[PATCH v2 0/3] arm64: dts: qcom: sc7280: Add property for sc7280

[PATCH v2 0/3] arm64: dts: qcom: sc7280: Add property for sc7280

[Ling Xu]

This patch series add memory region, nsessions and dma-coherent property
for sc7280.
Patch [v1]:https://lore.kernel.org/linux-arm-msm/20250514052323.3881600-1-quic_lxu5@quicinc.com/

Changes in v2:
  - Add compatible.

Ling Xu (3):
  arm64: dts: qcom: sc7280: Add memory region for audiopd
  arm64: dts: qcom: sc7280: Add nsessions property for adsp
  arm64: dts: qcom: sc7280: Add dma-coherent property for fastrpc nodes

 arch/arm64/boot/dts/qcom/sc7280.dtsi | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

-- 
2.34.1

[PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Ling Xu]

Add reserved memory region and VMIDs for audio PD dynamic loading and
remote heap memory requirements.

Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
---
 arch/arm64/boot/dts/qcom/sc7280.dtsi | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
index 8e86d75cc6b4..d9af79ff8c4e 100644
--- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
+++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
@@ -188,6 +188,14 @@ rmtfs_mem: rmtfs@9c900000 {
 			qcom,client-id = <1>;
 			qcom,vmid = <QCOM_SCM_VMID_MSS_MSA>;
 		};
+
+		adsp_rpc_remote_heap_mem: adsp-rpc-remote-heap {
+			compatible = "shared-dma-pool";
+			size = <0x0 0x800000>;
+			alignment = <0x0 0x100000>;
+			alloc-ranges = <0x0 0x80000000 0x0 0x40000000>;
+			no-map;
+		};
 	};
 
 	cpus {
@@ -3863,6 +3871,9 @@ fastrpc {
 					qcom,glink-channels = "fastrpcglink-apps-dsp";
 					label = "adsp";
 					qcom,non-secure-domain;
+					memory-region = <&adsp_rpc_remote_heap_mem>;
+					qcom,vmids = <QCOM_SCM_VMID_LPASS>,
+							  <QCOM_SCM_VMID_ADSP_HEAP>;
 					#address-cells = <1>;
 					#size-cells = <0>;
 
-- 
2.34.1

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Dmitry Baryshkov]

On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
> Add reserved memory region and VMIDs for audio PD dynamic loading and
> remote heap memory requirements.

Why? Was it not working without this heap?

> 
> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
> ---
>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> index 8e86d75cc6b4..d9af79ff8c4e 100644
> --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
> +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> @@ -188,6 +188,14 @@ rmtfs_mem: rmtfs@9c900000 {
>  			qcom,client-id = <1>;
>  			qcom,vmid = <QCOM_SCM_VMID_MSS_MSA>;
>  		};
> +
> +		adsp_rpc_remote_heap_mem: adsp-rpc-remote-heap {
> +			compatible = "shared-dma-pool";
> +			size = <0x0 0x800000>;
> +			alignment = <0x0 0x100000>;
> +			alloc-ranges = <0x0 0x80000000 0x0 0x40000000>;
> +			no-map;
> +		};
>  	};
>  
>  	cpus {
> @@ -3863,6 +3871,9 @@ fastrpc {
>  					qcom,glink-channels = "fastrpcglink-apps-dsp";
>  					label = "adsp";
>  					qcom,non-secure-domain;
> +					memory-region = <&adsp_rpc_remote_heap_mem>;
> +					qcom,vmids = <QCOM_SCM_VMID_LPASS>,
> +							  <QCOM_SCM_VMID_ADSP_HEAP>;

Please align '<' vertically.

