2019-06-03 23:23:31 +08:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
|
|
|
//
|
|
|
|
// Ingenic JZ47xx KMS driver
|
|
|
|
//
|
|
|
|
// Copyright (C) 2019, Paul Cercueil <paul@crapouillou.net>
|
|
|
|
|
2020-07-17 00:38:41 +08:00
|
|
|
#include "ingenic-drm.h"
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
#include <linux/component.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <linux/clk.h>
|
|
|
|
#include <linux/dma-mapping.h>
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
#include <linux/dma-noncoherent.h>
|
|
|
|
#include <linux/io.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <linux/module.h>
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
#include <linux/mutex.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <linux/of_device.h>
|
2020-09-27 01:04:56 +08:00
|
|
|
#include <linux/of_reserved_mem.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <linux/platform_device.h>
|
|
|
|
#include <linux/regmap.h>
|
|
|
|
|
|
|
|
#include <drm/drm_atomic.h>
|
|
|
|
#include <drm/drm_atomic_helper.h>
|
2019-08-26 23:26:29 +08:00
|
|
|
#include <drm/drm_bridge.h>
|
2020-09-28 03:36:45 +08:00
|
|
|
#include <drm/drm_color_mgmt.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <drm/drm_crtc.h>
|
|
|
|
#include <drm/drm_crtc_helper.h>
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
#include <drm/drm_damage_helper.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <drm/drm_drv.h>
|
|
|
|
#include <drm/drm_gem_cma_helper.h>
|
|
|
|
#include <drm/drm_fb_cma_helper.h>
|
|
|
|
#include <drm/drm_fb_helper.h>
|
|
|
|
#include <drm/drm_fourcc.h>
|
|
|
|
#include <drm/drm_gem_framebuffer_helper.h>
|
|
|
|
#include <drm/drm_irq.h>
|
2020-03-23 22:49:15 +08:00
|
|
|
#include <drm/drm_managed.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <drm/drm_of.h>
|
|
|
|
#include <drm/drm_panel.h>
|
|
|
|
#include <drm/drm_plane.h>
|
|
|
|
#include <drm/drm_plane_helper.h>
|
|
|
|
#include <drm/drm_probe_helper.h>
|
2020-03-05 23:59:37 +08:00
|
|
|
#include <drm/drm_simple_kms_helper.h>
|
2019-06-03 23:23:31 +08:00
|
|
|
#include <drm/drm_vblank.h>
|
|
|
|
|
|
|
|
struct ingenic_dma_hwdesc {
|
|
|
|
u32 next;
|
|
|
|
u32 addr;
|
|
|
|
u32 id;
|
|
|
|
u32 cmd;
|
2020-09-27 01:04:57 +08:00
|
|
|
} __aligned(16);
|
|
|
|
|
|
|
|
struct ingenic_dma_hwdescs {
|
|
|
|
struct ingenic_dma_hwdesc hwdesc_f0;
|
|
|
|
struct ingenic_dma_hwdesc hwdesc_f1;
|
2020-09-28 03:36:45 +08:00
|
|
|
struct ingenic_dma_hwdesc hwdesc_pal;
|
|
|
|
u16 palette[256] __aligned(16);
|
2020-09-27 01:04:57 +08:00
|
|
|
};
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
struct jz_soc_info {
|
|
|
|
bool needs_dev_clk;
|
2020-07-17 00:38:43 +08:00
|
|
|
bool has_osd;
|
2019-12-10 22:41:41 +08:00
|
|
|
unsigned int max_width, max_height;
|
2020-09-27 01:04:58 +08:00
|
|
|
const u32 *formats_f0, *formats_f1;
|
|
|
|
unsigned int num_formats_f0, num_formats_f1;
|
2019-06-03 23:23:31 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
struct ingenic_drm {
|
|
|
|
struct drm_device drm;
|
2020-07-17 00:38:43 +08:00
|
|
|
/*
|
|
|
|
* f1 (aka. foreground1) is our primary plane, on top of which
|
|
|
|
* f0 (aka. foreground0) can be overlayed. Z-order is fixed in
|
|
|
|
* hardware and cannot be changed.
|
|
|
|
*/
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
struct drm_plane f0, f1, *ipu_plane;
|
2019-06-03 23:23:31 +08:00
|
|
|
struct drm_crtc crtc;
|
|
|
|
|
|
|
|
struct device *dev;
|
|
|
|
struct regmap *map;
|
|
|
|
struct clk *lcd_clk, *pix_clk;
|
2019-12-10 22:41:41 +08:00
|
|
|
const struct jz_soc_info *soc_info;
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2020-09-27 01:04:57 +08:00
|
|
|
struct ingenic_dma_hwdescs *dma_hwdescs;
|
|
|
|
dma_addr_t dma_hwdescs_phys;
|
2019-06-28 02:21:13 +08:00
|
|
|
|
|
|
|
bool panel_is_sharp;
|
2020-07-17 00:38:43 +08:00
|
|
|
bool no_vblank;
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* clk_mutex is used to synchronize the pixel clock rate update with
|
|
|
|
* the VBLANK. When the pixel clock's parent clock needs to be updated,
|
|
|
|
* clock_nb's notifier function will lock the mutex, then wait until the
|
|
|
|
* next VBLANK. At that point, the parent clock's rate can be updated,
|
|
|
|
* and the mutex is then unlocked. If an atomic commit happens in the
|
|
|
|
* meantime, it will lock on the mutex, effectively waiting until the
|
|
|
|
* clock update process finishes. Finally, the pixel clock's rate will
|
|
|
|
* be recomputed when the mutex has been released, in the pending atomic
|
|
|
|
* commit, or a future one.
|
|
|
|
*/
|
|
|
|
struct mutex clk_mutex;
|
|
|
|
bool update_clk_rate;
|
|
|
|
struct notifier_block clock_nb;
|
2019-06-03 23:23:31 +08:00
|
|
|
};
|
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
static bool ingenic_drm_cached_gem_buf;
|
|
|
|
module_param_named(cached_gem_buffers, ingenic_drm_cached_gem_buf, bool, 0400);
|
|
|
|
MODULE_PARM_DESC(cached_gem_buffers,
|
|
|
|
"Enable fully cached GEM buffers [default=false]");
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static bool ingenic_drm_writeable_reg(struct device *dev, unsigned int reg)
|
|
|
|
{
|
|
|
|
switch (reg) {
|
|
|
|
case JZ_REG_LCD_IID:
|
|
|
|
case JZ_REG_LCD_SA0:
|
|
|
|
case JZ_REG_LCD_FID0:
|
|
|
|
case JZ_REG_LCD_CMD0:
|
|
|
|
case JZ_REG_LCD_SA1:
|
|
|
|
case JZ_REG_LCD_FID1:
|
|
|
|
case JZ_REG_LCD_CMD1:
|
|
|
|
return false;
|
|
|
|
default:
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct regmap_config ingenic_drm_regmap_config = {
|
|
|
|
.reg_bits = 32,
|
|
|
|
.val_bits = 32,
|
|
|
|
.reg_stride = 4,
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
.max_register = JZ_REG_LCD_SIZE1,
|
2019-06-03 23:23:31 +08:00
|
|
|
.writeable_reg = ingenic_drm_writeable_reg,
|
|
|
|
};
|
|
|
|
|
|
|
|
static inline struct ingenic_drm *drm_device_get_priv(struct drm_device *drm)
|
|
|
|
{
|
|
|
|
return container_of(drm, struct ingenic_drm, drm);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct ingenic_drm *drm_crtc_get_priv(struct drm_crtc *crtc)
|
|
|
|
{
|
|
|
|
return container_of(crtc, struct ingenic_drm, crtc);
|
|
|
|
}
|
|
|
|
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
static inline struct ingenic_drm *drm_nb_get_priv(struct notifier_block *nb)
|
|
|
|
{
|
|
|
|
return container_of(nb, struct ingenic_drm, clock_nb);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_update_pixclk(struct notifier_block *nb,
|
|
|
|
unsigned long action,
|
|
|
|
void *data)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_nb_get_priv(nb);
|
|
|
|
|
|
|
|
switch (action) {
|
|
|
|
case PRE_RATE_CHANGE:
|
|
|
|
mutex_lock(&priv->clk_mutex);
|
|
|
|
priv->update_clk_rate = true;
|
|
|
|
drm_crtc_wait_one_vblank(&priv->crtc);
|
|
|
|
return NOTIFY_OK;
|
|
|
|
default:
|
|
|
|
mutex_unlock(&priv->clk_mutex);
|
|
|
|
return NOTIFY_OK;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static void ingenic_drm_crtc_atomic_enable(struct drm_crtc *crtc,
|
|
|
|
struct drm_crtc_state *state)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_STATE, 0);
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_CTRL,
|
|
|
|
JZ_LCD_CTRL_ENABLE | JZ_LCD_CTRL_DISABLE,
|
|
|
|
JZ_LCD_CTRL_ENABLE);
|
|
|
|
|
|
|
|
drm_crtc_vblank_on(crtc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ingenic_drm_crtc_atomic_disable(struct drm_crtc *crtc,
|
|
|
|
struct drm_crtc_state *state)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
unsigned int var;
|
|
|
|
|
|
|
|
drm_crtc_vblank_off(crtc);
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_CTRL,
|
|
|
|
JZ_LCD_CTRL_DISABLE, JZ_LCD_CTRL_DISABLE);
|
|
|
|
|
|
|
|
regmap_read_poll_timeout(priv->map, JZ_REG_LCD_STATE, var,
|
|
|
|
var & JZ_LCD_STATE_DISABLED,
|
|
|
|
1000, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ingenic_drm_crtc_update_timings(struct ingenic_drm *priv,
|
|
|
|
struct drm_display_mode *mode)
|
|
|
|
{
|
|
|
|
unsigned int vpe, vds, vde, vt, hpe, hds, hde, ht;
|
|
|
|
|
|
|
|
vpe = mode->vsync_end - mode->vsync_start;
|
|
|
|
vds = mode->vtotal - mode->vsync_start;
|
|
|
|
vde = vds + mode->vdisplay;
|
|
|
|
vt = vde + mode->vsync_start - mode->vdisplay;
|
|
|
|
|
|
|
|
hpe = mode->hsync_end - mode->hsync_start;
|
|
|
|
hds = mode->htotal - mode->hsync_start;
|
|
|
|
hde = hds + mode->hdisplay;
|
|
|
|
ht = hde + mode->hsync_start - mode->hdisplay;
|
|
|
|
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_VSYNC,
|
|
|
|
0 << JZ_LCD_VSYNC_VPS_OFFSET |
|
|
|
|
vpe << JZ_LCD_VSYNC_VPE_OFFSET);
|
|
|
|
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_HSYNC,
|
|
|
|
0 << JZ_LCD_HSYNC_HPS_OFFSET |
|
|
|
|
hpe << JZ_LCD_HSYNC_HPE_OFFSET);
|
|
|
|
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_VAT,
|
|
|
|
ht << JZ_LCD_VAT_HT_OFFSET |
|
|
|
|
vt << JZ_LCD_VAT_VT_OFFSET);
|
|
|
|
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_DAH,
|
|
|
|
hds << JZ_LCD_DAH_HDS_OFFSET |
|
|
|
|
hde << JZ_LCD_DAH_HDE_OFFSET);
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_DAV,
|
|
|
|
vds << JZ_LCD_DAV_VDS_OFFSET |
|
|
|
|
vde << JZ_LCD_DAV_VDE_OFFSET);
|
2019-06-28 02:21:13 +08:00
|
|
|
|
|
|
|
if (priv->panel_is_sharp) {
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_PS, hde << 16 | (hde + 1));
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_CLS, hde << 16 | (hde + 1));
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_SPL, hpe << 16 | (hpe + 1));
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_REV, mode->htotal << 16);
|
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
regmap_set_bits(priv->map, JZ_REG_LCD_CTRL,
|
|
|
|
JZ_LCD_CTRL_OFUP | JZ_LCD_CTRL_BURST_16);
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* IPU restart - specify how much time the LCDC will wait before
|
|
|
|
* transferring a new frame from the IPU. The value is the one
|
|
|
|
* suggested in the programming manual.
