Re: [PATCH V6 1/3] dt-bindings: sram: qcom,imem: Add Boot Stat region within IMEM

Re: [PATCH V6 1/3] dt-bindings: sram: qcom,imem: Add Boot Stat region within IMEM

[Arnd Bergmann]

On Tue, May 9, 2023, at 13:35, Dmitry Baryshkov wrote:
> On Tue, 9 May 2023 at 13:53, Souradeep Chowdhury
> <quic_schowdhu@quicinc.com> wrote:
>>
>> All Qualcomm bootloaders log useful timestamp information related
>> to bootloader stats in the IMEM region. Add the child node within
>> IMEM for the boot stat region containing register address and
>> compatible string.
>
> I might have a minor vote here. Is there any reason why you have to
> instantiate the device from DT?
> It looks like a software interface. Ideally software should not be
> described in DT (e.g. this can be instantiated from imem
> driver-to-be).

There is nothing wrong with describing firmware in DT, if that
firmware is part of the platform, we do that for a lot of
other bits of firmware.

However, in this specific case, many things are wrong with the
implementation, and neither the DT binding nor the driver
makes sense to me in its current state.

>> +  "^stats@[0-9a-f]+$":
>> +    type: object
>> +    description:
>> +      Imem region dedicated for storing timestamps related
>> +      information regarding bootstats.
>> +
>> +    additionalProperties: false
>> +
>> +    properties:
>> +      compatible:
>> +        items:
>> +          - enum:
>> +              - qcom,sm8450-bootstats
>> +          - const: qcom,imem-bootstats
>> +
>> +      reg:
>> +        maxItems: 1

If I understand this right, this "qcom,imem-bootstats"
device serves as an indirection to store additional
properties of the system in a memory area, but the description
of that area is more complex than its contents, which
makes no sense to me.

Just create a binding for a firmware node in the devicetree
itself, and put the values in properties of that. The first
stage firmware can still use the same interface, but the
actual loader that assembles the DT can get it out of that
and store it in the properties. With that done, there is also
no need for a kernel driver, as userspace can just get the
values from /sys/firmware/devicetree/ directly.

      Arnd

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V6 1/3] dt-bindings: sram: qcom,imem: Add Boot Stat region within IMEM

[Trilok Soni]

On 5/16/2023 1:16 AM, Arnd Bergmann wrote:
> On Tue, May 9, 2023, at 13:35, Dmitry Baryshkov wrote:
>> On Tue, 9 May 2023 at 13:53, Souradeep Chowdhury
>> <quic_schowdhu@quicinc.com> wrote:
>>>
>>> All Qualcomm bootloaders log useful timestamp information related
>>> to bootloader stats in the IMEM region. Add the child node within
>>> IMEM for the boot stat region containing register address and
>>> compatible string.
>>
>> I might have a minor vote here. Is there any reason why you have to
>> instantiate the device from DT?
>> It looks like a software interface. Ideally software should not be
>> described in DT (e.g. this can be instantiated from imem
>> driver-to-be).
> 
> There is nothing wrong with describing firmware in DT, if that
> firmware is part of the platform, we do that for a lot of
> other bits of firmware.
> 
> However, in this specific case, many things are wrong with the
> implementation, and neither the DT binding nor the driver
> makes sense to me in its current state.
> 
>>> +  "^stats@[0-9a-f]+$":
>>> +    type: object
>>> +    description:
>>> +      Imem region dedicated for storing timestamps related
>>> +      information regarding bootstats.
>>> +
>>> +    additionalProperties: false
>>> +
>>> +    properties:
>>> +      compatible:
>>> +        items:
>>> +          - enum:
>>> +              - qcom,sm8450-bootstats
>>> +          - const: qcom,imem-bootstats
>>> +
>>> +      reg:
>>> +        maxItems: 1
> 
> If I understand this right, this "qcom,imem-bootstats"
> device serves as an indirection to store additional
> properties of the system in a memory area, but the description
> of that area is more complex than its contents, which
> makes no sense to me.
> 
> Just create a binding for a firmware node in the devicetree
> itself, and put the values in properties of that. The first
> stage firmware can still use the same interface, but the
> actual loader that assembles the DT can get it out of that
> and store it in the properties. With that done, there is also
> no need for a kernel driver, as userspace can just get the
> values from /sys/firmware/devicetree/ directly.
> 

To understand bit better here, what you are suggesting here that
application bootloader (e.g., UEFI app for Linux) can read these
boot values from the IMEM and encode them into the devicetree properties
which can be later retrieved directly from the userspace.

I am fine with this approach as well and in this case we just
need to submit the bindings document, right?

---Trilok Soni


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V6 1/3] dt-bindings: sram: qcom,imem: Add Boot Stat region within IMEM

[Arnd Bergmann]

On Tue, May 16, 2023, at 20:34, Trilok Soni wrote:
> On 5/16/2023 1:16 AM, Arnd Bergmann wrote:
>
> To understand bit better here, what you are suggesting here that
> application bootloader (e.g., UEFI app for Linux) can read these
> boot values from the IMEM and encode them into the devicetree properties
> which can be later retrieved directly from the userspace.