>  					#address-cells = <1>;
>  					#size-cells = <0>;
>  
> -- 
> 2.34.1
> 

-- 
With best wishes
Dmitry

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Ling Xu]

在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
>> Add reserved memory region and VMIDs for audio PD dynamic loading and
>> remote heap memory requirements.
> 
> Why? Was it not working without this heap?

yes, it will not working without this heap.
Memory region is required for audio PD for dynamic loading and remote heap memory
requirements. For more info, please refer below patches, it has provided a more
detailed explanation.
https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/

>>
>> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
>> ---
>>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 11 +++++++++++
>>  1 file changed, 11 insertions(+)
>>
>> diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
>> index 8e86d75cc6b4..d9af79ff8c4e 100644
>> --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
>> +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
>> @@ -188,6 +188,14 @@ rmtfs_mem: rmtfs@9c900000 {
>>  			qcom,client-id = <1>;
>>  			qcom,vmid = <QCOM_SCM_VMID_MSS_MSA>;
>>  		};
>> +
>> +		adsp_rpc_remote_heap_mem: adsp-rpc-remote-heap {
>> +			compatible = "shared-dma-pool";
>> +			size = <0x0 0x800000>;
>> +			alignment = <0x0 0x100000>;
>> +			alloc-ranges = <0x0 0x80000000 0x0 0x40000000>;
>> +			no-map;
>> +		};
>>  	};
>>  
>>  	cpus {
>> @@ -3863,6 +3871,9 @@ fastrpc {
>>  					qcom,glink-channels = "fastrpcglink-apps-dsp";
>>  					label = "adsp";
>>  					qcom,non-secure-domain;
>> +					memory-region = <&adsp_rpc_remote_heap_mem>;
>> +					qcom,vmids = <QCOM_SCM_VMID_LPASS>,
>> +							  <QCOM_SCM_VMID_ADSP_HEAP>;

Thanks. Will modify this.

> Please align '<' vertically.
> 
>>  					#address-cells = <1>;
>>  					#size-cells = <0>;
>>  
>> -- 
>> 2.34.1
>>
> 

-- 
Thx and BRs,
Ling Xu

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Dmitry Baryshkov]

On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
> > On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
> >> Add reserved memory region and VMIDs for audio PD dynamic loading and
> >> remote heap memory requirements.
> > 
> > Why? Was it not working without this heap?
> 
> yes, it will not working without this heap.
> Memory region is required for audio PD for dynamic loading and remote heap memory
> requirements. For more info, please refer below patches, it has provided a more
> detailed explanation.
> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/

This triggers a bunch of questions:

- What is audio PD?

- Is it a required component?

- For which platforms ?

- What happens if there is a memory region, but it is not used by the
  driver (SM8650, SM8550).

> 
> >>
> >> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
> >> ---
> >>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 11 +++++++++++
> >>  1 file changed, 11 insertions(+)
> >>
> >> diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> >> index 8e86d75cc6b4..d9af79ff8c4e 100644
> >> --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
> >> +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> >> @@ -188,6 +188,14 @@ rmtfs_mem: rmtfs@9c900000 {
> >>  			qcom,client-id = <1>;
> >>  			qcom,vmid = <QCOM_SCM_VMID_MSS_MSA>;
> >>  		};
> >> +
> >> +		adsp_rpc_remote_heap_mem: adsp-rpc-remote-heap {
> >> +			compatible = "shared-dma-pool";
> >> +			size = <0x0 0x800000>;
> >> +			alignment = <0x0 0x100000>;
> >> +			alloc-ranges = <0x0 0x80000000 0x0 0x40000000>;
> >> +			no-map;
> >> +		};
> >>  	};
> >>  
> >>  	cpus {
> >> @@ -3863,6 +3871,9 @@ fastrpc {
> >>  					qcom,glink-channels = "fastrpcglink-apps-dsp";
> >>  					label = "adsp";
> >>  					qcom,non-secure-domain;
> >> +					memory-region = <&adsp_rpc_remote_heap_mem>;
> >> +					qcom,vmids = <QCOM_SCM_VMID_LPASS>,
> >> +							  <QCOM_SCM_VMID_ADSP_HEAP>;
> 
> Thanks. Will modify this.
> 
> > Please align '<' vertically.
> > 
> >>  					#address-cells = <1>;
> >>  					#size-cells = <0>;
> >>  
> >> -- 
> >> 2.34.1
> >>
> > 
> 
> -- 
> Thx and BRs,
> Ling Xu
> 

-- 
With best wishes
Dmitry

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Ling Xu]

