linux-kernel - Re: [PATCH v2] phy: tegra: depend on COMMON

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <d280c555-5ee4-7803-ada1-88257f809816@gmail.com>
Date:   Tue, 16 Mar 2021 19:14:45 +0300
From:   Dmitry Osipenko <digetx@...il.com>
To:     Krzysztof Kozlowski <krzysztof.kozlowski@...onical.com>
Cc:     Thierry Reding <thierry.reding@...il.com>,
        Jonathan Hunter <jonathanh@...dia.com>,
        linux-tegra@...r.kernel.org, Felipe Balbi <balbi@...nel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org,
        Krzysztof Kozlowski <krzk@...nel.org>,
        kernel test robot <lkp@...el.com>
Subject: Re: [PATCH v2] phy: tegra: depend on COMMON_CLK to fix compile tests

16.03.2021 18:51, Krzysztof Kozlowski пишет:
> On Tue, 16 Mar 2021 at 16:47, Krzysztof Kozlowski
> <krzysztof.kozlowski@...onical.com> wrote:
>>
>> On Tue, 16 Mar 2021 at 16:43, Dmitry Osipenko <digetx@...il.com> wrote:
>>>
>>> 16.03.2021 10:55, Krzysztof Kozlowski пишет:
>>>> From: Krzysztof Kozlowski <krzk@...nel.org>
>>>>
>>>> The Tegra USB PHY driver uses Common Clock Framework thus it cannot be
>>>> built on platforms without it (e.g. compile test on MIPS with RALINK and
>>>> SOC_RT305X):
>>>>
>>>>     /usr/bin/mips-linux-gnu-ld: drivers/usb/phy/phy-tegra-usb.o: in function `tegra_usb_phy_init':
>>>>     phy-tegra-usb.c:(.text+0x1dd4): undefined reference to `clk_get_parent'
>>>>
>>>> Reported-by: kernel test robot <lkp@...el.com>
>>>> Signed-off-by: Krzysztof Kozlowski <krzk@...nel.org>
>>>>
>>>> ---
>>>>
>>>> Changes since v1:
>>>> 1. Depend on COMMON_CLK always, not only for compile test.
>>>> ---
>>>>  drivers/usb/phy/Kconfig | 1 +
>>>>  1 file changed, 1 insertion(+)
>>>>
>>>> diff --git a/drivers/usb/phy/Kconfig b/drivers/usb/phy/Kconfig
>>>> index 52eebcb88c1f..7500e77a7d01 100644
>>>> --- a/drivers/usb/phy/Kconfig
>>>> +++ b/drivers/usb/phy/Kconfig
>>>> @@ -163,6 +163,7 @@ config USB_MXS_PHY
>>>>  config USB_TEGRA_PHY
>>>>       tristate "NVIDIA Tegra USB PHY Driver"
>>>>       depends on ARCH_TEGRA || COMPILE_TEST
>>>> +     depends on COMMON_CLK
>>>>       select USB_COMMON
>>>>       select USB_PHY
>>>>       select USB_ULPI
>>>>
>>>
>>> But if COMMON_CLK is disabled, then include/linux/clk.h provides a stub
>>> for clk_get_parent(), meaning that MIPS has its own COMMON_CLK, no?
>>
>> Hi,
>>
>> It depends on the platform. Not all of them implement every clk API,
>> so you can have failures:
>> https://lore.kernel.org/lkml/202102170017.MgPVy7aZ-lkp@intel.com/
> 
> Ah, you mentioned the stub, so let me clarify more. The common clk
> stubs are not used for cases like !COMMON_CLK && HAVE_LEGACY_CLK (or
> HAVE_CLK, I don't remember). This is why you can have a MIPS platform
> defining some of the clock operations thus not using COMMON_CLK at all
> (and neither the stubs).

I see now that the stubs depend on CONFIG_HAVE_CLK and not COMMON_CLK,
thanks.

This raises question about why those platforms select CONFIG_HAVE_CLK,
while not implementing it fully. Sounds like a better fix should be to
add the missing stubs to the MIPS clk implementation, which should avoid
the need to patch each driver individually.

https://elixir.bootlin.com/linux/v5.12-rc3/source/arch/mips/ar7/clock.c#L489