I'm not entirely sure I understand the concept of "application bootloader",
but in principle this can be anything that runs before the devicetree
is handed off from the final boot loader to the kernel.

> I am fine with this approach as well and in this case we just
> need to submit the bindings document, right?

Yes, exactly.

     Arnd

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V6 1/3] dt-bindings: sram: qcom,imem: Add Boot Stat region within IMEM

[Dmitry Baryshkov]

On Tue, 16 May 2023 at 11:16, Arnd Bergmann <arnd@arndb.de> wrote:
>
> On Tue, May 9, 2023, at 13:35, Dmitry Baryshkov wrote:
> > On Tue, 9 May 2023 at 13:53, Souradeep Chowdhury
> > <quic_schowdhu@quicinc.com> wrote:
> >>
> >> All Qualcomm bootloaders log useful timestamp information related
> >> to bootloader stats in the IMEM region. Add the child node within
> >> IMEM for the boot stat region containing register address and
> >> compatible string.
> >
> > I might have a minor vote here. Is there any reason why you have to
> > instantiate the device from DT?
> > It looks like a software interface. Ideally software should not be
> > described in DT (e.g. this can be instantiated from imem
> > driver-to-be).
>
> There is nothing wrong with describing firmware in DT, if that
> firmware is part of the platform, we do that for a lot of
> other bits of firmware.
>
> However, in this specific case, many things are wrong with the
> implementation, and neither the DT binding nor the driver
> makes sense to me in its current state.
>
> >> +  "^stats@[0-9a-f]+$":
> >> +    type: object
> >> +    description:
> >> +      Imem region dedicated for storing timestamps related
> >> +      information regarding bootstats.
> >> +
> >> +    additionalProperties: false
> >> +
> >> +    properties:
> >> +      compatible:
> >> +        items:
> >> +          - enum:
> >> +              - qcom,sm8450-bootstats
> >> +          - const: qcom,imem-bootstats
> >> +
> >> +      reg:
> >> +        maxItems: 1
>
> If I understand this right, this "qcom,imem-bootstats"
> device serves as an indirection to store additional
> properties of the system in a memory area, but the description
> of that area is more complex than its contents, which
> makes no sense to me.
>
> Just create a binding for a firmware node in the devicetree
> itself, and put the values in properties of that. The first
> stage firmware can still use the same interface, but the
> actual loader that assembles the DT can get it out of that
> and store it in the properties. With that done, there is also
> no need for a kernel driver, as userspace can just get the
> values from /sys/firmware/devicetree/ directly.

This sounds good, except the always-present issue of the devices which
have already been released. Usually we can not expect a bootloader
update for these devices.

-- 
With best wishes
Dmitry

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V6 1/3] dt-bindings: sram: qcom,imem: Add Boot Stat region within IMEM

[Trilok Soni]

On 5/16/2023 12:17 PM, Dmitry Baryshkov wrote:
> On Tue, 16 May 2023 at 11:16, Arnd Bergmann <arnd@arndb.de> wrote:
>>
>> On Tue, May 9, 2023, at 13:35, Dmitry Baryshkov wrote:
>>> On Tue, 9 May 2023 at 13:53, Souradeep Chowdhury
>>> <quic_schowdhu@quicinc.com> wrote:
>>>>
>>>> All Qualcomm bootloaders log useful timestamp information related
>>>> to bootloader stats in the IMEM region. Add the child node within
>>>> IMEM for the boot stat region containing register address and
>>>> compatible string.
>>>
>>> I might have a minor vote here. Is there any reason why you have to
>>> instantiate the device from DT?
>>> It looks like a software interface. Ideally software should not be
>>> described in DT (e.g. this can be instantiated from imem
>>> driver-to-be).
>>
>> There is nothing wrong with describing firmware in DT, if that
>> firmware is part of the platform, we do that for a lot of
>> other bits of firmware.
>>
>> However, in this specific case, many things are wrong with the
>> implementation, and neither the DT binding nor the driver
>> makes sense to me in its current state.
>>
>>>> +  "^stats@[0-9a-f]+$":
>>>> +    type: object
>>>> +    description:
>>>> +      Imem region dedicated for storing timestamps related
>>>> +      information regarding bootstats.
>>>> +
>>>> +    additionalProperties: false
>>>> +
>>>> +    properties:
>>>> +      compatible:
>>>> +        items:
>>>> +          - enum:
>>>> +              - qcom,sm8450-bootstats
>>>> +          - const: qcom,imem-bootstats
>>>> +
>>>> +      reg:
>>>> +        maxItems: 1
>>
>> If I understand this right, this "qcom,imem-bootstats"
>> device serves as an indirection to store additional
>> properties of the system in a memory area, but the description
>> of that area is more complex than its contents, which
>> makes no sense to me.
>>
>> Just create a binding for a firmware node in the devicetree
>> itself, and put the values in properties of that. The first
>> stage firmware can still use the same interface, but the
>> actual loader that assembles the DT can get it out of that
>> and store it in the properties. With that done, there is also
>> no need for a kernel driver, as userspace can just get the
>> values from /sys/firmware/devicetree/ directly.
> 
> This sounds good, except the always-present issue of the devices which
> have already been released. Usually we can not expect a bootloader
> update for these devices.

