[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d09482d0-5169-4a2a-b00b-e492928abe5b@ti.com>
Date: Fri, 16 May 2025 09:30:56 -0500
From: Andrew Davis <afd@...com>
To: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>,
Peter Rosin
<peda@...ntia.se>, Andrew Lunn <andrew+netdev@...n.ch>,
"David S. Miller"
<davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski
<kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>, <linux-kernel@...r.kernel.org>,
<netdev@...r.kernel.org>, Heiner Kallweit
<hkallweit1@...il.com>
CC: kernel test robot <lkp@...el.com>,
Greg Kroah-Hartman
<gregkh@...uxfoundation.org>,
Samuel Holland <samuel@...lland.org>, "Arnd
Bergmann" <arnd@...db.de>
Subject: Re: [PATCH] mux: mmio: Fix missing CONFIG_REGMAP_MMIO
On 5/16/25 3:58 AM, Krzysztof Kozlowski wrote:
> On 16/05/2025 10:26, Krzysztof Kozlowski wrote:
>> On 16/05/2025 10:16, Krzysztof Kozlowski wrote:
>>>
>>> On Thu, 15 May 2025 16:05:56 +0200, Krzysztof Kozlowski wrote:
>>>> MMIO mux uses now regmap_init_mmio(), so one way or another
>>>> CONFIG_REGMAP_MMIO should be enabled, because there are no stubs for
>>>> !REGMAP_MMIO case:
>>>>
>>>> ERROR: modpost: "__regmap_init_mmio_clk" [drivers/mux/mux-mmio.ko] undefined!
>>>>
>>>> REGMAP_MMIO should be, because it is a non-visible symbol, but this
>>>> causes a circular dependency:
>>>>
>>>> [...]
>>>
>>> Applied, thanks!
>>>
>>> [1/1] mux: mmio: Fix missing CONFIG_REGMAP_MMIO
>>> https://git.kernel.org/krzk/linux/c/39bff565b40d26cc51f6e85b3b224c86a563367e
>>>
>> And dropped. More recursive dependencies are detected, so my fix here is
>> incomplete.
>>
>> error: recursive dependency detected!
>> symbol REGMAP default is visible depending on REGMAP_MMIO
>> symbol REGMAP_MMIO is selected by MUX_MMIO
>> symbol MUX_MMIO depends on MULTIPLEXER
>> symbol MULTIPLEXER is selected by MDIO_BUS_MUX_MULTIPLEXER
>> symbol MDIO_BUS_MUX_MULTIPLEXER depends on MDIO_BUS
>> symbol MDIO_BUS is selected by REGMAP
>>
>> https://krzk.eu/#/builders/43/builds/4855
>>
>> That's a mess, I need work on this a bit more.
>
>
> My branch fails with above error because I do not have Heiner's commit
> a3e1c0ad8357 ("net: phy: factor out provider part from mdio_bus.c").
> Will it reach current RC (rc7) at some point? I could merge the tag to
> my next branch to solve it.
>
>
I took a shot at breaking this recursive dependency at the REGMAP config[0].
That should allow REGMAP_MMIO to be selected by MUX_MMIO without touching
stuff here in netdev.
Kconfig is a huge mess at this point, the golden rule of "select should be
used with care" is completely ignored, there are now more uses of select
in Linux Kconfig than the normal "depends on"! Might be time to add that
SAT solver..
Andrew
[0] https://www.spinics.net/lists/kernel/msg5687922.html
Powered by blists - more mailing lists