|
|
|
|
*/
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_IPUR, JZ_LCD_IPUR_IPUREN |
|
|
|
|
(ht * vpe / 3) << JZ_LCD_IPUR_IPUR_LSB);
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_crtc_atomic_check(struct drm_crtc *crtc,
|
|
|
|
struct drm_crtc_state *state)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
2020-07-19 17:38:34 +08:00
|
|
|
struct drm_plane_state *f1_state, *f0_state, *ipu_state = NULL;
|
2019-12-10 22:41:41 +08:00
|
|
|
|
2020-09-28 03:36:45 +08:00
|
|
|
if (state->gamma_lut &&
|
|
|
|
drm_color_lut_size(state->gamma_lut) != ARRAY_SIZE(priv->dma_hwdescs->palette)) {
|
|
|
|
dev_dbg(priv->dev, "Invalid palette size\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2020-07-28 23:16:41 +08:00
|
|
|
if (drm_atomic_crtc_needs_modeset(state) && priv->soc_info->has_osd) {
|
2020-07-17 00:38:43 +08:00
|
|
|
f1_state = drm_atomic_get_plane_state(state->state, &priv->f1);
|
2020-07-28 23:16:40 +08:00
|
|
|
if (IS_ERR(f1_state))
|
|
|
|
return PTR_ERR(f1_state);
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
f0_state = drm_atomic_get_plane_state(state->state, &priv->f0);
|
2020-07-28 23:16:40 +08:00
|
|
|
if (IS_ERR(f0_state))
|
|
|
|
return PTR_ERR(f0_state);
|
2020-07-17 00:38:43 +08:00
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
if (IS_ENABLED(CONFIG_DRM_INGENIC_IPU) && priv->ipu_plane) {
|
|
|
|
ipu_state = drm_atomic_get_plane_state(state->state, priv->ipu_plane);
|
2020-07-28 23:16:40 +08:00
|
|
|
if (IS_ERR(ipu_state))
|
|
|
|
return PTR_ERR(ipu_state);
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
|
|
|
/* IPU and F1 planes cannot be enabled at the same time. */
|
|
|
|
if (f1_state->fb && ipu_state->fb) {
|
|
|
|
dev_dbg(priv->dev, "Cannot enable both F1 and IPU\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
/* If all the planes are disabled, we won't get a VBLANK IRQ */
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
priv->no_vblank = !f1_state->fb && !f0_state->fb &&
|
2020-07-19 17:38:34 +08:00
|
|
|
!(ipu_state && ipu_state->fb);
|
2020-07-17 00:38:43 +08:00
|
|
|
}
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2020-07-28 23:16:41 +08:00
|
|
|
static enum drm_mode_status
|
|
|
|
ingenic_drm_crtc_mode_valid(struct drm_crtc *crtc, const struct drm_display_mode *mode)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
long rate;
|
|
|
|
|
|
|
|
if (mode->hdisplay > priv->soc_info->max_width)
|
|
|
|
return MODE_BAD_HVALUE;
|
|
|
|
if (mode->vdisplay > priv->soc_info->max_height)
|
|
|
|
return MODE_BAD_VVALUE;
|
|
|
|
|
|
|
|
rate = clk_round_rate(priv->pix_clk, mode->clock * 1000);
|
|
|
|
if (rate < 0)
|
|
|
|
return MODE_CLOCK_RANGE;
|
|
|
|
|
|
|
|
return MODE_OK;
|
|
|
|
}
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
static void ingenic_drm_crtc_atomic_begin(struct drm_crtc *crtc,
|
|
|
|
struct drm_crtc_state *oldstate)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
u32 ctrl = 0;
|
|
|
|
|
|
|
|
if (priv->soc_info->has_osd &&
|
|
|
|
drm_atomic_crtc_needs_modeset(crtc->state)) {
|
|
|
|
/*
|
|
|
|
* If IPU plane is enabled, enable IPU as source for the F1
|
|
|
|
* plane; otherwise use regular DMA.
|
|
|
|
*/
|
|
|
|
if (priv->ipu_plane && priv->ipu_plane->state->fb)
|
|
|
|
ctrl |= JZ_LCD_OSDCTRL_IPU;
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_OSDCTRL,
|
|
|
|
JZ_LCD_OSDCTRL_IPU, ctrl);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static void ingenic_drm_crtc_atomic_flush(struct drm_crtc *crtc,
|
|
|
|
struct drm_crtc_state *oldstate)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
struct drm_crtc_state *state = crtc->state;
|
|
|
|
struct drm_pending_vblank_event *event = state->event;
|
|
|
|
|
|
|
|
if (drm_atomic_crtc_needs_modeset(state)) {
|
|
|
|
ingenic_drm_crtc_update_timings(priv, &state->mode);
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
priv->update_clk_rate = true;
|
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
if (priv->update_clk_rate) {
|
|
|
|
mutex_lock(&priv->clk_mutex);
|
2019-06-03 23:23:31 +08:00
|
|
|
clk_set_rate(priv->pix_clk, state->adjusted_mode.clock * 1000);
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
priv->update_clk_rate = false;
|
|
|
|
mutex_unlock(&priv->clk_mutex);
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (event) {
|
|
|
|
state->event = NULL;
|
|
|
|
|
|
|
|
spin_lock_irq(&crtc->dev->event_lock);
|
|
|
|
if (drm_crtc_vblank_get(crtc) == 0)
|
|
|
|
drm_crtc_arm_vblank_event(crtc, event);
|
|
|
|
else
|
|
|
|
drm_crtc_send_vblank_event(crtc, event);
|
|
|
|
spin_unlock_irq(&crtc->dev->event_lock);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
static int ingenic_drm_plane_atomic_check(struct drm_plane *plane,
|
|
|
|
struct drm_plane_state *state)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_device_get_priv(plane->dev);
|
|
|
|
struct drm_crtc_state *crtc_state;
|
|
|
|
struct drm_crtc *crtc = state->crtc ?: plane->state->crtc;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!crtc)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
crtc_state = drm_atomic_get_existing_crtc_state(state->state, crtc);
|
|
|
|
if (WARN_ON(!crtc_state))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
ret = drm_atomic_helper_check_plane_state(state, crtc_state,
|
|
|
|
DRM_PLANE_HELPER_NO_SCALING,
|
|
|
|
DRM_PLANE_HELPER_NO_SCALING,
|
|
|
|
priv->soc_info->has_osd,
|
|
|
|
true);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If OSD is not available, check that the width/height match.
|
|
|
|
* Note that state->src_* are in 16.16 fixed-point format.
|
|
|
|
*/
|
|
|
|
if (!priv->soc_info->has_osd &&
|
|
|
|
(state->src_x != 0 ||
|
|
|
|
(state->src_w >> 16) != state->crtc_w ||
|
|
|
|
(state->src_h >> 16) != state->crtc_h))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Require full modeset if enabling or disabling a plane, or changing
|
|
|
|
* its position, size or depth.
