_REG(VEC_SCHPH),
+ VEC_REG(VEC_CLMP0_START),
+ VEC_REG(VEC_CLMP0_END),
+ VEC_REG(VEC_FREQ3_2),
+ VEC_REG(VEC_FREQ1_0),
+ VEC_REG(VEC_CONFIG1),
+ VEC_REG(VEC_CONFIG2),
+ VEC_REG(VEC_INTERRUPT_CONTROL),
+ VEC_REG(VEC_INTERRUPT_STATUS),
+ VEC_REG(VEC_FCW_SECAM_B),
+ VEC_REG(VEC_SECAM_GAIN_VAL),
+ VEC_REG(VEC_CONFIG3),
+ VEC_REG(VEC_STATUS0),
+ VEC_REG(VEC_MASK0),
+ VEC_REG(VEC_CFG),
+ VEC_REG(VEC_DAC_TEST),
+ VEC_REG(VEC_DAC_CONFIG),
+ VEC_REG(VEC_DAC_MISC),
+};
+
+#ifdef CONFIG_DEBUG_FS
+int vc4_vec_debugfs_regs(struct seq_file *m, void *unused)
+{
+ struct drm_info_node *node = (struct drm_info_node *)m->private;
+ struct drm_device *dev = node->minor->dev;
+ struct vc4_dev *vc4 = to_vc4_dev(dev);
+ struct vc4_vec *vec = vc4->vec;
+ int i;
+
+ if (!vec)
+ return 0;
+
+ for (i = 0; i < ARRAY_SIZE(vec_regs); i++) {
+ seq_printf(m, "%s (0x%04x): 0x%08x\n",
+ vec_regs[i].name, vec_regs[i].reg,
+ VEC_READ(vec_regs[i].reg));
+ }
+
+ return 0;
+}
+#endif
+
+static void vc4_vec_ntsc_mode_set(struct vc4_vec *vec)
+{
+ VEC_WRITE(VEC_CONFIG0, VEC_CONFIG0_NTSC_STD | VEC_CONFIG0_PDEN);
+ VEC_WRITE(VEC_CONFIG1, VEC_CONFIG1_C_CVBS_CVBS);
+}
+
+static void vc4_vec_ntsc_j_mode_set(struct vc4_vec *vec)
+{
+ VEC_WRITE(VEC_CONFIG0, VEC_CONFIG0_NTSC_STD);
+ VEC_WRITE(VEC_CONFIG1, VEC_CONFIG1_C_CVBS_CVBS);
+}
+
+static const struct drm_display_mode ntsc_mode = {
+ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 13500,
+ 720, 720 + 14, 720 + 14 + 64, 720 + 14 + 64 + 60, 0,
+ 480, 480 + 3, 480 + 3 + 3, 480 + 3 + 3 + 16, 0,
+ DRM_MODE_FLAG_INTERLACE)
+};
+
+static void vc4_vec_pal_mode_set(struct vc4_vec *vec)
+{
+ VEC_WRITE(VEC_CONFIG0, VEC_CONFIG0_PAL_BDGHI_STD);
+ VEC_WRITE(VEC_CONFIG1, VEC_CONFIG1_C_CVBS_CVBS);
+}
+
+static void vc4_vec_pal_m_mode_set(struct vc4_vec *vec)
+{
+ VEC_WRITE(VEC_CONFIG0, VEC_CONFIG0_PAL_BDGHI_STD);
+ VEC_WRITE(VEC_CONFIG1,
+ VEC_CONFIG1_C_CVBS_CVBS | VEC_CONFIG1_CUSTOM_FREQ);
+ VEC_WRITE(VEC_FREQ3_2, 0x223b);
+ VEC_WRITE(VEC_FREQ1_0, 0x61d1);
+}
+
+static const struct drm_display_mode pal_mode = {
+ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 13500,
+ 720, 720 + 20, 720 + 20 + 64, 720 + 20 + 64 + 60, 0,
+ 576, 576 + 2, 576 + 2 + 3, 576 + 2 + 3 + 20, 0,
+ DRM_MODE_FLAG_INTERLACE)
+};
+
+static const struct vc4_vec_tv_mode vc4_vec_tv_modes[] = {
+ [VC4_VEC_TV_MODE_NTSC] = {
+ .mode = &ntsc_mode,
+ .mode_set = vc4_vec_ntsc_mode_set,
+ },
+ [VC4_VEC_TV_MODE_NTSC_J] = {
+ .mode = &ntsc_mode,
+ .mode_set = vc4_vec_ntsc_j_mode_set,
+ },
+ [VC4_VEC_TV_MODE_PAL] = {
+ .mode = &pal_mode,
+ .mode_set = vc4_vec_pal_mode_set,
+ },
+ [VC4_VEC_TV_MODE_PAL_M] = {
+ .mode = &pal_mode,
+ .mode_set = vc4_vec_pal_m_mode_set,
+ },
+};
+
+static enum drm_connector_status
+vc4_vec_connector_detect(struct drm_connector *connector, bool force)
+{
+ return connector_status_unknown;
+}
+
+static void vc4_vec_connector_destroy(struct drm_connector *connector)
+{
+ drm_connector_unregister(connector);
+ drm_connector_cleanup(connector);
+}
+
+static int vc4_vec_connector_get_modes(struct drm_connector *connector)
