2019-06-03 07:44:50 +02:00
|
|
|
// SPDX-License-Identifier: GPL-2.0-only
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
/*
|
2015-07-06 16:35:31 -04:00
|
|
|
* Copyright (C) 2014-2015 The Linux Foundation. All rights reserved.
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
* Copyright (C) 2013 Red Hat
|
|
|
|
* Author: Rob Clark <robdclark@gmail.com>
|
|
|
|
*/
|
|
|
|
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
#include <drm/drm_atomic.h>
|
2022-06-13 23:03:12 +03:00
|
|
|
#include <drm/drm_blend.h>
|
2019-05-31 05:46:14 -04:00
|
|
|
#include <drm/drm_damage_helper.h>
|
2019-08-04 08:55:51 +02:00
|
|
|
#include <drm/drm_fourcc.h>
|
2022-06-14 12:54:49 +03:00
|
|
|
#include <drm/drm_framebuffer.h>
|
2022-02-23 11:11:08 -08:00
|
|
|
#include <drm/drm_gem_atomic_helper.h>
|
2016-11-05 11:08:11 -04:00
|
|
|
#include <drm/drm_print.h>
|
2019-08-04 08:55:51 +02:00
|
|
|
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
#include "mdp5_kms.h"
|
|
|
|
|
|
|
|
struct mdp5_plane {
|
|
|
|
struct drm_plane base;
|
|
|
|
};
|
|
|
|
#define to_mdp5_plane(x) container_of(x, struct mdp5_plane, base)
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
static int mdp5_plane_mode_set(struct drm_plane *plane,
|
|
|
|
struct drm_crtc *crtc, struct drm_framebuffer *fb,
|
2017-01-16 11:46:17 +05:30
|
|
|
struct drm_rect *src, struct drm_rect *dest);
|
2015-07-06 16:35:31 -04:00
|
|
|
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
static struct mdp5_kms *get_kms(struct drm_plane *plane)
|
|
|
|
{
|
|
|
|
struct msm_drm_private *priv = plane->dev->dev_private;
|
|
|
|
return to_mdp5_kms(to_mdp_kms(priv->kms));
|
|
|
|
}
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
static bool plane_enabled(struct drm_plane_state *state)
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
{
|
2017-01-16 11:46:17 +05:30
|
|
|
return state->visible;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
}
|
|
|
|
|
2021-05-25 16:13:14 +03:00
|
|
|
/* helper to install properties which are common to planes and crtcs */
|
|
|
|
static void mdp5_plane_install_properties(struct drm_plane *plane,
|
|
|
|
struct drm_mode_object *obj)
|
2015-07-31 10:13:26 -04:00
|
|
|
{
|
2022-02-21 10:59:08 +01:00
|
|
|
unsigned int zpos;
|
|
|
|
|
2016-10-21 22:22:43 +03:00
|
|
|
drm_plane_create_rotation_property(plane,
|
2017-05-19 16:50:17 -04:00
|
|
|
DRM_MODE_ROTATE_0,
|
|
|
|
DRM_MODE_ROTATE_0 |
|
|
|
|
DRM_MODE_ROTATE_180 |
|
|
|
|
DRM_MODE_REFLECT_X |
|
|
|
|
DRM_MODE_REFLECT_Y);
|
2021-05-25 16:13:13 +03:00
|
|
|
drm_plane_create_alpha_property(plane);
|
|
|
|
drm_plane_create_blend_mode_property(plane,
|
|
|
|
BIT(DRM_MODE_BLEND_PIXEL_NONE) |
|
|
|
|
BIT(DRM_MODE_BLEND_PREMULTI) |
|
|
|
|
BIT(DRM_MODE_BLEND_COVERAGE));
|
2022-02-21 10:59:08 +01:00
|
|
|
|
|
|
|
if (plane->type == DRM_PLANE_TYPE_PRIMARY)
|
|
|
|
zpos = STAGE_BASE;
|
|
|
|
else
|
|
|
|
zpos = STAGE0 + drm_plane_index(plane);
|
|
|
|
drm_plane_create_zpos_property(plane, zpos, 1, 255);
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
}
|
|
|
|
|
2016-11-05 11:08:11 -04:00
|
|
|
static void
|
|
|
|
mdp5_plane_atomic_print_state(struct drm_printer *p,
|
|
|
|
const struct drm_plane_state *state)
|
|
|
|
{
|
|
|
|
struct mdp5_plane_state *pstate = to_mdp5_plane_state(state);
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
struct mdp5_kms *mdp5_kms = get_kms(state->plane);
|
2016-11-05 11:08:11 -04:00
|
|
|
|
2016-11-01 11:56:54 -04:00
|
|
|
drm_printf(p, "\thwpipe=%s\n", pstate->hwpipe ?
|
|
|
|
pstate->hwpipe->name : "(null)");
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
if (mdp5_kms->caps & MDP_CAP_SRC_SPLIT)
|
|
|
|
drm_printf(p, "\tright-hwpipe=%s\n",
|
|
|
|
pstate->r_hwpipe ? pstate->r_hwpipe->name :
|
|
|
|
"(null)");
|
2021-05-25 16:13:12 +03:00
|
|
|
drm_printf(p, "\tblend_mode=%u\n", pstate->base.pixel_blend_mode);
|
2021-05-25 16:13:14 +03:00
|
|
|
drm_printf(p, "\tzpos=%u\n", pstate->base.zpos);
|
|
|
|
drm_printf(p, "\tnormalized_zpos=%u\n", pstate->base.normalized_zpos);
|
2021-05-25 16:13:11 +03:00
|
|
|
drm_printf(p, "\talpha=%u\n", pstate->base.alpha);
|
2016-11-05 11:08:11 -04:00
|
|
|
drm_printf(p, "\tstage=%s\n", stage2name(pstate->stage));
|
|
|
|
}
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
static void mdp5_plane_reset(struct drm_plane *plane)
|
|
|
|
{
|
|
|
|
struct mdp5_plane_state *mdp5_state;
|
|
|
|
|
2021-05-25 16:13:10 +03:00
|
|
|
if (plane->state)
|
|
|
|
__drm_atomic_helper_plane_destroy_state(plane->state);
|
2014-11-19 12:31:03 -05:00
|
|
|
|
|
|
|
kfree(to_mdp5_plane_state(plane->state));
|
2022-04-07 10:31:51 +08:00
|
|
|
plane->state = NULL;
|
2014-11-19 12:31:03 -05:00
|
|
|
mdp5_state = kzalloc(sizeof(*mdp5_state), GFP_KERNEL);
|
2022-04-07 10:31:51 +08:00
|
|
|
if (!mdp5_state)
|
|
|
|
return;
|
2021-05-25 16:13:10 +03:00
|
|
|
__drm_atomic_helper_plane_reset(plane, &mdp5_state->base);
|
2014-11-19 12:31:03 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
static struct drm_plane_state *
|
|
|
|
mdp5_plane_duplicate_state(struct drm_plane *plane)
|
|
|
|
{
|
|
|
|
struct mdp5_plane_state *mdp5_state;
|
|
|
|
|
|
|
|
if (WARN_ON(!plane->state))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
mdp5_state = kmemdup(to_mdp5_plane_state(plane->state),
|
|
|
|
sizeof(*mdp5_state), GFP_KERNEL);
|
2017-05-03 10:04:48 -04:00
|
|
|
if (!mdp5_state)
|
|
|
|
return NULL;
|
2014-11-19 12:31:03 -05:00
|
|
|
|
2017-05-03 10:04:48 -04:00
|
|
|
__drm_atomic_helper_plane_duplicate_state(plane, &mdp5_state->base);
|
2014-11-19 12:31:03 -05:00
|
|
|
|
|
|
|
return &mdp5_state->base;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mdp5_plane_destroy_state(struct drm_plane *plane,
|
|
|
|
struct drm_plane_state *state)
|
|
|
|
{
|
2016-11-01 11:56:54 -04:00
|
|
|
struct mdp5_plane_state *pstate = to_mdp5_plane_state(state);
|
|
|
|
|
2023-08-03 22:45:21 +02:00
|
|
|
__drm_atomic_helper_plane_destroy_state(state);
|
2014-11-19 12:31:03 -05:00
|
|
|
|
2016-11-01 11:56:54 -04:00
|
|
|
kfree(pstate);
|
2014-11-19 12:31:03 -05:00
|
|
|
}
|
|
|
|
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
static const struct drm_plane_funcs mdp5_plane_funcs = {
|
2014-11-19 12:31:03 -05:00
|
|
|
.update_plane = drm_atomic_helper_update_plane,
|
|
|
|
.disable_plane = drm_atomic_helper_disable_plane,
|
|
|
|
.reset = mdp5_plane_reset,
|
|
|
|
.atomic_duplicate_state = mdp5_plane_duplicate_state,
|
|
|
|
.atomic_destroy_state = mdp5_plane_destroy_state,
|
2016-11-05 11:08:11 -04:00
|
|
|
.atomic_print_state = mdp5_plane_atomic_print_state,
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
};
|
|
|
|
|
2022-02-23 11:11:08 -08:00
|
|
|
static int mdp5_plane_prepare_fb(struct drm_plane *plane,
|
|
|
|
struct drm_plane_state *new_state)
|
|
|
|
{
|
|
|
|
bool needs_dirtyfb = to_mdp5_plane_state(new_state)->needs_dirtyfb;
|
|
|
|
|
|
|
|
if (!new_state->fb)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
drm_gem_plane_helper_prepare_fb(plane, new_state);
|
|
|
|
|
2025-06-29 13:12:54 -07:00
|
|
|
return msm_framebuffer_prepare(new_state->fb, needs_dirtyfb);
|
2022-02-23 11:11:08 -08:00
|
|
|
}
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
static void mdp5_plane_cleanup_fb(struct drm_plane *plane,
|
2016-08-18 19:00:16 +01:00
|
|
|
struct drm_plane_state *old_state)
|
2014-11-18 12:49:49 -05:00
|
|
|
{
|
2015-09-02 10:42:40 +02:00
|
|
|
struct drm_framebuffer *fb = old_state->fb;
|
2022-02-23 11:11:08 -08:00
|
|
|
bool needed_dirtyfb = to_mdp5_plane_state(old_state)->needs_dirtyfb;
|
2015-09-02 10:42:40 +02:00
|
|
|
|
|
|
|
if (!fb)
|
|
|
|
return;
|
2014-11-18 12:49:49 -05:00
|
|
|
|
2016-11-01 09:31:21 -04:00
|
|
|
DBG("%s: cleanup: FB[%u]", plane->name, fb->base.id);
|
2025-06-29 13:12:54 -07:00
|
|
|
msm_framebuffer_cleanup(fb, needed_dirtyfb);
|
2014-11-19 12:31:03 -05:00
|
|
|
}
|
2014-11-18 12:49:49 -05:00
|
|
|
|
2017-01-16 12:16:34 +05:30
|
|
|
static int mdp5_plane_atomic_check_with_state(struct drm_crtc_state *crtc_state,
|
|
|
|
struct drm_plane_state *state)
|
2014-11-19 12:31:03 -05:00
|
|
|
{
|
2016-11-01 11:56:54 -04:00
|
|
|
struct mdp5_plane_state *mdp5_state = to_mdp5_plane_state(state);
|
2017-01-16 12:16:34 +05:30
|
|
|
struct drm_plane *plane = state->plane;
|
2014-11-19 12:31:03 -05:00
|
|
|
struct drm_plane_state *old_state = plane->state;
|
2016-11-05 10:43:55 -04:00
|
|
|
struct mdp5_cfg *config = mdp5_cfg_get_config(get_kms(plane)->cfg);
|
2016-11-01 11:56:54 -04:00
|
|
|
bool new_hwpipe = false;
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
bool need_right_hwpipe = false;
|
2016-11-05 10:43:55 -04:00
|
|
|
uint32_t max_width, max_height;
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
bool out_of_bounds = false;
|
2016-11-01 11:56:54 -04:00
|
|
|
uint32_t caps = 0;
|
2017-01-16 11:46:17 +05:30
|
|
|
int min_scale, max_scale;
|
|
|
|
int ret;
|
2014-11-19 12:31:03 -05:00
|
|
|
|
2016-11-01 09:31:21 -04:00
|
|
|
DBG("%s: check (%d -> %d)", plane->name,
|
2014-11-19 12:31:03 -05:00
|
|
|
plane_enabled(old_state), plane_enabled(state));
|
|
|
|
|
2016-11-05 10:43:55 -04:00
|
|
|
max_width = config->hw->lm.max_width << 16;
|
|
|
|
max_height = config->hw->lm.max_height << 16;
|
|
|
|
|
|
|
|
/* Make sure source dimensions are within bounds. */
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
if (state->src_h > max_height)
|
|
|
|
out_of_bounds = true;
|
|
|
|
|
|
|
|
if (state->src_w > max_width) {
|
|
|
|
/* If source split is supported, we can go up to 2x
|
|
|
|
* the max LM width, but we'd need to stage another
|
|
|
|
* hwpipe to the right LM. So, the drm_plane would
|
|
|
|
* consist of 2 hwpipes.
