[PATCH 0/7] Enable most media drivers to build on ARM

[PATCH 0/7] Enable most media drivers to build on ARM

[Mauro Carvalho Chehab]

Right now, all media drivers build successfully with COMPILE_TEST on x86,
on both i386 and x86_64. Yet, several drivers there don't build on other
archs.

I don't need myself to build all drivers outside x86, but others could
find it useful. It also relps spreading COMPILE_TEST builds, with sounds
a good idea, as more developers may be seeing issues and submiting 
us patches.

So, this patch series makes most of them to be built elsewhere (tested
only with ARM with allyesconfig). The only two media drivers that don't build 
on such conditions are:

1) media/staging/atomisp: it uses several ACPI bits that no other media
driver requires (including Intel IPU3);

2) radio-miropcm20: This device depnds on ISA_DMA_API, with is available only
for a few non-Intel architectures.

In other words, the following symbols aren't enabled with allyesconfig:

	INTEL_ATOMISP VIDEO_ATOMISP
	VIDEO_ATOMISP_MSRLIST_HELPER VIDEO_ATOMISP_MT9M114
	VIDEO_ATOMISP_GC0310  VIDEO_ATOMISP_GC2235 
	VIDEO_ATOMISP_OV2722 VIDEO_ATOMISP_OV5693
	VIDEO_ATOMISP_OV2680 VIDEO_ATOMISP_LM3554
	RADIO_MIROPCM20

All patches in this series are available at:

	https://git.linuxtv.org/mchehab/experimental.git/log/?h=compile_test_v7

Mauro Carvalho Chehab (7):
  asm-generic, media: allow COMPILE_TEST with virt_to_bus
  media: meye: allow building it with COMPILE_TEST on non-x86
  media: rc: allow build pnp-dependent drivers with COMPILE_TEST
  media: ipu3: allow building it with COMPILE_TEST on non-x86 archs
  omapfb: omapfb_dss.h: add stubs to build with COMPILE_TEST && DRM_OMAP
  media: omap2: allow building it with COMPILE_TEST && DRM_OMAP
  media: via-camera: allow build on non-x86 archs with COMPILE_TEST

 drivers/media/pci/intel/ipu3/Kconfig |  3 +-
 drivers/media/pci/meye/Kconfig       |  3 +-
 drivers/media/pci/sta2x11/Kconfig    |  4 +--
 drivers/media/pci/zoran/Kconfig      |  3 +-
 drivers/media/platform/Kconfig       |  2 +-
 drivers/media/platform/omap/Kconfig  |  3 +-
 drivers/media/platform/via-camera.c  | 10 ++++++-
 drivers/media/rc/Kconfig             | 10 +++----
 include/asm-generic/io.h             |  2 +-
 include/linux/sony-laptop.h          |  4 +++
 include/linux/via-core.h             | 17 ++++++++++++
 include/linux/via-gpio.h             |  4 +++
 include/linux/via_i2c.h              |  5 ++++
 include/video/omapfb_dss.h           | 54 ++++++++++++++++++++++++++++++++++--
 14 files changed, 107 insertions(+), 17 deletions(-)

-- 
2.14.3


_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH 0/7] Enable most media drivers to build on ARM

[Mauro Carvalho Chehab]

Right now, all media drivers build successfully with COMPILE_TEST on x86,
on both i386 and x86_64. Yet, several drivers there don't build on other
archs.

I don't need myself to build all drivers outside x86, but others could
find it useful. It also relps spreading COMPILE_TEST builds, with sounds
a good idea, as more developers may be seeing issues and submiting 
us patches.

So, this patch series makes most of them to be built elsewhere (tested
only with ARM with allyesconfig). The only two media drivers that don't build 
on such conditions are:

1) media/staging/atomisp: it uses several ACPI bits that no other media
driver requires (including Intel IPU3);

2) radio-miropcm20: This device depnds on ISA_DMA_API, with is available only
for a few non-Intel architectures.

