[PATCH 0/4] Bring the IP0 resource from ICC to clk-rpmh on QCS615

[PATCH 0/4] Bring the IP0 resource from ICC to clk-rpmh on QCS615

[Konrad Dybcio]

This is how it should have been right from the beginning, but it seems
like nobody took a good look at the initial submission.

This series is essentially the same as the previous [1] again.

Compile-tested only.

[1] https://lore.kernel.org/all/20230109002935.244320-1-dmitry.baryshkov@linaro.org/

Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
Konrad Dybcio (4):
      arm64: dts: qcom: qcs615: Drop IPA interconnects
      interconnect: qcom: qcs615: Drop IP0 interconnects
      dt-bindings: interconnect: qcom,qcs615-rpmh: Drop IPA interconnects
      clk: qcom: rpmh: Define RPMH_IPA_CLK on QCS615

 .../bindings/interconnect/qcom,qcs615-rpmh.yaml    |  2 --
 arch/arm64/boot/dts/qcom/qcs615.dtsi               |  6 ----
 drivers/clk/qcom/clk-rpmh.c                        |  1 +
 drivers/interconnect/qcom/qcs615.c                 | 42 ----------------------
 4 files changed, 1 insertion(+), 50 deletions(-)
---
base-commit: 2aeda9592360c200085898a258c4754bfe879921
change-id: 20250627-topic-qcs615_icc_ipa-919124eeff22

Best regards,
-- 
Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

[PATCH 1/4] arm64: dts: qcom: qcs615: Drop IPA interconnects

[Konrad Dybcio]

From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

In the same spirit as e.g. Commit 6314184be391 ("arm64: dts: qcom:
sc8180x: Drop ipa-virt interconnect") drop the resources that should
be taken care of through the clk-rpmh driver.

Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
 arch/arm64/boot/dts/qcom/qcs615.dtsi | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/arch/arm64/boot/dts/qcom/qcs615.dtsi b/arch/arm64/boot/dts/qcom/qcs615.dtsi
index bfbb210354922766a03fe05e6d117ea21d118081..b554471175de88fd96d12dd5183ee9bf61532bbf 100644
--- a/arch/arm64/boot/dts/qcom/qcs615.dtsi
+++ b/arch/arm64/boot/dts/qcom/qcs615.dtsi
@@ -320,12 +320,6 @@ camnoc_virt: interconnect-0 {
 		qcom,bcm-voters = <&apps_bcm_voter>;
 	};
 
-	ipa_virt: interconnect-1 {
-		compatible = "qcom,qcs615-ipa-virt";
-		#interconnect-cells = <2>;
-		qcom,bcm-voters = <&apps_bcm_voter>;
-	};
-
 	mc_virt: interconnect-2 {
 		compatible = "qcom,qcs615-mc-virt";
 		#interconnect-cells = <2>;

-- 
2.50.0

Re: [PATCH 1/4] arm64: dts: qcom: qcs615: Drop IPA interconnects

[Dmitry Baryshkov]

On Fri, Jun 27, 2025 at 09:37:55PM +0200, Konrad Dybcio wrote:
> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> In the same spirit as e.g. Commit 6314184be391 ("arm64: dts: qcom:
> sc8180x: Drop ipa-virt interconnect") drop the resources that should
> be taken care of through the clk-rpmh driver.
> 
> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> ---
>  arch/arm64/boot/dts/qcom/qcs615.dtsi | 6 ------
>  1 file changed, 6 deletions(-)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry

Re: [PATCH 1/4] arm64: dts: qcom: qcs615: Drop IPA interconnects

[Krzysztof Kozlowski]

On Fri, Jun 27, 2025 at 09:37:55PM +0200, Konrad Dybcio wrote:
> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> In the same spirit as e.g. Commit 6314184be391 ("arm64: dts: qcom:
> sc8180x: Drop ipa-virt interconnect") drop the resources that should
> be taken care of through the clk-rpmh driver.
> 
> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> ---
>  arch/arm64/boot/dts/qcom/qcs615.dtsi | 6 ------

This is non-bisectable and breaks the users. Should be the last patch,
not the first, with explicit notice that it must wait after the next
merge window.