|
|
|
|
*/
|
|
|
|
if (config->hw->mdp.caps & MDP_CAP_SRC_SPLIT &&
|
|
|
|
(state->src_w <= 2 * max_width))
|
|
|
|
need_right_hwpipe = true;
|
|
|
|
else
|
|
|
|
out_of_bounds = true;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (out_of_bounds) {
|
2016-11-05 10:43:55 -04:00
|
|
|
struct drm_rect src = drm_plane_state_src(state);
|
|
|
|
DBG("Invalid source size "DRM_RECT_FP_FMT,
|
|
|
|
DRM_RECT_FP_ARG(&src));
|
|
|
|
return -ERANGE;
|
|
|
|
}
|
|
|
|
|
2017-01-16 11:46:17 +05:30
|
|
|
min_scale = FRAC_16_16(1, 8);
|
|
|
|
max_scale = FRAC_16_16(8, 1);
|
|
|
|
|
2018-01-23 19:08:57 +02:00
|
|
|
ret = drm_atomic_helper_check_plane_state(state, crtc_state,
|
2017-11-01 22:16:19 +02:00
|
|
|
min_scale, max_scale,
|
|
|
|
true, true);
|
2017-01-16 11:46:17 +05:30
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2015-07-08 18:12:40 -04:00
|
|
|
if (plane_enabled(state)) {
|
2016-10-21 22:22:44 +03:00
|
|
|
unsigned int rotation;
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct msm_format *format;
|
2016-11-01 16:35:32 -04:00
|
|
|
struct mdp5_kms *mdp5_kms = get_kms(plane);
|
|
|
|
uint32_t blkcfg = 0;
|
2016-10-21 22:22:44 +03:00
|
|
|
|
2024-04-20 07:01:02 +03:00
|
|
|
format = msm_framebuffer_format(state->fb);
|
|
|
|
if (MSM_FORMAT_IS_YUV(format))
|
2016-11-01 11:56:54 -04:00
|
|
|
caps |= MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_CSC;
|
2015-07-08 18:12:40 -04:00
|
|
|
|
2016-11-01 11:56:54 -04:00
|
|
|
if (((state->src_w >> 16) != state->crtc_w) ||
|
|
|
|
((state->src_h >> 16) != state->crtc_h))
|
|
|
|
caps |= MDP_PIPE_CAP_SCALE;
|
2015-07-31 10:13:26 -04:00
|
|
|
|
2016-10-21 22:22:44 +03:00
|
|
|
rotation = drm_rotation_simplify(state->rotation,
|
2017-05-19 16:50:17 -04:00
|
|
|
DRM_MODE_ROTATE_0 |
|
|
|
|
DRM_MODE_REFLECT_X |
|
|
|
|
DRM_MODE_REFLECT_Y);
|
2016-11-01 09:56:51 -04:00
|
|
|
|
2017-05-19 16:50:17 -04:00
|
|
|
if (rotation & DRM_MODE_REFLECT_X)
|
2016-11-01 11:56:54 -04:00
|
|
|
caps |= MDP_PIPE_CAP_HFLIP;
|
|
|
|
|
2017-05-19 16:50:17 -04:00
|
|
|
if (rotation & DRM_MODE_REFLECT_Y)
|
2016-11-01 11:56:54 -04:00
|
|
|
caps |= MDP_PIPE_CAP_VFLIP;
|
|
|
|
|
2017-01-16 11:57:04 +05:30
|
|
|
if (plane->type == DRM_PLANE_TYPE_CURSOR)
|
|
|
|
caps |= MDP_PIPE_CAP_CURSOR;
|
|
|
|
|
2016-11-01 11:56:54 -04:00
|
|
|
/* (re)allocate hw pipe if we don't have one or caps-mismatch: */
|
|
|
|
if (!mdp5_state->hwpipe || (caps & ~mdp5_state->hwpipe->caps))
|
|
|
|
new_hwpipe = true;
|
|
|
|
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
/*
|
|
|
|
* (re)allocte hw pipe if we're either requesting for 2 hw pipes
|
|
|
|
* or we're switching from 2 hw pipes to 1 hw pipe because the
|
|
|
|
* new src_w can be supported by 1 hw pipe itself.
|
|
|
|
*/
|
|
|
|
if ((need_right_hwpipe && !mdp5_state->r_hwpipe) ||
|
|
|
|
(!need_right_hwpipe && mdp5_state->r_hwpipe))
|
|
|
|
new_hwpipe = true;
|
|
|
|
|
2016-11-01 16:35:32 -04:00
|
|
|
if (mdp5_kms->smp) {
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct msm_format *format =
|
|
|
|
msm_framebuffer_format(state->fb);
|
2016-11-01 16:35:32 -04:00
|
|
|
|
|
|
|
blkcfg = mdp5_smp_calculate(mdp5_kms->smp, format,
|
|
|
|
state->src_w >> 16, false);
|
|
|
|
|
|
|
|
if (mdp5_state->hwpipe && (mdp5_state->hwpipe->blkcfg != blkcfg))
|
|
|
|
new_hwpipe = true;
|
2015-07-31 10:13:26 -04:00
|
|
|
}
|
2015-07-08 18:12:40 -04:00
|
|
|
|
2016-11-01 11:56:54 -04:00
|
|
|
/* (re)assign hwpipe if needed, otherwise keep old one: */
|
|
|
|
if (new_hwpipe) {
|
|
|
|
/* TODO maybe we want to re-assign hwpipe sometimes
|
|
|
|
* in cases when we no-longer need some caps to make
|
|
|
|
* it available for other planes?
|
|
|
|
*/
|
|
|
|
struct mdp5_hw_pipe *old_hwpipe = mdp5_state->hwpipe;
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
struct mdp5_hw_pipe *old_right_hwpipe =
|
|
|
|
mdp5_state->r_hwpipe;
|
2017-10-27 16:27:30 +05:30
|
|
|
struct mdp5_hw_pipe *new_hwpipe = NULL;
|
|
|
|
struct mdp5_hw_pipe *new_right_hwpipe = NULL;
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
|
2017-10-27 16:27:29 +05:30
|
|
|
ret = mdp5_pipe_assign(state->state, plane, caps,
|
2017-10-27 16:27:30 +05:30
|
|
|
blkcfg, &new_hwpipe,
|
|
|
|
need_right_hwpipe ?
|
|
|
|
&new_right_hwpipe : NULL);
|
2017-10-27 16:27:29 +05:30
|
|
|
if (ret) {
|
2017-10-27 16:27:30 +05:30
|
|
|
DBG("%s: failed to assign hwpipe(s)!",
|
|
|
|
plane->name);
|
2017-10-27 16:27:29 +05:30
|
|
|
return ret;
|
2016-11-01 11:56:54 -04:00
|
|
|
}
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
|
2017-10-27 16:27:30 +05:30
|
|
|
mdp5_state->hwpipe = new_hwpipe;
|
|
|
|
if (need_right_hwpipe)
|
|
|
|
mdp5_state->r_hwpipe = new_right_hwpipe;
|
|
|
|
else
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
/*
|
|
|
|
* set it to NULL so that the driver knows we
|
|
|
|
* don't have a right hwpipe when committing a
|
|
|
|
* new state
|
|
|
|
*/
|
|
|
|
mdp5_state->r_hwpipe = NULL;
|
2017-10-27 16:27:30 +05:30
|
|
|
|
drm/msm/mdp5: Assign a 'right hwpipe' to plane state
If the drm_plane has a source width that's greater than the max width
supported by a SSPP (2560 pixels on 8x96), then we assign a 'r_hwpipe'
to it in mdp5_plane_atomic_check().
