[v1] dax/hmem, cxl: Coordinate Soft Reserved handling with CXL

[PATCH 2/6] dax/hmem: Request cxl_acpi and cxl_pci before walking Soft Reserved ranges

Posted by Smita Koralahalli 5 months, 3 weeks ago

Ensure that cxl_acpi has published CXL Window resources before dax_hmem
walks Soft Reserved ranges.

Replace MODULE_SOFTDEP("pre: cxl_acpi") with an explicit, synchronous
request_module("cxl_acpi"). MODULE_SOFTDEP() only guarantees eventual
loading, it does not enforce that the dependency has finished init
before the current module runs. This can cause dax_hmem to start before
cxl_acpi has populated the resource tree, breaking detection of overlaps
between Soft Reserved and CXL Windows.

Also, request cxl_pci before dax_hmem walks Soft Reserved ranges. Unlike
cxl_acpi, cxl_pci attach is asynchronous and creates dependent devices
that trigger further module loads. Asynchronous probe flushing
(wait_for_device_probe()) is added later in the series in a deferred
context before dax_hmem makes ownership decisions for Soft Reserved
ranges.

Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 drivers/dax/hmem/hmem.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/dax/hmem/hmem.c b/drivers/dax/hmem/hmem.c
index d5b8f06d531e..9277e5ea0019 100644
--- a/drivers/dax/hmem/hmem.c
+++ b/drivers/dax/hmem/hmem.c
@@ -146,6 +146,16 @@ static __init int dax_hmem_init(void)
 {
 	int rc;
 
+	/*
+	 * Ensure that cxl_acpi and cxl_pci have a chance to kick off
+	 * CXL topology discovery at least once before scanning the
+	 * iomem resource tree for IORES_DESC_CXL resources.
+	 */
+	if (IS_ENABLED(CONFIG_DEV_DAX_CXL)) {
+		request_module("cxl_acpi");
+		request_module("cxl_pci");
+	}
+
 	rc = platform_driver_register(&dax_hmem_platform_driver);
 	if (rc)
 		return rc;
@@ -166,13 +176,6 @@ static __exit void dax_hmem_exit(void)
 module_init(dax_hmem_init);
 module_exit(dax_hmem_exit);
 
-/* Allow for CXL to define its own dax regions */
-#if IS_ENABLED(CONFIG_CXL_REGION)
-#if IS_MODULE(CONFIG_CXL_ACPI)
-MODULE_SOFTDEP("pre: cxl_acpi");
-#endif
-#endif
-
 MODULE_ALIAS("platform:hmem*");
 MODULE_ALIAS("platform:hmem_platform*");
 MODULE_DESCRIPTION("HMEM DAX: direct access to 'specific purpose' memory");
-- 
2.17.1

Re: [PATCH 2/6] dax/hmem: Request cxl_acpi and cxl_pci before walking Soft Reserved ranges

Posted by Dave Jiang 5 months, 1 week ago


On 8/21/25 8:41 PM, Smita Koralahalli wrote:
> Ensure that cxl_acpi has published CXL Window resources before dax_hmem
> walks Soft Reserved ranges.
> 
> Replace MODULE_SOFTDEP("pre: cxl_acpi") with an explicit, synchronous
> request_module("cxl_acpi"). MODULE_SOFTDEP() only guarantees eventual
> loading, it does not enforce that the dependency has finished init
> before the current module runs. This can cause dax_hmem to start before
> cxl_acpi has populated the resource tree, breaking detection of overlaps
> between Soft Reserved and CXL Windows.
> 
> Also, request cxl_pci before dax_hmem walks Soft Reserved ranges. Unlike
> cxl_acpi, cxl_pci attach is asynchronous and creates dependent devices
> that trigger further module loads. Asynchronous probe flushing
> (wait_for_device_probe()) is added later in the series in a deferred
> context before dax_hmem makes ownership decisions for Soft Reserved
> ranges.
> 
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>

Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> ---
>  drivers/dax/hmem/hmem.c | 17 ++++++++++-------
>  1 file changed, 10 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/dax/hmem/hmem.c b/drivers/dax/hmem/hmem.c
> index d5b8f06d531e..9277e5ea0019 100644
> --- a/drivers/dax/hmem/hmem.c
> +++ b/drivers/dax/hmem/hmem.c
> @@ -146,6 +146,16 @@ static __init int dax_hmem_init(void)
>  {
>  	int rc;
>  
> +	/*
> +	 * Ensure that cxl_acpi and cxl_pci have a chance to kick off
> +	 * CXL topology discovery at least once before scanning the
> +	 * iomem resource tree for IORES_DESC_CXL resources.
> +	 */
> +	if (IS_ENABLED(CONFIG_DEV_DAX_CXL)) {
> +		request_module("cxl_acpi");
> +		request_module("cxl_pci");
> +	}
> +
>  	rc = platform_driver_register(&dax_hmem_platform_driver);
>  	if (rc)
>  		return rc;
> @@ -166,13 +176,6 @@ static __exit void dax_hmem_exit(void)
>  module_init(dax_hmem_init);
>  module_exit(dax_hmem_exit);
>  
> -/* Allow for CXL to define its own dax regions */
> -#if IS_ENABLED(CONFIG_CXL_REGION)
> -#if IS_MODULE(CONFIG_CXL_ACPI)
> -MODULE_SOFTDEP("pre: cxl_acpi");
> -#endif
> -#endif
> -
>  MODULE_ALIAS("platform:hmem*");
>  MODULE_ALIAS("platform:hmem_platform*");
>  MODULE_DESCRIPTION("HMEM DAX: direct access to 'specific purpose' memory");

Re: [PATCH 2/6] dax/hmem: Request cxl_acpi and cxl_pci before walking Soft Reserved ranges

Posted by Zhijian Li (Fujitsu) 5 months, 1 week ago

I personally recommend that this patch be moved to position 1/6 in the series, as it specifically addresses the existing issue.


Thanks
Zhijian


On 22/08/2025 11:41, Smita Koralahalli wrote:
> Ensure that cxl_acpi has published CXL Window resources before dax_hmem
> walks Soft Reserved ranges.
> 
> Replace MODULE_SOFTDEP("pre: cxl_acpi") with an explicit, synchronous
> request_module("cxl_acpi"). MODULE_SOFTDEP() only guarantees eventual
> loading, it does not enforce that the dependency has finished init
> before the current module runs. This can cause dax_hmem to start before
> cxl_acpi has populated the resource tree, breaking detection of overlaps
> between Soft Reserved and CXL Windows.
> 
> Also, request cxl_pci before dax_hmem walks Soft Reserved ranges. Unlike
> cxl_acpi, cxl_pci attach is asynchronous and creates dependent devices
> that trigger further module loads. Asynchronous probe flushing
> (wait_for_device_probe()) is added later in the series in a deferred
> context before dax_hmem makes ownership decisions for Soft Reserved
> ranges.
> 
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> ---
>   drivers/dax/hmem/hmem.c | 17 ++++++++++-------
>   1 file changed, 10 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/dax/hmem/hmem.c b/drivers/dax/hmem/hmem.c
> index d5b8f06d531e..9277e5ea0019 100644
> --- a/drivers/dax/hmem/hmem.c
> +++ b/drivers/dax/hmem/hmem.c
> @@ -146,6 +146,16 @@ static __init int dax_hmem_init(void)
>   {
>   	int rc;
>   
> +	/*
> +	 * Ensure that cxl_acpi and cxl_pci have a chance to kick off
> +	 * CXL topology discovery at least once before scanning the
> +	 * iomem resource tree for IORES_DESC_CXL resources.
> +	 */
> +	if (IS_ENABLED(CONFIG_DEV_DAX_CXL)) {
> +		request_module("cxl_acpi");
> +		request_module("cxl_pci");
> +	}
> +
>   	rc = platform_driver_register(&dax_hmem_platform_driver);
>   	if (rc)
>   		return rc;
> @@ -166,13 +176,6 @@ static __exit void dax_hmem_exit(void)
>   module_init(dax_hmem_init);
>   module_exit(dax_hmem_exit);
>   
> -/* Allow for CXL to define its own dax regions */
> -#if IS_ENABLED(CONFIG_CXL_REGION)
> -#if IS_MODULE(CONFIG_CXL_ACPI)
> -MODULE_SOFTDEP("pre: cxl_acpi");
> -#endif
> -#endif
> -
>   MODULE_ALIAS("platform:hmem*");
>   MODULE_ALIAS("platform:hmem_platform*");
>   MODULE_DESCRIPTION("HMEM DAX: direct access to 'specific purpose' memory");

[PATCH 1/6] dax/hmem, e820, resource: Defer Soft Reserved registration until hmem is ready
[PATCH 2/6] dax/hmem: Request cxl_acpi and cxl_pci before walking Soft Reserved ranges
[PATCH 3/6] dax/hmem, cxl: Tighten dependencies on DEV_DAX_CXL and dax_hmem
[PATCH 4/6] dax/hmem: Defer Soft Reserved overlap handling until CXL region assembly completes
[PATCH 5/6] dax/hmem: Reintroduce Soft Reserved ranges back into the iomem tree
[RFC PATCH 6/6] cxl/region, dax/hmem: Guard CXL DAX region creation and tighten HMEM deps