This patch series adds support for limiting the maximum high-speed
gear and rate used by the UFS controller via device tree properties.

Some platforms may have signal integrity, clock configuration, or
layout issues that prevent reliable operation at higher gears or rates.
This is especially critical in automotive and other platforms where
stability is prioritized over peak performance.

The series follows this logical progression:
1. Document the new DT properties in the common UFS binding
2. Clean up existing redundant code in the qcom driver
3. Add platform-level parsing support for the new properties
4. Integrate the platform support in the qcom driver

This approach makes the functionality available to other UFS host
drivers and provides a cleaner, more maintainable implementation.

Changes from V1: 
- Restructured patch series for better logical flow and maintainability.
- Moved DT bindings to ufs-common.yaml making it available for all UFS 
  controllers.
- Added platform-level support in ufshcd-pltfrm.c for code reusability.
- Separated the cleanup patch to remove redundant hs_rate assignment in
  qcom driver.
- Removed SA8155 DTS changes to keep the series focused on core
  functionality.
- Improved commit messages with better technical rationale.

Changes from V2: 
- Documented default values of limit-rate and limit-hs-gear in DT bindings
  as per Krzysztof's suggestion.

Changes from V3: 
- Changed limit-rate property from numeric values 1 and 2 to string values
  Rate-A and Rate-B for better readability and clarity as suggested by
  Bart and Krzysztof.
- Added Co-developed-by tag for Nitin Rawat in 3rd patch. 

Changes from V4:
- Added the missing argument to the error message while parsing
  limit-rate property.
- Updated the maximum supported value and default for limit-gear
  property to gear 6, as per Krzysztof's and Bart's recommendation.
- Renamed Rate-A and Rate-B to lowercase (rate-a, rate-b) as suggested
  by Krzysztof.
 
Ram Kumar Dwivedi (4):
  ufs: dt-bindings: Document gear and rate limit properties
  ufs: ufs-qcom: Remove redundant re-assignment to hs_rate
  ufs: pltfrm: Allow limiting HS gear and rate via DT
  ufs: ufs-qcom: Add support for limiting HS gear and rate

 .../devicetree/bindings/ufs/ufs-common.yaml   | 16 +++++++++
 drivers/ufs/host/ufs-qcom.c                   | 21 +++++++----
 drivers/ufs/host/ufshcd-pltfrm.c              | 36 +++++++++++++++++++
 drivers/ufs/host/ufshcd-pltfrm.h              |  1 +
 4 files changed, 68 insertions(+), 6 deletions(-)

-- 
2.50.1

Hi Ram

> -----Original Message-----
> From: Ram Kumar Dwivedi <quic_rdwivedi@quicinc.com>
> Sent: Tuesday, September 2, 2025 10:19 PM
> To: alim.akhtar@samsung.com; avri.altman@wdc.com;
> bvanassche@acm.org; robh@kernel.org; krzk+dt@kernel.org;
> conor+dt@kernel.org; mani@kernel.org;
> James.Bottomley@HansenPartnership.com; martin.petersen@oracle.com
> Cc: linux-scsi@vger.kernel.org; devicetree@vger.kernel.org; linux-
> kernel@vger.kernel.org; linux-arm-msm@vger.kernel.org
> Subject: [PATCH V5 0/4] Add DT-based gear and rate limiting support
> 
> This patch series adds support for limiting the maximum high-speed gear
and
> rate used by the UFS controller via device tree properties.
> 
> Some platforms may have signal integrity, clock configuration, or layout
> issues that prevent reliable operation at higher gears or rates.
> This is especially critical in automotive and other platforms where
stability is
> prioritized over peak performance.
> 
> The series follows this logical progression:
> 1. Document the new DT properties in the common UFS binding 2. Clean up
> existing redundant code in the qcom driver 3. Add platform-level parsing
> support for the new properties 4. Integrate the platform support in the
qcom
> driver
> 
> This approach makes the functionality available to other UFS host drivers
and
> provides a cleaner, more maintainable implementation.
> 
> Changes from V1:
> - Restructured patch series for better logical flow and maintainability.
> - Moved DT bindings to ufs-common.yaml making it available for all UFS
>   controllers.
> - Added platform-level support in ufshcd-pltfrm.c for code reusability.
> - Separated the cleanup patch to remove redundant hs_rate assignment in
>   qcom driver.
> - Removed SA8155 DTS changes to keep the series focused on core
>   functionality.
> - Improved commit messages with better technical rationale.
> 
> Changes from V2:
> - Documented default values of limit-rate and limit-hs-gear in DT bindings
>   as per Krzysztof's suggestion.
> 
> Changes from V3:
> - Changed limit-rate property from numeric values 1 and 2 to string values
>   Rate-A and Rate-B for better readability and clarity as suggested by
>   Bart and Krzysztof.
> - Added Co-developed-by tag for Nitin Rawat in 3rd patch.
> 
> Changes from V4:
> - Added the missing argument to the error message while parsing
>   limit-rate property.
> - Updated the maximum supported value and default for limit-gear
>   property to gear 6, as per Krzysztof's and Bart's recommendation.
> - Renamed Rate-A and Rate-B to lowercase (rate-a, rate-b) as suggested
>   by Krzysztof.
> 
Please allow minimum 4 ~ 5 days for reviewers to complete the review before
posting next version.
That will also help to reduce the number of iteration a patch goes through.
Thanks

On 03-Sep-25 10:18 AM, Alim Akhtar wrote:
> Hi Ram
> 
>> -----Original Message-----
>> From: Ram Kumar Dwivedi <quic_rdwivedi@quicinc.com>
>> Sent: Tuesday, September 2, 2025 10:19 PM
>> To: alim.akhtar@samsung.com; avri.altman@wdc.com;
>> bvanassche@acm.org; robh@kernel.org; krzk+dt@kernel.org;
>> conor+dt@kernel.org; mani@kernel.org;
>> James.Bottomley@HansenPartnership.com; martin.petersen@oracle.com
>> Cc: linux-scsi@vger.kernel.org; devicetree@vger.kernel.org; linux-
>> kernel@vger.kernel.org; linux-arm-msm@vger.kernel.org
>> Subject: [PATCH V5 0/4] Add DT-based gear and rate limiting support
>>
>> This patch series adds support for limiting the maximum high-speed gear
> and
>> rate used by the UFS controller via device tree properties.
>>
>> Some platforms may have signal integrity, clock configuration, or layout
>> issues that prevent reliable operation at higher gears or rates.
>> This is especially critical in automotive and other platforms where
> stability is
>> prioritized over peak performance.
>>
>> The series follows this logical progression:
>> 1. Document the new DT properties in the common UFS binding 2. Clean up
>> existing redundant code in the qcom driver 3. Add platform-level parsing
>> support for the new properties 4. Integrate the platform support in the
> qcom
>> driver
>>
>> This approach makes the functionality available to other UFS host drivers
> and
>> provides a cleaner, more maintainable implementation.
>>
>> Changes from V1:
>> - Restructured patch series for better logical flow and maintainability.
>> - Moved DT bindings to ufs-common.yaml making it available for all UFS
>>   controllers.
>> - Added platform-level support in ufshcd-pltfrm.c for code reusability.
>> - Separated the cleanup patch to remove redundant hs_rate assignment in
>>   qcom driver.
>> - Removed SA8155 DTS changes to keep the series focused on core
>>   functionality.
>> - Improved commit messages with better technical rationale.
>>
>> Changes from V2:
>> - Documented default values of limit-rate and limit-hs-gear in DT bindings
>>   as per Krzysztof's suggestion.
>>
>> Changes from V3:
>> - Changed limit-rate property from numeric values 1 and 2 to string values
>>   Rate-A and Rate-B for better readability and clarity as suggested by
>>   Bart and Krzysztof.
>> - Added Co-developed-by tag for Nitin Rawat in 3rd patch.
>>
>> Changes from V4:
>> - Added the missing argument to the error message while parsing
>>   limit-rate property.
>> - Updated the maximum supported value and default for limit-gear
>>   property to gear 6, as per Krzysztof's and Bart's recommendation.
>> - Renamed Rate-A and Rate-B to lowercase (rate-a, rate-b) as suggested
>>   by Krzysztof.
>>
> Please allow minimum 4 ~ 5 days for reviewers to complete the review before
> posting next version.
> That will also help to reduce the number of iteration a patch goes through.
> Thanks

Hi Alim,

Thanks for input. I’ll make sure to allow at least 4–5 days for the 
reviewers to complete their review before posting the next version.

Thanks,
Ram.> 
> 
>