[PATCH] ARM: config: Refresh mutli v7

[Hans Verkuil]

On 08/06/2021 18:14, Arnd Bergmann wrote:
> On Tue, Jun 8, 2021 at 12:38 PM Joel Stanley <joel at jms.id.au> wrote:
>>
>> This regenerates the v7 defconfig, with some new options added to ensure
>> drivers that were enabled in previous releases are once again enabled.
>>
>> This was generated atop next-20210607.
> 
> Thank you for going through this and updating all the settings. If you
> still have
> energy after writing up the long changelog, can I ask you to split it up into
> separate changes for options that are no longer available (sirf, efi_vars, ...),
> options that got moved within the file and those that need some other change?
> 
>> SPI_AX88796C was enabled in v5.11 for Exynos boards, but the driver is
>> not yet merged. There's still some work being doing to merge this
>> driver, so leave it alone for now.
> 
> Makes sense
> 
>> MICREL_PHY is selected by the ks8851 ethernet driver as of f0791b92d2b6 ("net:
>> ks8851: Select PHYLIB and MICREL_PHY in Kconfig").
>>
>> SMSC_PHY is selected by USB_NET_SMSC95XX as of commit 05b35e7eb9a1 ("smsc95xx:
>> add phylib support").
>>
>> PTP_1588_CLOCK is unknown. SOmething ot do with FEC or STMMAC_ETH?
> 
> No idea, in my copy of linux-next it still exists. There are a couple of drivers
> that use 'imply PTP_1588_CLOCK', but that is something we should probably
> change, as 'imply' usually doesn't do what was intended. Yes, I realize that
> PTP_1588_CLOCK was the original motivation for introducing 'imply'...
> 
>> The MSM pinctrl drivers fell out of the defconfig as of commit be117ca32261
>> ("pinctrl: qcom: Kconfig: Rework PINCTRL_MSM to be a depenency rather then a
>> selected config"). Add PINCTRL_MSM so these stay enabled.
> 
> Good catch
> 
>> POWER_AVS was renamed POWER_AVS_OMAP in commit bca815d62054 ("PM: AVS:
>> smartreflex Move driver to soc specific drivers"). Add the new symbol.
>>
>> STMPE_I2C is default y. It was explicitly added to the defconfig in commit
>> d8f0ddc83877 ("ARM: defconfig: add STMPE ADC driver for P4 Note") but it's not
>> required.
> 
> I suppose it shouldn't really use 'default y', but removing it here is
> easier, it's
> been this way for ten years now.
> 
>> VIVID needs FB=y, which became disabled at some point, so enable it.
>> This also ensure the FB_EFI and other drivers stay enabled.
> 
> There should probably be a separate effort to remove dependencies on
> CONFIG_FB and leave it turned off in defconfig. Hans, do you know the
> hard dependency in vivid can be avoided or if this is inherent to how the driver
> works?

Right now it is inherent to the driver. It is probably possible to drop support
for video overlay devices if CONFIG_FB=n, but it is not something I have time
for. It's just a test driver (albeit a very useful test driver), so it is no
big deal if it is disabled when CONFIG_FB=n.

Regards,

	Hans

> 
> I had previously tried to decouple DRM from FB, so I'm happy that Kees' patch
> made it into linux-next.
> 
>> CLK_ACTIONS and CLK_OWL_S500 was explicitly added in commit 26ba4a474aec ("ARM:
>> multi_v7_defconfig: Enable Actions Semi platform and drivers"), but it defaults
>> on when ARCH_ACTIONS is enabled. Drop them.
>>
>> QCOM_PM became ARM_QCOM_SPM_CPUIDLE in commit a871be6b8eee ("cpuidle: Convert
>> Qualcomm SPM driver to a generic CPUidle driver").
>>
>> SND_MMP_SOC became CONFIG_SND_MMP_SOC_SSPA in commit a97e384ba78f ("ASoC:
>> mmp-sspa: Add Device Tree support"). Enable that instead.
>>
>> MFD_CROS_EC_DEV defualts on when CROS_EC is enabled.
>>
>> TEGRA20_APB_DMA is selected by SOC_TEGRA_FUSE if ARCH_TEGRA_2x_SOC since commit
>> 19d41e5e9c68 ("soc/tegra: fuse: Add APB DMA dependency for Tegra20").
>>
>> USB_CHIPIDEA_TEGRA is enabled when USB_CHIPIDEA is enabled. It falls out
>> of the defconfig but is enabled, as intended by d50229cee69b ("ARM:
>> multi_v7_defconfig: Stop using deprecated USB_EHCI_TEGRA").
> 
> These all look good.
> 
>> EDAC depends on RAS, so enable it to ensure the EDAC drivers stay
>> enabled.
>>
>> SND_CTL_LED selects LEDS_TRIGGERS and NEW_LEDS so they fall out of the
>> defconfig.
> 
> Please skip this last change, I think we should change the Kconfig
> logic instead here,
> so SND_CTL_LED doesn't force-enable the subsystem.
> 
>       Arnd
>