在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
> On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
>> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
>>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
>>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
>>>> remote heap memory requirements.
>>>
>>> Why? Was it not working without this heap?
>>
>> yes, it will not working without this heap.
>> Memory region is required for audio PD for dynamic loading and remote heap memory
>> requirements. For more info, please refer below patches, it has provided a more
>> detailed explanation.
>> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
>> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
> 
> This triggers a bunch of questions:
> 
> - What is audio PD?
> 
Audio PD is a protection domain which is running on DSP for audio
specific processing.
> - Is it a required component?
yes. it is needed wherever there is some audio use cases where 
DSP audio PD dynamic loading is needed.
> - For which platforms ?
> 
For platforms that require dynamic loading of DSP audio PD.
> - What happens if there is a memory region, but it is not used by the
>   driver (SM8650, SM8550).
> 
It won't be used.
>>
>>>>
>>>> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
>>>> ---
>>>>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 11 +++++++++++
>>>>  1 file changed, 11 insertions(+)
>>>>
>>>> diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
>>>> index 8e86d75cc6b4..d9af79ff8c4e 100644
>>>> --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
>>>> +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
>>>> @@ -188,6 +188,14 @@ rmtfs_mem: rmtfs@9c900000 {
>>>>  			qcom,client-id = <1>;
>>>>  			qcom,vmid = <QCOM_SCM_VMID_MSS_MSA>;
>>>>  		};
>>>> +
>>>> +		adsp_rpc_remote_heap_mem: adsp-rpc-remote-heap {
>>>> +			compatible = "shared-dma-pool";
>>>> +			size = <0x0 0x800000>;
>>>> +			alignment = <0x0 0x100000>;
>>>> +			alloc-ranges = <0x0 0x80000000 0x0 0x40000000>;
>>>> +			no-map;
>>>> +		};
>>>>  	};
>>>>  
>>>>  	cpus {
>>>> @@ -3863,6 +3871,9 @@ fastrpc {
>>>>  					qcom,glink-channels = "fastrpcglink-apps-dsp";
>>>>  					label = "adsp";
>>>>  					qcom,non-secure-domain;
>>>> +					memory-region = <&adsp_rpc_remote_heap_mem>;
>>>> +					qcom,vmids = <QCOM_SCM_VMID_LPASS>,
>>>> +							  <QCOM_SCM_VMID_ADSP_HEAP>;
>>
>> Thanks. Will modify this.
>>
>>> Please align '<' vertically.
>>>
>>>>  					#address-cells = <1>;
>>>>  					#size-cells = <0>;
>>>>  
>>>> -- 
>>>> 2.34.1
>>>>
>>>
>>
>> -- 
>> Thx and BRs,
>> Ling Xu
>>
> 

-- 
Thx and BRs,
Ling Xu

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Dmitry Baryshkov]

On Thu, 22 May 2025 at 13:44, Ling Xu <quic_lxu5@quicinc.com> wrote:
>
> 在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
> > On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
> >> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
> >>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
> >>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
> >>>> remote heap memory requirements.
> >>>
> >>> Why? Was it not working without this heap?
> >>
> >> yes, it will not working without this heap.
> >> Memory region is required for audio PD for dynamic loading and remote heap memory
> >> requirements. For more info, please refer below patches, it has provided a more
> >> detailed explanation.
> >> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
> >> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
> >
> > This triggers a bunch of questions:
> >
> > - What is audio PD?
> >
> Audio PD is a protection domain which is running on DSP for audio
> specific processing.
> > - Is it a required component?
> yes. it is needed wherever there is some audio use cases where
> DSP audio PD dynamic loading is needed.

Which use cases?

> > - For which platforms ?
> >
> For platforms that require dynamic loading of DSP audio PD.

Please be more specific, we are not playing a guessing game here.

> > - What happens if there is a memory region, but it is not used by the
> >   driver (SM8650, SM8550).
> >
> It won't be used.

So, what happens? DSP crash? system reboot? blue smoke coming from the SoC?

I'm asking all those questions to understand:
- whether this is applicable to previous SC7280 platforms or just RB3Gen2 / IDP
- whether this needs to be backported or not
- whether a similar fix should be applied to other platforms.