|
|
|
|
*/
|
|
|
|
if (priv->soc_info->has_osd &&
|
|
|
|
(!plane->state->fb || !state->fb ||
|
|
|
|
plane->state->crtc_x != state->crtc_x ||
|
|
|
|
plane->state->crtc_y != state->crtc_y ||
|
|
|
|
plane->state->crtc_w != state->crtc_w ||
|
|
|
|
plane->state->crtc_h != state->crtc_h ||
|
|
|
|
plane->state->fb->format->format != state->fb->format->format))
|
|
|
|
crtc_state->mode_changed = true;
|
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
drm_atomic_helper_check_plane_damage(state->state, state);
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ingenic_drm_plane_enable(struct ingenic_drm *priv,
|
|
|
|
struct drm_plane *plane)
|
|
|
|
{
|
|
|
|
unsigned int en_bit;
|
|
|
|
|
|
|
|
if (priv->soc_info->has_osd) {
|
|
|
|
if (plane->type == DRM_PLANE_TYPE_PRIMARY)
|
|
|
|
en_bit = JZ_LCD_OSDC_F1EN;
|
|
|
|
else
|
|
|
|
en_bit = JZ_LCD_OSDC_F0EN;
|
|
|
|
|
|
|
|
regmap_set_bits(priv->map, JZ_REG_LCD_OSDC, en_bit);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
void ingenic_drm_plane_disable(struct device *dev, struct drm_plane *plane)
|
2020-07-17 00:38:43 +08:00
|
|
|
{
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
struct ingenic_drm *priv = dev_get_drvdata(dev);
|
2020-07-17 00:38:43 +08:00
|
|
|
unsigned int en_bit;
|
|
|
|
|
|
|
|
if (priv->soc_info->has_osd) {
|
|
|
|
if (plane->type == DRM_PLANE_TYPE_PRIMARY)
|
|
|
|
en_bit = JZ_LCD_OSDC_F1EN;
|
|
|
|
else
|
|
|
|
en_bit = JZ_LCD_OSDC_F0EN;
|
|
|
|
|
|
|
|
regmap_clear_bits(priv->map, JZ_REG_LCD_OSDC, en_bit);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
static void ingenic_drm_plane_atomic_disable(struct drm_plane *plane,
|
|
|
|
struct drm_plane_state *old_state)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_device_get_priv(plane->dev);
|
|
|
|
|
|
|
|
ingenic_drm_plane_disable(priv->dev, plane);
|
|
|
|
}
|
|
|
|
|
|
|
|
void ingenic_drm_plane_config(struct device *dev,
|
|
|
|
struct drm_plane *plane, u32 fourcc)
|
2020-07-17 00:38:43 +08:00
|
|
|
{
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
struct ingenic_drm *priv = dev_get_drvdata(dev);
|
2020-07-17 00:38:43 +08:00
|
|
|
struct drm_plane_state *state = plane->state;
|
|
|
|
unsigned int xy_reg, size_reg;
|
|
|
|
unsigned int ctrl = 0;
|
|
|
|
|
|
|
|
ingenic_drm_plane_enable(priv, plane);
|
|
|
|
|
|
|
|
if (priv->soc_info->has_osd &&
|
|
|
|
plane->type == DRM_PLANE_TYPE_PRIMARY) {
|
|
|
|
switch (fourcc) {
|
|
|
|
case DRM_FORMAT_XRGB1555:
|
|
|
|
ctrl |= JZ_LCD_OSDCTRL_RGB555;
|
|
|
|
fallthrough;
|
|
|
|
case DRM_FORMAT_RGB565:
|
|
|
|
ctrl |= JZ_LCD_OSDCTRL_BPP_15_16;
|
|
|
|
break;
|
2020-09-27 01:05:01 +08:00
|
|
|
case DRM_FORMAT_RGB888:
|
|
|
|
ctrl |= JZ_LCD_OSDCTRL_BPP_24_COMP;
|
|
|
|
break;
|
2020-07-17 00:38:43 +08:00
|
|
|
case DRM_FORMAT_XRGB8888:
|
|
|
|
ctrl |= JZ_LCD_OSDCTRL_BPP_18_24;
|
|
|
|
break;
|
2020-09-27 01:05:00 +08:00
|
|
|
case DRM_FORMAT_XRGB2101010:
|
|
|
|
ctrl |= JZ_LCD_OSDCTRL_BPP_30;
|
|
|
|
break;
|
2020-07-17 00:38:43 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_OSDCTRL,
|
|
|
|
JZ_LCD_OSDCTRL_BPP_MASK, ctrl);
|
|
|
|
} else {
|
|
|
|
switch (fourcc) {
|
2020-09-28 03:36:45 +08:00
|
|
|
case DRM_FORMAT_C8:
|
|
|
|
ctrl |= JZ_LCD_CTRL_BPP_8;
|
|
|
|
break;
|
2020-07-17 00:38:43 +08:00
|
|
|
case DRM_FORMAT_XRGB1555:
|
|
|
|
ctrl |= JZ_LCD_CTRL_RGB555;
|
|
|
|
fallthrough;
|
|
|
|
case DRM_FORMAT_RGB565:
|
|
|
|
ctrl |= JZ_LCD_CTRL_BPP_15_16;
|
|
|
|
break;
|
2020-09-27 01:05:01 +08:00
|
|
|
case DRM_FORMAT_RGB888:
|
|
|
|
ctrl |= JZ_LCD_CTRL_BPP_24_COMP;
|
|
|
|
break;
|
2020-07-17 00:38:43 +08:00
|
|
|
case DRM_FORMAT_XRGB8888:
|
|
|
|
ctrl |= JZ_LCD_CTRL_BPP_18_24;
|
|
|
|
break;
|
2020-09-27 01:05:00 +08:00
|
|
|
case DRM_FORMAT_XRGB2101010:
|
|
|
|
ctrl |= JZ_LCD_CTRL_BPP_30;
|
|
|
|
break;
|
2020-07-17 00:38:43 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_CTRL,
|
|
|
|
JZ_LCD_CTRL_BPP_MASK, ctrl);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (priv->soc_info->has_osd) {
|
|
|
|
if (plane->type == DRM_PLANE_TYPE_PRIMARY) {
|
|
|
|
xy_reg = JZ_REG_LCD_XYP1;
|
|
|
|
size_reg = JZ_REG_LCD_SIZE1;
|
|
|
|
} else {
|
|
|
|
xy_reg = JZ_REG_LCD_XYP0;
|
|
|
|
size_reg = JZ_REG_LCD_SIZE0;
|
|
|
|
}
|
|
|
|
|
|
|
|
regmap_write(priv->map, xy_reg,
|
|
|
|
state->crtc_x << JZ_LCD_XYP01_XPOS_LSB |
|
|
|
|
state->crtc_y << JZ_LCD_XYP01_YPOS_LSB);
|
|
|
|
regmap_write(priv->map, size_reg,
|
|
|
|
state->crtc_w << JZ_LCD_SIZE01_WIDTH_LSB |
|
|
|
|
state->crtc_h << JZ_LCD_SIZE01_HEIGHT_LSB);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
void ingenic_drm_sync_data(struct device *dev,
|
|
|
|
struct drm_plane_state *old_state,
|
|
|
|
struct drm_plane_state *state)
|
|
|
|
{
|
|
|
|
const struct drm_format_info *finfo = state->fb->format;
|
|
|
|
struct ingenic_drm *priv = dev_get_drvdata(dev);
|
|
|
|
struct drm_atomic_helper_damage_iter iter;
|
|
|
|
unsigned int offset, i;
|
|
|
|
struct drm_rect clip;
|
|
|
|
dma_addr_t paddr;
|
|
|
|
void *addr;
|
|
|
|
|
|
|
|
if (!ingenic_drm_cached_gem_buf)
|
|
|
|
return;
|
|
|
|
|
|
|
|
drm_atomic_helper_damage_iter_init(&iter, old_state, state);
|
|
|
|
|
|
|
|
drm_atomic_for_each_plane_damage(&iter, &clip) {
|
|
|
|
for (i = 0; i < finfo->num_planes; i++) {
|
|
|
|
paddr = drm_fb_cma_get_gem_addr(state->fb, state, i);
|
|
|
|
addr = phys_to_virt(paddr);
|
|
|
|
|
|
|
|
/* Ignore x1/x2 values, invalidate complete lines */
|
|
|
|
offset = clip.y1 * state->fb->pitches[i];
|
|
|
|
|
|
|
|
dma_cache_sync(priv->dev, addr + offset,
|
|
|
|
(clip.y2 - clip.y1) * state->fb->pitches[i],
|
|
|
|
DMA_TO_DEVICE);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-09-28 03:36:45 +08:00
|
|
|
static void ingenic_drm_update_palette(struct ingenic_drm *priv,
|
|
|
|
const struct drm_color_lut *lut)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(priv->dma_hwdescs->palette); i++) {
|
|
|
|
u16 color = drm_color_lut_extract(lut[i].red, 5) << 11
|
|
|
|
| drm_color_lut_extract(lut[i].green, 6) << 5
|
|
|
|
| drm_color_lut_extract(lut[i].blue, 5);
|
|
|
|
|
|
|
|
priv->dma_hwdescs->palette[i] = color;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static void ingenic_drm_plane_atomic_update(struct drm_plane *plane,
|
|
|
|
struct drm_plane_state *oldstate)
|
|
|
|
{
|
2020-07-17 00:38:43 +08:00
|
|
|
struct ingenic_drm *priv = drm_device_get_priv(plane->dev);
|
2019-06-03 23:23:31 +08:00
|
|
|
struct drm_plane_state *state = plane->state;
|
2020-09-28 03:36:45 +08:00
|
|
|
struct drm_crtc_state *crtc_state;
|
2020-07-17 00:38:43 +08:00
|
|
|
struct ingenic_dma_hwdesc *hwdesc;
|
2020-09-28 03:36:45 +08:00
|
|
|
unsigned int width, height, cpp, offset;
|
2019-12-10 22:41:38 +08:00
|
|
|
dma_addr_t addr;
|
2020-09-28 03:36:45 +08:00
|
|
|
u32 fourcc;
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2019-12-10 22:41:38 +08:00
|
|
|
if (state && state->fb) {
|
2020-09-28 03:36:45 +08:00
|
|
|
crtc_state = state->crtc->state;
|
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
ingenic_drm_sync_data(priv->dev, oldstate, state);
|
|
|
|
|
2019-12-10 22:41:38 +08:00
|
|
|
addr = drm_fb_cma_get_gem_addr(state->fb, state, 0);
|
2019-12-10 22:41:39 +08:00
|
|
|
width = state->src_w >> 16;
|
|
|
|
height = state->src_h >> 16;
|
2020-07-17 00:38:35 +08:00
|
|
|
cpp = state->fb->format->cpp[0];
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
if (priv->soc_info->has_osd && plane->type == DRM_PLANE_TYPE_OVERLAY)
|
2020-09-27 01:04:57 +08:00
|
|
|
hwdesc = &priv->dma_hwdescs->hwdesc_f0;
|
2020-07-17 00:38:43 +08:00
|
|
|
else
|
2020-09-27 01:04:57 +08:00
|
|
|
hwdesc = &priv->dma_hwdescs->hwdesc_f1;
|
2020-07-17 00:38:43 +08:00
|
|
|
|
|
|
|
hwdesc->addr = addr;
|
|
|
|
hwdesc->cmd = JZ_LCD_CMD_EOF_IRQ | (width * height * cpp / 4);
|
|
|
|
|
2020-09-28 03:36:45 +08:00
|
|
|
if (drm_atomic_crtc_needs_modeset(crtc_state)) {
|
|
|
|
fourcc = state->fb->format->format;
|
|
|
|
|
|
|
|
ingenic_drm_plane_config(priv->dev, plane, fourcc);
|
|
|
|
|
|
|
|
if (fourcc == DRM_FORMAT_C8)
|
|
|
|
offset = offsetof(struct ingenic_dma_hwdescs, hwdesc_pal);
|
|
|
|
else
|
|
|
|
offset = offsetof(struct ingenic_dma_hwdescs, hwdesc_f0);
|
|
|
|
|
|
|
|
priv->dma_hwdescs->hwdesc_f0.next = priv->dma_hwdescs_phys + offset;
|
|
|
|
|
|
|
|
crtc_state->color_mgmt_changed = fourcc == DRM_FORMAT_C8;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (crtc_state->color_mgmt_changed)
|
|
|
|
ingenic_drm_update_palette(priv, crtc_state->gamma_lut->data);
|
2019-12-10 22:41:38 +08:00
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void ingenic_drm_encoder_atomic_mode_set(struct drm_encoder *encoder,
|
|
|
|
struct drm_crtc_state *crtc_state,
|
|
|
|
struct drm_connector_state *conn_state)
|
|
|
|
{
|
2020-07-17 00:38:45 +08:00
|
|
|
struct ingenic_drm *priv = drm_device_get_priv(encoder->dev);
|
2019-06-03 23:23:31 +08:00
|
|
|
struct drm_display_mode *mode = &crtc_state->adjusted_mode;
|
2019-06-28 02:21:13 +08:00
|
|
|
struct drm_connector *conn = conn_state->connector;
|
|
|
|
struct drm_display_info *info = &conn->display_info;
|
|
|
|
unsigned int cfg;
|
|
|
|
|
|
|
|
priv->panel_is_sharp = info->bus_flags & DRM_BUS_FLAG_SHARP_SIGNALS;
|
|
|
|
|
|
|
|
if (priv->panel_is_sharp) {
|
|
|
|
cfg = JZ_LCD_CFG_MODE_SPECIAL_TFT_1 | JZ_LCD_CFG_REV_POLARITY;
|
|
|
|
} else {
|
|
|
|
cfg = JZ_LCD_CFG_PS_DISABLE | JZ_LCD_CFG_CLS_DISABLE
|
|
|
|
| JZ_LCD_CFG_SPL_DISABLE | JZ_LCD_CFG_REV_DISABLE;
|
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
if (mode->flags & DRM_MODE_FLAG_NHSYNC)
|
|
|
|
cfg |= JZ_LCD_CFG_HSYNC_ACTIVE_LOW;
|
|
|
|
if (mode->flags & DRM_MODE_FLAG_NVSYNC)
|
|
|
|
cfg |= JZ_LCD_CFG_VSYNC_ACTIVE_LOW;
|
|
|
|
if (info->bus_flags & DRM_BUS_FLAG_DE_LOW)
|
|
|
|
cfg |= JZ_LCD_CFG_DE_ACTIVE_LOW;
|
2020-07-01 02:05:40 +08:00
|
|
|
if (info->bus_flags & DRM_BUS_FLAG_PIXDATA_DRIVE_NEGEDGE)
|
2019-06-03 23:23:31 +08:00
|
|
|
cfg |= JZ_LCD_CFG_PCLK_FALLING_EDGE;
|
|
|
|
|
2019-06-28 02:21:13 +08:00
|
|
|
if (!priv->panel_is_sharp) {
|
|
|
|
if (conn->connector_type == DRM_MODE_CONNECTOR_TV) {
|
|
|
|
if (mode->flags & DRM_MODE_FLAG_INTERLACE)
|
|
|
|
cfg |= JZ_LCD_CFG_MODE_TV_OUT_I;
|
|
|
|
else
|
|
|
|
cfg |= JZ_LCD_CFG_MODE_TV_OUT_P;
|
|
|
|
} else {
|
|
|
|
switch (*info->bus_formats) {
|
|
|
|
case MEDIA_BUS_FMT_RGB565_1X16:
|
|
|
|
cfg |= JZ_LCD_CFG_MODE_GENERIC_16BIT;
|
|
|
|
break;
|
|
|
|
case MEDIA_BUS_FMT_RGB666_1X18:
|
|
|
|
cfg |= JZ_LCD_CFG_MODE_GENERIC_18BIT;
|
|
|
|
break;
|
|
|
|
case MEDIA_BUS_FMT_RGB888_1X24:
|
|
|
|
cfg |= JZ_LCD_CFG_MODE_GENERIC_24BIT;
|
|
|
|
break;
|
2019-06-28 02:21:14 +08:00
|
|
|
case MEDIA_BUS_FMT_RGB888_3X8:
|
|
|
|
cfg |= JZ_LCD_CFG_MODE_8BIT_SERIAL;
|
|
|
|
break;
|
2019-06-28 02:21:13 +08:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_CFG, cfg);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_encoder_atomic_check(struct drm_encoder *encoder,
|
|
|
|
struct drm_crtc_state *crtc_state,
|
|
|
|
struct drm_connector_state *conn_state)
|
|
|
|
{
|
|
|
|
struct drm_display_info *info = &conn_state->connector->display_info;
|
|
|
|
|
|
|
|
if (info->num_bus_formats != 1)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (conn_state->connector->connector_type == DRM_MODE_CONNECTOR_TV)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
switch (*info->bus_formats) {
|
|
|
|
case MEDIA_BUS_FMT_RGB565_1X16:
|
|
|
|
case MEDIA_BUS_FMT_RGB666_1X18:
|
|
|
|
case MEDIA_BUS_FMT_RGB888_1X24:
|
2019-06-28 02:21:14 +08:00
|
|
|
case MEDIA_BUS_FMT_RGB888_3X8:
|
2019-06-03 23:23:31 +08:00
|
|
|
return 0;
|
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
static void ingenic_drm_atomic_helper_commit_tail(struct drm_atomic_state *old_state)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Just your regular drm_atomic_helper_commit_tail(), but only calls
|
|
|
|
* drm_atomic_helper_wait_for_vblanks() if priv->no_vblank.
|
|
|
|
*/
|
|
|
|
struct drm_device *dev = old_state->dev;
|
|
|
|
struct ingenic_drm *priv = drm_device_get_priv(dev);
|
|
|
|
|
|
|
|
drm_atomic_helper_commit_modeset_disables(dev, old_state);
|
|
|
|
|
|
|
|
drm_atomic_helper_commit_planes(dev, old_state, 0);
|
|
|
|
|
|
|
|
drm_atomic_helper_commit_modeset_enables(dev, old_state);
|
|
|
|
|
|
|
|
drm_atomic_helper_commit_hw_done(old_state);
|
|
|
|
|
|
|
|
if (!priv->no_vblank)
|
|
|
|
drm_atomic_helper_wait_for_vblanks(dev, old_state);
|
|
|
|
|
|
|
|
drm_atomic_helper_cleanup_planes(dev, old_state);
|
|
|
|
}
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static irqreturn_t ingenic_drm_irq_handler(int irq, void *arg)
|
|
|
|
{
|
2020-05-17 05:50:50 +08:00
|
|
|
struct ingenic_drm *priv = drm_device_get_priv(arg);
|
2019-06-03 23:23:31 +08:00
|
|
|
unsigned int state;
|
|
|
|
|
|
|
|
regmap_read(priv->map, JZ_REG_LCD_STATE, &state);
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_STATE,
|
|
|
|
JZ_LCD_STATE_EOF_IRQ, 0);
|
|
|
|
|
|
|
|
if (state & JZ_LCD_STATE_EOF_IRQ)
|
|
|
|
drm_crtc_handle_vblank(&priv->crtc);
|
|
|
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_enable_vblank(struct drm_crtc *crtc)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_CTRL,
|
|
|
|
JZ_LCD_CTRL_EOF_IRQ, JZ_LCD_CTRL_EOF_IRQ);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ingenic_drm_disable_vblank(struct drm_crtc *crtc)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = drm_crtc_get_priv(crtc);
|
|
|
|
|
|
|
|
regmap_update_bits(priv->map, JZ_REG_LCD_CTRL, JZ_LCD_CTRL_EOF_IRQ, 0);
|
|
|
|
}
|
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
static struct drm_framebuffer *
|
|
|
|
ingenic_drm_gem_fb_create(struct drm_device *dev, struct drm_file *file,
|
|
|
|
const struct drm_mode_fb_cmd2 *mode_cmd)
|
|
|
|
{
|
|
|
|
if (ingenic_drm_cached_gem_buf)
|
|
|
|
return drm_gem_fb_create_with_dirty(dev, file, mode_cmd);
|
|
|
|
|
|
|
|
return drm_gem_fb_create(dev, file, mode_cmd);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_gem_mmap(struct drm_gem_object *obj,
|
|
|
|
struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
|
|
|
|
struct device *dev = cma_obj->base.dev->dev;
|
|
|
|
unsigned long attrs;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (ingenic_drm_cached_gem_buf)
|
|
|
|
attrs = DMA_ATTR_NON_CONSISTENT;
|
|
|
|
else
|
|
|
|
attrs = DMA_ATTR_WRITE_COMBINE;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Clear the VM_PFNMAP flag that was set by drm_gem_mmap(), and set the
|
|
|
|
* vm_pgoff (used as a fake buffer offset by DRM) to 0 as we want to map
|
|
|
|
* the whole buffer.
|
|
|
|
*/
|
|
|
|
vma->vm_flags &= ~VM_PFNMAP;
|
|
|
|
vma->vm_pgoff = 0;
|
|
|
|
vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
|
|
|
|
|
|
|
|
ret = dma_mmap_attrs(dev, vma, cma_obj->vaddr, cma_obj->paddr,
|
|
|
|
vma->vm_end - vma->vm_start, attrs);
|
|
|
|
if (ret)
|
|
|
|
drm_gem_vm_close(vma);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_gem_cma_mmap(struct file *filp,
|
|
|
|
struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = drm_gem_mmap(filp, vma);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
return ingenic_drm_gem_mmap(vma->vm_private_data, vma);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations ingenic_drm_fops = {
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.open = drm_open,
|
|
|
|
.release = drm_release,
|
|
|
|
.unlocked_ioctl = drm_ioctl,
|
|
|
|
.compat_ioctl = drm_compat_ioctl,
|
|
|
|
.poll = drm_poll,
|
|
|
|
.read = drm_read,
|
|
|
|
.llseek = noop_llseek,
|
|
|
|
.mmap = ingenic_drm_gem_cma_mmap,
|
|
|
|
};
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
static struct drm_driver ingenic_drm_driver_data = {
|
2019-06-17 23:39:24 +08:00
|
|
|
.driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_ATOMIC,
|
2019-06-03 23:23:31 +08:00
|
|
|
.name = "ingenic-drm",
|
|
|
|
.desc = "DRM module for Ingenic SoCs",
|
2020-07-17 00:38:46 +08:00
|
|
|
.date = "20200716",
|
2019-06-03 23:23:31 +08:00
|
|
|
.major = 1,
|
2020-07-17 00:38:46 +08:00
|
|
|
.minor = 1,
|
2019-06-03 23:23:31 +08:00
|
|
|
.patchlevel = 0,
|
|
|
|
|
|
|
|
.fops = &ingenic_drm_fops,
|
2020-06-05 15:32:22 +08:00
|
|
|
DRM_GEM_CMA_DRIVER_OPS,
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
.irq_handler = ingenic_drm_irq_handler,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct drm_plane_funcs ingenic_drm_primary_plane_funcs = {
|
|
|
|
.update_plane = drm_atomic_helper_update_plane,
|
|
|
|
.disable_plane = drm_atomic_helper_disable_plane,
|
|
|
|
.reset = drm_atomic_helper_plane_reset,
|
|
|
|
.destroy = drm_plane_cleanup,
|
|
|
|
|
|
|
|
.atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state,
|
|
|
|
.atomic_destroy_state = drm_atomic_helper_plane_destroy_state,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct drm_crtc_funcs ingenic_drm_crtc_funcs = {
|
|
|
|
.set_config = drm_atomic_helper_set_config,
|
|
|
|
.page_flip = drm_atomic_helper_page_flip,
|
|
|
|
.reset = drm_atomic_helper_crtc_reset,
|
|
|
|
.destroy = drm_crtc_cleanup,
|
|
|
|
|
|
|
|
.atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state,
|
|
|
|
.atomic_destroy_state = drm_atomic_helper_crtc_destroy_state,
|
|
|
|
|
|
|
|
.enable_vblank = ingenic_drm_enable_vblank,
|
|
|
|
.disable_vblank = ingenic_drm_disable_vblank,
|
|
|
|
|
|
|
|
.gamma_set = drm_atomic_helper_legacy_gamma_set,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct drm_plane_helper_funcs ingenic_drm_plane_helper_funcs = {
|
|
|
|
.atomic_update = ingenic_drm_plane_atomic_update,
|
2020-07-17 00:38:43 +08:00
|
|
|
.atomic_check = ingenic_drm_plane_atomic_check,
|
|
|
|
.atomic_disable = ingenic_drm_plane_atomic_disable,
|
2019-06-03 23:23:31 +08:00
|
|
|
.prepare_fb = drm_gem_fb_prepare_fb,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct drm_crtc_helper_funcs ingenic_drm_crtc_helper_funcs = {
|
|
|
|
.atomic_enable = ingenic_drm_crtc_atomic_enable,
|
|
|
|
.atomic_disable = ingenic_drm_crtc_atomic_disable,
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
.atomic_begin = ingenic_drm_crtc_atomic_begin,
|
2019-06-03 23:23:31 +08:00
|
|
|
.atomic_flush = ingenic_drm_crtc_atomic_flush,
|
|
|
|
.atomic_check = ingenic_drm_crtc_atomic_check,
|
2020-07-28 23:16:41 +08:00
|
|
|
.mode_valid = ingenic_drm_crtc_mode_valid,
|
2019-06-03 23:23:31 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static const struct drm_encoder_helper_funcs ingenic_drm_encoder_helper_funcs = {
|
|
|
|
.atomic_mode_set = ingenic_drm_encoder_atomic_mode_set,
|
|
|
|
.atomic_check = ingenic_drm_encoder_atomic_check,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct drm_mode_config_funcs ingenic_drm_mode_config_funcs = {
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
.fb_create = ingenic_drm_gem_fb_create,
|
2019-06-03 23:23:31 +08:00
|
|
|
.output_poll_changed = drm_fb_helper_output_poll_changed,
|
|
|
|
.atomic_check = drm_atomic_helper_check,
|
|
|
|
.atomic_commit = drm_atomic_helper_commit,
|
|
|
|
};
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
static struct drm_mode_config_helper_funcs ingenic_drm_mode_config_helpers = {
|
|
|
|
.atomic_commit_tail = ingenic_drm_atomic_helper_commit_tail,
|
|
|
|
};
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
static void ingenic_drm_unbind_all(void *d)
|
2019-06-03 23:23:31 +08:00
|
|
|
{
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
struct ingenic_drm *priv = d;
|
|
|
|
|
|
|
|
component_unbind_all(priv->dev, &priv->drm);
|
|
|
|
}
|
|
|
|
|
2020-09-27 01:04:56 +08:00
|
|
|
static void __maybe_unused ingenic_drm_release_rmem(void *d)
|
|
|
|
{
|
|
|
|
of_reserved_mem_device_release(d);
|
|
|
|
}
|
|
|
|
|
2020-08-27 19:44:04 +08:00
|
|
|
static int ingenic_drm_bind(struct device *dev, bool has_components)
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
{
|
|
|
|
struct platform_device *pdev = to_platform_device(dev);
|
2019-06-03 23:23:31 +08:00
|
|
|
const struct jz_soc_info *soc_info;
|
|
|
|
struct ingenic_drm *priv;
|
|
|
|
struct clk *parent_clk;
|
|
|
|
struct drm_bridge *bridge;
|
|
|
|
struct drm_panel *panel;
|
2020-07-17 00:38:45 +08:00
|
|
|