Plus this affects all outside users, which also should be somehow
explained. "Spirit" of some other commit does not apply here. These are
different products, different users. It needs its own rationale and
justification.

Best regards,
Krzysztof

Re: [PATCH 1/4] arm64: dts: qcom: qcs615: Drop IPA interconnects

[Konrad Dybcio]

On 7/8/25 10:51 AM, Krzysztof Kozlowski wrote:
> On Fri, Jun 27, 2025 at 09:37:55PM +0200, Konrad Dybcio wrote:
>> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
>>
>> In the same spirit as e.g. Commit 6314184be391 ("arm64: dts: qcom:
>> sc8180x: Drop ipa-virt interconnect") drop the resources that should
>> be taken care of through the clk-rpmh driver.
>>
>> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
>> ---
>>  arch/arm64/boot/dts/qcom/qcs615.dtsi | 6 ------
> 
> This is non-bisectable and breaks the users.

Of which there are clearly none, as can be seen by the node not
being referenced by any other one. I can add that to the commit
message.

This is actually as bisectable as it gets, since it goes out of
the way to satisfy both compilation and dt-bindings checks

> Should be the last patch,> not the first, with explicit notice that it must wait after the next
> merge window.

Patches to all three subsystems can go in independently to their
respective trees (except maybe icc bindings going in before dts
could spawn a dt-checker warning)

> Plus this affects all outside users, which also should be somehow
> explained. "Spirit" of some other commit does not apply here. These are
> different products, different users. It needs its own rationale and
> justification.

Yes and no, you're focusing on me referencing a different commit
while I included the same rationale it presents later in the
sentence.

Konrad

[PATCH 2/4] interconnect: qcom: qcs615: Drop IP0 interconnects

[Konrad Dybcio]

From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

In the same spirit as e.g. Commit b136d257ee0b ("interconnect: qcom:
sc8280xp: Drop IP0 interconnects"), drop the resources that should be
taken care of through the clk-rpmh driver.

Fixes: 77d79677b04b ("interconnect: qcom: add QCS615 interconnect provider driver")
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
 drivers/interconnect/qcom/qcs615.c | 42 --------------------------------------
 1 file changed, 42 deletions(-)

diff --git a/drivers/interconnect/qcom/qcs615.c b/drivers/interconnect/qcom/qcs615.c
index 7e59e91ce886d641599a780b0f0d56a9e64b7de4..0549cfcbac64e4284ae5df1de3683985fffdcac8 100644
--- a/drivers/interconnect/qcom/qcs615.c
+++ b/drivers/interconnect/qcom/qcs615.c
@@ -342,15 +342,6 @@ static struct qcom_icc_node qnm_snoc_sf = {
 	.links = { QCS615_SLAVE_LLCC },
 };
 
-static struct qcom_icc_node ipa_core_master = {
-	.name = "ipa_core_master",
-	.id = QCS615_MASTER_IPA_CORE,
-	.channels = 1,
-	.buswidth = 8,
-	.num_links = 1,
-	.links = { QCS615_SLAVE_IPA_CORE },
-};
-
 static struct qcom_icc_node llcc_mc = {
 	.name = "llcc_mc",
 	.id = QCS615_MASTER_LLCC,
@@ -942,14 +933,6 @@ static struct qcom_icc_node srvc_gemnoc = {
 	.num_links = 0,
 };
 
-static struct qcom_icc_node ipa_core_slave = {
-	.name = "ipa_core_slave",
-	.id = QCS615_SLAVE_IPA_CORE,
-	.channels = 1,
-	.buswidth = 8,
-	.num_links = 0,
-};
-
 static struct qcom_icc_node ebi = {
 	.name = "ebi",
 	.id = QCS615_SLAVE_EBI1,
@@ -1113,12 +1096,6 @@ static struct qcom_icc_bcm bcm_cn1 = {
 		   &qhs_sdc1, &qhs_sdc2 },
 };
 