> >>
> >>>>
> >>>> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
> >>>> ---
> >>>>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 11 +++++++++++
> >>>>  1 file changed, 11 insertions(+)
> >>>>
> >>>> diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> >>>> index 8e86d75cc6b4..d9af79ff8c4e 100644
> >>>> --- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
> >>>> +++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
> >>>> @@ -188,6 +188,14 @@ rmtfs_mem: rmtfs@9c900000 {
> >>>>                    qcom,client-id = <1>;
> >>>>                    qcom,vmid = <QCOM_SCM_VMID_MSS_MSA>;
> >>>>            };
> >>>> +
> >>>> +          adsp_rpc_remote_heap_mem: adsp-rpc-remote-heap {
> >>>> +                  compatible = "shared-dma-pool";
> >>>> +                  size = <0x0 0x800000>;
> >>>> +                  alignment = <0x0 0x100000>;
> >>>> +                  alloc-ranges = <0x0 0x80000000 0x0 0x40000000>;
> >>>> +                  no-map;
> >>>> +          };
> >>>>    };
> >>>>
> >>>>    cpus {
> >>>> @@ -3863,6 +3871,9 @@ fastrpc {
> >>>>                                    qcom,glink-channels = "fastrpcglink-apps-dsp";
> >>>>                                    label = "adsp";
> >>>>                                    qcom,non-secure-domain;
> >>>> +                                  memory-region = <&adsp_rpc_remote_heap_mem>;
> >>>> +                                  qcom,vmids = <QCOM_SCM_VMID_LPASS>,
> >>>> +                                                    <QCOM_SCM_VMID_ADSP_HEAP>;
> >>
> >> Thanks. Will modify this.
> >>
> >>> Please align '<' vertically.
> >>>
> >>>>                                    #address-cells = <1>;
> >>>>                                    #size-cells = <0>;
> >>>>
> >>>> --
> >>>> 2.34.1
> >>>>
> >>>
> >>
> >> --
> >> Thx and BRs,
> >> Ling Xu
> >>
> >
>
> --
> Thx and BRs,
> Ling Xu
>


-- 
With best wishes
Dmitry

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Konrad Dybcio]

On 5/22/25 2:16 PM, Dmitry Baryshkov wrote:
> On Thu, 22 May 2025 at 13:44, Ling Xu <quic_lxu5@quicinc.com> wrote:
>>
>> 在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
>>> On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
>>>> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
>>>>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
>>>>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
>>>>>> remote heap memory requirements.
>>>>>
>>>>> Why? Was it not working without this heap?
>>>>
>>>> yes, it will not working without this heap.
>>>> Memory region is required for audio PD for dynamic loading and remote heap memory
>>>> requirements. For more info, please refer below patches, it has provided a more
>>>> detailed explanation.
>>>> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
>>>> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
>>>
>>> This triggers a bunch of questions:
>>>
>>> - What is audio PD?
>>>
>> Audio PD is a protection domain which is running on DSP for audio
>> specific processing.
>>> - Is it a required component?
>> yes. it is needed wherever there is some audio use cases where
>> DSP audio PD dynamic loading is needed.
> 
> Which use cases?
> 
>>> - For which platforms ?
>>>
>> For platforms that require dynamic loading of DSP audio PD.
> 
> Please be more specific, we are not playing a guessing game here.
> 
>>> - What happens if there is a memory region, but it is not used by the
>>>   driver (SM8650, SM8550).
>>>
>> It won't be used.
> 
> So, what happens? DSP crash? system reboot? blue smoke coming from the SoC?
> 
> I'm asking all those questions to understand:
> - whether this is applicable to previous SC7280 platforms or just RB3Gen2 / IDP
> - whether this needs to be backported or not
> - whether a similar fix should be applied to other platforms.

I wouldn't worry about FP5 too much (albeit it's apparently not easy
to test it out and even having spent some time trying to, I couldn't
figure out how these things connect), but the chromebooks may be
potentially problematic. That said, we can overwrite it in sc7280-chrome
if necessary

Konrad

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Dmitry Baryshkov]

On Thu, May 22, 2025 at 08:27:40PM +0200, Konrad Dybcio wrote:
> On 5/22/25 2:16 PM, Dmitry Baryshkov wrote:
> > On Thu, 22 May 2025 at 13:44, Ling Xu <quic_lxu5@quicinc.com> wrote:
> >>
> >> 在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
> >>> On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
> >>>> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
> >>>>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
> >>>>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
> >>>>>> remote heap memory requirements.
> >>>>>
> >>>>> Why? Was it not working without this heap?
> >>>>
> >>>> yes, it will not working without this heap.
> >>>> Memory region is required for audio PD for dynamic loading and remote heap memory
> >>>> requirements. For more info, please refer below patches, it has provided a more
> >>>> detailed explanation.
> >>>> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
> >>>> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
> >>>
> >>> This triggers a bunch of questions:
> >>>
> >>> - What is audio PD?
> >>>
> >> Audio PD is a protection domain which is running on DSP for audio
> >> specific processing.
> >>> - Is it a required component?
> >> yes. it is needed wherever there is some audio use cases where
> >> DSP audio PD dynamic loading is needed.
> > 
> > Which use cases?
> > 
> >>> - For which platforms ?
> >>>
> >> For platforms that require dynamic loading of DSP audio PD.
> > 
> > Please be more specific, we are not playing a guessing game here.
> > 
> >>> - What happens if there is a memory region, but it is not used by the
> >>>   driver (SM8650, SM8550).
> >>>
> >> It won't be used.
> > 
> > So, what happens? DSP crash? system reboot? blue smoke coming from the SoC?
> > 
> > I'm asking all those questions to understand:
> > - whether this is applicable to previous SC7280 platforms or just RB3Gen2 / IDP
> > - whether this needs to be backported or not
> > - whether a similar fix should be applied to other platforms.
> 
> I wouldn't worry about FP5 too much (albeit it's apparently not easy
> to test it out and even having spent some time trying to, I couldn't
> figure out how these things connect), but the chromebooks may be
> potentially problematic. That said, we can overwrite it in sc7280-chrome
> if necessary

What about NothingPhone?

-- 
With best wishes
Dmitry

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Konrad Dybcio]

On 5/22/25 8:51 PM, Dmitry Baryshkov wrote:
> On Thu, May 22, 2025 at 08:27:40PM +0200, Konrad Dybcio wrote:
>> On 5/22/25 2:16 PM, Dmitry Baryshkov wrote:
>>> On Thu, 22 May 2025 at 13:44, Ling Xu <quic_lxu5@quicinc.com> wrote:
>>>>
>>>> 在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
>>>>> On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
>>>>>> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
>>>>>>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
>>>>>>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
>>>>>>>> remote heap memory requirements.
>>>>>>>
>>>>>>> Why? Was it not working without this heap?
>>>>>>
>>>>>> yes, it will not working without this heap.
>>>>>> Memory region is required for audio PD for dynamic loading and remote heap memory
>>>>>> requirements. For more info, please refer below patches, it has provided a more
>>>>>> detailed explanation.
>>>>>> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
>>>>>> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
>>>>>
>>>>> This triggers a bunch of questions:
>>>>>
>>>>> - What is audio PD?
>>>>>
>>>> Audio PD is a protection domain which is running on DSP for audio
>>>> specific processing.
>>>>> - Is it a required component?
>>>> yes. it is needed wherever there is some audio use cases where
>>>> DSP audio PD dynamic loading is needed.
>>>
>>> Which use cases?
>>>
>>>>> - For which platforms ?
>>>>>
>>>> For platforms that require dynamic loading of DSP audio PD.
>>>
>>> Please be more specific, we are not playing a guessing game here.
>>>
>>>>> - What happens if there is a memory region, but it is not used by the
>>>>>   driver (SM8650, SM8550).
>>>>>
>>>> It won't be used.
>>>
>>> So, what happens? DSP crash? system reboot? blue smoke coming from the SoC?
>>>
>>> I'm asking all those questions to understand:
>>> - whether this is applicable to previous SC7280 platforms or just RB3Gen2 / IDP
>>> - whether this needs to be backported or not
>>> - whether a similar fix should be applied to other platforms.
>>
>> I wouldn't worry about FP5 too much (albeit it's apparently not easy
>> to test it out and even having spent some time trying to, I couldn't
>> figure out how these things connect), but the chromebooks may be
>> potentially problematic. That said, we can overwrite it in sc7280-chrome
>> if necessary
> 
> What about NothingPhone?