TODO: There are a few scenarios where the hwpipe assignments aren't
recommended by HW. For example, an assignment which results in a
drm_plane to of two different types of hwpipes (say RGB0 on left
and DMA1 on right) is not recommended.
Also, hwpipes have a priority mapping, where the higher priority pipe
needs to be staged on left LM, and the lower priority needs to be
staged on the right LM. For example, the priority order for VIG pipes
in decreasing order of priority is VIG0, VIG1, VIG2, and VIG3. So, VIG0
on left and VIG1 on right is a correct configuration, but VIG1 on left
and VIG0 on right isn't. These scenarios are ignored for now for the
sake of simplicity.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
2017-03-23 15:58:10 +05:30
|
|
|
|
2022-05-05 14:40:50 -07:00
|
|
|
ret = mdp5_pipe_release(state->state, old_hwpipe);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret = mdp5_pipe_release(state->state, old_right_hwpipe);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
}
|
2017-05-21 12:05:07 -04:00
|
|
|
} else {
|
2022-05-05 14:40:50 -07:00
|
|
|
ret = mdp5_pipe_release(state->state, mdp5_state->hwpipe);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret = mdp5_pipe_release(state->state, mdp5_state->r_hwpipe);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2017-05-21 12:05:07 -04:00
|
|
|
mdp5_state->hwpipe = mdp5_state->r_hwpipe = NULL;
|
2014-11-19 12:31:03 -05:00
|
|
|
}
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2017-01-16 12:16:34 +05:30
|
|
|
static int mdp5_plane_atomic_check(struct drm_plane *plane,
|
drm/atomic: Pass the full state to planes atomic_check
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_check.
The conversion was done using the coccinelle script below plus some
manual changes for vmwgfx, built tested on all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_check = func,
...,
};
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_check(plane, plane_state)
+ FUNCS->atomic_check(plane, state)
...+>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *new_plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-4-maxime@cerno.tech
2021-02-19 13:00:24 +01:00
|
|
|
struct drm_atomic_state *state)
|
2017-01-16 12:16:34 +05:30
|
|
|
{
|
drm: Use state helper instead of plane state pointer in atomic_check
Many drivers reference the plane->state pointer in order to get the
current plane state in their atomic_check hook, which would be the old
plane state in the global atomic state since _swap_state hasn't happened
when atomic_check is run.
Use the drm_atomic_get_old_plane_state helper to get that state to make
it more obvious.
This was made using the coccinelle script below:
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
static struct drm_plane_helper_funcs helpers = {
...,
.atomic_check = func,
...,
};
@ replaces_old_state @
identifier plane_atomic_func.func;
identifier plane, state, plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
- struct drm_plane_state *plane_state = plane->state;
+ struct drm_plane_state *plane_state = drm_atomic_get_old_plane_state(state, plane);
...
}
@@
identifier plane_atomic_func.func;
identifier plane, state, plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
struct drm_plane_state *plane_state = drm_atomic_get_old_plane_state(state, plane);
<...
- plane->state
+ plane_state
...>
}
@ adds_old_state @
identifier plane_atomic_func.func;
identifier plane, state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
+ struct drm_plane_state *old_plane_state = drm_atomic_get_old_plane_state(state, plane);
<...
- plane->state
+ old_plane_state
...>
}
@ include depends on adds_old_state || replaces_old_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && (adds_old_state || replaces_old_state) @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-6-maxime@cerno.tech
2021-02-19 13:00:26 +01:00
|
|
|
struct drm_plane_state *old_plane_state = drm_atomic_get_old_plane_state(state,
|
|
|
|
plane);
|
drm/atomic: Pass the full state to planes atomic_check
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_check.
The conversion was done using the coccinelle script below plus some
manual changes for vmwgfx, built tested on all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_check = func,
...,
};
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_check(plane, plane_state)
+ FUNCS->atomic_check(plane, state)
...+>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *new_plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-4-maxime@cerno.tech
2021-02-19 13:00:24 +01:00
|
|
|
struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
|
|
|
|
plane);
|
2017-01-16 12:16:34 +05:30
|
|
|
struct drm_crtc *crtc;
|
|
|
|
struct drm_crtc_state *crtc_state;
|
|
|
|
|
drm: Use state helper instead of plane state pointer in atomic_check
Many drivers reference the plane->state pointer in order to get the
current plane state in their atomic_check hook, which would be the old
plane state in the global atomic state since _swap_state hasn't happened
when atomic_check is run.
Use the drm_atomic_get_old_plane_state helper to get that state to make
it more obvious.
This was made using the coccinelle script below:
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
static struct drm_plane_helper_funcs helpers = {
...,
.atomic_check = func,
...,
};
@ replaces_old_state @
identifier plane_atomic_func.func;
identifier plane, state, plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
- struct drm_plane_state *plane_state = plane->state;
+ struct drm_plane_state *plane_state = drm_atomic_get_old_plane_state(state, plane);
...
}
@@
identifier plane_atomic_func.func;
identifier plane, state, plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
struct drm_plane_state *plane_state = drm_atomic_get_old_plane_state(state, plane);
<...
- plane->state
+ plane_state
...>
}
@ adds_old_state @
identifier plane_atomic_func.func;
identifier plane, state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
+ struct drm_plane_state *old_plane_state = drm_atomic_get_old_plane_state(state, plane);
<...
- plane->state
+ old_plane_state
...>
}
@ include depends on adds_old_state || replaces_old_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && (adds_old_state || replaces_old_state) @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-6-maxime@cerno.tech
2021-02-19 13:00:26 +01:00
|
|
|
crtc = new_plane_state->crtc ? new_plane_state->crtc : old_plane_state->crtc;
|
2017-01-16 12:16:34 +05:30
|
|
|
if (!crtc)
|
|
|
|
return 0;
|
|
|
|
|
drm: Use the state pointer directly in planes atomic_check
Now that atomic_check takes the global atomic state as a parameter, we
don't need to go through the pointer in the plane state.
This was done using the following coccinelle script:
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
static struct drm_plane_helper_funcs helpers = {
...,
.atomic_check = func,
...,
};
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
- struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<... when != plane_state
- plane_state->state
+ state
...>
}
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<...
- plane_state->state
+ state
...>
}
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-5-maxime@cerno.tech
2021-02-19 13:00:25 +01:00
|
|
|
crtc_state = drm_atomic_get_existing_crtc_state(state,
|
2021-02-19 13:00:22 +01:00
|
|
|
crtc);
|
2017-01-16 12:16:34 +05:30
|
|
|
if (WARN_ON(!crtc_state))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2021-02-19 13:00:22 +01:00
|
|
|
return mdp5_plane_atomic_check_with_state(crtc_state, new_plane_state);
|
2017-01-16 12:16:34 +05:30
|
|
|
}
|
|
|
|
|
2014-11-25 12:09:44 +01:00
|
|
|
static void mdp5_plane_atomic_update(struct drm_plane *plane,
|
drm/atomic: Pass the full state to planes atomic disable and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's convert the remaining helpers to provide a consistent interface,
this time with the planes atomic_update and atomic_disable.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_disable)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_update = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_disable = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier crtc_state;
identifier plane, plane_state, state;
expression e;
@@
f(struct drm_crtc_state *crtc_state)
{
...
struct drm_atomic_state *state = e;
<+...
(
- FUNCS->atomic_disable(plane, plane_state)
+ FUNCS->atomic_disable(plane, state)
|
- FUNCS->atomic_update(plane, plane_state)
+ FUNCS->atomic_update(plane, state)
)
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *old_plane_state)
{
<...
- state
+ old_plane_state
...>
}
@ ignores_old_state @
identifier plane_atomic_func.func;
identifier plane, old_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
... when != old_state
}
@ adds_old_state depends on plane_atomic_func && !ignores_old_state @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *plane_state)
{
+ struct drm_plane_state *plane_state = drm_atomic_get_old_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_old_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_old_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_old_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-9-maxime@cerno.tech
2021-02-19 13:00:29 +01:00
|
|
|
struct drm_atomic_state *state)
|
2014-11-19 12:31:03 -05:00
|
|
|
{
|
drm: Use state helper instead of the plane state pointer
Many drivers reference the plane->state pointer in order to get the
current plane state in their atomic_update or atomic_disable hooks,
which would be the new plane state in the global atomic state since
_swap_state happened when those hooks are run.
Use the drm_atomic_get_new_plane_state helper to get that state to make it
more obvious.
This was made using the coccinelle script below:
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_disable = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_update = func,
...,
};
)
@ adds_new_state @
identifier plane_atomic_func.func;
identifier plane, state;
identifier new_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state)
{
...
- struct drm_plane_state *new_state = plane->state;
+ struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://lore.kernel.org/r/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:30 +01:00
|
|
|
struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
|
|
|
|
plane);
|
2014-11-19 12:31:03 -05:00
|
|
|
|
2016-11-01 09:31:21 -04:00
|
|
|
DBG("%s: update", plane->name);
|
2014-11-18 12:49:49 -05:00
|
|
|
|
drm: Rename plane->state variables in atomic update and disable
Some drivers are storing the plane->state pointer in atomic_update and
atomic_disable in a variable simply called state, while the state passed
as an argument is called old_state.
In order to ease subsequent reworks and to avoid confusing or
inconsistent names, let's rename those variables to new_state.
This was done using the following coccinelle script, plus some manual
changes for mtk and tegra.
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_disable = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_update = func,
...,
};
)
@ moves_new_state_old_state @
identifier plane_atomic_func.func;
identifier plane;
symbol old_state;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
...
- struct drm_plane_state *state = plane->state;
+ struct drm_plane_state *new_state = plane->state;
...
}
@ depends on moves_new_state_old_state @
identifier plane_atomic_func.func;
identifier plane;
identifier old_state;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
<...
- state
+ new_state
...>
}
@ moves_new_state_oldstate @
identifier plane_atomic_func.func;
identifier plane;
symbol oldstate;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *oldstate)
{
...
- struct drm_plane_state *state = plane->state;
+ struct drm_plane_state *newstate = plane->state;
...
}
@ depends on moves_new_state_oldstate @
identifier plane_atomic_func.func;
identifier plane;
identifier old_state;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
<...
