[PATCH 0/2] Shikra: Fix qce configuration

[PATCH 0/2] Shikra: Fix qce configuration

[Kuldeep Singh]

This patch series fixes QCE on Shikra.
Shikra is derived from agatti target and QCE was using different
compatible configuration.
Also, one more reason to update QCE is, clk-smd-rpm was doing proxy
voting and crypto was consuming it. With proxy vote removal, qce need to
vote for it's own clocks explicitly which is already being done for
agatti target. Adapt same mechanism to Shikra QCE too.

The current qce configuration will still work but this one is more
appropriately describing the hardware.

Validations:
- Qce probe and kcapi_convenience.sh tests.
- Correct clk_summary output as prepare_count and enable_count are 0.

The previous patch-seriest[1] was sent to an older thread and with feedback
received, submitting a new series.

[1] https://lore.kernel.org/lkml/20260907-b4-shikra_crypto_changse-v6-0-0676f61894b3@oss.qualcomm.com/

Signed-off-by: Kuldeep Singh <kuldeep.singh@oss.qualcomm.com>
---
Kuldeep Singh (2):
      dt-bindings: crypto: Adjust Shikra compatible hierarchy
      arm64: dts: qcom: shikra: Correct QCE node configuration

 Documentation/devicetree/bindings/crypto/qcom-qce.yaml | 3 ++-
 arch/arm64/boot/dts/qcom/shikra.dtsi                   | 7 +++----
 2 files changed, 5 insertions(+), 5 deletions(-)
---
base-commit: bf1ee2bd5c2da8992f33c8950ef194aaff74ed6c
change-id: 20261006-shikra_qce_fix-270b735479ad

Best regards,
--  
Kuldeep Singh <kuldeep.singh@oss.qualcomm.com>

[PATCH 1/2] dt-bindings: crypto: Adjust Shikra compatible hierarchy

[Kuldeep Singh]

Shikra reuses the same QCE integration as Agatti and should follow the
corresponding binding requirements too.

Move the Shikra compatible to the appropriate binding hierarchy
(matching qcom,qcm2290-qce) to reflect the underlying hardware
description.

Fixes: 45834ff95a6f ("dt-bindings: crypto: qcom-qce: Document the Shikra crypto engine")
Signed-off-by: Kuldeep Singh <kuldeep.singh@oss.qualcomm.com>
Reviewed-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
---
 Documentation/devicetree/bindings/crypto/qcom-qce.yaml | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/crypto/qcom-qce.yaml b/Documentation/devicetree/bindings/crypto/qcom-qce.yaml
index 5a653757ee75..fcf725bfa36b 100644
--- a/Documentation/devicetree/bindings/crypto/qcom-qce.yaml
+++ b/Documentation/devicetree/bindings/crypto/qcom-qce.yaml
@@ -39,6 +39,7 @@ properties:
               - qcom,msm8996-qce
               - qcom,qcm2290-qce
               - qcom,sdm845-qce
+              - qcom,shikra-qce
               - qcom,sm6115-qce
           - const: qcom,ipq4019-qce
           - const: qcom,qce
@@ -54,7 +55,6 @@ properties:
               - qcom,qcs8300-qce
               - qcom,sa8775p-qce
               - qcom,sc7280-qce
-              - qcom,shikra-qce
               - qcom,sm6350-qce
               - qcom,sm8250-qce
               - qcom,sm8350-qce
@@ -133,6 +133,7 @@ allOf:
           contains:
             enum:
               - qcom,qcm2290-qce
+              - qcom,shikra-qce
               - qcom,sm6115-qce
     then:
       properties:

-- 
2.34.1

Re: [PATCH 1/2] dt-bindings: crypto: Adjust Shikra compatible hierarchy

[Krzysztof Kozlowski]

On Tue, Oct 06, 2026 at 04:44:03PM +0530, Kuldeep Singh wrote:
> Shikra reuses the same QCE integration as Agatti and should follow the

`git grep Agatti` gives zero.

> corresponding binding requirements too.
> 
> Move the Shikra compatible to the appropriate binding hierarchy
> (matching qcom,qcm2290-qce) to reflect the underlying hardware
> description.

You do not change any binding hierarchy. You are changing fallback
compatible which is ABI break without any reason.

> 
> Fixes: 45834ff95a6f ("dt-bindings: crypto: qcom-qce: Document the Shikra crypto engine")
> Signed-off-by: Kuldeep Singh <kuldeep.singh@oss.qualcomm.com>
> Reviewed-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>

Drop, that's v1, so this likely did not happen. Anyway, your SoB is
supposed to be last.

Best regards,
Krzysztof

Re: [PATCH 1/2] dt-bindings: crypto: Adjust Shikra compatible hierarchy

[Kuldeep Singh]

On 13-10-2026 12:14, Krzysztof Kozlowski wrote:
> On Tue, Oct 06, 2026 at 04:44:03PM +0530, Kuldeep Singh wrote:
>> Shikra reuses the same QCE integration as Agatti and should follow the
> 
> `git grep Agatti` gives zero.

Agatti is chipset name and compatible name is "qcom,qcm2290-qce".
Maybe I'll mention both next time.

> 
>> corresponding binding requirements too.
>>
>> Move the Shikra compatible to the appropriate binding hierarchy
>> (matching qcom,qcm2290-qce) to reflect the underlying hardware
>> description.
> 
> You do not change any binding hierarchy. You are changing fallback
> compatible which is ABI break without any reason.

Agatti and shikra both consume RPMCC managed clocks and clk-smd-rpm.c
does proxy clocks voting as of now even when no consumer is voting for
it. The required consumer i.e QCE need to vote for it explicitly as
proxy voting being dropped now in patch [1].

As agatti(qcom,qcm2290-qce) specifies explicit clock entry as per the
spec[2] and shikra is derivative of agatti hence, follow the same
convention.

Coming to updating fallback compatible introducing ABI break,
"qcom,shikra-qce" compatible in bindings got accepted in v7.3[3] but not
the corresponding DTS change[4](which consume this compatible), the DTS
change is queued for v7.4[5].

So we can straightaway update the hierarchy then?
> 
>>
>> Fixes: 45834ff95a6f ("dt-bindings: crypto: qcom-qce: Document the Shikra crypto engine")
>> Signed-off-by: Kuldeep Singh <kuldeep.singh@oss.qualcomm.com>
>> Reviewed-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
> 
> Drop, that's v1, so this likely did not happen. Anyway, your SoB is
> supposed to be last.

Kindly check cover letter for more details, it's respin after receiving
feedback on previous series.
Quoting here again:
"
The previous patch-series[1] was sent to an older thread and with feedback
received, submitting a new series.

[1]
https://lore.kernel.org/lkml/20260907-b4-shikra_crypto_changse-v6-0-0676f61894b3@oss.qualcomm.com/
"

Seems using b4 shazam automatically add tags below SoB.
Let me fix it next time.

-- 
Regards
Kuldeep

[1]
https://lore.kernel.org/linux-arm-msm/20260910-clk-smd-rpm-skip-proxy-v1-1-1cb5694a99d9@oss.qualcomm.com/
[2]
https://github.com/torvalds/linux/blob/master/Documentation/devicetree/bindings/crypto/qcom-qce.yaml#L130
[3]
https://github.com/torvalds/linux/blob/v7.3-rc7/Documentation/devicetree/bindings/crypto/qcom-qce.yaml#L57
[4]
https://github.com/torvalds/linux/blob/v7.3-rc7/arch/arm64/boot/dts/qcom/shikra.dtsi
[5]
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/arch/arm64/boot/dts/qcom/shikra.dtsi?id=099cd05d200ba241bcc7497c4384b8d657a5632e

[PATCH 2/2] arm64: dts: qcom: shikra: Correct QCE node configuration

[Kuldeep Singh]

Shikra is derived from Agatti and uses the same QCE integration.
Update the QCE node to match the underlying hardware implementation by
adjusting the fallback compatible and related properties, including
clocks and clock-names.

Since cryptobam is marked qcom,controlled-remotely, its clocks are
managed remotely and hence don't specify explicitly.

Fixes: 099cd05d200b ("arm64: dts: qcom: shikra: Add ICE, TRNG and QCE nodes")
Tested-by: Wenjia Zhang <wenjia.zhang@oss.qualcomm.com> #on shikra device
Signed-off-by: Kuldeep Singh <kuldeep.singh@oss.qualcomm.com>
---
 arch/arm64/boot/dts/qcom/shikra.dtsi | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/arch/arm64/boot/dts/qcom/shikra.dtsi b/arch/arm64/boot/dts/qcom/shikra.dtsi
index 205e70ce0748..b20e4ce13d08 100644
--- a/arch/arm64/boot/dts/qcom/shikra.dtsi
+++ b/arch/arm64/boot/dts/qcom/shikra.dtsi
@@ -1266,7 +1266,7 @@ cryptobam: dma-controller@1b04000 {
 		};
 
 		crypto: crypto@1b3a000 {
-			compatible = "qcom,shikra-qce", "qcom,sm8150-qce", "qcom,qce";
+			compatible = "qcom,shikra-qce", "qcom,ipq4019-qce", "qcom,qce";
 			reg = <0x0 0x01b3a000 0x0 0x6000>;
 			dmas = <&cryptobam 4>, <&cryptobam 5>;
 			dma-names = "rx", "tx";
@@ -1277,9 +1277,8 @@ crypto: crypto@1b3a000 {
 				 <&apps_smmu 0x96 0x0011>,
 				 <&apps_smmu 0x98 0x0001>,
 				 <&apps_smmu 0x9f 0x0>;
-			interconnects = <&system_noc MASTER_CRYPTO_CORE0 0
-					 &mc_virt SLAVE_EBI_CH0 0>;
-			interconnect-names = "memory";
+			clocks = <&rpmcc RPM_SMD_CE1_CLK>;
+			clock-names = "core";
 		};
 
 		qfprom: efuse@1b44000 {

-- 
2.34.1

Re: [PATCH 2/2] arm64: dts: qcom: shikra: Correct QCE node configuration

[Konrad Dybcio]

On 10/6/26 1:14 PM, Kuldeep Singh wrote:
> Shikra is derived from Agatti and uses the same QCE integration.

This is useful, but..

> Update the QCE node to match the underlying hardware implementation by
> adjusting the fallback compatible and related properties, including
> clocks and clock-names.

because the compatible you switch to says ipq4019 and not "agatti" or
"qcm2290", the argument falls apart.

The real problem here is "the hardware needs a clock and it's currently
missing". The similarity to Agatti is irrelevant in this context, even
if true.

Konrad

Re: [PATCH 2/2] arm64: dts: qcom: shikra: Correct QCE node configuration

[Kuldeep Singh]

On 16-10-2026 18:01, Konrad Dybcio wrote:
> On 10/6/26 1:14 PM, Kuldeep Singh wrote:
>> Shikra is derived from Agatti and uses the same QCE integration.
> 
> This is useful, but..
> 
>> Update the QCE node to match the underlying hardware implementation by
>> adjusting the fallback compatible and related properties, including
>> clocks and clock-names.
> 
> because the compatible you switch to says ipq4019 and not "agatti" or
> "qcm2290", the argument falls apart.

Ok, then I'll update the commit message and avoid adding agatti context
for shikra changes.

-- 
Regards
Kuldeep