I got Nothing to say, we can ask +Eugene to test it

But I suppose you should be able to check if 845/8150 needs it
and we can probably assume the base-7280's config matches

Konrad

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Srinivas Kandagatla]

On 5/22/25 7:51 PM, Dmitry Baryshkov wrote:
> On Thu, May 22, 2025 at 08:27:40PM +0200, Konrad Dybcio wrote:
>> On 5/22/25 2:16 PM, Dmitry Baryshkov wrote:
>>> On Thu, 22 May 2025 at 13:44, Ling Xu <quic_lxu5@quicinc.com> wrote:
>>>>
>>>> 在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
>>>>> On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
>>>>>> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
>>>>>>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
>>>>>>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
>>>>>>>> remote heap memory requirements.
>>>>>>>
>>>>>>> Why? Was it not working without this heap?
>>>>>>
>>>>>> yes, it will not working without this heap.
>>>>>> Memory region is required for audio PD for dynamic loading and remote heap memory
>>>>>> requirements. For more info, please refer below patches, it has provided a more
>>>>>> detailed explanation.
>>>>>> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
>>>>>> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
>>>>>
>>>>> This triggers a bunch of questions:
>>>>>
>>>>> - What is audio PD?
>>>>>
>>>> Audio PD is a protection domain which is running on DSP for audio
>>>> specific processing.
>>>>> - Is it a required component?
>>>> yes. it is needed wherever there is some audio use cases where
>>>> DSP audio PD dynamic loading is needed.
>>>
>>> Which use cases?
>>>

Usecase is for audio compress offload, ex: when we play mp3 or any other
compress format, we need audio pd support.



thanks,
Srini
>>>>> - For which platforms ?
>>>>>
>>>> For platforms that require dynamic loading of DSP audio PD.
>>>
>>> Please be more specific, we are not playing a guessing game here.
>>>
>>>>> - What happens if there is a memory region, but it is not used by the
>>>>>   driver (SM8650, SM8550).
>>>>>
>>>> It won't be used.
>>>
>>> So, what happens? DSP crash? system reboot? blue smoke coming from the SoC?
>>>
>>> I'm asking all those questions to understand:
>>> - whether this is applicable to previous SC7280 platforms or just RB3Gen2 / IDP
>>> - whether this needs to be backported or not
>>> - whether a similar fix should be applied to other platforms.
>>
>> I wouldn't worry about FP5 too much (albeit it's apparently not easy
>> to test it out and even having spent some time trying to, I couldn't
>> figure out how these things connect), but the chromebooks may be
>> potentially problematic. That said, we can overwrite it in sc7280-chrome
>> if necessary
> 
> What about NothingPhone?
> 

Re: [PATCH v2 1/3] arm64: dts: qcom: sc7280: Add memory region for audiopd

[Dmitry Baryshkov]

On 27/06/2025 15:16, Srinivas Kandagatla wrote:
> On 5/22/25 7:51 PM, Dmitry Baryshkov wrote:
>> On Thu, May 22, 2025 at 08:27:40PM +0200, Konrad Dybcio wrote:
>>> On 5/22/25 2:16 PM, Dmitry Baryshkov wrote:
>>>> On Thu, 22 May 2025 at 13:44, Ling Xu <quic_lxu5@quicinc.com> wrote:
>>>>>
>>>>> 在 5/20/2025 9:53 PM, Dmitry Baryshkov 写道:
>>>>>> On Tue, May 20, 2025 at 02:41:22PM +0800, Ling Xu wrote:
>>>>>>> 在 5/17/2025 5:47 AM, Dmitry Baryshkov 写道:
>>>>>>>> On Fri, May 16, 2025 at 04:30:27PM +0530, Ling Xu wrote:
>>>>>>>>> Add reserved memory region and VMIDs for audio PD dynamic loading and
>>>>>>>>> remote heap memory requirements.
>>>>>>>>
>>>>>>>> Why? Was it not working without this heap?
>>>>>>>
>>>>>>> yes, it will not working without this heap.
>>>>>>> Memory region is required for audio PD for dynamic loading and remote heap memory
>>>>>>> requirements. For more info, please refer below patches, it has provided a more
>>>>>>> detailed explanation.
>>>>>>> https://lore.kernel.org/all/bb68da04-ef52-4172-8b6e-f4027bcc2786@oss.qualcomm.com/
>>>>>>> https://lore.kernel.org/all/effea02f-6ffb-42e9-87df-081caafab728@oss.qualcomm.com/
>>>>>>
>>>>>> This triggers a bunch of questions:
>>>>>>
>>>>>> - What is audio PD?
>>>>>>
>>>>> Audio PD is a protection domain which is running on DSP for audio
>>>>> specific processing.
>>>>>> - Is it a required component?
>>>>> yes. it is needed wherever there is some audio use cases where
>>>>> DSP audio PD dynamic loading is needed.
>>>>
>>>> Which use cases?
>>>>
> 
> Usecase is for audio compress offload, ex: when we play mp3 or any other
> compress format, we need audio pd support.