+{
+ struct drm_connector_state *state = connector->state;
+ struct drm_display_mode *mode;
+
+ mode = drm_mode_duplicate(connector->dev,
+ vc4_vec_tv_modes[state->tv.mode].mode);
+ if (!mode) {
+ DRM_ERROR("Failed to create a new display mode\n");
+ return -ENOMEM;
+ }
+
+ drm_mode_probed_add(connector, mode);
+
+ return 1;
+}
+
+static const struct drm_connector_funcs vc4_vec_connector_funcs = {
+ .dpms = drm_atomic_helper_connector_dpms,
+ .detect = vc4_vec_connector_detect,
+ .fill_modes = drm_helper_probe_single_connector_modes,
+ .set_property = drm_atomic_helper_connector_set_property,
+ .destroy = vc4_vec_connector_destroy,
+ .reset = drm_atomic_helper_connector_reset,
+ .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+ .atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+};
+
+static const struct drm_connector_helper_funcs vc4_vec_connector_helper_funcs = {
+ .get_modes = vc4_vec_connector_get_modes,
+};
+
+static struct drm_connector *vc4_vec_connector_init(struct drm_device *dev,
+ struct vc4_vec *vec)
+{
+ struct drm_connector *connector = NULL;
+ struct vc4_vec_connector *vec_connector;
+
+ vec_connector = devm_kzalloc(dev->dev, sizeof(*vec_connector),
+ GFP_KERNEL);
+ if (!vec_connector)
+ return ERR_PTR(-ENOMEM);
+
+ connector = &vec_connector->base;
+ connector->interlace_allowed = true;
+
+ vec_connector->encoder = vec->encoder;
+ vec_connector->vec = vec;
+
+ drm_connector_init(dev, connector, &vc4_vec_connector_funcs,
+ DRM_MODE_CONNECTOR_Composite);
+ drm_connector_helper_add(connector, &vc4_vec_connector_helper_funcs);
+
+ drm_object_attach_property(&connector->base,
+ dev->mode_config.tv_mode_property,
+ VC4_VEC_TV_MODE_NTSC);
+ vec->tv_mode = &vc4_vec_tv_modes[VC4_VEC_TV_MODE_NTSC];
+
+ drm_mode_connector_attach_encoder(connector, vec->encoder);
+
+ return connector;
+}
+
+static const struct drm_encoder_funcs vc4_vec_encoder_funcs = {
+ .destroy = drm_encoder_cleanup,
+};
+
+static void vc4_vec_encoder_disable(struct drm_encoder *encoder)
+{
+ struct vc4_vec_encoder *vc4_vec_encoder = to_vc4_vec_encoder(encoder);
+ struct vc4_vec *vec = vc4_vec_encoder->vec;
+ int ret;
+
+ VEC_WRITE(VEC_CFG, 0);
+ VEC_WRITE(VEC_DAC_MISC,
+ VEC_DAC_MISC_VCD_PWRDN |
+ VEC_DAC_MISC_BIAS_PWRDN |
+ VEC_DAC_MISC_DAC_PWRDN |
+ VEC_DAC_MISC_LDO_PWRDN);
+
+ clk_disable_unprepare(vec->clock);
+
+ ret = pm_runtime_put(&vec->pdev->dev);
+ if (ret < 0) {
+ DRM_ERROR("Failed to release power domain: %d\n", ret);
+ return;
+ }
+}
+
+static void vc4_vec_encoder_enable(struct drm_encoder *encoder)
+{
+ struct vc4_vec_encoder *vc4_vec_encoder = to_vc4_vec_encoder(encoder);
+ struct vc4_vec *vec = vc4_vec_encoder->vec;
+ int ret;
+
+ ret = pm_runtime_get_sync(&vec->pdev->dev);
+ if (ret < 0) {
+ DRM_ERROR("Failed to retain power domain: %d\n", ret);
+ return;
+ }
+
+ /*
+ * We need to set the clock rate each time we enable the encoder
+ * because there's a chance we share the same parent with the HDMI
+ * clock, and both drivers are requesting different rates.