- state
+ newstate
...>
}
@ moves_new_state_old_pstate @
identifier plane_atomic_func.func;
identifier plane;
symbol old_pstate;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_pstate)
{
...
- struct drm_plane_state *state = plane->state;
+ struct drm_plane_state *new_pstate = plane->state;
...
}
@ depends on moves_new_state_old_pstate @
identifier plane_atomic_func.func;
identifier plane;
identifier old_pstate;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_pstate)
{
<...
- state
+ new_pstate
...>
}
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-8-maxime@cerno.tech
2021-02-19 13:00:28 +01:00
|
|
|
if (plane_enabled(new_state)) {
|
2014-11-19 12:31:03 -05:00
|
|
|
int ret;
|
2016-11-04 17:12:44 -04:00
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
ret = mdp5_plane_mode_set(plane,
|
drm: Rename plane->state variables in atomic update and disable
Some drivers are storing the plane->state pointer in atomic_update and
atomic_disable in a variable simply called state, while the state passed
as an argument is called old_state.
In order to ease subsequent reworks and to avoid confusing or
inconsistent names, let's rename those variables to new_state.
This was done using the following coccinelle script, plus some manual
changes for mtk and tegra.
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_disable = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_update = func,
...,
};
)
@ moves_new_state_old_state @
identifier plane_atomic_func.func;
identifier plane;
symbol old_state;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
...
- struct drm_plane_state *state = plane->state;
+ struct drm_plane_state *new_state = plane->state;
...
}
@ depends on moves_new_state_old_state @
identifier plane_atomic_func.func;
identifier plane;
identifier old_state;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
<...
- state
+ new_state
...>
}
@ moves_new_state_oldstate @
identifier plane_atomic_func.func;
identifier plane;
symbol oldstate;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *oldstate)
{
...
- struct drm_plane_state *state = plane->state;
+ struct drm_plane_state *newstate = plane->state;
...
}
@ depends on moves_new_state_oldstate @
identifier plane_atomic_func.func;
identifier plane;
identifier old_state;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_state)
{
<...
- state
+ newstate
...>
}
@ moves_new_state_old_pstate @
identifier plane_atomic_func.func;
identifier plane;
symbol old_pstate;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_pstate)
{
...
- struct drm_plane_state *state = plane->state;
+ struct drm_plane_state *new_pstate = plane->state;
...
}
@ depends on moves_new_state_old_pstate @
identifier plane_atomic_func.func;
identifier plane;
identifier old_pstate;
symbol state;
@@
func(struct drm_plane *plane, struct drm_plane_state *old_pstate)
{
<...
- state
+ new_pstate
...>
}
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-8-maxime@cerno.tech
2021-02-19 13:00:28 +01:00
|
|
|
new_state->crtc, new_state->fb,
|
|
|
|
&new_state->src, &new_state->dst);
|
2014-11-19 12:31:03 -05:00
|
|
|
/* atomic_check should have ensured that this doesn't fail */
|
|
|
|
WARN_ON(ret < 0);
|
|
|
|
}
|
2014-11-18 12:49:49 -05:00
|
|
|
}
|
|
|
|
|
2017-06-30 15:03:20 -03:00
|
|
|
static int mdp5_plane_atomic_async_check(struct drm_plane *plane,
|
2025-01-27 16:59:39 -03:00
|
|
|
struct drm_atomic_state *state, bool flip)
|
2017-06-30 15:03:20 -03:00
|
|
|
{
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
|
|
|
|
plane);
|
|
|
|
struct mdp5_plane_state *mdp5_state = to_mdp5_plane_state(new_plane_state);
|
2017-06-30 15:03:20 -03:00
|
|
|
struct drm_crtc_state *crtc_state;
|
2017-10-27 16:27:32 +05:30
|
|
|
int min_scale, max_scale;
|
|
|
|
int ret;
|
2017-06-30 15:03:20 -03:00
|
|
|
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
crtc_state = drm_atomic_get_existing_crtc_state(state,
|
|
|
|
new_plane_state->crtc);
|
2017-06-30 15:03:20 -03:00
|
|
|
if (WARN_ON(!crtc_state))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (!crtc_state->active)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/* don't use fast path if we don't have a hwpipe allocated yet */
|
|
|
|
if (!mdp5_state->hwpipe)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/* only allow changing of position(crtc x/y or src x/y) in fast path */
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
if (plane->state->crtc != new_plane_state->crtc ||
|
|
|
|
plane->state->src_w != new_plane_state->src_w ||
|
|
|
|
plane->state->src_h != new_plane_state->src_h ||
|
|
|
|
plane->state->crtc_w != new_plane_state->crtc_w ||
|
|
|
|
plane->state->crtc_h != new_plane_state->crtc_h ||
|
2017-06-30 15:03:20 -03:00
|
|
|
!plane->state->fb ||
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
plane->state->fb != new_plane_state->fb)
|
2017-06-30 15:03:20 -03:00
|
|
|
return -EINVAL;
|
|
|
|
|
2017-10-27 16:27:32 +05:30
|
|
|
min_scale = FRAC_16_16(1, 8);
|
|
|
|
max_scale = FRAC_16_16(8, 1);
|
|
|
|
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
ret = drm_atomic_helper_check_plane_state(new_plane_state, crtc_state,
|
2017-11-21 11:33:10 +01:00
|
|
|
min_scale, max_scale,
|
|
|
|
true, true);
|
2017-10-27 16:27:32 +05:30
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* if the visibility of the plane changes (i.e, if the cursor is
|
|
|
|
* clipped out completely, we can't take the async path because
|
|
|
|
* we need to stage/unstage the plane from the Layer Mixer(s). We
|
|
|
|
* also assign/unassign the hwpipe(s) tied to the plane. We avoid
|
|
|
|
* taking the fast path for both these reasons.
|
|
|
|
*/
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
if (new_plane_state->visible != plane->state->visible)
|
2017-10-27 16:27:32 +05:30
|
|
|
return -EINVAL;
|
|
|
|
|
2017-06-30 15:03:20 -03:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mdp5_plane_atomic_async_update(struct drm_plane *plane,
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
struct drm_atomic_state *state)
|
2017-06-30 15:03:20 -03:00
|
|
|
{
|
drm/atomic: Pass the full state to planes async atomic check and update
The current atomic helpers have either their object state being passed as
an argument or the full atomic state.
The former is the pattern that was done at first, before switching to the
latter for new hooks or when it was needed.
Let's start convert all the remaining helpers to provide a consistent
interface, starting with the planes atomic_async_check and
atomic_async_update.
The conversion was done using the coccinelle script below, built tested on
all the drivers.
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
int (*atomic_async_check)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@@
identifier plane, plane_state;
symbol state;
@@
struct drm_plane_helper_funcs {
...
void (*atomic_async_update)(struct drm_plane *plane,
- struct drm_plane_state *plane_state);
+ struct drm_atomic_state *state);
...
}
@ plane_atomic_func @
identifier helpers;
identifier func;
@@
(
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_check = func,
...,
};
|
static const struct drm_plane_helper_funcs helpers = {
...,
.atomic_async_update = func,
...,
};
)
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_check(plane, plane_state)
+ FUNCS->atomic_async_check(plane, state)
...+>
}
@@
struct drm_plane_helper_funcs *FUNCS;
identifier f;
identifier dev;
identifier plane, plane_state, state;
@@
f(struct drm_device *dev, struct drm_atomic_state *state)
{
<+...
- FUNCS->atomic_async_update(plane, plane_state)
+ FUNCS->atomic_async_update(plane, state)
...+>
}
@@
identifier mtk_plane_atomic_async_update;
identifier plane;
symbol new_state, state;
expression e;
@@
void mtk_plane_atomic_async_update(struct drm_plane *plane, struct drm_plane_state *new_state)
{
...
- struct mtk_plane_state *state = e;
+ struct mtk_plane_state *new_plane_state = e;
<+...
- state
+ new_plane_state
...+>
}
@@
identifier plane_atomic_func.func;
identifier plane;
symbol state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *state)
+ struct drm_plane_state *new_plane_state)
{
<...
- state
+ new_plane_state
...>
}
@ ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
... when != new_plane_state
}
@ adds_new_state depends on plane_atomic_func && !ignores_new_state @
identifier plane_atomic_func.func;
identifier plane, new_plane_state;
@@
func(struct drm_plane *plane, struct drm_plane_state *new_plane_state)
{
+ struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state, plane);
...
}
@ depends on plane_atomic_func @
identifier plane_atomic_func.func;
identifier plane, plane_state;
@@
func(struct drm_plane *plane,
- struct drm_plane_state *plane_state
+ struct drm_atomic_state *state
)
{ ... }
@ include depends on adds_new_state @
@@
#include <drm/drm_atomic.h>
@ no_include depends on !include && adds_new_state @
@@
+ #include <drm/drm_atomic.h>
#include <drm/...>
@@
identifier plane_atomic_func.func;
identifier plane, state;
identifier plane_state;
@@
func(struct drm_plane *plane, struct drm_atomic_state *state) {
...
struct drm_plane_state *plane_state = drm_atomic_get_new_plane_state(state, plane);
<+...
