[PATCH 0/3] soc: qcom: ocmem: let the driver work again

[PATCH 0/3] soc: qcom: ocmem: let the driver work again

[Dmitry Baryshkov]

Commit d6edc31f3a68 ("clk: qcom: smd-rpm: Separate out interconnect bus
clocks") moved control over several RPM resources from the clk-smd-rpm
driver to the icc-rpm.c interconnect helpers. However it was unnoticed
that the OCMEM device also uses that clock. The clock is not required as
all OCMEM users (currently only the GPU) will vote on the OCMEM NOC,
bringing up the clock. Make it optional (and also apply several other
small fixes).

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
Dmitry Baryshkov (3):
      soc: qcom: ocmem: make the core clock optional
      soc: qcom: ocmem: register reasons for probe deferrals
      soc: qcom: ocmem: return -EPROBE_DEFER is ocmem is not available

 drivers/soc/qcom/ocmem.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)
---
base-commit: 8e5a478b6d6a5bb0a3d52147862b15e4d826af19
change-id: 20260323-ocmem-dfe2d207d0e9

Best regards,
-- 
With best wishes
Dmitry

[PATCH 1/3] soc: qcom: ocmem: make the core clock optional

[Dmitry Baryshkov]

OCMEM's core clock (aka RPM bus 2 clock) is being handled internally by
the interconnect driver. Corresponding clock has been dropped from the
SMD RPM clock driver. The users of the ocmem will vote on the ocmemnoc
interconnect paths, making sure that ocmem is on. Make the clock
optional, keeping it for compatibility with older DT.

Fixes: d6edc31f3a68 ("clk: qcom: smd-rpm: Separate out interconnect bus clocks")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 drivers/soc/qcom/ocmem.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
index 6a23f18b0281..dd46bb14b7be 100644
--- a/drivers/soc/qcom/ocmem.c
+++ b/drivers/soc/qcom/ocmem.c
@@ -308,7 +308,7 @@ static int ocmem_dev_probe(struct platform_device *pdev)
 	ocmem->dev = dev;
 	ocmem->config = device_get_match_data(dev);
 
-	ocmem->core_clk = devm_clk_get(dev, "core");
+	ocmem->core_clk = devm_clk_get_optional(dev, "core");
 	if (IS_ERR(ocmem->core_clk))
 		return dev_err_probe(dev, PTR_ERR(ocmem->core_clk),
 				     "Unable to get core clock\n");

-- 
2.47.3

Re: [PATCH 1/3] soc: qcom: ocmem: make the core clock optional

[Konrad Dybcio]

On 3/23/26 2:20 AM, Dmitry Baryshkov wrote:
> OCMEM's core clock (aka RPM bus 2 clock) is being handled internally by
> the interconnect driver. Corresponding clock has been dropped from the
> SMD RPM clock driver. The users of the ocmem will vote on the ocmemnoc
> interconnect paths, making sure that ocmem is on. Make the clock
> optional, keeping it for compatibility with older DT.
> 
> Fixes: d6edc31f3a68 ("clk: qcom: smd-rpm: Separate out interconnect bus clocks")
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

[PATCH 2/3] soc: qcom: ocmem: register reasons for probe deferrals

[Dmitry Baryshkov]

Instead of printing messages to the dmesg, let the message be recorded
as a reason for the OCMEM client deferral.

Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 drivers/soc/qcom/ocmem.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
index dd46bb14b7be..d57baa9cfa03 100644
--- a/drivers/soc/qcom/ocmem.c
+++ b/drivers/soc/qcom/ocmem.c
@@ -196,10 +196,10 @@ struct ocmem *of_get_ocmem(struct device *dev)
 	}
 
 	pdev = of_find_device_by_node(devnode->parent);
-	if (!pdev) {
-		dev_err(dev, "Cannot find device node %s\n", devnode->name);
-		return ERR_PTR(-EPROBE_DEFER);
-	}
+	if (!pdev)
+		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER,
+					     "Cannot find device node %s\n",
+					     devnode->name));
 
 	ocmem = platform_get_drvdata(pdev);
 	put_device(&pdev->dev);

-- 
2.47.3

Re: [PATCH 2/3] soc: qcom: ocmem: register reasons for probe deferrals

[Konrad Dybcio]