struct drm_encoder *encoder;
|
2019-06-03 23:23:31 +08:00
|
|
|
struct drm_device *drm;
|
|
|
|
void __iomem *base;
|
|
|
|
long parent_rate;
|
2020-07-17 00:38:45 +08:00
|
|
|
unsigned int i, clone_mask = 0;
|
2020-09-27 01:04:57 +08:00
|
|
|
dma_addr_t dma_hwdesc_phys_f0, dma_hwdesc_phys_f1;
|
2019-06-03 23:23:31 +08:00
|
|
|
int ret, irq;
|
|
|
|
|
|
|
|
soc_info = of_device_get_match_data(dev);
|
|
|
|
if (!soc_info) {
|
|
|
|
dev_err(dev, "Missing platform data\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2020-09-27 01:04:56 +08:00
|
|
|
if (IS_ENABLED(CONFIG_OF_RESERVED_MEM)) {
|
|
|
|
ret = of_reserved_mem_device_init(dev);
|
|
|
|
|
|
|
|
if (ret && ret != -ENODEV)
|
|
|
|
dev_warn(dev, "Failed to get reserved memory: %d\n", ret);
|
|
|
|
|
|
|
|
if (!ret) {
|
|
|
|
ret = devm_add_action_or_reset(dev, ingenic_drm_release_rmem, dev);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-04-15 15:40:05 +08:00
|
|
|
priv = devm_drm_dev_alloc(dev, &ingenic_drm_driver_data,
|
|
|
|
struct ingenic_drm, drm);
|
|
|
|
if (IS_ERR(priv))
|
|
|
|
return PTR_ERR(priv);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2019-12-10 22:41:41 +08:00
|
|
|
priv->soc_info = soc_info;
|
2019-06-03 23:23:31 +08:00
|
|
|
priv->dev = dev;
|
|
|
|
drm = &priv->drm;
|
|
|
|
|
|
|
|
platform_set_drvdata(pdev, priv);
|
|
|
|
|
2020-03-23 22:49:30 +08:00
|
|
|
ret = drmm_mode_config_init(drm);
|
|
|
|
if (ret)
|
2019-06-03 23:23:31 +08:00
|
|
|
return ret;
|
|
|
|
|
|
|
|
drm->mode_config.min_width = 0;
|
|
|
|
drm->mode_config.min_height = 0;
|
2019-12-10 22:41:41 +08:00
|
|
|
drm->mode_config.max_width = soc_info->max_width;
|
2019-12-10 22:41:40 +08:00
|
|
|
drm->mode_config.max_height = 4095;
|
2019-06-03 23:23:31 +08:00
|
|
|
drm->mode_config.funcs = &ingenic_drm_mode_config_funcs;
|
2020-07-17 00:38:43 +08:00
|
|
|
drm->mode_config.helper_private = &ingenic_drm_mode_config_helpers;
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2019-06-28 02:21:12 +08:00
|
|
|
base = devm_platform_ioremap_resource(pdev, 0);
|
2019-06-03 23:23:31 +08:00
|
|
|
if (IS_ERR(base)) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed to get memory resource\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return PTR_ERR(base);
|
|
|
|
}
|
|
|
|
|
|
|
|
priv->map = devm_regmap_init_mmio(dev, base,
|
|
|
|
&ingenic_drm_regmap_config);
|
|
|
|
if (IS_ERR(priv->map)) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed to create regmap\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return PTR_ERR(priv->map);
|
|
|
|
}
|
|
|
|
|
|
|
|
irq = platform_get_irq(pdev, 0);
|
2020-04-05 17:25:30 +08:00
|
|
|
if (irq < 0)
|
2019-06-03 23:23:31 +08:00
|
|
|
return irq;
|
|
|
|
|
|
|
|
if (soc_info->needs_dev_clk) {
|
|
|
|
priv->lcd_clk = devm_clk_get(dev, "lcd");
|
|
|
|
if (IS_ERR(priv->lcd_clk)) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed to get lcd clock\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return PTR_ERR(priv->lcd_clk);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
priv->pix_clk = devm_clk_get(dev, "lcd_pclk");
|
|
|
|
if (IS_ERR(priv->pix_clk)) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed to get pixel clock\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return PTR_ERR(priv->pix_clk);
|
|
|
|
}
|
|
|
|
|
2020-09-27 01:04:57 +08:00
|
|
|
priv->dma_hwdescs = dmam_alloc_coherent(dev,
|
|
|
|
sizeof(*priv->dma_hwdescs),
|
|
|
|
&priv->dma_hwdescs_phys,
|
|
|
|
GFP_KERNEL);
|
|
|
|
if (!priv->dma_hwdescs)
|
2019-06-03 23:23:31 +08:00
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
2020-09-27 01:04:57 +08:00
|
|
|
/* Configure DMA hwdesc for foreground0 plane */
|
|
|
|
dma_hwdesc_phys_f0 = priv->dma_hwdescs_phys
|
|
|
|
+ offsetof(struct ingenic_dma_hwdescs, hwdesc_f0);
|
|
|
|
priv->dma_hwdescs->hwdesc_f0.next = dma_hwdesc_phys_f0;
|
|
|
|
priv->dma_hwdescs->hwdesc_f0.id = 0xf0;
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2020-09-27 01:04:57 +08:00
|
|
|
/* Configure DMA hwdesc for foreground1 plane */
|
|
|
|
dma_hwdesc_phys_f1 = priv->dma_hwdescs_phys
|
|
|
|
+ offsetof(struct ingenic_dma_hwdescs, hwdesc_f1);
|
|
|
|
priv->dma_hwdescs->hwdesc_f1.next = dma_hwdesc_phys_f1;
|
|
|
|
priv->dma_hwdescs->hwdesc_f1.id = 0xf1;
|
2020-07-17 00:38:43 +08:00
|
|
|
|
2020-09-28 03:36:45 +08:00
|
|
|
/* Configure DMA hwdesc for palette */
|
|
|
|
priv->dma_hwdescs->hwdesc_pal.next = priv->dma_hwdescs_phys
|
|
|
|
+ offsetof(struct ingenic_dma_hwdescs, hwdesc_f0);
|
|
|
|
priv->dma_hwdescs->hwdesc_pal.id = 0xc0;
|
|
|
|
priv->dma_hwdescs->hwdesc_pal.addr = priv->dma_hwdescs_phys
|
|
|
|
+ offsetof(struct ingenic_dma_hwdescs, palette);
|
|
|
|
priv->dma_hwdescs->hwdesc_pal.cmd = JZ_LCD_CMD_ENABLE_PAL
|
|
|
|
| (sizeof(priv->dma_hwdescs->palette) / 4);
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
if (soc_info->has_osd)
|
|
|
|
priv->ipu_plane = drm_plane_from_index(drm, 0);
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
drm_plane_helper_add(&priv->f1, &ingenic_drm_plane_helper_funcs);
|
|
|
|
|
|
|
|
ret = drm_universal_plane_init(drm, &priv->f1, 1,
|
|
|
|
&ingenic_drm_primary_plane_funcs,
|
2020-09-27 01:04:58 +08:00
|
|
|
priv->soc_info->formats_f1,
|
|
|
|
priv->soc_info->num_formats_f1,
|
2019-06-03 23:23:31 +08:00
|
|
|
NULL, DRM_PLANE_TYPE_PRIMARY, NULL);
|
|
|
|
if (ret) {
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
dev_err(dev, "Failed to register plane: %i\n", ret);
|
2019-06-03 23:23:31 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
drm_plane_enable_fb_damage_clips(&priv->f1);
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
drm_crtc_helper_add(&priv->crtc, &ingenic_drm_crtc_helper_funcs);
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
ret = drm_crtc_init_with_planes(drm, &priv->crtc, &priv->f1,
|
2019-06-03 23:23:31 +08:00
|
|
|
NULL, &ingenic_drm_crtc_funcs, NULL);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed to init CRTC: %i\n", ret);
|
2019-06-03 23:23:31 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2020-09-28 03:36:45 +08:00
|
|
|
drm_crtc_enable_color_mgmt(&priv->crtc, 0, false,
|
|
|
|
ARRAY_SIZE(priv->dma_hwdescs->palette));
|
|
|
|
|
2020-07-17 00:38:43 +08:00
|
|
|
if (soc_info->has_osd) {
|
|
|
|
drm_plane_helper_add(&priv->f0,
|
|
|
|
&ingenic_drm_plane_helper_funcs);
|
|
|
|
|
|
|
|
ret = drm_universal_plane_init(drm, &priv->f0, 1,
|
|
|
|
&ingenic_drm_primary_plane_funcs,
|
2020-09-27 01:04:58 +08:00
|
|
|
priv->soc_info->formats_f0,
|
|
|
|
priv->soc_info->num_formats_f0,
|
2020-07-17 00:38:43 +08:00
|
|
|
NULL, DRM_PLANE_TYPE_OVERLAY,
|
|
|
|
NULL);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(dev, "Failed to register overlay plane: %i\n",
|
|
|
|
ret);
|
|
|
|
return ret;
|
|
|
|
}
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
gpu/drm: ingenic: Add option to mmap GEM buffers cached
Ingenic SoCs are most notably used in cheap chinese handheld gaming
consoles. There, the games and applications generally render in software
directly into GEM buffers.
Traditionally, GEM buffers are mapped write-combine. Writes to the
buffer are accelerated, and reads are slow. Application doing lots of
alpha-blending paint inside shadow buffers, which is then memcpy'd into
the final GEM buffer.
On recent Ingenic SoCs however, it is much faster to have a fully cached
GEM buffer, in which applications paint directly, and whose data is
invalidated before scanout, than having a write-combine GEM buffer, even
when alpha blending is not used.
Add an optional 'cached_gem_buffers' parameter to the ingenic-drm driver
to allow GEM buffers to be mapped fully-cached, in order to speed up
software rendering.
v2: Use standard noncoherent DMA APIs
v3: Use damage clips instead of invalidating full frames
v4: Avoid dma_pgprot() which is not exported. Using vm_get_page_prot()
is enough in this case.
v5:
- Avoid calling drm_gem_cma_prime_mmap(). It has the side effect that an
extra object reference is obtained, which causes our dumb buffers to
never be freed. It should have been drm_gem_cma_mmap_obj(). However,
our custom mmap function only differs with one flag, so we can cleanly
handle both modes in ingenic_drm_gem_mmap().