+ * The good news is, these 2 encoders cannot be enabled at the same
+ * time, thus preventing incompatible rate requests.
+ */
+ ret = clk_set_rate(vec->clock, 108000000);
+ if (ret) {
+ DRM_ERROR("Failed to set clock rate: %d\n", ret);
+ return;
+ }
+
+ ret = clk_prepare_enable(vec->clock);
+ if (ret) {
+ DRM_ERROR("Failed to turn on core clock: %d\n", ret);
+ return;
+ }
+
+ /* Reset the different blocks */
+ VEC_WRITE(VEC_WSE_RESET, 1);
+ VEC_WRITE(VEC_SOFT_RESET, 1);
+
+ /* Disable the CGSM-A and WSE blocks */
+ VEC_WRITE(VEC_WSE_CONTROL, 0);
+
+ /* Write config common to all modes. */
+
+ /*
+ * Color subcarrier phase: phase = 360 * SCHPH / 256.
+ * 0x28 <=> 39.375 deg.
+ */
+ VEC_WRITE(VEC_SCHPH, 0x28);
+
+ /*
+ * Reset to default values.
+ */
+ 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_WRITE(VEC_CONFIG3, VEC_CONFIG3_HORIZ_LEN_STD);
+ VEC_WRITE(VEC_DAC_CONFIG,
+ VEC_DAC_CONFIG_DAC_CTRL(0xc) |
+ VEC_DAC_CONFIG_DRIVER_CTRL(0xc) |
+ VEC_DAC_CONFIG_LDO_BIAS_CTRL(0x46));
+
+ /* Mask all interrupts. */
+ VEC_WRITE(VEC_MASK0, 0);
+
+ vec->tv_mode->mode_set(vec);
+
+ VEC_WRITE(VEC_DAC_MISC,
+ VEC_DAC_MISC_VID_ACT | VEC_DAC_MISC_DAC_RST_N);
+ VEC_WRITE(VEC_CFG, VEC_CFG_VEC_EN);
+}
+
+
+static bool vc4_vec_encoder_mode_fixup(struct drm_encoder *encoder,
+ const struct drm_display_mode *mode,
+ struct drm_display_mode *adjusted_mode)
+{
+ return true;
+}
+
+static void vc4_vec_encoder_atomic_mode_set(struct drm_encoder *encoder,
+ struct drm_crtc_state *crtc_state,
+ struct drm_connector_state *conn_state)
+{
+ struct vc4_vec_encoder *vc4_vec_encoder = to_vc4_vec_encoder(encoder);
+ struct vc4_vec *vec = vc4_vec_encoder->vec;
+
+ vec->tv_mode = &vc4_vec_tv_modes[conn_state->tv.mode];
+}
+
+static int vc4_vec_encoder_atomic_check(struct drm_encoder *encoder,
+ struct drm_crtc_state *crtc_state,
+ struct drm_connector_state *conn_state)
+{
+ const struct vc4_vec_tv_mode *vec_mode;
+
+ vec_mode = &vc4_vec_tv_modes[conn_state->tv.mode];
+
+ if (conn_state->crtc &&
+ !drm_mode_equal(vec_mode->mode, &crtc_state->adjusted_mode))
+ return -EINVAL;
+
+ return 0;
+}
+
+static const struct drm_encoder_helper_funcs vc4_vec_encoder_helper_funcs = {
+ .disable = vc4_vec_encoder_disable,
+ .enable = vc4_vec_encoder_enable,
+ .mode_fixup = vc4_vec_encoder_mode_fixup,
+ .atomic_check = vc4_vec_encoder_atomic_check,
+ .atomic_mode_set = vc4_vec_encoder_atomic_mode_set,
+};
+
+static const struct of_device_id vc4_vec_dt_match[] = {
+ { .compatible = "brcm,bcm2835-vec", .data = NULL },
+ { /* sentinel */ },
+};
+
+static const char * const tv_mode_names[] = {
+ [VC4_VEC_TV_MODE_NTSC] = "NTSC",
+ [VC4_VEC_TV_MODE_NTSC_J] = "NTSC-J",