-static struct qcom_icc_bcm bcm_ip0 = {
-	.name = "IP0",
-	.num_nodes = 1,
-	.nodes = { &ipa_core_slave },
-};
-
 static struct qcom_icc_bcm bcm_mc0 = {
 	.name = "MC0",
 	.keepalive = true,
@@ -1260,7 +1237,6 @@ static struct qcom_icc_bcm * const aggre1_noc_bcms[] = {
 	&bcm_qup0,
 	&bcm_sn3,
 	&bcm_sn14,
-	&bcm_ip0,
 };
 
 static struct qcom_icc_node * const aggre1_noc_nodes[] = {
@@ -1411,22 +1387,6 @@ static const struct qcom_icc_desc qcs615_gem_noc = {
 	.num_bcms = ARRAY_SIZE(gem_noc_bcms),
 };
 
-static struct qcom_icc_bcm * const ipa_virt_bcms[] = {
-	&bcm_ip0,
-};
-
-static struct qcom_icc_node * const ipa_virt_nodes[] = {
-	[MASTER_IPA_CORE] = &ipa_core_master,
-	[SLAVE_IPA_CORE] = &ipa_core_slave,
-};
-
-static const struct qcom_icc_desc qcs615_ipa_virt = {
-	.nodes = ipa_virt_nodes,
-	.num_nodes = ARRAY_SIZE(ipa_virt_nodes),
-	.bcms = ipa_virt_bcms,
-	.num_bcms = ARRAY_SIZE(ipa_virt_bcms),
-};
-
 static struct qcom_icc_bcm * const mc_virt_bcms[] = {
 	&bcm_acv,
 	&bcm_mc0,
@@ -1525,8 +1485,6 @@ static const struct of_device_id qnoc_of_match[] = {
 	  .data = &qcs615_dc_noc},
 	{ .compatible = "qcom,qcs615-gem-noc",
 	  .data = &qcs615_gem_noc},
-	{ .compatible = "qcom,qcs615-ipa-virt",
-	  .data = &qcs615_ipa_virt},
 	{ .compatible = "qcom,qcs615-mc-virt",
 	  .data = &qcs615_mc_virt},
 	{ .compatible = "qcom,qcs615-mmss-noc",

-- 
2.50.0

Re: [PATCH 2/4] interconnect: qcom: qcs615: Drop IP0 interconnects

[Dmitry Baryshkov]

On Fri, Jun 27, 2025 at 09:37:56PM +0200, Konrad Dybcio wrote:
> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> In the same spirit as e.g. Commit b136d257ee0b ("interconnect: qcom:
> sc8280xp: Drop IP0 interconnects"), drop the resources that should be
> taken care of through the clk-rpmh driver.
> 
> Fixes: 77d79677b04b ("interconnect: qcom: add QCS615 interconnect provider driver")
> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> ---
>  drivers/interconnect/qcom/qcs615.c | 42 --------------------------------------
>  1 file changed, 42 deletions(-)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry

[PATCH 3/4] dt-bindings: interconnect: qcom,qcs615-rpmh: Drop IPA interconnects

[Konrad Dybcio]

From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

This has been agreed to be characterized as a clock resource, not an
interconnect provider. Bring QCS615 in line with the expectation.

Fixes: 6c5e948f1fff ("dt-bindings: interconnect: document the RPMh Network-On-Chip interconnect in QCS615 SoC")
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
 Documentation/devicetree/bindings/interconnect/qcom,qcs615-rpmh.yaml | 2 --
 1 file changed, 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/interconnect/qcom,qcs615-rpmh.yaml b/Documentation/devicetree/bindings/interconnect/qcom,qcs615-rpmh.yaml
index 9d762b2a1fcf87b638a0b9060e7b90579e329cef..e0640482882481408095ee1acac5c824fa8820e0 100644
--- a/Documentation/devicetree/bindings/interconnect/qcom,qcs615-rpmh.yaml
+++ b/Documentation/devicetree/bindings/interconnect/qcom,qcs615-rpmh.yaml
@@ -27,7 +27,6 @@ properties:
       - qcom,qcs615-config-noc
       - qcom,qcs615-dc-noc
       - qcom,qcs615-gem-noc
-      - qcom,qcs615-ipa-virt
       - qcom,qcs615-mc-virt
       - qcom,qcs615-mmss-noc
       - qcom,qcs615-system-noc
@@ -46,7 +45,6 @@ allOf:
           contains:
             enum:
               - qcom,qcs615-camnoc-virt
-              - qcom,qcs615-ipa-virt
               - qcom,qcs615-mc-virt
     then:
       properties:

-- 
2.50.0

Re: [PATCH 3/4] dt-bindings: interconnect: qcom,qcs615-rpmh: Drop IPA interconnects

[Krzysztof Kozlowski]

On Fri, Jun 27, 2025 at 09:37:57PM +0200, Konrad Dybcio wrote:
> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> This has been agreed to be characterized as a clock resource, not an
> interconnect provider. Bring QCS615 in line with the expectation.
> 
> Fixes: 6c5e948f1fff ("dt-bindings: interconnect: document the RPMh Network-On-Chip interconnect in QCS615 SoC")

You did not explain any bug here to fix. Difference in agreements in
characterization does not feel like a bug.

Best regards,
Krzysztof

[PATCH 4/4] clk: qcom: rpmh: Define RPMH_IPA_CLK on QCS615

[Konrad Dybcio]

From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

This was previously (mis)represented in the interconnect driver, move
the resource under the clk-rpmh driver control, just like we did for
all platforms in the past, see e.g. Commit aa055bf158cd ("clk: qcom:
rpmh: define IPA clocks where required")

Fixes: 42a1905a10d6 ("clk: qcom: rpmhcc: Add support for QCS615 Clocks")
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
 drivers/clk/qcom/clk-rpmh.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/clk/qcom/clk-rpmh.c b/drivers/clk/qcom/clk-rpmh.c
index 00fb3e53a388ed24ed76622983eb5bd81a6b7002..e02346bccdc6f26b1a832ed62dbfdc35f271c858 100644
--- a/drivers/clk/qcom/clk-rpmh.c
+++ b/drivers/clk/qcom/clk-rpmh.c
@@ -825,6 +825,7 @@ static struct clk_hw *qcs615_rpmh_clocks[] = {
 	[RPMH_RF_CLK1_A]	= &clk_rpmh_rf_clk1_a_ao.hw,
 	[RPMH_RF_CLK2]		= &clk_rpmh_rf_clk2_a.hw,
 	[RPMH_RF_CLK2_A]	= &clk_rpmh_rf_clk2_a_ao.hw,
+	[RPMH_IPA_CLK]		= &clk_rpmh_ipa.hw,
 };
 
 static const struct clk_rpmh_desc clk_rpmh_qcs615 = {

-- 
2.50.0

Re: [PATCH 4/4] clk: qcom: rpmh: Define RPMH_IPA_CLK on QCS615

[Dmitry Baryshkov]

On Fri, Jun 27, 2025 at 09:37:58PM +0200, Konrad Dybcio wrote:
> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> This was previously (mis)represented in the interconnect driver, move
> the resource under the clk-rpmh driver control, just like we did for
> all platforms in the past, see e.g. Commit aa055bf158cd ("clk: qcom:
> rpmh: define IPA clocks where required")
> 
> Fixes: 42a1905a10d6 ("clk: qcom: rpmhcc: Add support for QCS615 Clocks")
> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> ---
>  drivers/clk/qcom/clk-rpmh.c | 1 +
>  1 file changed, 1 insertion(+)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry