Hi Laurent, Kieran, et al,
This patch series adds missing dependencies to Kconfig symbols related to the R-Car Display Unit. These dependencies prevent asking the user about R-Car display drivers when configuring a kernel without Renesas SoC support, or when the answer wouldn't have any impact on the kernel build.
Thanks for your comments!
Geert Uytterhoeven (3): drm: rcar-du: DRM_RCAR_DW_HDMI should depend on ARCH_RENESAS drm: rcar-du: DRM_RCAR_USE_LVDS should depend on DRM_RCAR_DU drm: rcar-du: DRM_RCAR_MIPI_DSI should depend on ARCH_RENESAS
drivers/gpu/drm/rcar-du/Kconfig | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
The Renesas R-Car Gen3 and RZ/G2 internal HDMI encoder is only present on Renesas R-Car Gen3 and RZ/G2 SoCs. Hence add a dependency on ARCH_RENESAS, to prevent asking the user about this driver when configuring a kernel without Renesas SoC support.
Fixes: 40d0fa7095d06c73 ("drm: rcar-du: Add Gen3 HDMI encoder support") Signed-off-by: Geert Uytterhoeven geert+renesas@glider.be --- drivers/gpu/drm/rcar-du/Kconfig | 1 + 1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig index f6e6a6d5d987bf95..65d72be50f46f19e 100644 --- a/drivers/gpu/drm/rcar-du/Kconfig +++ b/drivers/gpu/drm/rcar-du/Kconfig @@ -25,6 +25,7 @@ config DRM_RCAR_CMM config DRM_RCAR_DW_HDMI tristate "R-Car Gen3 and RZ/G2 DU HDMI Encoder Support" depends on DRM && OF + depends on ARCH_RENESAS || COMPILE_TEST select DRM_DW_HDMI help Enable support for R-Car Gen3 or RZ/G2 internal HDMI encoder.
Hi Geert,
Thank you for the patch.
On Wed, Dec 15, 2021 at 10:27:45AM +0100, Geert Uytterhoeven wrote:
Let's move this entry first to keep alphabetical order.
Reviewed-by: Laurent Pinchart laurent.pinchart@ideasonboard.com
Hi Laurent,
On Wed, Dec 15, 2021 at 11:24 AM Laurent Pinchart laurent.pinchart@ideasonboard.com wrote:
I tend to keep the "hard" and "soft" dependencies separate, with the "hard" dependencies listed first.
Reviewed-by: Laurent Pinchart laurent.pinchart@ideasonboard.com
Thank you!
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
The Renesas R-Car LVDS encoder driver is a subdriver of the R-Car Display Unit driver, and enabling DRM_RCAR_USE_LVDS while DRM_RCAR_DU is disabled doesn't have any impact on the kernel built. Hence add a dependency on DRM_RCAR_DU, to prevent asking the user about this driver when configuring a kernel without R-Car Display Unit support, like is already done for DRM_RCAR_CMM.
Fixes: 42d95d1b3a9c649b ("drm/rcar: stop using 'imply' for dependencies") Signed-off-by: Geert Uytterhoeven geert+renesas@glider.be --- The problem pre-existed before commit 42d95d1b3a9c649b, as the dependency of DRM_RCAR_LVDS on DRM_RCAR_DU was accidentally removed before. Fixes: c6a27fa41fabb35f ("drm: rcar-du: Convert LVDS encoder code to bridge driver") --- drivers/gpu/drm/rcar-du/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig index 65d72be50f46f19e..a7aa556e301d1087 100644 --- a/drivers/gpu/drm/rcar-du/Kconfig +++ b/drivers/gpu/drm/rcar-du/Kconfig @@ -32,7 +32,7 @@ config DRM_RCAR_DW_HDMI
config DRM_RCAR_USE_LVDS bool "R-Car DU LVDS Encoder Support" - depends on DRM_BRIDGE && OF + depends on DRM_BRIDGE && OF && DRM_RCAR_DU default DRM_RCAR_DU help Enable support for the R-Car Display Unit embedded LVDS encoders.
Hi Geert,
Thank you for the patch.
On Wed, Dec 15, 2021 at 10:27:46AM +0100, Geert Uytterhoeven wrote:
Shouldn't the same be done for DRM_RCAR_DW_HDMI ? Even better, we could wrap all the entries for the subdrivers in a 'if DRM_RCAR_DU'.
Hi Laurent,
On Wed, Dec 15, 2021 at 11:12 AM Laurent Pinchart laurent.pinchart@ideasonboard.com wrote:
DRM_RCAR_DW_HDMI can be enabled and built with CONFIG_COMPILE_TEST=y and CONFIG_DRM_RCAR_DU=n (yes I've tried on RISC-V ;-)
wrap all the entries for the subdrivers in a 'if DRM_RCAR_DU'.
That might work. It can be tricky with bool/tristate, as sometimes m is not properly propagated.
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
Hi Geert,
On Wed, Dec 15, 2021 at 11:17:37AM +0100, Geert Uytterhoeven wrote:
It would seem so indeed, my question is whether that shouldn't be fixed as well.
Would you give it a try for a v2 ?
Hi Geert,
On Wed, Dec 15, 2021 at 12:23:39PM +0200, Laurent Pinchart wrote:
Another option is to introduce DRM_RCAR_USE_HDMI and DRM_RCAR_USE_DSI. I'd like to keep Kconfig consistent, with the same method to handle all subdrivers if no specific reason requires doing otherwise.
Hi Laurent,
On Wed, Dec 15, 2021 at 11:30 AM Laurent Pinchart laurent.pinchart@ideasonboard.com wrote:
What is there to fix? You can build the HDMI fine without the DU driver, when compile-testing.
The HDMI and DSI drivers are separate drivers that can be (test)compiled, regardless of DRM_RCAR_DU is enabled or not.
The DRM_RCAR_USE_LVDS symbol is different: enabling it does not have any impact on the kernel build when DRM_RCAR_DU=y.
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
Hi Geert,
On Wed, Dec 15, 2021 at 11:47:27AM +0100, Geert Uytterhoeven wrote:
But it polutes the kernel config by asking a question that is irrelevant to users who don't have DRM_RCAR_DU enabled. Other patches in this series address this relevance issue with a dependency on ARCH_RENESAS, while for CMM and LVDS the dependency is on DRM_RCAR_DU.
That's only caused by how Kconfig is structured, the LVDS encoder is also a separate kernel module, like the HDMI and DSI encoders. The difference is caused by the fact that the DU driver needs to call functions exported by the LVDS driver, which caused requires preventing a configuration where DRM_RCAR_DU would be built-in and DRM_RCAR_LVDS would be built as a module.
Patches 1/3 and 3/3 in this series address the fact that the user shouldn't be asked about the DSI and HDMI encoders if ARCH_RENESAS isn't set, as those options are meaningless in that case (except when compile-testing, which is enabled through COMPILE_TEST). The same happens with this patch due to the dependency on DRM_RCAR_DU, which itself depends on ARCH_RENESAS. I don't like how we end up doing similar things in different ways, that makes Kconfig harder to understand and maintain. I'd like to see all this fixed in a cleaner way, and I think a 'if DRM_RCAR_DU' would address the config "polution" issue in one go for all subdrivers.
Hi Laurent,
On Wed, Dec 15, 2021 at 12:02 PM Laurent Pinchart laurent.pinchart@ideasonboard.com wrote:
If you prefer the big "if", fine for me.
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
The Renesas R-Car Display Unit embedded MIPI DSI encoder is only present on Renesas R-Car V3U SoCs. Hence add a dependency on ARCH_RENESAS, to prevent asking the user about this driver when configuring a kernel without Renesas SoC support.
Fixes: 155358310f013c23 ("drm: rcar-du: Add R-Car DSI driver") Signed-off-by: Geert Uytterhoeven geert+renesas@glider.be --- drivers/gpu/drm/rcar-du/Kconfig | 1 + 1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig index a7aa556e301d1087..2f4f6ac5fd4e66b1 100644 --- a/drivers/gpu/drm/rcar-du/Kconfig +++ b/drivers/gpu/drm/rcar-du/Kconfig @@ -48,6 +48,7 @@ config DRM_RCAR_LVDS config DRM_RCAR_MIPI_DSI tristate "R-Car DU MIPI DSI Encoder Support" depends on DRM && DRM_BRIDGE && OF + depends on ARCH_RENESAS || COMPILE_TEST select DRM_MIPI_DSI help Enable support for the R-Car Display Unit embedded MIPI DSI encoders.
Hi Geert,
Thank you for the patch.
On Wed, Dec 15, 2021 at 10:27:47AM +0100, Geert Uytterhoeven wrote:
Same as for 1/3, let's move this entry first. I can handle this when applying the patches, I'll take 1/3 and 3/3 in my tree already.
Reviewed-by: Laurent Pinchart laurent.pinchart@ideasonboard.com
On Wed, Dec 15, 2021 at 12:25:26PM +0200, Laurent Pinchart wrote:
Actually, if we wrap all subdrivers in 'if DRM_RCAR_DU', we won't need 1/3 and 3/3. Let's reach a conclusion on that topic first.
Hi Geert,
Thank you for the patches.
On Wed, Dec 15, 2021 at 10:27:44AM +0100, Geert Uytterhoeven wrote:
Won't this have an impact when someone will regenerate an ARM64 defconfig, given that we want to keep those options enabled in arch/arm64/configs/defconfig ?
Hi Laurent,
On Wed, Dec 15, 2021 at 10:43 AM Laurent Pinchart laurent.pinchart@ideasonboard.com wrote:
arch/arm64/configs/defconfig has CONFIG_ARCH_RENESAS=y, so it is not impacted.
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
dri-devel@lists.freedesktop.org
-
Geert Uytterhoeven -
Geert Uytterhoeven -
Laurent Pinchart