On 3/23/26 2:20 AM, Dmitry Baryshkov wrote:
> Instead of printing messages to the dmesg, let the message be recorded
> as a reason for the OCMEM client deferral.
> 
> Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---
>  drivers/soc/qcom/ocmem.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
> index dd46bb14b7be..d57baa9cfa03 100644
> --- a/drivers/soc/qcom/ocmem.c
> +++ b/drivers/soc/qcom/ocmem.c
> @@ -196,10 +196,10 @@ struct ocmem *of_get_ocmem(struct device *dev)
>  	}
>  
>  	pdev = of_find_device_by_node(devnode->parent);
> -	if (!pdev) {
> -		dev_err(dev, "Cannot find device node %s\n", devnode->name);
> -		return ERR_PTR(-EPROBE_DEFER);
> -	}
> +	if (!pdev)
> +		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER,
> +					     "Cannot find device node %s\n",
> +					     devnode->name));

dev_err_ptr_probe()

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

Re: [PATCH 2/3] soc: qcom: ocmem: register reasons for probe deferrals

[Dmitry Baryshkov]

On Mon, Mar 23, 2026 at 11:16:54AM +0100, Konrad Dybcio wrote:
> On 3/23/26 2:20 AM, Dmitry Baryshkov wrote:
> > Instead of printing messages to the dmesg, let the message be recorded
> > as a reason for the OCMEM client deferral.
> > 
> > Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
> > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> > ---
> >  drivers/soc/qcom/ocmem.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
> > index dd46bb14b7be..d57baa9cfa03 100644
> > --- a/drivers/soc/qcom/ocmem.c
> > +++ b/drivers/soc/qcom/ocmem.c
> > @@ -196,10 +196,10 @@ struct ocmem *of_get_ocmem(struct device *dev)
> >  	}
> >  
> >  	pdev = of_find_device_by_node(devnode->parent);
> > -	if (!pdev) {
> > -		dev_err(dev, "Cannot find device node %s\n", devnode->name);
> > -		return ERR_PTR(-EPROBE_DEFER);
> > -	}
> > +	if (!pdev)
> > +		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER,
> > +					     "Cannot find device node %s\n",
> > +					     devnode->name));
> 
> dev_err_ptr_probe()

Cool!

> 
> Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> Konrad

-- 
With best wishes
Dmitry

Re: [PATCH 2/3] soc: qcom: ocmem: register reasons for probe deferrals

[Brian Masney]

On Mon, Mar 23, 2026 at 03:20:58AM +0200, Dmitry Baryshkov wrote:
> Instead of printing messages to the dmesg, let the message be recorded
> as a reason for the OCMEM client deferral.
> 
> Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---
>  drivers/soc/qcom/ocmem.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
> index dd46bb14b7be..d57baa9cfa03 100644
> --- a/drivers/soc/qcom/ocmem.c
> +++ b/drivers/soc/qcom/ocmem.c
> @@ -196,10 +196,10 @@ struct ocmem *of_get_ocmem(struct device *dev)
>  	}
>  
>  	pdev = of_find_device_by_node(devnode->parent);
> -	if (!pdev) {
> -		dev_err(dev, "Cannot find device node %s\n", devnode->name);
> -		return ERR_PTR(-EPROBE_DEFER);
> -	}
> +	if (!pdev)
> +		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER,
> +					     "Cannot find device node %s\n",
> +					     devnode->name));

Reviewed-by: Brian Masney <bmasney@redhat.com>

dev_err_probe() was introduced in 2020 (a787e5400a1c) and this driver
was added to the tree prior to that.

Re: [PATCH 2/3] soc: qcom: ocmem: register reasons for probe deferrals

[Dmitry Baryshkov]

On Mon, Mar 23, 2026 at 07:07:40AM -0400, Brian Masney wrote:
> On Mon, Mar 23, 2026 at 03:20:58AM +0200, Dmitry Baryshkov wrote:
> > Instead of printing messages to the dmesg, let the message be recorded
> > as a reason for the OCMEM client deferral.
> > 
> > Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
> > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> > ---
> >  drivers/soc/qcom/ocmem.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
> > index dd46bb14b7be..d57baa9cfa03 100644
> > --- a/drivers/soc/qcom/ocmem.c
> > +++ b/drivers/soc/qcom/ocmem.c
> > @@ -196,10 +196,10 @@ struct ocmem *of_get_ocmem(struct device *dev)
> >  	}
> >  
> >  	pdev = of_find_device_by_node(devnode->parent);
> > -	if (!pdev) {
> > -		dev_err(dev, "Cannot find device node %s\n", devnode->name);
> > -		return ERR_PTR(-EPROBE_DEFER);
> > -	}
> > +	if (!pdev)
> > +		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER,
> > +					     "Cannot find device node %s\n",
> > +					     devnode->name));
> 
> Reviewed-by: Brian Masney <bmasney@redhat.com>
> 
> dev_err_probe() was introduced in 2020 (a787e5400a1c) and this driver
> was added to the tree prior to that.

Well... Fixes doesnt' require porting (or 1:1 porting).

-- 
With best wishes
Dmitry

[PATCH 3/3] soc: qcom: ocmem: return -EPROBE_DEFER is ocmem is not available

[Dmitry Baryshkov]

If OCMEM is declared in DT, it is expected that it is present and
handled by the driver. The GPU driver will ignore -ENODEV error, which
typically means that OCMEM isn't defined in DT. Let ocmem return
-EPROBE_DEFER if it supposed to be used, but it is not probed (yet).

Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 drivers/soc/qcom/ocmem.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
index d57baa9cfa03..c35a98984acc 100644
--- a/drivers/soc/qcom/ocmem.c
+++ b/drivers/soc/qcom/ocmem.c
@@ -203,10 +203,9 @@ struct ocmem *of_get_ocmem(struct device *dev)
 
 	ocmem = platform_get_drvdata(pdev);
 	put_device(&pdev->dev);
-	if (!ocmem) {
-		dev_err(dev, "Cannot get ocmem\n");
-		return ERR_PTR(-ENODEV);
-	}
+	if (!ocmem)
+		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER, "Cannot get ocmem\n"));
+
 	return ocmem;
 }
 EXPORT_SYMBOL_GPL(of_get_ocmem);

-- 
2.47.3

Re: [PATCH 3/3] soc: qcom: ocmem: return -EPROBE_DEFER is ocmem is not available

[Konrad Dybcio]

On 3/23/26 2:20 AM, Dmitry Baryshkov wrote:
> If OCMEM is declared in DT, it is expected that it is present and
> handled by the driver. The GPU driver will ignore -ENODEV error, which
> typically means that OCMEM isn't defined in DT. Let ocmem return
> -EPROBE_DEFER if it supposed to be used, but it is not probed (yet).
> 
> Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---
>  drivers/soc/qcom/ocmem.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/soc/qcom/ocmem.c b/drivers/soc/qcom/ocmem.c
> index d57baa9cfa03..c35a98984acc 100644
> --- a/drivers/soc/qcom/ocmem.c
> +++ b/drivers/soc/qcom/ocmem.c
> @@ -203,10 +203,9 @@ struct ocmem *of_get_ocmem(struct device *dev)
>  
>  	ocmem = platform_get_drvdata(pdev);
>  	put_device(&pdev->dev);
> -	if (!ocmem) {
> -		dev_err(dev, "Cannot get ocmem\n");
> -		return ERR_PTR(-ENODEV);
> -	}
> +	if (!ocmem)
> +		return ERR_PTR(dev_err_probe(dev, -EPROBE_DEFER, "Cannot get ocmem\n"));

dev_err_ptr_probe()

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

Re: [PATCH 0/3] soc: qcom: ocmem: let the driver work again

[Bjorn Andersson]

On Mon, 23 Mar 2026 03:20:56 +0200, Dmitry Baryshkov wrote:
> Commit d6edc31f3a68 ("clk: qcom: smd-rpm: Separate out interconnect bus
> clocks") moved control over several RPM resources from the clk-smd-rpm
> driver to the icc-rpm.c interconnect helpers. However it was unnoticed
> that the OCMEM device also uses that clock. The clock is not required as
> all OCMEM users (currently only the GPU) will vote on the OCMEM NOC,
> bringing up the clock. Make it optional (and also apply several other
> small fixes).
> 
> [...]

Applied, thanks!

[1/3] soc: qcom: ocmem: make the core clock optional
      commit: e8a61c51417c679d1a599fb36695e9d3b8d95514
[2/3] soc: qcom: ocmem: register reasons for probe deferrals
      commit: 9dfd69cd89cd6afa4723be9098979abeef3bb8c6
[3/3] soc: qcom: ocmem: return -EPROBE_DEFER is ocmem is not available
      commit: 91b59009c7d48b58dbc50fecb27f2ad20749a05a

Best regards,
-- 
Bjorn Andersson <andersson@kernel.org>