[0/2] Remove QDF2xxx pinctrl drivers

Message ID 20240122-topic-qdf_cleanup_pinctrl-v1-0-0c619ea25091@linaro.org
Headers
Series Remove QDF2xxx pinctrl drivers |

Message

Konrad Dybcio Jan. 22, 2024, 11:57 a.m. UTC
  The SoC line was never productized, remove the maintenance burden.

Compile-tested only.

Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
---
Konrad Dybcio (2):
      pinctrl: qcom: Remove QDF2xxx support
      arm64: defconfig: Remove QDF24XX pinctrl

 arch/arm64/configs/defconfig           |   1 -
 drivers/pinctrl/qcom/Kconfig.msm       |   7 --
 drivers/pinctrl/qcom/Makefile          |   1 -
 drivers/pinctrl/qcom/pinctrl-qdf2xxx.c | 164 ---------------------------------
 4 files changed, 173 deletions(-)
---
base-commit: 319fbd8fc6d339e0a1c7b067eed870c518a13a02
change-id: 20240122-topic-qdf_cleanup_pinctrl-98e17cdb375b

Best regards,
  

Comments

Jeffrey Hugo Jan. 22, 2024, 5:04 p.m. UTC | #1
On 1/22/2024 4:57 AM, Konrad Dybcio wrote:
> The SoC line was never productized, remove the maintenance burden.
> 
> Compile-tested only.
> 
> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
> ---
> Konrad Dybcio (2):
>        pinctrl: qcom: Remove QDF2xxx support
>        arm64: defconfig: Remove QDF24XX pinctrl
> 
>   arch/arm64/configs/defconfig           |   1 -
>   drivers/pinctrl/qcom/Kconfig.msm       |   7 --
>   drivers/pinctrl/qcom/Makefile          |   1 -
>   drivers/pinctrl/qcom/pinctrl-qdf2xxx.c | 164 ---------------------------------
>   4 files changed, 173 deletions(-)
> ---
> base-commit: 319fbd8fc6d339e0a1c7b067eed870c518a13a02
> change-id: 20240122-topic-qdf_cleanup_pinctrl-98e17cdb375b
> 
> Best regards,

NACK.

This was productized, there are some out in the wild, and the platform 
is still in (limited) use.

I'd like to see support hang around for a few more years yet.
  
Dmitry Baryshkov Jan. 22, 2024, 5:56 p.m. UTC | #2
On Mon, 22 Jan 2024 at 19:43, Jeffrey Hugo <quic_jhugo@quicinc.com> wrote:
>
> On 1/22/2024 4:57 AM, Konrad Dybcio wrote:
> > The SoC line was never productized, remove the maintenance burden.
> >
> > Compile-tested only.
> >
> > Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
> > ---
> > Konrad Dybcio (2):
> >        pinctrl: qcom: Remove QDF2xxx support
> >        arm64: defconfig: Remove QDF24XX pinctrl
> >
> >   arch/arm64/configs/defconfig           |   1 -
> >   drivers/pinctrl/qcom/Kconfig.msm       |   7 --
> >   drivers/pinctrl/qcom/Makefile          |   1 -
> >   drivers/pinctrl/qcom/pinctrl-qdf2xxx.c | 164 ---------------------------------
> >   4 files changed, 173 deletions(-)
> > ---
> > base-commit: 319fbd8fc6d339e0a1c7b067eed870c518a13a02
> > change-id: 20240122-topic-qdf_cleanup_pinctrl-98e17cdb375b
> >
> > Best regards,
>
> NACK.
>
> This was productized, there are some out in the wild, and the platform
> is still in (limited) use.
>
> I'd like to see support hang around for a few more years yet.

The problem is that... its support is pretty strange. I can see
pinctrl, ethernet and quirks for the platform in GIC-ITS and PL011
drivers. Is this enough to get the platform into the useful state? I
can imagine that "QCOM2430" ACPI handle was used for USB hosts on that
platform, but I don't remember when we last tested DWC3 with the ACPI.

So, all this boils down to the question whether mainline (or something
close by, LTS for example) is actually used and tested on these
devices?
  
Jeffrey Hugo Jan. 22, 2024, 6:44 p.m. UTC | #3
On 1/22/2024 10:56 AM, Dmitry Baryshkov wrote:
> On Mon, 22 Jan 2024 at 19:43, Jeffrey Hugo <quic_jhugo@quicinc.com> wrote:
>>
>> On 1/22/2024 4:57 AM, Konrad Dybcio wrote:
>>> The SoC line was never productized, remove the maintenance burden.
>>>
>>> Compile-tested only.
>>>
>>> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
>>> ---
>>> Konrad Dybcio (2):
>>>         pinctrl: qcom: Remove QDF2xxx support
>>>         arm64: defconfig: Remove QDF24XX pinctrl
>>>
>>>    arch/arm64/configs/defconfig           |   1 -
>>>    drivers/pinctrl/qcom/Kconfig.msm       |   7 --
>>>    drivers/pinctrl/qcom/Makefile          |   1 -
>>>    drivers/pinctrl/qcom/pinctrl-qdf2xxx.c | 164 ---------------------------------
>>>    4 files changed, 173 deletions(-)
>>> ---
>>> base-commit: 319fbd8fc6d339e0a1c7b067eed870c518a13a02
>>> change-id: 20240122-topic-qdf_cleanup_pinctrl-98e17cdb375b
>>>
>>> Best regards,
>>
>> NACK.
>>
>> This was productized, there are some out in the wild, and the platform
>> is still in (limited) use.
>>
>> I'd like to see support hang around for a few more years yet.
> 
> The problem is that... its support is pretty strange. I can see
> pinctrl, ethernet and quirks for the platform in GIC-ITS and PL011
> drivers. Is this enough to get the platform into the useful state? I
> can imagine that "QCOM2430" ACPI handle was used for USB hosts on that
> platform, but I don't remember when we last tested DWC3 with the ACPI.
> 
> So, all this boils down to the question whether mainline (or something
> close by, LTS for example) is actually used and tested on these
> devices?

Its an ACPI system, so you won't see all of the fun DTisms of a MSM chip.

The platform was fully functional upstream, and had an Ubuntu 
certification.  I run Ubuntu on the two that I have in my office.  I 
haven't strictly checked out mainline in a while, but I could.  I still 
have access to the documentation.

There is a small, but active set of users including myself.  From what 
I've seen, they've been happy with things.

-Jeff
  
Dmitry Baryshkov Jan. 22, 2024, 7:23 p.m. UTC | #4
On Mon, 22 Jan 2024 at 20:44, Jeffrey Hugo <quic_jhugo@quicinc.com> wrote:
>
> On 1/22/2024 10:56 AM, Dmitry Baryshkov wrote:
> > On Mon, 22 Jan 2024 at 19:43, Jeffrey Hugo <quic_jhugo@quicinc.com> wrote:
> >>
> >> On 1/22/2024 4:57 AM, Konrad Dybcio wrote:
> >>> The SoC line was never productized, remove the maintenance burden.
> >>>
> >>> Compile-tested only.
> >>>
> >>> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
> >>> ---
> >>> Konrad Dybcio (2):
> >>>         pinctrl: qcom: Remove QDF2xxx support
> >>>         arm64: defconfig: Remove QDF24XX pinctrl
> >>>
> >>>    arch/arm64/configs/defconfig           |   1 -
> >>>    drivers/pinctrl/qcom/Kconfig.msm       |   7 --
> >>>    drivers/pinctrl/qcom/Makefile          |   1 -
> >>>    drivers/pinctrl/qcom/pinctrl-qdf2xxx.c | 164 ---------------------------------
> >>>    4 files changed, 173 deletions(-)
> >>> ---
> >>> base-commit: 319fbd8fc6d339e0a1c7b067eed870c518a13a02
> >>> change-id: 20240122-topic-qdf_cleanup_pinctrl-98e17cdb375b
> >>>
> >>> Best regards,
> >>
> >> NACK.
> >>
> >> This was productized, there are some out in the wild, and the platform
> >> is still in (limited) use.
> >>
> >> I'd like to see support hang around for a few more years yet.
> >
> > The problem is that... its support is pretty strange. I can see
> > pinctrl, ethernet and quirks for the platform in GIC-ITS and PL011
> > drivers. Is this enough to get the platform into the useful state? I
> > can imagine that "QCOM2430" ACPI handle was used for USB hosts on that
> > platform, but I don't remember when we last tested DWC3 with the ACPI.
> >
> > So, all this boils down to the question whether mainline (or something
> > close by, LTS for example) is actually used and tested on these
> > devices?
>
> Its an ACPI system, so you won't see all of the fun DTisms of a MSM chip.
>
> The platform was fully functional upstream, and had an Ubuntu
> certification.  I run Ubuntu on the two that I have in my office.  I
> haven't strictly checked out mainline in a while, but I could.  I still
> have access to the documentation.
>
> There is a small, but active set of users including myself.  From what
> I've seen, they've been happy with things.

Thanks for the information! It looks like it has a small but stable
user base. I think we should keep it, maybe ensuring that we are able
to test the kernel.
  
Konrad Dybcio Jan. 23, 2024, 6 p.m. UTC | #5
On 1/22/24 20:23, Dmitry Baryshkov wrote:
> On Mon, 22 Jan 2024 at 20:44, Jeffrey Hugo <quic_jhugo@quicinc.com> wrote:
>>
>> On 1/22/2024 10:56 AM, Dmitry Baryshkov wrote:
>>> On Mon, 22 Jan 2024 at 19:43, Jeffrey Hugo <quic_jhugo@quicinc.com> wrote:
>>>>
>>>> On 1/22/2024 4:57 AM, Konrad Dybcio wrote:
>>>>> The SoC line was never productized, remove the maintenance burden.
>>>>>
>>>>> Compile-tested only.
>>>>>
>>>>> Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
>>>>> ---
>>>>> Konrad Dybcio (2):
>>>>>          pinctrl: qcom: Remove QDF2xxx support
>>>>>          arm64: defconfig: Remove QDF24XX pinctrl
>>>>>
>>>>>     arch/arm64/configs/defconfig           |   1 -
>>>>>     drivers/pinctrl/qcom/Kconfig.msm       |   7 --
>>>>>     drivers/pinctrl/qcom/Makefile          |   1 -
>>>>>     drivers/pinctrl/qcom/pinctrl-qdf2xxx.c | 164 ---------------------------------
>>>>>     4 files changed, 173 deletions(-)
>>>>> ---
>>>>> base-commit: 319fbd8fc6d339e0a1c7b067eed870c518a13a02
>>>>> change-id: 20240122-topic-qdf_cleanup_pinctrl-98e17cdb375b
>>>>>
>>>>> Best regards,
>>>>
>>>> NACK.
>>>>
>>>> This was productized, there are some out in the wild, and the platform
>>>> is still in (limited) use.
>>>>
>>>> I'd like to see support hang around for a few more years yet.
>>>
>>> The problem is that... its support is pretty strange. I can see
>>> pinctrl, ethernet and quirks for the platform in GIC-ITS and PL011
>>> drivers. Is this enough to get the platform into the useful state? I
>>> can imagine that "QCOM2430" ACPI handle was used for USB hosts on that
>>> platform, but I don't remember when we last tested DWC3 with the ACPI.
>>>
>>> So, all this boils down to the question whether mainline (or something
>>> close by, LTS for example) is actually used and tested on these
>>> devices?
>>
>> Its an ACPI system, so you won't see all of the fun DTisms of a MSM chip.
>>
>> The platform was fully functional upstream, and had an Ubuntu
>> certification.  I run Ubuntu on the two that I have in my office.  I
>> haven't strictly checked out mainline in a while, but I could.  I still
>> have access to the documentation.
>>
>> There is a small, but active set of users including myself.  From what
>> I've seen, they've been happy with things.
> 
> Thanks for the information! It looks like it has a small but stable
> user base. I think we should keep it, maybe ensuring that we are able
> to test the kernel.

Alright, please make sure it still boots etc. then!

Konrad