+ [VC4_VEC_TV_MODE_PAL] = "PAL",
+ [VC4_VEC_TV_MODE_PAL_M] = "PAL-M",
+};
+
+static int vc4_vec_bind(struct device *dev, struct device *master, void *data)
+{
+ struct platform_device *pdev = to_platform_device(dev);
+ struct drm_device *drm = dev_get_drvdata(master);
+ struct vc4_dev *vc4 = to_vc4_dev(drm);
+ struct vc4_vec *vec;
+ struct vc4_vec_encoder *vc4_vec_encoder;
+ int ret;
+
+ ret = drm_mode_create_tv_properties(drm, ARRAY_SIZE(tv_mode_names),
+ tv_mode_names);
+ if (ret)
+ return ret;
+
+ vec = devm_kzalloc(dev, sizeof(*vec), GFP_KERNEL);
+ if (!vec)
+ return -ENOMEM;
+
+ vc4_vec_encoder = devm_kzalloc(dev, sizeof(*vc4_vec_encoder),
+ GFP_KERNEL);
+ if (!vc4_vec_encoder)
+ return -ENOMEM;
+ vc4_vec_encoder->base.type = VC4_ENCODER_TYPE_VEC;
+ vc4_vec_encoder->vec = vec;
+ vec->encoder = &vc4_vec_encoder->base.base;
+
+ vec->pdev = pdev;
+ vec->regs = vc4_ioremap_regs(pdev, 0);
+ if (IS_ERR(vec->regs))
+ return PTR_ERR(vec->regs);
+
+ vec->clock = devm_clk_get(dev, NULL);
+ if (IS_ERR(vec->clock)) {
+ ret = PTR_ERR(vec->clock);
+ if (ret != -EPROBE_DEFER)
+ DRM_ERROR("Failed to get clock: %d\n", ret);
+ return ret;
+ }
+
+ pm_runtime_enable(dev);
+
+ drm_encoder_init(drm, vec->encoder, &vc4_vec_encoder_funcs,
+ DRM_MODE_ENCODER_TVDAC, NULL);
+ drm_encoder_helper_add(vec->encoder, &vc4_vec_encoder_helper_funcs);
+
+ vec->connector = vc4_vec_connector_init(drm, vec);
+ if (IS_ERR(vec->connector)) {
+ ret = PTR_ERR(vec->connector);
+ goto err_destroy_encoder;
+ }
+
+ dev_set_drvdata(dev, vec);
+
+ vc4->vec = vec;
+
+ return 0;
+
+err_destroy_encoder:
+ drm_encoder_cleanup(vec->encoder);
+ pm_runtime_disable(dev);
+
+ return ret;
+}
+
+static void vc4_vec_unbind(struct device *dev, struct device *master,
+ void *data)
+{
+ struct drm_device *drm = dev_get_drvdata(master);
+ struct vc4_dev *vc4 = to_vc4_dev(drm);
+ struct vc4_vec *vec = dev_get_drvdata(dev);
+
+ vc4_vec_connector_destroy(vec->connector);
+ drm_encoder_cleanup(vec->encoder);
+ pm_runtime_disable(dev);
+
+ vc4->vec = NULL;
+}
+
+static const struct component_ops vc4_vec_ops = {
+ .bind = vc4_vec_bind,
+ .unbind = vc4_vec_unbind,
+};
+
+static int vc4_vec_dev_probe(struct platform_device *pdev)
+{
+ return component_add(&pdev->dev, &vc4_vec_ops);
+}
+
+static int vc4_vec_dev_remove(struct platform_device *pdev)
+{
+ component_del(&pdev->dev, &vc4_vec_ops);
+ return 0;
+}
+
+struct platform_driver vc4_vec_driver = {
+ .probe = vc4_vec_dev_probe,
+ .remove = vc4_vec_dev_remove,
+ .driver = {
+ .name = "vc4_vec",
+ .of_match_table = vc4_vec_dt_match,
+ },
+};
generated by cgit v1.2.3 (git 2.46.0) at 2025-09-07 11:02:55 +0000