In other words, the following symbols aren't enabled with allyesconfig:

	INTEL_ATOMISP VIDEO_ATOMISP
	VIDEO_ATOMISP_MSRLIST_HELPER VIDEO_ATOMISP_MT9M114
	VIDEO_ATOMISP_GC0310  VIDEO_ATOMISP_GC2235 
	VIDEO_ATOMISP_OV2722 VIDEO_ATOMISP_OV5693
	VIDEO_ATOMISP_OV2680 VIDEO_ATOMISP_LM3554
	RADIO_MIROPCM20

All patches in this series are available at:

	https://git.linuxtv.org/mchehab/experimental.git/log/?h=compile_test_v7

Mauro Carvalho Chehab (7):
  asm-generic, media: allow COMPILE_TEST with virt_to_bus
  media: meye: allow building it with COMPILE_TEST on non-x86
  media: rc: allow build pnp-dependent drivers with COMPILE_TEST
  media: ipu3: allow building it with COMPILE_TEST on non-x86 archs
  omapfb: omapfb_dss.h: add stubs to build with COMPILE_TEST && DRM_OMAP
  media: omap2: allow building it with COMPILE_TEST && DRM_OMAP
  media: via-camera: allow build on non-x86 archs with COMPILE_TEST

 drivers/media/pci/intel/ipu3/Kconfig |  3 +-
 drivers/media/pci/meye/Kconfig       |  3 +-
 drivers/media/pci/sta2x11/Kconfig    |  4 +--
 drivers/media/pci/zoran/Kconfig      |  3 +-
 drivers/media/platform/Kconfig       |  2 +-
 drivers/media/platform/omap/Kconfig  |  3 +-
 drivers/media/platform/via-camera.c  | 10 ++++++-
 drivers/media/rc/Kconfig             | 10 +++----
 include/asm-generic/io.h             |  2 +-
 include/linux/sony-laptop.h          |  4 +++
 include/linux/via-core.h             | 17 ++++++++++++
 include/linux/via-gpio.h             |  4 +++
 include/linux/via_i2c.h              |  5 ++++
 include/video/omapfb_dss.h           | 54 ++++++++++++++++++++++++++++++++++--
 14 files changed, 107 insertions(+), 17 deletions(-)

-- 
2.14.3

[PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Mauro Carvalho Chehab]

The virt_to_bus/bus_to_virt macros are arch-specific. Some
archs don't support it. Yet, as it is interesting to allow
doing compilation tests on non-ia32/ia64 archs, provide a
fallback for such archs.

While here, enable COMPILE_TEST for two media drivers that
depends on it.

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
---
 drivers/media/pci/sta2x11/Kconfig | 4 ++--
 drivers/media/pci/zoran/Kconfig   | 3 ++-
 include/asm-generic/io.h          | 2 +-
 3 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/media/pci/sta2x11/Kconfig b/drivers/media/pci/sta2x11/Kconfig
index 7af3f1cbcea8..fb4b4c8ac430 100644
--- a/drivers/media/pci/sta2x11/Kconfig
+++ b/drivers/media/pci/sta2x11/Kconfig
@@ -1,10 +1,10 @@
 config STA2X11_VIP
 	tristate "STA2X11 VIP Video For Linux"
-	depends on STA2X11 || COMPILE_TEST
+	depends on (STA2X11 && VIRT_TO_BUS) || COMPILE_TEST
 	depends on HAS_DMA
 	select VIDEO_ADV7180 if MEDIA_SUBDRV_AUTOSELECT
 	select VIDEOBUF2_DMA_CONTIG
-	depends on PCI && VIDEO_V4L2 && VIRT_TO_BUS
+	depends on PCI && VIDEO_V4L2
 	depends on VIDEO_V4L2_SUBDEV_API
 	depends on I2C
 	help
diff --git a/drivers/media/pci/zoran/Kconfig b/drivers/media/pci/zoran/Kconfig
index 39ec35bd21a5..5d2678a9e310 100644
--- a/drivers/media/pci/zoran/Kconfig
+++ b/drivers/media/pci/zoran/Kconfig
@@ -1,6 +1,7 @@
 config VIDEO_ZORAN
 	tristate "Zoran ZR36057/36067 Video For Linux"
-	depends on PCI && I2C_ALGOBIT && VIDEO_V4L2 && VIRT_TO_BUS
+	depends on PCI && I2C_ALGOBIT && VIDEO_V4L2
+	depends on VIRT_TO_BUS || COMPILE_TEST
 	depends on !ALPHA
 	help
 	  Say Y for support for MJPEG capture cards based on the Zoran
diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
index 66d1d45fa2e1..f448129ad15c 100644
--- a/include/asm-generic/io.h
+++ b/include/asm-generic/io.h
@@ -1068,7 +1068,7 @@ static inline void unxlate_dev_mem_ptr(phys_addr_t phys, void *addr)
 }
 #endif
 