- plane_state->state
+ state
...+>
}
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20210219120032.260676-1-maxime@cerno.tech
2021-02-19 13:00:21 +01:00
|
|
|
struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
|
|
|
|
plane);
|
2019-06-03 13:56:08 -03:00
|
|
|
struct drm_framebuffer *old_fb = plane->state->fb;
|
|
|
|
|
2017-06-30 15:03:20 -03:00
|
|
|
plane->state->src_x = new_state->src_x;
|
|
|
|
plane->state->src_y = new_state->src_y;
|
|
|
|
plane->state->crtc_x = new_state->crtc_x;
|
|
|
|
plane->state->crtc_y = new_state->crtc_y;
|
|
|
|
|
|
|
|
if (plane_enabled(new_state)) {
|
|
|
|
struct mdp5_ctl *ctl;
|
|
|
|
struct mdp5_pipeline *pipeline =
|
2018-04-05 18:13:48 +03:00
|
|
|
mdp5_crtc_get_pipeline(new_state->crtc);
|
2017-06-30 15:03:20 -03:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = mdp5_plane_mode_set(plane, new_state->crtc, new_state->fb,
|
|
|
|
&new_state->src, &new_state->dst);
|
|
|
|
WARN_ON(ret < 0);
|
|
|
|
|
|
|
|
ctl = mdp5_crtc_get_ctl(new_state->crtc);
|
|
|
|
|
2018-02-19 08:17:06 -05:00
|
|
|
mdp5_ctl_commit(ctl, pipeline, mdp5_plane_get_flush(plane), true);
|
2017-06-30 15:03:20 -03:00
|
|
|
}
|
|
|
|
|
|
|
|
*to_mdp5_plane_state(plane->state) =
|
|
|
|
*to_mdp5_plane_state(new_state);
|
2019-06-03 13:56:08 -03:00
|
|
|
|
|
|
|
new_state->fb = old_fb;
|
2017-06-30 15:03:20 -03:00
|
|
|
}
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
static const struct drm_plane_helper_funcs mdp5_plane_helper_funcs = {
|
2022-02-23 11:11:08 -08:00
|
|
|
.prepare_fb = mdp5_plane_prepare_fb,
|
2014-11-19 12:31:03 -05:00
|
|
|
.cleanup_fb = mdp5_plane_cleanup_fb,
|
|
|
|
.atomic_check = mdp5_plane_atomic_check,
|
|
|
|
.atomic_update = mdp5_plane_atomic_update,
|
2017-06-30 15:03:20 -03:00
|
|
|
.atomic_async_check = mdp5_plane_atomic_async_check,
|
|
|
|
.atomic_async_update = mdp5_plane_atomic_async_update,
|
2014-11-19 12:31:03 -05:00
|
|
|
};
|
|
|
|
|
2017-03-23 15:58:09 +05:30
|
|
|
static void set_scanout_locked(struct mdp5_kms *mdp5_kms,
|
|
|
|
enum mdp5_pipe pipe,
|
|
|
|
struct drm_framebuffer *fb)
|
2014-11-18 12:49:49 -05:00
|
|
|
{
|
2014-11-19 12:31:03 -05:00
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_STRIDE_A(pipe),
|
|
|
|
MDP5_PIPE_SRC_STRIDE_A_P0(fb->pitches[0]) |
|
|
|
|
MDP5_PIPE_SRC_STRIDE_A_P1(fb->pitches[1]));
|
2014-11-18 12:49:49 -05:00
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_STRIDE_B(pipe),
|
|
|
|
MDP5_PIPE_SRC_STRIDE_B_P2(fb->pitches[2]) |
|
|
|
|
MDP5_PIPE_SRC_STRIDE_B_P3(fb->pitches[3]));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC0_ADDR(pipe),
|
2025-06-29 13:12:54 -07:00
|
|
|
msm_framebuffer_iova(fb, 0));
|
2014-11-19 12:31:03 -05:00
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC1_ADDR(pipe),
|
2025-06-29 13:12:54 -07:00
|
|
|
msm_framebuffer_iova(fb, 1));
|
2014-11-19 12:31:03 -05:00
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC2_ADDR(pipe),
|
2025-06-29 13:12:54 -07:00
|
|
|
msm_framebuffer_iova(fb, 2));
|
2014-11-19 12:31:03 -05:00
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC3_ADDR(pipe),
|
2025-06-29 13:12:54 -07:00
|
|
|
msm_framebuffer_iova(fb, 3));
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
}
|
|
|
|
|
2014-12-08 10:48:58 -05:00
|
|
|
/* Note: mdp5_plane->pipe_lock must be locked */
|
|
|
|
static void csc_disable(struct mdp5_kms *mdp5_kms, enum mdp5_pipe pipe)
|
|
|
|
{
|
|
|
|
uint32_t value = mdp5_read(mdp5_kms, REG_MDP5_PIPE_OP_MODE(pipe)) &
|
|
|
|
~MDP5_PIPE_OP_MODE_CSC_1_EN;
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_OP_MODE(pipe), value);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Note: mdp5_plane->pipe_lock must be locked */
|
|
|
|
static void csc_enable(struct mdp5_kms *mdp5_kms, enum mdp5_pipe pipe,
|
|
|
|
struct csc_cfg *csc)
|
|
|
|
{
|
|
|
|
uint32_t i, mode = 0; /* RGB, no CSC */
|
|
|
|
uint32_t *matrix;
|
|
|
|
|
|
|
|
if (unlikely(!csc))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if ((csc->type == CSC_YUV2RGB) || (CSC_YUV2YUV == csc->type))
|
|
|
|
mode |= MDP5_PIPE_OP_MODE_CSC_SRC_DATA_FORMAT(DATA_FORMAT_YUV);
|
|
|
|
if ((csc->type == CSC_RGB2YUV) || (CSC_YUV2YUV == csc->type))
|
|
|
|
mode |= MDP5_PIPE_OP_MODE_CSC_DST_DATA_FORMAT(DATA_FORMAT_YUV);
|
|
|
|
mode |= MDP5_PIPE_OP_MODE_CSC_1_EN;
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_OP_MODE(pipe), mode);
|
|
|
|
|
|
|
|
matrix = csc->matrix;
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_MATRIX_COEFF_0(pipe),
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_0_COEFF_11(matrix[0]) |
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_0_COEFF_12(matrix[1]));
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_MATRIX_COEFF_1(pipe),
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_1_COEFF_13(matrix[2]) |
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_1_COEFF_21(matrix[3]));
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_MATRIX_COEFF_2(pipe),
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_2_COEFF_22(matrix[4]) |
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_2_COEFF_23(matrix[5]));
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_MATRIX_COEFF_3(pipe),
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_3_COEFF_31(matrix[6]) |
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_3_COEFF_32(matrix[7]));
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_MATRIX_COEFF_4(pipe),
|
|
|
|
MDP5_PIPE_CSC_1_MATRIX_COEFF_4_COEFF_33(matrix[8]));
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(csc->pre_bias); i++) {
|
|
|
|
uint32_t *pre_clamp = csc->pre_clamp;
|
|
|
|
uint32_t *post_clamp = csc->post_clamp;
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_PRE_CLAMP(pipe, i),
|
|
|
|
MDP5_PIPE_CSC_1_PRE_CLAMP_REG_HIGH(pre_clamp[2*i+1]) |
|
|
|
|
MDP5_PIPE_CSC_1_PRE_CLAMP_REG_LOW(pre_clamp[2*i]));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_POST_CLAMP(pipe, i),
|
|
|
|
MDP5_PIPE_CSC_1_POST_CLAMP_REG_HIGH(post_clamp[2*i+1]) |
|
|
|
|
MDP5_PIPE_CSC_1_POST_CLAMP_REG_LOW(post_clamp[2*i]));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_PRE_BIAS(pipe, i),
|
|
|
|
MDP5_PIPE_CSC_1_PRE_BIAS_REG_VALUE(csc->pre_bias[i]));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_CSC_1_POST_BIAS(pipe, i),
|
|
|
|
MDP5_PIPE_CSC_1_POST_BIAS_REG_VALUE(csc->post_bias[i]));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#define PHASE_STEP_SHIFT 21
|
|
|
|
#define DOWN_SCALE_RATIO_MAX 32 /* 2^(26-21) */
|
|
|
|
|
|
|
|
static int calc_phase_step(uint32_t src, uint32_t dst, uint32_t *out_phase)
|
|
|
|
{
|
|
|
|
uint32_t unit;
|
|
|
|
|
|
|
|
if (src == 0 || dst == 0)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* PHASE_STEP_X/Y is coded on 26 bits (25:0),
|
|
|
|
* where 2^21 represents the unity "1" in fixed-point hardware design.
|
|
|
|
* This leaves 5 bits for the integer part (downscale case):
|
|
|
|
* -> maximum downscale ratio = 0b1_1111 = 31
|
|
|
|
*/
|
|
|
|
if (src > (dst * DOWN_SCALE_RATIO_MAX))
|
|
|
|
return -EOVERFLOW;
|
|
|
|
|
|
|
|
unit = 1 << PHASE_STEP_SHIFT;
|
|
|
|
*out_phase = mult_frac(unit, src, dst);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-07-06 16:35:31 -04:00
|
|
|
static int calc_scalex_steps(struct drm_plane *plane,
|
|
|
|
uint32_t pixel_format, uint32_t src, uint32_t dest,
|
2015-09-15 08:41:51 -04:00
|
|
|
uint32_t phasex_steps[COMP_MAX])
|
2014-12-08 10:48:58 -05:00
|
|
|
{
|
2019-05-16 12:31:48 +02:00
|
|
|
const struct drm_format_info *info = drm_format_info(pixel_format);
|
2015-07-06 16:35:31 -04:00
|
|
|
struct mdp5_kms *mdp5_kms = get_kms(plane);
|
|
|
|
struct device *dev = mdp5_kms->dev->dev;
|
2014-12-08 10:48:58 -05:00
|
|
|
uint32_t phasex_step;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = calc_phase_step(src, dest, &phasex_step);
|
2015-07-06 16:35:31 -04:00
|
|
|
if (ret) {
|
2018-10-20 23:19:26 +05:30
|
|
|
DRM_DEV_ERROR(dev, "X scaling (%d->%d) failed: %d\n", src, dest, ret);
|
2014-12-08 10:48:58 -05:00
|
|
|
return ret;
|
2015-07-06 16:35:31 -04:00
|
|
|
}
|
2014-12-08 10:48:58 -05:00
|
|
|
|
2015-09-15 08:41:51 -04:00
|
|
|
phasex_steps[COMP_0] = phasex_step;
|
|
|
|
phasex_steps[COMP_3] = phasex_step;
|
2019-05-16 12:31:48 +02:00
|
|
|
phasex_steps[COMP_1_2] = phasex_step / info->hsub;
|
2014-12-08 10:48:58 -05:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-07-06 16:35:31 -04:00
|
|
|
static int calc_scaley_steps(struct drm_plane *plane,
|
|
|
|
uint32_t pixel_format, uint32_t src, uint32_t dest,
|
2015-09-15 08:41:51 -04:00
|
|
|
uint32_t phasey_steps[COMP_MAX])
|
2014-12-08 10:48:58 -05:00
|
|
|
{
|
2019-05-16 12:31:48 +02:00
|
|
|
const struct drm_format_info *info = drm_format_info(pixel_format);
|
2015-07-06 16:35:31 -04:00
|
|
|
struct mdp5_kms *mdp5_kms = get_kms(plane);
|
|
|
|
struct device *dev = mdp5_kms->dev->dev;
|
2014-12-08 10:48:58 -05:00
|
|
|
uint32_t phasey_step;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = calc_phase_step(src, dest, &phasey_step);
|
2015-07-06 16:35:31 -04:00
|
|
|
if (ret) {
|
2018-10-20 23:19:26 +05:30
|
|
|
DRM_DEV_ERROR(dev, "Y scaling (%d->%d) failed: %d\n", src, dest, ret);
|
2014-12-08 10:48:58 -05:00
|
|
|
return ret;
|
2015-07-06 16:35:31 -04:00
|
|
|
}
|
2014-12-08 10:48:58 -05:00
|
|
|
|
2015-09-15 08:41:51 -04:00
|
|
|
phasey_steps[COMP_0] = phasey_step;
|
|
|
|
phasey_steps[COMP_3] = phasey_step;
|
2019-05-16 12:31:48 +02:00
|
|
|
phasey_steps[COMP_1_2] = phasey_step / info->vsub;
|
2014-12-08 10:48:58 -05:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2024-04-20 07:01:02 +03:00
|
|
|
static uint32_t get_scale_config(const struct msm_format *format,
|
2015-09-15 08:41:52 -04:00
|
|
|
uint32_t src, uint32_t dst, bool horz)
|
2014-12-08 10:48:58 -05:00
|
|
|
{
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct drm_format_info *info = drm_format_info(format->pixel_format);
|
|
|
|
bool yuv = MSM_FORMAT_IS_YUV(format);
|
2024-04-20 07:01:01 +03:00
|
|
|
bool scaling = yuv ? true : (src != dst);
|
2019-05-16 12:31:48 +02:00
|
|
|
uint32_t sub;
|
2015-09-15 08:41:52 -04:00
|
|
|
uint32_t ya_filter, uv_filter;
|
|
|
|
|
|
|
|
if (!scaling)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (yuv) {
|
2019-05-16 12:31:48 +02:00
|
|
|
sub = horz ? info->hsub : info->vsub;
|
2015-09-15 08:41:52 -04:00
|
|
|
uv_filter = ((src / sub) <= dst) ?