- Call drm_gem_vm_close() if drm_mmap_attrs() failed, just like in
drm_gem_cma_mmap_obj().
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200912195639.176001-1-paul@crapouillou.net
2020-09-13 03:56:39 +08:00
|
|
|
drm_plane_enable_fb_damage_clips(&priv->f0);
|
|
|
|
|
2020-08-27 19:44:04 +08:00
|
|
|
if (IS_ENABLED(CONFIG_DRM_INGENIC_IPU) && has_components) {
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
ret = component_bind_all(dev, drm);
|
|
|
|
if (ret) {
|
|
|
|
if (ret != -EPROBE_DEFER)
|
|
|
|
dev_err(dev, "Failed to bind components: %i\n", ret);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = devm_add_action_or_reset(dev, ingenic_drm_unbind_all, priv);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
priv->ipu_plane = drm_plane_from_index(drm, 2);
|
|
|
|
if (!priv->ipu_plane) {
|
|
|
|
dev_err(dev, "Failed to retrieve IPU plane\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
2020-07-17 00:38:43 +08:00
|
|
|
}
|
|
|
|
|
2020-07-17 00:38:45 +08:00
|
|
|
for (i = 0; ; i++) {
|
|
|
|
ret = drm_of_find_panel_or_bridge(dev->of_node, 0, i, &panel, &bridge);
|
|
|
|
if (ret) {
|
|
|
|
if (ret == -ENODEV)
|
|
|
|
break; /* we're done */
|
|
|
|
if (ret != -EPROBE_DEFER)
|
|
|
|
dev_err(dev, "Failed to get bridge handle\n");
|
|
|
|
return ret;
|
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2020-07-17 00:38:45 +08:00
|
|
|
if (panel)
|
|
|
|
bridge = devm_drm_panel_bridge_add_typed(dev, panel,
|
|
|
|
DRM_MODE_CONNECTOR_DPI);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
2020-07-17 00:38:45 +08:00
|
|
|
encoder = devm_kzalloc(dev, sizeof(*encoder), GFP_KERNEL);
|
|
|
|
if (!encoder)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
encoder->possible_crtcs = 1;
|
|
|
|
|
|
|
|
drm_encoder_helper_add(encoder, &ingenic_drm_encoder_helper_funcs);
|
|
|
|
|
|
|
|
ret = drm_simple_encoder_init(drm, encoder, DRM_MODE_ENCODER_DPI);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(dev, "Failed to init encoder: %d\n", ret);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = drm_bridge_attach(encoder, bridge, NULL, 0);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(dev, "Unable to attach bridge\n");
|
|
|
|
return ret;
|
|
|
|
}
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
|
2020-07-17 00:38:45 +08:00
|
|
|
drm_for_each_encoder(encoder, drm) {
|
|
|
|
clone_mask |= BIT(drm_encoder_index(encoder));
|
|
|
|
}
|
|
|
|
|
|
|
|
drm_for_each_encoder(encoder, drm) {
|
|
|
|
encoder->possible_clones = clone_mask;
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
ret = drm_irq_install(drm, irq);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Unable to install IRQ handler\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = drm_vblank_init(drm, 1);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed calling drm_vblank_init()\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
drm_mode_config_reset(drm);
|
|
|
|
|
|
|
|
ret = clk_prepare_enable(priv->pix_clk);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Unable to start pixel clock\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (priv->lcd_clk) {
|
|
|
|
parent_clk = clk_get_parent(priv->lcd_clk);
|
|
|
|
parent_rate = clk_get_rate(parent_clk);
|
|
|
|
|
|
|
|
/* LCD Device clock must be 3x the pixel clock for STN panels,
|
|
|
|
* or 1.5x the pixel clock for TFT panels. To avoid having to
|
|
|
|
* check for the LCD device clock everytime we do a mode change,
|
|
|
|
* we set the LCD device clock to the highest rate possible.
|
|
|
|
*/
|
|
|
|
ret = clk_set_rate(priv->lcd_clk, parent_rate);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Unable to set LCD clock rate\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
goto err_pixclk_disable;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = clk_prepare_enable(priv->lcd_clk);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Unable to start lcd clock\n");
|
2019-06-03 23:23:31 +08:00
|
|
|
goto err_pixclk_disable;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-07-17 00:38:40 +08:00
|
|
|
/* Set address of our DMA descriptor chain */
|
2020-09-27 01:04:57 +08:00
|
|
|
regmap_write(priv->map, JZ_REG_LCD_DA0, dma_hwdesc_phys_f0);
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_DA1, dma_hwdesc_phys_f1);
|
2020-07-17 00:38:43 +08:00
|
|
|
|
|
|
|
/* Enable OSD if available */
|
|
|
|
if (soc_info->has_osd)
|
|
|
|
regmap_write(priv->map, JZ_REG_LCD_OSDC, JZ_LCD_OSDC_OSDEN);
|
2020-07-17 00:38:40 +08:00
|
|
|
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
mutex_init(&priv->clk_mutex);
|
|
|
|
priv->clock_nb.notifier_call = ingenic_drm_update_pixclk;
|
|
|
|
|
|
|
|
parent_clk = clk_get_parent(priv->pix_clk);
|
|
|
|
ret = clk_notifier_register(parent_clk, &priv->clock_nb);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(dev, "Unable to register clock notifier\n");
|
|
|
|
goto err_devclk_disable;
|
|
|
|
}
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
ret = drm_dev_register(drm, 0);
|
|
|
|
if (ret) {
|
2020-07-17 00:38:39 +08:00
|
|
|
dev_err(dev, "Failed to register DRM driver\n");
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
goto err_clk_notifier_unregister;
|
2019-06-03 23:23:31 +08:00
|
|
|
}
|
|
|
|
|
2020-04-08 16:26:35 +08:00
|
|
|
drm_fbdev_generic_setup(drm, 32);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
err_clk_notifier_unregister:
|
|
|
|
clk_notifier_unregister(parent_clk, &priv->clock_nb);
|
2019-06-03 23:23:31 +08:00
|
|
|
err_devclk_disable:
|
|
|
|
if (priv->lcd_clk)
|
|
|
|
clk_disable_unprepare(priv->lcd_clk);
|
|
|
|
err_pixclk_disable:
|
|
|
|
clk_disable_unprepare(priv->pix_clk);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2020-08-27 19:44:04 +08:00
|
|
|
static int ingenic_drm_bind_with_components(struct device *dev)
|
|
|
|
{
|
|
|
|
return ingenic_drm_bind(dev, true);
|
|
|
|
}
|
|
|
|
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
static int compare_of(struct device *dev, void *data)
|
2019-06-03 23:23:31 +08:00
|
|
|
{
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
return dev->of_node == data;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ingenic_drm_unbind(struct device *dev)
|
|
|
|
{
|
|
|
|
struct ingenic_drm *priv = dev_get_drvdata(dev);
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
struct clk *parent_clk = clk_get_parent(priv->pix_clk);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
drm/ingenic: Reset pixclock rate when parent clock rate changes
Old Ingenic SoCs can overclock very well, up to +50% of their nominal
clock rate, whithout requiring overvolting or anything like that, just
by changing the rate of the main PLL. Unfortunately, all clocks on the
system are derived from that PLL, and when the PLL rate is updated, so
is our pixel clock.
To counter that issue, we make sure that the panel is in VBLANK before
the rate change happens, and we will then re-set the pixel clock rate
afterwards, once the PLL has been changed, to be as close as possible to
the pixel rate requested by the encoder.
v2: Add comment about mutex usage
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200926170501.1109197-2-paul@crapouillou.net
2020-09-27 01:04:55 +08:00
|
|
|
clk_notifier_unregister(parent_clk, &priv->clock_nb);
|
2019-06-03 23:23:31 +08:00
|
|
|
if (priv->lcd_clk)
|
|
|
|
clk_disable_unprepare(priv->lcd_clk);
|
|
|
|
clk_disable_unprepare(priv->pix_clk);
|
|
|
|
|
|
|
|
drm_dev_unregister(&priv->drm);
|
|
|
|
drm_atomic_helper_shutdown(&priv->drm);
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static const struct component_master_ops ingenic_master_ops = {
|
2020-08-27 19:44:04 +08:00
|
|
|
.bind = ingenic_drm_bind_with_components,
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
.unbind = ingenic_drm_unbind,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int ingenic_drm_probe(struct platform_device *pdev)
|
|
|
|
{
|
|
|
|
struct device *dev = &pdev->dev;
|
|
|
|
struct component_match *match = NULL;
|
|
|
|
struct device_node *np;
|
|
|
|
|
|
|
|
if (!IS_ENABLED(CONFIG_DRM_INGENIC_IPU))
|
2020-08-27 19:44:04 +08:00
|
|
|
return ingenic_drm_bind(dev, false);
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
|
|
|
/* IPU is at port address 8 */
|
|
|
|
np = of_graph_get_remote_node(dev->of_node, 8, 0);
|
2020-08-27 19:44:04 +08:00
|
|
|
if (!np)
|
|
|
|
return ingenic_drm_bind(dev, false);
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
|
|
|
drm_of_component_match_add(dev, &match, compare_of, np);
|
2020-08-27 19:44:03 +08:00
|
|
|
of_node_put(np);
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
|
|
|
return component_master_add_with_match(dev, &ingenic_master_ops, match);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ingenic_drm_remove(struct platform_device *pdev)
|
|
|
|
{
|
|
|
|
struct device *dev = &pdev->dev;
|
|
|
|
|
|
|
|
if (!IS_ENABLED(CONFIG_DRM_INGENIC_IPU))
|
|
|
|
ingenic_drm_unbind(dev);
|
|
|
|
else
|
|
|
|
component_master_del(dev, &ingenic_master_ops);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2020-09-27 01:04:58 +08:00
|
|
|
static const u32 jz4740_formats[] = {
|
|
|
|
DRM_FORMAT_XRGB1555,
|
|
|
|
DRM_FORMAT_RGB565,
|
|
|
|
DRM_FORMAT_XRGB8888,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const u32 jz4725b_formats_f1[] = {
|
|
|
|
DRM_FORMAT_XRGB1555,
|
|
|
|
DRM_FORMAT_RGB565,
|
|
|
|
DRM_FORMAT_XRGB8888,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const u32 jz4725b_formats_f0[] = {
|
2020-09-28 03:36:45 +08:00
|
|
|
DRM_FORMAT_C8,
|
2020-09-27 01:04:58 +08:00
|
|
|
DRM_FORMAT_XRGB1555,
|
|
|
|
DRM_FORMAT_RGB565,
|
|
|
|
DRM_FORMAT_XRGB8888,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const u32 jz4770_formats_f1[] = {
|
|
|
|
DRM_FORMAT_XRGB1555,
|
|
|
|
DRM_FORMAT_RGB565,
|
2020-09-27 01:05:01 +08:00
|
|
|
DRM_FORMAT_RGB888,
|
2020-09-27 01:04:58 +08:00
|
|
|
DRM_FORMAT_XRGB8888,
|
2020-09-27 01:05:00 +08:00
|
|
|
DRM_FORMAT_XRGB2101010,
|
2020-09-27 01:04:58 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static const u32 jz4770_formats_f0[] = {
|
2020-09-28 03:36:45 +08:00
|
|
|
DRM_FORMAT_C8,
|
2020-09-27 01:04:58 +08:00
|
|
|
DRM_FORMAT_XRGB1555,
|
|
|
|
DRM_FORMAT_RGB565,
|
2020-09-27 01:05:01 +08:00
|
|
|
DRM_FORMAT_RGB888,
|
2020-09-27 01:04:58 +08:00
|
|
|
DRM_FORMAT_XRGB8888,
|
2020-09-27 01:05:00 +08:00
|
|
|
DRM_FORMAT_XRGB2101010,
|
2020-09-27 01:04:58 +08:00
|
|
|
};
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static const struct jz_soc_info jz4740_soc_info = {
|
|
|
|
.needs_dev_clk = true,
|
2020-07-17 00:38:43 +08:00
|
|
|
.has_osd = false,
|
2019-12-10 22:41:41 +08:00
|
|
|
.max_width = 800,
|
|
|
|
.max_height = 600,
|
2020-09-27 01:04:58 +08:00
|
|
|
.formats_f1 = jz4740_formats,
|
|
|
|
.num_formats_f1 = ARRAY_SIZE(jz4740_formats),
|
|
|
|
/* JZ4740 has only one plane */
|
2019-06-03 23:23:31 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static const struct jz_soc_info jz4725b_soc_info = {
|
|
|
|
.needs_dev_clk = false,
|
2020-07-17 00:38:43 +08:00
|
|
|
.has_osd = true,
|
2019-12-10 22:41:41 +08:00
|
|
|
.max_width = 800,
|
|
|
|
.max_height = 600,
|
2020-09-27 01:04:58 +08:00
|
|
|
.formats_f1 = jz4725b_formats_f1,
|
|
|
|
.num_formats_f1 = ARRAY_SIZE(jz4725b_formats_f1),
|
|
|
|
.formats_f0 = jz4725b_formats_f0,
|
|
|
|
.num_formats_f0 = ARRAY_SIZE(jz4725b_formats_f0),
|
2019-06-03 23:23:31 +08:00
|
|
|
};
|
|
|
|
|
2019-12-10 22:41:42 +08:00
|
|
|
static const struct jz_soc_info jz4770_soc_info = {
|
|
|
|
.needs_dev_clk = false,
|
2020-07-17 00:38:43 +08:00
|
|
|
.has_osd = true,
|
2019-12-10 22:41:42 +08:00
|
|
|
.max_width = 1280,
|
|
|
|
.max_height = 720,
|
2020-09-27 01:04:58 +08:00
|
|
|
.formats_f1 = jz4770_formats_f1,
|
|
|
|
.num_formats_f1 = ARRAY_SIZE(jz4770_formats_f1),
|
|
|
|
.formats_f0 = jz4770_formats_f0,
|
|
|
|
.num_formats_f0 = ARRAY_SIZE(jz4770_formats_f0),
|
2019-12-10 22:41:42 +08:00
|
|
|
};
|
|
|
|
|
2019-06-03 23:23:31 +08:00
|
|
|
static const struct of_device_id ingenic_drm_of_match[] = {
|
|
|
|
{ .compatible = "ingenic,jz4740-lcd", .data = &jz4740_soc_info },
|
|
|
|
{ .compatible = "ingenic,jz4725b-lcd", .data = &jz4725b_soc_info },
|
2019-12-10 22:41:42 +08:00
|
|
|
{ .compatible = "ingenic,jz4770-lcd", .data = &jz4770_soc_info },
|
2019-06-03 23:23:31 +08:00
|
|
|
{ /* sentinel */ },
|
|
|
|
};
|
2020-05-04 14:35:12 +08:00
|
|
|
MODULE_DEVICE_TABLE(of, ingenic_drm_of_match);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
static struct platform_driver ingenic_drm_driver = {
|
|
|
|
.driver = {
|
|
|
|
.name = "ingenic-drm",
|
|
|
|
.of_match_table = of_match_ptr(ingenic_drm_of_match),
|
|
|
|
},
|
|
|
|
.probe = ingenic_drm_probe,
|
|
|
|
.remove = ingenic_drm_remove,
|
|
|
|
};
|
drm/ingenic: Add support for the IPU
Add support for the Image Processing Unit (IPU) found in all Ingenic
SoCs.
The IPU can upscale and downscale a source frame of arbitrary size
ranging from 4x4 to 4096x4096 on newer SoCs, with bicubic filtering
on newer SoCs, bilinear filtering on older SoCs. Nearest-neighbour can
also be obtained with proper coefficients.
Starting from the JZ4725B, the IPU supports a mode where its output is
sent directly to the LCDC, without having to be written to RAM first.
This makes it possible to use the IPU as a DRM plane on the compatible
SoCs, and have it convert and scale anything the userspace asks for to
what's available for the display.
Regarding pixel formats, older SoCs support packed YUV 4:2:2 and various
planar YUV formats. Newer SoCs introduced support for RGB.
Since the IPU is a separate hardware block, to make it work properly the
Ingenic DRM driver will now register itself as a component master in
case the IPU driver has been enabled in the config.
When enabled in the config, the CRTC will see the IPU as a second primary
plane. It cannot be enabled at the same time as the regular primary
plane. It has the same priority, which means that it will also display
below the overlay plane.
v2: - ingenic-ipu is no longer its own module. It will be built
into the ingenic-drm module.
- If enabled in the config, both the core driver and the IPU
driver will register as components; otherwise the core
driver will bypass that and call the ingenic_drm_bind()
function directly.
- Since both files now build into the same module, the
symbols previously exported as GPL are not exported anymore,
since they are only used internally.
- Fix SPDX license header in ingenic-ipu.h
- Avoid using 'for(;;);' loops without trailing statement(s)
v3: - Pass priv structure to IRQ handler; that way we don't hardcode
the expectation that the IPU plane is at index #0.
- Rework osd_changed() to account for src_* changes
- Add multiplanar YUV 4:4:4 support
- Commit fb addresses to HW at vblank, since addr registers are
not shadow registers
- Probe IPU component later so that IPU plane is last
- Fix driver not working on IPU-less hardware
- Use IPU driver's name as the IRQ name to avoid having two
'ingenic-drm' in /proc/interrupts
- Fix IPU only working for still images on JZ4725B
- Add a bit more code comments
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20200716163846.174790-10-paul@crapouillou.net
2020-07-17 00:38:44 +08:00
|
|
|
|
|
|
|
static int ingenic_drm_init(void)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
|
|
|
|
if (IS_ENABLED(CONFIG_DRM_INGENIC_IPU)) {
|
|
|
|
err = platform_driver_register(ingenic_ipu_driver_ptr);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
return platform_driver_register(&ingenic_drm_driver);
|
|
|
|
}
|
|
|
|
module_init(ingenic_drm_init);
|
|
|
|
|
|
|
|
static void ingenic_drm_exit(void)
|
|
|
|
{
|
|
|
|
platform_driver_unregister(&ingenic_drm_driver);
|
|
|
|
|
|
|
|
if (IS_ENABLED(CONFIG_DRM_INGENIC_IPU))
|
|
|
|
platform_driver_unregister(ingenic_ipu_driver_ptr);
|
|
|
|
}
|
|
|
|
module_exit(ingenic_drm_exit);
|
2019-06-03 23:23:31 +08:00
|
|
|
|
|
|
|
MODULE_AUTHOR("Paul Cercueil <paul@crapouillou.net>");
|
|
|
|
MODULE_DESCRIPTION("DRM driver for the Ingenic SoCs\n");
|
|
|
|
MODULE_LICENSE("GPL v2");
|