-#ifdef CONFIG_VIRT_TO_BUS
+#if defined(CONFIG_VIRT_TO_BUS) || defined(CONFIG_COMPILE_TEST)
 #ifndef virt_to_bus
 static inline unsigned long virt_to_bus(void *address)
 {
-- 
2.14.3

Re: [PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Geert Uytterhoeven]

Hi Mauro,

On Fri, Apr 20, 2018 at 7:42 PM, Mauro Carvalho Chehab
<mchehab@s-opensource.com> wrote:
> The virt_to_bus/bus_to_virt macros are arch-specific. Some
> archs don't support it. Yet, as it is interesting to allow
> doing compilation tests on non-ia32/ia64 archs, provide a
> fallback for such archs.
>
> While here, enable COMPILE_TEST for two media drivers that
> depends on it.
>
> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>

Thanks for your patch!

I guess this will cause the kernel to crash if I insert an STA2X11 or Zoran PCI
card in a non-VIRT_TO_BUS machine and boot an all{mod,yes}config kernel?

IMHO dummies are fine if they return error codes, not it they can cause
crashes.

> ---
>  drivers/media/pci/sta2x11/Kconfig | 4 ++--
>  drivers/media/pci/zoran/Kconfig   | 3 ++-
>  include/asm-generic/io.h          | 2 +-
>  3 files changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/media/pci/sta2x11/Kconfig b/drivers/media/pci/sta2x11/Kconfig
> index 7af3f1cbcea8..fb4b4c8ac430 100644
> --- a/drivers/media/pci/sta2x11/Kconfig
> +++ b/drivers/media/pci/sta2x11/Kconfig
> @@ -1,10 +1,10 @@
>  config STA2X11_VIP
>         tristate "STA2X11 VIP Video For Linux"
> -       depends on STA2X11 || COMPILE_TEST
> +       depends on (STA2X11 && VIRT_TO_BUS) || COMPILE_TEST
>         depends on HAS_DMA
>         select VIDEO_ADV7180 if MEDIA_SUBDRV_AUTOSELECT
>         select VIDEOBUF2_DMA_CONTIG
> -       depends on PCI && VIDEO_V4L2 && VIRT_TO_BUS
> +       depends on PCI && VIDEO_V4L2
>         depends on VIDEO_V4L2_SUBDEV_API
>         depends on I2C
>         help
> diff --git a/drivers/media/pci/zoran/Kconfig b/drivers/media/pci/zoran/Kconfig
> index 39ec35bd21a5..5d2678a9e310 100644
> --- a/drivers/media/pci/zoran/Kconfig
> +++ b/drivers/media/pci/zoran/Kconfig
> @@ -1,6 +1,7 @@
>  config VIDEO_ZORAN
>         tristate "Zoran ZR36057/36067 Video For Linux"
> -       depends on PCI && I2C_ALGOBIT && VIDEO_V4L2 && VIRT_TO_BUS
> +       depends on PCI && I2C_ALGOBIT && VIDEO_V4L2
> +       depends on VIRT_TO_BUS || COMPILE_TEST
>         depends on !ALPHA
>         help
>           Say Y for support for MJPEG capture cards based on the Zoran
> diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
> index 66d1d45fa2e1..f448129ad15c 100644
> --- a/include/asm-generic/io.h
> +++ b/include/asm-generic/io.h
> @@ -1068,7 +1068,7 @@ static inline void unxlate_dev_mem_ptr(phys_addr_t phys, void *addr)
>  }
>  #endif
>
> -#ifdef CONFIG_VIRT_TO_BUS
> +#if defined(CONFIG_VIRT_TO_BUS) || defined(CONFIG_COMPILE_TEST)
>  #ifndef virt_to_bus
>  static inline unsigned long virt_to_bus(void *address)
>  {

