[PATCH] remoteproc: qcom_wcnss: Fix reserved region mapping failure

[PATCH] remoteproc: qcom_wcnss: Fix reserved region mapping failure

[Rob Herring (Arm)]

Commit c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_*
functions for "memory-region"") switched from devm_ioremap_wc() to
devm_ioremap_resource_wc(). The difference is devm_ioremap_resource_wc()
also requests the resource which fails. Testing of both fixed and
dynamic reserved regions indicates that requesting the resource should
work, so I'm not sure why it doesn't work in this case. Fix the issue by
reverting back to devm_ioremap_wc().

Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reported-by: André Apitzsch <git@apitzsch.eu>
Fixes: c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_* functions for "memory-region"")
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
---
I created some dummy regions matching the flags used here and did
request_mem_region() on them successfully, so I'm not sure why it
doesn't work for this driver. Something else is requesting the region?

Perhaps there's a deferred probe and somehow request_mem_region() is not
getting undone correctly?
---
 drivers/remoteproc/qcom_wcnss.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/remoteproc/qcom_wcnss.c b/drivers/remoteproc/qcom_wcnss.c
index ee18bf2e8054..4add9037dbd5 100644
--- a/drivers/remoteproc/qcom_wcnss.c
+++ b/drivers/remoteproc/qcom_wcnss.c
@@ -537,7 +537,7 @@ static int wcnss_alloc_memory_region(struct qcom_wcnss *wcnss)
 
 	wcnss->mem_phys = wcnss->mem_reloc = res.start;
 	wcnss->mem_size = resource_size(&res);
-	wcnss->mem_region = devm_ioremap_resource_wc(wcnss->dev, &res);
+	wcnss->mem_region = devm_ioremap_wc(wcnss->dev, wcnss->mem_phys, wcnss->mem_size);
 	if (IS_ERR(wcnss->mem_region)) {
 		dev_err(wcnss->dev, "unable to map memory region: %pR\n", &res);
 		return PTR_ERR(wcnss->mem_region);
-- 
2.51.0

Re: [PATCH] remoteproc: qcom_wcnss: Fix reserved region mapping failure

[André Apitzsch]

Am Mittwoch, dem 28.01.2026 um 16:02 -0600 schrieb Rob Herring (Arm):
> Commit c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_*
> functions for "memory-region"") switched from devm_ioremap_wc() to
> devm_ioremap_resource_wc(). The difference is
> devm_ioremap_resource_wc() also requests the resource which fails.
> Testing of both fixed and dynamic reserved regions indicates that
> requesting the resource should work, so I'm not sure why it doesn't
> work in this case. Fix the issue by reverting back to
> devm_ioremap_wc().
> 
> Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
> Reported-by: André Apitzsch <git@apitzsch.eu>
> Fixes: c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_*
> functions for "memory-region"")
> Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
> ---
> I created some dummy regions matching the flags used here and did
> request_mem_region() on them successfully, so I'm not sure why it
> doesn't work for this driver. Something else is requesting the
> region?
> 
> Perhaps there's a deferred probe and somehow request_mem_region() is
> not getting undone correctly?
> ---
>  drivers/remoteproc/qcom_wcnss.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/remoteproc/qcom_wcnss.c
> b/drivers/remoteproc/qcom_wcnss.c
> index ee18bf2e8054..4add9037dbd5 100644
> --- a/drivers/remoteproc/qcom_wcnss.c
> +++ b/drivers/remoteproc/qcom_wcnss.c
> @@ -537,7 +537,7 @@ static int wcnss_alloc_memory_region(struct
> qcom_wcnss *wcnss)
>  
>  	wcnss->mem_phys = wcnss->mem_reloc = res.start;
>  	wcnss->mem_size = resource_size(&res);
> -	wcnss->mem_region = devm_ioremap_resource_wc(wcnss->dev,
> &res);
> +	wcnss->mem_region = devm_ioremap_wc(wcnss->dev, wcnss-
> >mem_phys, wcnss->mem_size);
>  	if (IS_ERR(wcnss->mem_region)) {
>  		dev_err(wcnss->dev, "unable to map memory region:
> %pR\n", &res);
>  		return PTR_ERR(wcnss->mem_region);

Thanks.

Tested-by: André Apitzsch <git@apitzsch.eu> # on BQ Aquaris M5

Re: [PATCH] remoteproc: qcom_wcnss: Fix reserved region mapping failure

[Marek Szyprowski]

On 28.01.2026 23:02, Rob Herring (Arm) wrote:
> Commit c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_*
> functions for "memory-region"") switched from devm_ioremap_wc() to
> devm_ioremap_resource_wc(). The difference is devm_ioremap_resource_wc()
> also requests the resource which fails. Testing of both fixed and
> dynamic reserved regions indicates that requesting the resource should
> work, so I'm not sure why it doesn't work in this case. Fix the issue by
> reverting back to devm_ioremap_wc().
>
> Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
> Reported-by: André Apitzsch <git@apitzsch.eu>
> Fixes: c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_* functions for "memory-region"")
> Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
> ---
> I created some dummy regions matching the flags used here and did
> request_mem_region() on them successfully, so I'm not sure why it
> doesn't work for this driver. Something else is requesting the region?
>
> Perhaps there's a deferred probe and somehow request_mem_region() is not
> getting undone correctly?

I didn't manage to track what is the root cause of the 
request_mem_region() failure, but this is definitely not related to 
deferred probe. This function is called only once from the qcom_wcnss 
driver and fails, because the region is already assigned to 'reserved' 
owner.

> ---
>   drivers/remoteproc/qcom_wcnss.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/remoteproc/qcom_wcnss.c b/drivers/remoteproc/qcom_wcnss.c
> index ee18bf2e8054..4add9037dbd5 100644
> --- a/drivers/remoteproc/qcom_wcnss.c
> +++ b/drivers/remoteproc/qcom_wcnss.c
> @@ -537,7 +537,7 @@ static int wcnss_alloc_memory_region(struct qcom_wcnss *wcnss)
>   
>   	wcnss->mem_phys = wcnss->mem_reloc = res.start;
>   	wcnss->mem_size = resource_size(&res);
> -	wcnss->mem_region = devm_ioremap_resource_wc(wcnss->dev, &res);
> +	wcnss->mem_region = devm_ioremap_wc(wcnss->dev, wcnss->mem_phys, wcnss->mem_size);
>   	if (IS_ERR(wcnss->mem_region)) {
>   		dev_err(wcnss->dev, "unable to map memory region: %pR\n", &res);
>   		return PTR_ERR(wcnss->mem_region);

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland

Re: [PATCH] remoteproc: qcom_wcnss: Fix reserved region mapping failure

[Bjorn Andersson]

On Wed, 28 Jan 2026 16:02:42 -0600, Rob Herring (Arm) wrote:
> Commit c70b9d5fdcd7 ("remoteproc: qcom: Use of_reserved_mem_region_*
> functions for "memory-region"") switched from devm_ioremap_wc() to
> devm_ioremap_resource_wc(). The difference is devm_ioremap_resource_wc()
> also requests the resource which fails. Testing of both fixed and
> dynamic reserved regions indicates that requesting the resource should
> work, so I'm not sure why it doesn't work in this case. Fix the issue by
> reverting back to devm_ioremap_wc().
> 
> [...]

Applied, thanks!

[1/1] remoteproc: qcom_wcnss: Fix reserved region mapping failure
      commit: f9b888599418951b8229bbb28851ed4da50c58e9

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