This needs to be in the commit message.


-- 
With best wishes
Dmitry

[PATCH v2 2/3] arm64: dts: qcom: sc7280: Add nsessions property for adsp

[Ling Xu]

Add nsessions property for adsp context bank to make sessions
available for ADSP offload.

Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
---
 arch/arm64/boot/dts/qcom/sc7280.dtsi | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
index d9af79ff8c4e..86b51cffcd54 100644
--- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
+++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
@@ -3893,6 +3893,7 @@ compute-cb@5 {
 						compatible = "qcom,fastrpc-compute-cb";
 						reg = <5>;
 						iommus = <&apps_smmu 0x1805 0x0>;
+						qcom,nsessions = <5>;
 					};
 				};
 			};
-- 
2.34.1

[PATCH v2 3/3] arm64: dts: qcom: sc7280: Add dma-coherent property for fastrpc nodes

[Ling Xu]

Add dma-cherent property to fastrpc context bank nodes to ensure that
the DMA operations for these nodes are coherent.

Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
---
 arch/arm64/boot/dts/qcom/sc7280.dtsi | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/arch/arm64/boot/dts/qcom/sc7280.dtsi b/arch/arm64/boot/dts/qcom/sc7280.dtsi
index 86b51cffcd54..b40e114d2ca8 100644
--- a/arch/arm64/boot/dts/qcom/sc7280.dtsi
+++ b/arch/arm64/boot/dts/qcom/sc7280.dtsi
@@ -3881,12 +3881,14 @@ compute-cb@3 {
 						compatible = "qcom,fastrpc-compute-cb";
 						reg = <3>;
 						iommus = <&apps_smmu 0x1803 0x0>;
+						dma-coherent;
 					};
 
 					compute-cb@4 {
 						compatible = "qcom,fastrpc-compute-cb";
 						reg = <4>;
 						iommus = <&apps_smmu 0x1804 0x0>;
+						dma-coherent;
 					};
 
 					compute-cb@5 {
@@ -3894,6 +3896,7 @@ compute-cb@5 {
 						reg = <5>;
 						iommus = <&apps_smmu 0x1805 0x0>;
 						qcom,nsessions = <5>;
+						dma-coherent;
 					};
 				};
 			};
@@ -4121,6 +4124,7 @@ compute-cb@1 {
 						reg = <1>;
 						iommus = <&apps_smmu 0x11a1 0x0420>,
 							 <&apps_smmu 0x1181 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@2 {
@@ -4128,6 +4132,7 @@ compute-cb@2 {
 						reg = <2>;
 						iommus = <&apps_smmu 0x11a2 0x0420>,
 							 <&apps_smmu 0x1182 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@3 {
@@ -4135,6 +4140,7 @@ compute-cb@3 {
 						reg = <3>;
 						iommus = <&apps_smmu 0x11a3 0x0420>,
 							 <&apps_smmu 0x1183 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@4 {
@@ -4142,6 +4148,7 @@ compute-cb@4 {
 						reg = <4>;
 						iommus = <&apps_smmu 0x11a4 0x0420>,
 							 <&apps_smmu 0x1184 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@5 {
@@ -4149,6 +4156,7 @@ compute-cb@5 {
 						reg = <5>;
 						iommus = <&apps_smmu 0x11a5 0x0420>,
 							 <&apps_smmu 0x1185 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@6 {
@@ -4156,6 +4164,7 @@ compute-cb@6 {
 						reg = <6>;
 						iommus = <&apps_smmu 0x11a6 0x0420>,
 							 <&apps_smmu 0x1186 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@7 {
@@ -4163,6 +4172,7 @@ compute-cb@7 {
 						reg = <7>;
 						iommus = <&apps_smmu 0x11a7 0x0420>,
 							 <&apps_smmu 0x1187 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@8 {
@@ -4170,6 +4180,7 @@ compute-cb@8 {
 						reg = <8>;
 						iommus = <&apps_smmu 0x11a8 0x0420>,
 							 <&apps_smmu 0x1188 0x0420>;
+						dma-coherent;
 					};
 