Valid point. I don't expect current SOCs supported at upstream to update 
with the newer bootloader having this feature done through bootloader.

---Trilok Soni


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V6 2/3] soc: qcom: boot_stat: Add Driver Support for Boot Stats

[Arnd Bergmann]

On Tue, May 9, 2023, at 12:52, Souradeep Chowdhury wrote:
> All of Qualcomm's proprietary Android boot-loaders capture boot time
> stats, like the time when the bootloader started execution and at what
> point the bootloader handed over control to the kernel etc. in the IMEM
> region. This information is captured in a specific format by this driver
> by mapping a structure to the IMEM memory region and then accessing the
> members of the structure to show the information within debugfs file.
> This information is useful in verifying if the existing boot KPIs have
> regressed or not. The information is shown in milliseconds, a sample
> log from sm8450(waipio) device is as follows:-
>
> /sys/kernel/debug/qcom_boot_stats # cat abl_time
> 17898 ms
> /sys/kernel/debug/qcom_boot_stats # cat pre_abl_time
> 2879 ms
>
> The Module Power Manager(MPM) sleep counter starts ticking at the PBL
> stage and the timestamp generated by the sleep counter is logged by
> the Qualcomm proprietary bootloader(ABL) at two points-> First when it
> starts execution which is logged here as "pre_abl_time" and the second
> when it is about to load the kernel logged as "abl_time". Documentation
> details are also added in Documentation/ABI/testing/debugfs-driver-bootstat
>
> Signed-off-by: Souradeep Chowdhury <quic_schowdhu@quicinc.com>
> ---
>  .../ABI/testing/debugfs-driver-bootstat       |  17 +++
>  drivers/soc/qcom/Kconfig                      |  10 ++
>  drivers/soc/qcom/Makefile                     |   1 +
>  drivers/soc/qcom/boot_stats.c                 | 100 ++++++++++++++++++

As mentioned in my reply to the binding, I don't think this should
be a driver at all. On top of that, even if it was a driver, it is
clearly not a "soc" driver since nothing in it has any relevance to
the hardware, rather than the first-stage loader, and drivers/soc/
drivers should never have their own user space interface either.

       Arnd

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH V6 2/3] soc: qcom: boot_stat: Add Driver Support for Boot Stats

[Dmitry Baryshkov]

On 16/05/2023 11:19, Arnd Bergmann wrote:
> On Tue, May 9, 2023, at 12:52, Souradeep Chowdhury wrote:
>> All of Qualcomm's proprietary Android boot-loaders capture boot time
>> stats, like the time when the bootloader started execution and at what
>> point the bootloader handed over control to the kernel etc. in the IMEM
>> region. This information is captured in a specific format by this driver
>> by mapping a structure to the IMEM memory region and then accessing the
>> members of the structure to show the information within debugfs file.
>> This information is useful in verifying if the existing boot KPIs have
>> regressed or not. The information is shown in milliseconds, a sample
>> log from sm8450(waipio) device is as follows:-
>>
>> /sys/kernel/debug/qcom_boot_stats # cat abl_time
>> 17898 ms
>> /sys/kernel/debug/qcom_boot_stats # cat pre_abl_time
>> 2879 ms
>>
>> The Module Power Manager(MPM) sleep counter starts ticking at the PBL
>> stage and the timestamp generated by the sleep counter is logged by
>> the Qualcomm proprietary bootloader(ABL) at two points-> First when it
>> starts execution which is logged here as "pre_abl_time" and the second
>> when it is about to load the kernel logged as "abl_time". Documentation
>> details are also added in Documentation/ABI/testing/debugfs-driver-bootstat
>>
>> Signed-off-by: Souradeep Chowdhury <quic_schowdhu@quicinc.com>
>> ---
>>   .../ABI/testing/debugfs-driver-bootstat       |  17 +++
>>   drivers/soc/qcom/Kconfig                      |  10 ++
>>   drivers/soc/qcom/Makefile                     |   1 +
>>   drivers/soc/qcom/boot_stats.c                 | 100 ++++++++++++++++++
> 
> As mentioned in my reply to the binding, I don't think this should
> be a driver at all. On top of that, even if it was a driver, it is
> clearly not a "soc" driver since nothing in it has any relevance to
> the hardware, rather than the first-stage loader, and drivers/soc/
> drivers should never have their own user space interface either.

I suppose that we should add a proper driver for imem rather than always 
using it through syscon.

-- 
With best wishes
Dmitry


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel