Message ID | 20191016192807.1278987-1-jernej.skrabec@siol.net |
---|---|
Headers | show |
Series | media: Introduce Allwinner H3 deinterlace driver | expand |
On 10/16/19 9:28 PM, Jernej Skrabec wrote: > Allwinner H3 SoC contains deinterlace unit, which has several modes of > operation - bypass, weave, bob and mixed (advanced) mode. I don't know > how mixed mode works, but according to Allwinner it gives best results, > so they use it exclusively. Currently this mode is also hardcoded here. > > For each interleaved frame queued, this driver produces 2 deinterlaced > frames. Deinterlaced frames are based on 2 consequtive output buffers, > except for the first 2, where same output buffer is given to peripheral > as current and previous. > > There is no documentation for this core, so register layout and fixed > values were taken from BSP driver. > > I'm not sure if maximum size of the image unit is capable to process is > governed by size of "flag" buffers, frequency or it really is some HW > limitation. Currently driver can process full HD image in ~15ms (7.5ms > for each capture buffer), which allows to process 1920x1080@60i video > smoothly in real time. > > Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net> > --- > MAINTAINERS | 7 + > drivers/media/platform/sunxi/Kconfig | 1 + > drivers/media/platform/sunxi/Makefile | 1 + > drivers/media/platform/sunxi/sun8i-di/Kconfig | 11 + > .../media/platform/sunxi/sun8i-di/Makefile | 2 + > .../media/platform/sunxi/sun8i-di/sun8i-di.c | 1020 +++++++++++++++++ > .../media/platform/sunxi/sun8i-di/sun8i-di.h | 237 ++++ > 7 files changed, 1279 insertions(+) > create mode 100644 drivers/media/platform/sunxi/sun8i-di/Kconfig > create mode 100644 drivers/media/platform/sunxi/sun8i-di/Makefile > create mode 100644 drivers/media/platform/sunxi/sun8i-di/sun8i-di.c > create mode 100644 drivers/media/platform/sunxi/sun8i-di/sun8i-di.h > > diff --git a/MAINTAINERS b/MAINTAINERS > index c7b48525822a..c375455125fb 100644 > --- a/MAINTAINERS > +++ b/MAINTAINERS > @@ -4646,6 +4646,13 @@ M: "Maciej W. Rozycki" <macro@linux-mips.org> > S: Maintained > F: drivers/net/fddi/defxx.* > > +DEINTERLACE DRIVERS FOR ALLWINNER H3 > +M: Jernej Skrabec <jernej.skrabec@siol.net> > +L: linux-media@vger.kernel.org > +T: git git://linuxtv.org/media_tree.git > +S: Maintained > +F: drivers/media/platform/sunxi/sun8i-di/ > + > DELL SMBIOS DRIVER > M: Pali Rohár <pali.rohar@gmail.com> > M: Mario Limonciello <mario.limonciello@dell.com> > diff --git a/drivers/media/platform/sunxi/Kconfig b/drivers/media/platform/sunxi/Kconfig > index 71808e93ac2e..d7a5621bf327 100644 > --- a/drivers/media/platform/sunxi/Kconfig > +++ b/drivers/media/platform/sunxi/Kconfig > @@ -1,2 +1,3 @@ > source "drivers/media/platform/sunxi/sun4i-csi/Kconfig" > source "drivers/media/platform/sunxi/sun6i-csi/Kconfig" > +source "drivers/media/platform/sunxi/sun8i-di/Kconfig" This is a m2m driver, so this belongs in drivers/media/platform/Kconfig in the Memory-to-memory section. > diff --git a/drivers/media/platform/sunxi/Makefile b/drivers/media/platform/sunxi/Makefile > index a05127529006..3878cb4efdc2 100644 > --- a/drivers/media/platform/sunxi/Makefile > +++ b/drivers/media/platform/sunxi/Makefile > @@ -1,2 +1,3 @@ > obj-y += sun4i-csi/ > obj-y += sun6i-csi/ > +obj-y += sun8i-di/ > diff --git a/drivers/media/platform/sunxi/sun8i-di/Kconfig b/drivers/media/platform/sunxi/sun8i-di/Kconfig > new file mode 100644 > index 000000000000..dbd77a61e3b3 > --- /dev/null > +++ b/drivers/media/platform/sunxi/sun8i-di/Kconfig > @@ -0,0 +1,11 @@ > +# SPDX-License-Identifier: GPL-2.0-only > +config VIDEO_SUN8I_DEINTERLACE > + tristate "Allwinner Deinterlace driver" > + depends on VIDEO_DEV && VIDEO_V4L2 > + depends on HAS_DMA > + depends on OF > + depends on PM > + select VIDEOBUF2_DMA_CONTIG > + select V4L2_MEM2MEM_DEV > + help > + Support for the Allwinner Deinterlace unit found on some SoCs. Shouldn't this depend on ARCH_SUNXI || COMPILE_TEST? And also on COMMON_CLK? Regards, Hans