 					/* note: secure cb9 in downstream */
@@ -4179,6 +4190,7 @@ compute-cb@11 {
 						reg = <11>;
 						iommus = <&apps_smmu 0x11ab 0x0420>,
 							 <&apps_smmu 0x118b 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@12 {
@@ -4186,6 +4198,7 @@ compute-cb@12 {
 						reg = <12>;
 						iommus = <&apps_smmu 0x11ac 0x0420>,
 							 <&apps_smmu 0x118c 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@13 {
@@ -4193,6 +4206,7 @@ compute-cb@13 {
 						reg = <13>;
 						iommus = <&apps_smmu 0x11ad 0x0420>,
 							 <&apps_smmu 0x118d 0x0420>;
+						dma-coherent;
 					};
 
 					compute-cb@14 {
@@ -4200,6 +4214,7 @@ compute-cb@14 {
 						reg = <14>;
 						iommus = <&apps_smmu 0x11ae 0x0420>,
 							 <&apps_smmu 0x118e 0x0420>;
+						dma-coherent;
 					};
 				};
 			};
-- 
2.34.1

Re: [PATCH v2 3/3] arm64: dts: qcom: sc7280: Add dma-coherent property for fastrpc nodes

[Dmitry Baryshkov]

On Fri, May 16, 2025 at 04:30:29PM +0530, Ling Xu wrote:
> Add dma-cherent property to fastrpc context bank nodes to ensure that
> the DMA operations for these nodes are coherent.

Does it apply to all firmware versions?

> 
> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
> ---
>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 15 +++++++++++++++
>  1 file changed, 15 insertions(+)
> 

-- 
With best wishes
Dmitry

Re: [PATCH v2 3/3] arm64: dts: qcom: sc7280: Add dma-coherent property for fastrpc nodes

[Ling Xu]

在 5/17/2025 5:48 AM, Dmitry Baryshkov 写道:
> On Fri, May 16, 2025 at 04:30:29PM +0530, Ling Xu wrote:
>> Add dma-cherent property to fastrpc context bank nodes to ensure that
>> the DMA operations for these nodes are coherent.
> 
> Does it apply to all firmware versions?

No, it's not applicable for all the firmwares, especifically for older one.
This is only applicable where IO coherency is enabled.
>>
>> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
>> ---
>>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 15 +++++++++++++++
>>  1 file changed, 15 insertions(+)
>>
> 

-- 
Thx and BRs,
Ling Xu

Re: [PATCH v2 3/3] arm64: dts: qcom: sc7280: Add dma-coherent property for fastrpc nodes

[Dmitry Baryshkov]

On Tue, May 20, 2025 at 02:47:45PM +0800, Ling Xu wrote:
> 在 5/17/2025 5:48 AM, Dmitry Baryshkov 写道:
> > On Fri, May 16, 2025 at 04:30:29PM +0530, Ling Xu wrote:
> >> Add dma-cherent property to fastrpc context bank nodes to ensure that
> >> the DMA operations for these nodes are coherent.
> > 
> > Does it apply to all firmware versions?
> 
> No, it's not applicable for all the firmwares, especifically for older one.
> This is only applicable where IO coherency is enabled.

There are enough platforms using sc7280.dtsi. Which of those have new
firmware? Will this break FP5 phone? Herobrine laptops? Nothing Phone?

> >>
> >> Signed-off-by: Ling Xu <quic_lxu5@quicinc.com>
> >> ---
> >>  arch/arm64/boot/dts/qcom/sc7280.dtsi | 15 +++++++++++++++
> >>  1 file changed, 15 insertions(+)
> >>
> > 
> 
> -- 
> Thx and BRs,
> Ling Xu
> 

-- 
With best wishes
Dmitry