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

Re: [PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Geert Uytterhoeven]

Hi Mauro,

On Fri, Apr 20, 2018 at 7:42 PM, Mauro Carvalho Chehab
<mchehab@s-opensource.com> wrote:
> The virt_to_bus/bus_to_virt macros are arch-specific. Some
> archs don't support it. Yet, as it is interesting to allow
> doing compilation tests on non-ia32/ia64 archs, provide a
> fallback for such archs.
>
> While here, enable COMPILE_TEST for two media drivers that
> depends on it.
>
> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>

Thanks for your patch!

I guess this will cause the kernel to crash if I insert an STA2X11 or Zoran PCI
card in a non-VIRT_TO_BUS machine and boot an all{mod,yes}config kernel?

IMHO dummies are fine if they return error codes, not it they can cause
crashes.

> ---
>  drivers/media/pci/sta2x11/Kconfig | 4 ++--
>  drivers/media/pci/zoran/Kconfig   | 3 ++-
>  include/asm-generic/io.h          | 2 +-
>  3 files changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/media/pci/sta2x11/Kconfig b/drivers/media/pci/sta2x11/Kconfig
> index 7af3f1cbcea8..fb4b4c8ac430 100644
> --- a/drivers/media/pci/sta2x11/Kconfig
> +++ b/drivers/media/pci/sta2x11/Kconfig
> @@ -1,10 +1,10 @@
>  config STA2X11_VIP
>         tristate "STA2X11 VIP Video For Linux"
> -       depends on STA2X11 || COMPILE_TEST
> +       depends on (STA2X11 && VIRT_TO_BUS) || COMPILE_TEST
>         depends on HAS_DMA
>         select VIDEO_ADV7180 if MEDIA_SUBDRV_AUTOSELECT
>         select VIDEOBUF2_DMA_CONTIG
> -       depends on PCI && VIDEO_V4L2 && VIRT_TO_BUS
> +       depends on PCI && VIDEO_V4L2
>         depends on VIDEO_V4L2_SUBDEV_API
>         depends on I2C
>         help
> diff --git a/drivers/media/pci/zoran/Kconfig b/drivers/media/pci/zoran/Kconfig
> index 39ec35bd21a5..5d2678a9e310 100644
> --- a/drivers/media/pci/zoran/Kconfig
> +++ b/drivers/media/pci/zoran/Kconfig
> @@ -1,6 +1,7 @@
>  config VIDEO_ZORAN
>         tristate "Zoran ZR36057/36067 Video For Linux"
> -       depends on PCI && I2C_ALGOBIT && VIDEO_V4L2 && VIRT_TO_BUS
> +       depends on PCI && I2C_ALGOBIT && VIDEO_V4L2
> +       depends on VIRT_TO_BUS || COMPILE_TEST
>         depends on !ALPHA
>         help
>           Say Y for support for MJPEG capture cards based on the Zoran
> diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
> index 66d1d45fa2e1..f448129ad15c 100644
> --- a/include/asm-generic/io.h
> +++ b/include/asm-generic/io.h
> @@ -1068,7 +1068,7 @@ static inline void unxlate_dev_mem_ptr(phys_addr_t phys, void *addr)
>  }
>  #endif
>
> -#ifdef CONFIG_VIRT_TO_BUS
> +#if defined(CONFIG_VIRT_TO_BUS) || defined(CONFIG_COMPILE_TEST)
>  #ifndef virt_to_bus
>  static inline unsigned long virt_to_bus(void *address)
>  {

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Arnd Bergmann]

On Fri, Apr 20, 2018 at 7:42 PM, Mauro Carvalho Chehab
<mchehab@s-opensource.com> wrote:
> The virt_to_bus/bus_to_virt macros are arch-specific. Some
> archs don't support it. Yet, as it is interesting to allow
> doing compilation tests on non-ia32/ia64 archs, provide a
> fallback for such archs.
>
> While here, enable COMPILE_TEST for two media drivers that
> depends on it.
>
> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>

I'd prefer not to do this: virt_to_bus() is deprecated for good reasons,
and I'd rather see the drivers fixed to use dma-mapping.h correctly.

One problem with your patch is that not all architectures include
asm-generic/io.h, so it likely breaks allmodconfig builds on architectures
that don't use that file and don't provide virt_to_bus() either.

      Arnd

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

Re: [PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Arnd Bergmann]

On Fri, Apr 20, 2018 at 7:42 PM, Mauro Carvalho Chehab
<mchehab@s-opensource.com> wrote:
> The virt_to_bus/bus_to_virt macros are arch-specific. Some
> archs don't support it. Yet, as it is interesting to allow
> doing compilation tests on non-ia32/ia64 archs, provide a
> fallback for such archs.
>
> While here, enable COMPILE_TEST for two media drivers that
> depends on it.
>
> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>

I'd prefer not to do this: virt_to_bus() is deprecated for good reasons,
and I'd rather see the drivers fixed to use dma-mapping.h correctly.

One problem with your patch is that not all architectures include
asm-generic/io.h, so it likely breaks allmodconfig builds on architectures
that don't use that file and don't provide virt_to_bus() either.

      Arnd

Re: [PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Mauro Carvalho Chehab]

Em Tue, 24 Apr 2018 09:55:55 +0200
Arnd Bergmann <arnd@arndb.de> escreveu:

> On Fri, Apr 20, 2018 at 7:42 PM, Mauro Carvalho Chehab
> <mchehab@s-opensource.com> wrote:
> > The virt_to_bus/bus_to_virt macros are arch-specific. Some
> > archs don't support it. Yet, as it is interesting to allow
> > doing compilation tests on non-ia32/ia64 archs, provide a
> > fallback for such archs.
> >
> > While here, enable COMPILE_TEST for two media drivers that
> > depends on it.
> >
> > Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>  
> 
> I'd prefer not to do this: virt_to_bus() is deprecated for good reasons,
> and I'd rather see the drivers fixed to use dma-mapping.h correctly.
> 
> One problem with your patch is that not all architectures include
> asm-generic/io.h, so it likely breaks allmodconfig builds on architectures
> that don't use that file and don't provide virt_to_bus() either.

I see. I will drop this patch then. 

I might eventually try to fix the Zoran driver some day, with is the
only media driver that uses virt_to_bus(), if I switch some day my 
builds to default to ARM, but such task has very low priority to me.

Thanks,
Mauro

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

Re: [PATCH 1/7] asm-generic, media: allow COMPILE_TEST with virt_to_bus

[Mauro Carvalho Chehab]

Em Tue, 24 Apr 2018 09:55:55 +0200
Arnd Bergmann <arnd@arndb.de> escreveu:

> On Fri, Apr 20, 2018 at 7:42 PM, Mauro Carvalho Chehab
> <mchehab@s-opensource.com> wrote:
> > The virt_to_bus/bus_to_virt macros are arch-specific. Some
> > archs don't support it. Yet, as it is interesting to allow
> > doing compilation tests on non-ia32/ia64 archs, provide a
> > fallback for such archs.
> >
> > While here, enable COMPILE_TEST for two media drivers that
> > depends on it.
> >
> > Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>  
> 
> I'd prefer not to do this: virt_to_bus() is deprecated for good reasons,
> and I'd rather see the drivers fixed to use dma-mapping.h correctly.
> 
> One problem with your patch is that not all architectures include
> asm-generic/io.h, so it likely breaks allmodconfig builds on architectures
> that don't use that file and don't provide virt_to_bus() either.

I see. I will drop this patch then. 

I might eventually try to fix the Zoran driver some day, with is the
only media driver that uses virt_to_bus(), if I switch some day my 
builds to default to ARM, but such task has very low priority to me.

Thanks,
Mauro