Hi, I have a small comment that can definitely be addressed in a subsequent patch On Wed, Oct 16, 2019 at 09:28:06PM +0200, Jernej Skrabec wrote: > + dev->bus_clk = devm_clk_get(dev->dev, "bus"); > + if (IS_ERR(dev->bus_clk)) { > + dev_err(dev->dev, "Failed to get bus clock\n"); > + > + return PTR_ERR(dev->bus_clk); > + } > + > + dev->mod_clk = devm_clk_get(dev->dev, "mod"); > + if (IS_ERR(dev->mod_clk)) { > + dev_err(dev->dev, "Failed to get mod clock\n"); > + > + return PTR_ERR(dev->mod_clk); > + } > + > + dev->ram_clk = devm_clk_get(dev->dev, "ram"); > + if (IS_ERR(dev->ram_clk)) { > + dev_err(dev->dev, "Failed to get ram clock\n"); > + > + return PTR_ERR(dev->ram_clk); > + } > + > + dev->rstc = devm_reset_control_get(dev->dev, NULL); > + if (IS_ERR(dev->rstc)) { > + dev_err(dev->dev, "Failed to get reset control\n"); > + > + return PTR_ERR(dev->rstc); > + } > + > + clk_set_rate_exclusive(dev->mod_clk, 300000000); clk_set_rate_exclusive puts a pretty big constraint on the clock tree, and we shouldn't really enforce it if the device is unused. I guess we should move it to the runtime_pm resume hook (with the put_exclusive call in suspend). Otherwise, that patch is Acked-by: Maxime Ripard <mripard@kernel.org> Maxime
Dne četrtek, 17. oktober 2019 ob 09:51:28 CEST je Hans Verkuil napisal(a): > On 10/16/19 9:28 PM, Jernej Skrabec wrote: > > Allwinner H3 SoC contains deinterlace unit, which has several modes of > > operation - bypass, weave, bob and mixed (advanced) mode. I don't know > > how mixed mode works, but according to Allwinner it gives best results, > > so they use it exclusively. Currently this mode is also hardcoded here. > > > > For each interleaved frame queued, this driver produces 2 deinterlaced > > frames. Deinterlaced frames are based on 2 consequtive output buffers, > > except for the first 2, where same output buffer is given to peripheral > > as current and previous. > > > > There is no documentation for this core, so register layout and fixed > > values were taken from BSP driver. > > > > I'm not sure if maximum size of the image unit is capable to process is > > governed by size of "flag" buffers, frequency or it really is some HW > > limitation. Currently driver can process full HD image in ~15ms (7.5ms > > for each capture buffer), which allows to process 1920x1080@60i video > > smoothly in real time. > > > > Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net> > > --- > > > > MAINTAINERS | 7 + > > drivers/media/platform/sunxi/Kconfig | 1 + > > drivers/media/platform/sunxi/Makefile | 1 + > > drivers/media/platform/sunxi/sun8i-di/Kconfig | 11 + > > .../media/platform/sunxi/sun8i-di/Makefile | 2 + > > .../media/platform/sunxi/sun8i-di/sun8i-di.c | 1020 +++++++++++++++++ > > .../media/platform/sunxi/sun8i-di/sun8i-di.h | 237 ++++ > > 7 files changed, 1279 insertions(+) > > create mode 100644 drivers/media/platform/sunxi/sun8i-di/Kconfig > > create mode 100644 drivers/media/platform/sunxi/sun8i-di/Makefile > > create mode 100644 drivers/media/platform/sunxi/sun8i-di/sun8i-di.c > > create mode 100644 drivers/media/platform/sunxi/sun8i-di/sun8i-di.h > > > > diff --git a/MAINTAINERS b/MAINTAINERS > > index c7b48525822a..c375455125fb 100644 > > --- a/MAINTAINERS > > +++ b/MAINTAINERS > > @@ -4646,6 +4646,13 @@ M: "Maciej W. Rozycki" <macro@linux-mips.org> > > > > S: Maintained > > F: drivers/net/fddi/defxx.