|
|
|
|
SCALE_FILTER_BIL : SCALE_FILTER_PCMN;
|
|
|
|
}
|
|
|
|
ya_filter = (src <= dst) ? SCALE_FILTER_BIL : SCALE_FILTER_PCMN;
|
|
|
|
|
|
|
|
if (horz)
|
|
|
|
return MDP5_PIPE_SCALE_CONFIG_SCALEX_EN |
|
|
|
|
MDP5_PIPE_SCALE_CONFIG_SCALEX_FILTER_COMP_0(ya_filter) |
|
|
|
|
MDP5_PIPE_SCALE_CONFIG_SCALEX_FILTER_COMP_3(ya_filter) |
|
|
|
|
COND(yuv, MDP5_PIPE_SCALE_CONFIG_SCALEX_FILTER_COMP_1_2(uv_filter));
|
|
|
|
else
|
|
|
|
return MDP5_PIPE_SCALE_CONFIG_SCALEY_EN |
|
|
|
|
MDP5_PIPE_SCALE_CONFIG_SCALEY_FILTER_COMP_0(ya_filter) |
|
|
|
|
MDP5_PIPE_SCALE_CONFIG_SCALEY_FILTER_COMP_3(ya_filter) |
|
|
|
|
COND(yuv, MDP5_PIPE_SCALE_CONFIG_SCALEY_FILTER_COMP_1_2(uv_filter));
|
|
|
|
}
|
|
|
|
|
2024-04-20 07:01:02 +03:00
|
|
|
static void calc_pixel_ext(const struct msm_format *format,
|
2015-09-15 08:41:52 -04:00
|
|
|
uint32_t src, uint32_t dst, uint32_t phase_step[2],
|
|
|
|
int pix_ext_edge1[COMP_MAX], int pix_ext_edge2[COMP_MAX],
|
|
|
|
bool horz)
|
|
|
|
{
|
2024-04-20 07:01:02 +03:00
|
|
|
bool scaling = MSM_FORMAT_IS_YUV(format) ? true : (src != dst);
|
2015-09-15 08:41:52 -04:00
|
|
|
int i;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Note:
|
|
|
|
* We assume here that:
|
|
|
|
* 1. PCMN filter is used for downscale
|
|
|
|
* 2. bilinear filter is used for upscale
|
|
|
|
* 3. we are in a single pipe configuration
|
|
|
|
*/
|
|
|
|
|
|
|
|
for (i = 0; i < COMP_MAX; i++) {
|
|
|
|
pix_ext_edge1[i] = 0;
|
|
|
|
pix_ext_edge2[i] = scaling ? 1 : 0;
|
2015-07-06 16:35:31 -04:00
|
|
|
}
|
2015-09-15 08:41:52 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
static void mdp5_write_pixel_ext(struct mdp5_kms *mdp5_kms, enum mdp5_pipe pipe,
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct msm_format *format,
|
2015-09-15 08:41:52 -04:00
|
|
|
uint32_t src_w, int pe_left[COMP_MAX], int pe_right[COMP_MAX],
|
|
|
|
uint32_t src_h, int pe_top[COMP_MAX], int pe_bottom[COMP_MAX])
|
|
|
|
{
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct drm_format_info *info = drm_format_info(format->pixel_format);
|
2015-09-15 08:41:52 -04:00
|
|
|
uint32_t lr, tb, req;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < COMP_MAX; i++) {
|
|
|
|
uint32_t roi_w = src_w;
|
|
|
|
uint32_t roi_h = src_h;
|
|
|
|
|
2024-04-20 07:01:02 +03:00
|
|
|
if (MSM_FORMAT_IS_YUV(format) && i == COMP_1_2) {
|
2019-05-16 12:31:48 +02:00
|
|
|
roi_w /= info->hsub;
|
|
|
|
roi_h /= info->vsub;
|
2015-09-15 08:41:52 -04:00
|
|
|
}
|
2014-12-08 10:48:58 -05:00
|
|
|
|
2015-09-15 08:41:52 -04:00
|
|
|
lr = (pe_left[i] >= 0) ?
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_LR_LEFT_RPT(pe_left[i]) :
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_LR_LEFT_OVF(pe_left[i]);
|
|
|
|
|
|
|
|
lr |= (pe_right[i] >= 0) ?
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_LR_RIGHT_RPT(pe_right[i]) :
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_LR_RIGHT_OVF(pe_right[i]);
|
|
|
|
|
|
|
|
tb = (pe_top[i] >= 0) ?
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_TB_TOP_RPT(pe_top[i]) :
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_TB_TOP_OVF(pe_top[i]);
|
|
|
|
|
|
|
|
tb |= (pe_bottom[i] >= 0) ?
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_TB_BOTTOM_RPT(pe_bottom[i]) :
|
|
|
|
MDP5_PIPE_SW_PIX_EXT_TB_BOTTOM_OVF(pe_bottom[i]);
|
|
|
|
|
|
|
|
req = MDP5_PIPE_SW_PIX_EXT_REQ_PIXELS_LEFT_RIGHT(roi_w +
|
|
|
|
pe_left[i] + pe_right[i]);
|
|
|
|
|
|
|
|
req |= MDP5_PIPE_SW_PIX_EXT_REQ_PIXELS_TOP_BOTTOM(roi_h +
|
|
|
|
pe_top[i] + pe_bottom[i]);
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SW_PIX_EXT_LR(pipe, i), lr);
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SW_PIX_EXT_TB(pipe, i), tb);
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SW_PIX_EXT_REQ_PIXELS(pipe, i), req);
|
|
|
|
|
|
|
|
DBG("comp-%d (L/R): rpt=%d/%d, ovf=%d/%d, req=%d", i,
|
|
|
|
FIELD(lr, MDP5_PIPE_SW_PIX_EXT_LR_LEFT_RPT),
|
|
|
|
FIELD(lr, MDP5_PIPE_SW_PIX_EXT_LR_RIGHT_RPT),
|
|
|
|
FIELD(lr, MDP5_PIPE_SW_PIX_EXT_LR_LEFT_OVF),
|
|
|
|
FIELD(lr, MDP5_PIPE_SW_PIX_EXT_LR_RIGHT_OVF),
|
|
|
|
FIELD(req, MDP5_PIPE_SW_PIX_EXT_REQ_PIXELS_LEFT_RIGHT));
|
|
|
|
|
|
|
|
DBG("comp-%d (T/B): rpt=%d/%d, ovf=%d/%d, req=%d", i,
|
|
|
|
FIELD(tb, MDP5_PIPE_SW_PIX_EXT_TB_TOP_RPT),
|
|
|
|
FIELD(tb, MDP5_PIPE_SW_PIX_EXT_TB_BOTTOM_RPT),
|
|
|
|
FIELD(tb, MDP5_PIPE_SW_PIX_EXT_TB_TOP_OVF),
|
|
|
|
FIELD(tb, MDP5_PIPE_SW_PIX_EXT_TB_BOTTOM_OVF),
|
|
|
|
FIELD(req, MDP5_PIPE_SW_PIX_EXT_REQ_PIXELS_TOP_BOTTOM));
|
|
|
|
}
|
2014-12-08 10:48:58 -05:00
|
|
|
}
|
|
|
|
|
2017-03-23 15:58:09 +05:30
|
|
|
struct pixel_ext {
|
|
|
|
int left[COMP_MAX];
|
|
|
|
int right[COMP_MAX];
|
|
|
|
int top[COMP_MAX];
|
|
|
|
int bottom[COMP_MAX];
|
|
|
|
};
|
|
|
|
|
|
|
|
struct phase_step {
|
|
|
|
u32 x[COMP_MAX];
|
|
|
|
u32 y[COMP_MAX];
|
|
|
|
};
|
|
|
|
|
|
|
|
static void mdp5_hwpipe_mode_set(struct mdp5_kms *mdp5_kms,
|
|
|
|
struct mdp5_hw_pipe *hwpipe,
|
|
|
|
struct drm_framebuffer *fb,
|
|
|
|
struct phase_step *step,
|
|
|
|
struct pixel_ext *pe,
|
|
|
|
u32 scale_config, u32 hdecm, u32 vdecm,
|
|
|
|
bool hflip, bool vflip,
|
|
|
|
int crtc_x, int crtc_y,
|
|
|
|
unsigned int crtc_w, unsigned int crtc_h,
|
|
|
|
u32 src_img_w, u32 src_img_h,
|
|
|
|
u32 src_x, u32 src_y,
|
|
|
|
u32 src_w, u32 src_h)
|
|
|
|
{
|
|
|
|
enum mdp5_pipe pipe = hwpipe->pipe;
|
|
|
|
bool has_pe = hwpipe->caps & MDP_PIPE_CAP_SW_PIX_EXT;
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct msm_format *format =
|
|
|
|
msm_framebuffer_format(fb);
|
2017-03-23 15:58:09 +05:30
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_IMG_SIZE(pipe),
|
|
|
|
MDP5_PIPE_SRC_IMG_SIZE_WIDTH(src_img_w) |
|
|
|
|
MDP5_PIPE_SRC_IMG_SIZE_HEIGHT(src_img_h));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_SIZE(pipe),
|
|
|
|
MDP5_PIPE_SRC_SIZE_WIDTH(src_w) |
|
|
|
|
MDP5_PIPE_SRC_SIZE_HEIGHT(src_h));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_XY(pipe),
|
|
|
|
MDP5_PIPE_SRC_XY_X(src_x) |
|
|
|
|
MDP5_PIPE_SRC_XY_Y(src_y));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_OUT_SIZE(pipe),
|
|
|
|
MDP5_PIPE_OUT_SIZE_WIDTH(crtc_w) |
|
|
|
|
MDP5_PIPE_OUT_SIZE_HEIGHT(crtc_h));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_OUT_XY(pipe),
|
|
|
|
MDP5_PIPE_OUT_XY_X(crtc_x) |
|
|
|
|
MDP5_PIPE_OUT_XY_Y(crtc_y));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_FORMAT(pipe),
|
|
|
|
MDP5_PIPE_SRC_FORMAT_A_BPC(format->bpc_a) |
|
2024-04-20 07:01:02 +03:00
|
|
|
MDP5_PIPE_SRC_FORMAT_R_BPC(format->bpc_r_cr) |
|
|
|
|
MDP5_PIPE_SRC_FORMAT_G_BPC(format->bpc_g_y) |
|
|
|
|
MDP5_PIPE_SRC_FORMAT_B_BPC(format->bpc_b_cb) |
|
2017-03-23 15:58:09 +05:30
|
|
|
COND(format->alpha_enable, MDP5_PIPE_SRC_FORMAT_ALPHA_ENABLE) |
|
2024-04-20 07:01:02 +03:00
|
|
|
MDP5_PIPE_SRC_FORMAT_CPP(format->bpp - 1) |
|
2017-03-23 15:58:09 +05:30
|
|
|
MDP5_PIPE_SRC_FORMAT_UNPACK_COUNT(format->unpack_count - 1) |
|
2024-04-20 07:01:03 +03:00
|
|
|
COND(format->flags & MSM_FORMAT_FLAG_UNPACK_TIGHT,
|
|
|
|
MDP5_PIPE_SRC_FORMAT_UNPACK_TIGHT) |
|
2017-03-23 15:58:09 +05:30
|
|
|
MDP5_PIPE_SRC_FORMAT_FETCH_TYPE(format->fetch_type) |
|
|
|
|
MDP5_PIPE_SRC_FORMAT_CHROMA_SAMP(format->chroma_sample));
|
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_UNPACK(pipe),
|
2024-04-20 07:01:02 +03:00
|
|
|
MDP5_PIPE_SRC_UNPACK_ELEM0(format->element[0]) |
|
|
|
|
MDP5_PIPE_SRC_UNPACK_ELEM1(format->element[1]) |
|
|
|
|
MDP5_PIPE_SRC_UNPACK_ELEM2(format->element[2]) |
|
|
|
|
MDP5_PIPE_SRC_UNPACK_ELEM3(format->element[3]));
|
2017-03-23 15:58:09 +05:30
|
|
|
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_OP_MODE(pipe),
|
|
|
|
(hflip ? MDP5_PIPE_SRC_OP_MODE_FLIP_LR : 0) |
|
|
|
|
(vflip ? MDP5_PIPE_SRC_OP_MODE_FLIP_UD : 0) |
|
|
|
|
COND(has_pe, MDP5_PIPE_SRC_OP_MODE_SW_PIX_EXT_OVERRIDE) |
|
|
|
|
MDP5_PIPE_SRC_OP_MODE_BWC(BWC_LOSSLESS));
|
|
|
|
|
|
|
|
/* not using secure mode: */
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SRC_ADDR_SW_STATUS(pipe), 0);
|
|
|
|
|
|
|
|
if (hwpipe->caps & MDP_PIPE_CAP_SW_PIX_EXT)
|
|
|
|
mdp5_write_pixel_ext(mdp5_kms, pipe, format,
|
|
|
|
src_w, pe->left, pe->right,
|
|
|
|
src_h, pe->top, pe->bottom);
|
|
|
|
|
|
|
|
if (hwpipe->caps & MDP_PIPE_CAP_SCALE) {
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SCALE_PHASE_STEP_X(pipe),
|
|
|
|
step->x[COMP_0]);
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SCALE_PHASE_STEP_Y(pipe),
|
|
|
|
step->y[COMP_0]);
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SCALE_CR_PHASE_STEP_X(pipe),
|
|
|
|
step->x[COMP_1_2]);
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SCALE_CR_PHASE_STEP_Y(pipe),
|
|
|
|
step->y[COMP_1_2]);
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_DECIMATION(pipe),
|
|
|
|
MDP5_PIPE_DECIMATION_VERT(vdecm) |
|
|
|
|
MDP5_PIPE_DECIMATION_HORZ(hdecm));
|
|
|
|
mdp5_write(mdp5_kms, REG_MDP5_PIPE_SCALE_CONFIG(pipe),
|
|
|
|
scale_config);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (hwpipe->caps & MDP_PIPE_CAP_CSC) {
|
2024-04-20 07:01:02 +03:00
|
|
|
if (MSM_FORMAT_IS_YUV(format))
|
2017-03-23 15:58:09 +05:30
|
|
|
csc_enable(mdp5_kms, pipe,
|
|
|
|
mdp_get_default_csc_cfg(CSC_YUV2RGB));
|
|
|
|
else
|
|
|
|
csc_disable(mdp5_kms, pipe);
|
|
|
|
}
|
|
|
|
|
|
|
|
set_scanout_locked(mdp5_kms, pipe, fb);
|
|
|
|
}
|
2015-09-15 08:41:52 -04:00
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
static int mdp5_plane_mode_set(struct drm_plane *plane,
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
struct drm_crtc *crtc, struct drm_framebuffer *fb,
|
2017-01-16 11:46:17 +05:30
|
|
|
struct drm_rect *src, struct drm_rect *dest)
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
{
|
2015-07-31 10:13:26 -04:00
|
|
|
struct drm_plane_state *pstate = plane->state;
|
2016-11-01 11:56:54 -04:00
|
|
|
struct mdp5_hw_pipe *hwpipe = to_mdp5_plane_state(pstate)->hwpipe;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
struct mdp5_kms *mdp5_kms = get_kms(plane);
|
2016-11-01 09:56:51 -04:00
|
|
|
enum mdp5_pipe pipe = hwpipe->pipe;
|
2017-03-23 15:58:11 +05:30
|
|
|
struct mdp5_hw_pipe *right_hwpipe;
|
2024-04-20 07:01:02 +03:00
|
|
|
const struct msm_format *format;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
uint32_t nplanes, config = 0;
|
2017-06-29 14:49:59 +05:30
|
|
|
struct phase_step step = { { 0 } };
|
|
|
|
struct pixel_ext pe = { { 0 } };
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
uint32_t hdecm = 0, vdecm = 0;
|
2014-12-08 10:48:58 -05:00
|
|
|
uint32_t pix_format;
|
2016-10-21 22:22:44 +03:00
|
|
|
unsigned int rotation;
|
2015-07-31 10:13:26 -04:00
|
|
|
bool vflip, hflip;
|
2017-01-16 11:46:17 +05:30
|
|
|
int crtc_x, crtc_y;
|
|
|
|
unsigned int crtc_w, crtc_h;
|
|
|
|
uint32_t src_x, src_y;
|
|
|
|
uint32_t src_w, src_h;
|
2017-03-23 15:58:09 +05:30
|
|
|
uint32_t src_img_w, src_img_h;
|
2014-11-18 12:49:47 -05:00
|
|
|
int ret;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
2016-12-14 23:30:22 +02:00
|
|
|
nplanes = fb->format->num_planes;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
|
|
|
/* bad formats should already be rejected: */
|
|
|
|
if (WARN_ON(nplanes > pipe2nclients(pipe)))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2024-04-20 07:01:02 +03:00
|
|
|
format = msm_framebuffer_format(fb);
|
|
|
|
pix_format = format->pixel_format;
|
2014-12-08 10:48:58 -05:00
|
|
|
|
2017-01-16 11:46:17 +05:30
|
|
|
src_x = src->x1;
|
|
|
|
src_y = src->y1;
|
|
|
|
src_w = drm_rect_width(src);
|
|
|
|
src_h = drm_rect_height(src);
|
|
|
|
|
|
|
|
crtc_x = dest->x1;
|
|
|
|
crtc_y = dest->y1;
|
|
|
|
crtc_w = drm_rect_width(dest);
|
|
|
|
crtc_h = drm_rect_height(dest);
|
|
|
|
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
/* src values are in Q16 fixed point, convert to integer: */
|
|
|
|
src_x = src_x >> 16;
|
|
|
|
src_y = src_y >> 16;
|
|
|
|
src_w = src_w >> 16;
|
|
|
|
src_h = src_h >> 16;
|
|
|
|
|
2017-03-23 15:58:09 +05:30
|
|
|
src_img_w = min(fb->width, src_w);
|
|
|
|
src_img_h = min(fb->height, src_h);
|
|
|
|
|
2016-11-01 09:31:21 -04:00
|
|
|
DBG("%s: FB[%u] %u,%u,%u,%u -> CRTC[%u] %d,%d,%u,%u", plane->name,
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
fb->base.id, src_x, src_y, src_w, src_h,
|
|
|
|
crtc->base.id, crtc_x, crtc_y, crtc_w, crtc_h);
|
|
|
|
|
2017-03-23 15:58:11 +05:30
|
|
|
right_hwpipe = to_mdp5_plane_state(pstate)->r_hwpipe;
|
|
|
|
if (right_hwpipe) {
|
|
|
|
/*
|
|
|
|
* if the plane comprises of 2 hw pipes, assume that the width
|
|
|
|
* is split equally across them. The only parameters that varies
|
|
|
|
* between the 2 pipes are src_x and crtc_x
|
|
|
|
*/
|
|
|
|
crtc_w /= 2;
|
|
|
|
src_w /= 2;
|
|
|
|
src_img_w /= 2;
|
|
|
|
}
|
|
|
|
|
2017-03-23 15:58:09 +05:30
|
|
|
ret = calc_scalex_steps(plane, pix_format, src_w, crtc_w, step.x);
|
2015-07-06 16:35:31 -04:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
2014-12-08 10:48:58 -05:00
|
|
|
|
2017-03-23 15:58:09 +05:30
|
|
|
ret = calc_scaley_steps(plane, pix_format, src_h, crtc_h, step.y);
|
2015-07-06 16:35:31 -04:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
2016-11-01 09:56:51 -04:00
|
|
|
if (hwpipe->caps & MDP_PIPE_CAP_SW_PIX_EXT) {
|
2017-03-23 15:58:09 +05:30
|
|
|
calc_pixel_ext(format, src_w, crtc_w, step.x,
|
|
|
|
pe.left, pe.right, true);
|
|
|
|
calc_pixel_ext(format, src_h, crtc_h, step.y,
|
|
|
|
pe.top, pe.bottom, false);
|
2015-09-15 08:41:52 -04:00
|
|
|
}
|
|
|
|
|
2015-07-06 16:35:31 -04:00
|
|
|
/* TODO calc hdecm, vdecm */
|
|
|
|
|
|
|
|
/* SCALE is used to both scale and up-sample chroma components */
|
2015-09-15 08:41:52 -04:00
|
|
|
config |= get_scale_config(format, src_w, crtc_w, true);
|
|
|
|
config |= get_scale_config(format, src_h, crtc_h, false);
|
2015-07-06 16:35:31 -04:00
|
|
|
DBG("scale config = %x", config);
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
2016-10-21 22:22:44 +03:00
|
|
|
rotation = drm_rotation_simplify(pstate->rotation,
|
2017-05-19 16:50:17 -04:00
|
|
|
DRM_MODE_ROTATE_0 |
|
|
|
|
DRM_MODE_REFLECT_X |
|
|
|
|
DRM_MODE_REFLECT_Y);
|
|
|
|
hflip = !!(rotation & DRM_MODE_REFLECT_X);
|
|
|
|
vflip = !!(rotation & DRM_MODE_REFLECT_Y);
|
2015-07-31 10:13:26 -04:00
|
|
|
|
2017-03-23 15:58:09 +05:30
|
|
|
mdp5_hwpipe_mode_set(mdp5_kms, hwpipe, fb, &step, &pe,
|
|
|
|
config, hdecm, vdecm, hflip, vflip,
|
|
|
|
crtc_x, crtc_y, crtc_w, crtc_h,
|
|
|
|
src_img_w, src_img_h,
|
|
|
|
src_x, src_y, src_w, src_h);
|
2017-03-23 15:58:11 +05:30
|
|
|
if (right_hwpipe)
|
|
|
|
mdp5_hwpipe_mode_set(mdp5_kms, right_hwpipe, fb, &step, &pe,
|
|
|
|
config, hdecm, vdecm, hflip, vflip,
|
2017-12-04 15:44:23 +01:00
|
|
|
crtc_x + crtc_w, crtc_y, crtc_w, crtc_h,
|
2017-03-23 15:58:11 +05:30
|
|
|
src_img_w, src_img_h,
|
2017-12-04 15:44:23 +01:00
|
|
|
src_x + src_w, src_y, src_w, src_h);
|
2014-11-18 12:49:49 -05:00
|
|
|
|
|
|
|
return ret;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
}
|
|
|
|
|
2017-03-23 15:58:11 +05:30
|
|
|
/*
|
|
|
|
* Use this func and the one below only after the atomic state has been
|
|
|
|
* successfully swapped
|
|
|
|
*/
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
enum mdp5_pipe mdp5_plane_pipe(struct drm_plane *plane)
|
|
|
|
{
|
2016-11-01 11:56:54 -04:00
|
|
|
struct mdp5_plane_state *pstate = to_mdp5_plane_state(plane->state);
|
|
|
|
|
|
|
|
if (WARN_ON(!pstate->hwpipe))
|
2016-12-12 15:17:56 +05:30
|
|
|
return SSPP_NONE;
|
2016-11-01 11:56:54 -04:00
|
|
|
|
|
|
|
return pstate->hwpipe->pipe;
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
}
|
|
|
|
|
2017-03-23 15:58:11 +05:30
|
|
|
enum mdp5_pipe mdp5_plane_right_pipe(struct drm_plane *plane)
|
|
|
|
{
|
|
|
|
struct mdp5_plane_state *pstate = to_mdp5_plane_state(plane->state);
|
|
|
|
|
|
|
|
if (!pstate->r_hwpipe)
|
|
|
|
return SSPP_NONE;
|
|
|
|
|
|
|
|
return pstate->r_hwpipe->pipe;
|
|
|
|
}
|
|
|
|
|
2014-11-18 12:49:49 -05:00
|
|
|
uint32_t mdp5_plane_get_flush(struct drm_plane *plane)
|
|
|
|
{
|
2016-11-01 11:56:54 -04:00
|
|
|
struct mdp5_plane_state *pstate = to_mdp5_plane_state(plane->state);
|
2017-03-23 15:58:11 +05:30
|
|
|
u32 mask;
|
2016-11-01 11:56:54 -04:00
|
|
|
|
|
|
|
if (WARN_ON(!pstate->hwpipe))
|
|
|
|
return 0;
|
2014-11-18 12:49:49 -05:00
|
|
|
|
2017-03-23 15:58:11 +05:30
|
|
|
mask = pstate->hwpipe->flush_mask;
|
|
|
|
|
|
|
|
if (pstate->r_hwpipe)
|
|
|
|
mask |= pstate->r_hwpipe->flush_mask;
|
|
|
|
|
|
|
|
return mask;
|
2014-11-18 12:49:49 -05:00
|
|
|
}
|
|
|
|
|
2024-04-20 07:00:59 +03:00
|
|
|
static const uint32_t mdp5_plane_formats[] = {
|
|
|
|
DRM_FORMAT_ARGB8888,
|
|
|
|
DRM_FORMAT_ABGR8888,
|
|
|
|
DRM_FORMAT_RGBA8888,
|
|
|
|
DRM_FORMAT_BGRA8888,
|
|
|
|
DRM_FORMAT_XRGB8888,
|
|
|
|
DRM_FORMAT_XBGR8888,
|
|
|
|
DRM_FORMAT_RGBX8888,
|
|
|
|
DRM_FORMAT_BGRX8888,
|
|
|
|
DRM_FORMAT_RGB888,
|
|
|
|
DRM_FORMAT_BGR888,
|
|
|
|
DRM_FORMAT_RGB565,
|
|
|
|
DRM_FORMAT_BGR565,
|
|
|
|
|
|
|
|
DRM_FORMAT_NV12,
|
|
|
|
DRM_FORMAT_NV21,
|
|
|
|
DRM_FORMAT_NV16,
|
|
|
|
DRM_FORMAT_NV61,
|
|
|
|
DRM_FORMAT_VYUY,
|
|
|
|
DRM_FORMAT_UYVY,
|
|
|
|
DRM_FORMAT_YUYV,
|
|
|
|
DRM_FORMAT_YVYU,
|
|
|
|
DRM_FORMAT_YUV420,
|
|
|
|
DRM_FORMAT_YVU420,
|
|
|
|
};
|
|
|
|
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
/* initialize plane */
|
2017-01-16 11:57:04 +05:30
|
|
|
struct drm_plane *mdp5_plane_init(struct drm_device *dev,
|
|
|
|
enum drm_plane_type type)
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
{
|
|
|
|
struct drm_plane *plane = NULL;
|
|
|
|
struct mdp5_plane *mdp5_plane;
|
|
|
|
|
2023-07-08 04:04:00 +03:00
|
|
|
mdp5_plane = drmm_universal_plane_alloc(dev, struct mdp5_plane, base,
|
|
|
|
0xff, &mdp5_plane_funcs,
|
|
|
|
mdp5_plane_formats, ARRAY_SIZE(mdp5_plane_formats),
|
|
|
|
NULL, type, NULL);
|
|
|
|
if (IS_ERR(mdp5_plane))
|
|
|
|
return ERR_CAST(mdp5_plane);
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
|
|
|
plane = &mdp5_plane->base;
|
|
|
|
|
2014-11-19 12:31:03 -05:00
|
|
|
drm_plane_helper_add(plane, &mdp5_plane_helper_funcs);
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
|
|
|
|
mdp5_plane_install_properties(plane, &plane->base);
|
|
|
|
|
2019-05-31 05:46:14 -04:00
|
|
|
drm_plane_enable_fb_damage_clips(plane);
|
|
|
|
|
drm/msm: add mdp5/apq8x74
Add support for the new MDP5 display controller block. The mapping
between parts of the display controller and KMS is:
plane -> PIPE{RGBn,VIGn} \
crtc -> LM (layer mixer) |-> MDP "device"
encoder -> INTF /
connector -> HDMI/DSI/eDP/etc --> other device(s)
Unlike MDP4, it appears we can get by with a single encoder, rather
than needing a different implementation for DTV, DSI, etc. (Ie. the
register interface is same, just different bases.)
Also unlike MDP4, all the IRQs for other blocks (HDMI, DSI, etc) are
routed through MDP.
And finally, MDP5 has this "Shared Memory Pool" (called "SMP"), from
which blocks need to be allocated to the active pipes based on fetch
stride.
Signed-off-by: Rob Clark <robdclark@gmail.com>
2013-11-30 17:51:47 -05:00
|
|
|
return plane;
|
|
|
|
}
|