* > > > > +DEINTERLACE DRIVERS FOR ALLWINNER H3 > > +M: Jernej Skrabec <jernej.skrabec@siol.net> > > +L: linux-media@vger.kernel.org > > +T: git git://linuxtv.org/media_tree.git > > +S: Maintained > > +F: drivers/media/platform/sunxi/sun8i-di/ > > + > > > > DELL SMBIOS DRIVER > > M: Pali Rohár <pali.rohar@gmail.com> > > M: Mario Limonciello <mario.limonciello@dell.com> > > > > diff --git a/drivers/media/platform/sunxi/Kconfig > > b/drivers/media/platform/sunxi/Kconfig index 71808e93ac2e..d7a5621bf327 > > 100644 > > --- a/drivers/media/platform/sunxi/Kconfig > > +++ b/drivers/media/platform/sunxi/Kconfig > > @@ -1,2 +1,3 @@ > > > > source "drivers/media/platform/sunxi/sun4i-csi/Kconfig" > > source "drivers/media/platform/sunxi/sun6i-csi/Kconfig" > > > > +source "drivers/media/platform/sunxi/sun8i-di/Kconfig" > > This is a m2m driver, so this belongs in drivers/media/platform/Kconfig in > the Memory-to-memory section. > > > diff --git a/drivers/media/platform/sunxi/Makefile > > b/drivers/media/platform/sunxi/Makefile index a05127529006..3878cb4efdc2 > > 100644 > > --- a/drivers/media/platform/sunxi/Makefile > > +++ b/drivers/media/platform/sunxi/Makefile > > @@ -1,2 +1,3 @@ > > > > obj-y += sun4i-csi/ > > obj-y += sun6i-csi/ > > > > +obj-y += sun8i-di/ > > diff --git a/drivers/media/platform/sunxi/sun8i-di/Kconfig > > b/drivers/media/platform/sunxi/sun8i-di/Kconfig new file mode 100644 > > index 000000000000..dbd77a61e3b3 > > --- /dev/null > > +++ b/drivers/media/platform/sunxi/sun8i-di/Kconfig > > @@ -0,0 +1,11 @@ > > +# SPDX-License-Identifier: GPL-2.0-only > > +config VIDEO_SUN8I_DEINTERLACE > > + tristate "Allwinner Deinterlace driver" > > + depends on VIDEO_DEV && VIDEO_V4L2 > > + depends on HAS_DMA > > + depends on OF > > + depends on PM > > + select VIDEOBUF2_DMA_CONTIG > > + select V4L2_MEM2MEM_DEV > > + help > > + Support for the Allwinner Deinterlace unit found on some SoCs. > > Shouldn't this depend on ARCH_SUNXI || COMPILE_TEST? > And also on COMMON_CLK? Yes to both. Also I don't see a reason why it would depend on HAS_DMA, so I will remove that. Best regards, Jernej > > Regards, > > Hans
Dne četrtek, 17. oktober 2019 ob 11:28:00 CEST je Maxime Ripard napisal(a): > Hi, > > I have a small comment that can definitely be addressed in a subsequent > patch > On Wed, Oct 16, 2019 at 09:28:06PM +0200, Jernej Skrabec wrote: > > + dev->bus_clk = devm_clk_get(dev->dev, "bus"); > > + if (IS_ERR(dev->bus_clk)) { > > + dev_err(dev->dev, "Failed to get bus clock\n"); > > + > > + return PTR_ERR(dev->bus_clk); > > + } > > + > > + dev->mod_clk = devm_clk_get(dev->dev, "mod"); > > + if (IS_ERR(dev->mod_clk)) { > > + dev_err(dev->dev, "Failed to get mod clock\n"); > > + > > + return PTR_ERR(dev->mod_clk); > > + } > > + > > + dev->ram_clk = devm_clk_get(dev->dev, "ram"); > > + if (IS_ERR(dev->ram_clk)) { > > + dev_err(dev->dev, "Failed to get ram clock\n"); > > + > > + return PTR_ERR(dev->ram_clk); > > + } > > + > > + dev->rstc = devm_reset_control_get(dev->dev, NULL); > > + if (IS_ERR(dev->rstc)) { > > + dev_err(dev->dev, "Failed to get reset control\n"); > > + > > + return PTR_ERR(dev->rstc); > > + } > > + > > + clk_set_rate_exclusive(dev->mod_clk, 300000000); > > clk_set_rate_exclusive puts a pretty big constraint on the clock tree, > and we shouldn't really enforce it if the device is unused. That is true in general, but as I said before, that is not really an issue for H3. Deinterlace clock default parent is peripheral clock, which is fixed to 600 MHz and doesn't change. > > I guess we should move it to the runtime_pm resume hook (with the > put_exclusive call in suspend). Ok, I'll move it in case that this deinterlace core is used on other SoCs with non-fixed parent clock. > > Otherwise, that patch is > Acked-by: Maxime Ripard <mripard@kernel.org> Thanks! Best regards, Jernej