[PATCH v3 0/3] PCI/AER, CXL: Fix appropriate _OSC check for CXL RAS Cap

[PATCH v3 0/3] PCI/AER, CXL: Fix appropriate _OSC check for CXL RAS Cap

[Smita Koralahalli]

This series of patches fixes the appropriate _OSC check for CXL RAS
registers.

First patch addresses the _OSC check.

Second patch moves around pcie_aer_is_native() function declaration to a
common location to be used by cxl/pci module and third patch reuses
pcie_aer_is_native() in cxl/pci module.

Link to v2:
https://lore.kernel.org/all/20230721214740.256602-1-Smita.KoralahalliChannabasappa@amd.com

Smita Koralahalli (3):
  cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS
    registers
  PCI/AER: Export pcie_aer_is_native()
  cxl/pci: Replace host_bridge->native_aer with pcie_aer_is_native()

 drivers/cxl/pci.c          | 7 +++----
 drivers/pci/pcie/aer.c     | 1 +
 drivers/pci/pcie/portdrv.h | 2 --
 include/linux/aer.h        | 2 ++
 4 files changed, 6 insertions(+), 6 deletions(-)

-- 
2.17.1

[PATCH v3 1/3] cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS registers

[Smita Koralahalli]

According to Section 9.17.2, Table 9-26 of CXL Specification [1], owner
of AER should also own CXL Protocol Error Management as there is no
explicit control of CXL Protocol error. And the CXL RAS Cap registers
reported on Protocol errors should check for AER _OSC rather than CXL
Memory Error Reporting Control _OSC.

The CXL Memory Error Reporting Control _OSC specifically highlights
handling Memory Error Logging and Signaling Enhancements. These kinds of
errors are reported through a device's mailbox and can be managed
independently from CXL Protocol Errors.

This change fixes handling and reporting CXL Protocol Errors and RAS
registers natively with native AER and FW-First CXL Memory Error Reporting
Control.

[1] Compute Express Link (CXL) Specification, Revision 3.1, Aug 1 2022.

Fixes: 248529edc86f ("cxl: add RAS status unmasking for CXL")
Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
Reviewed-by: Robert Richter <rrichter@amd.com>
---
v2:
	Added fixes tag.
	Included what the patch fixes in commit message.
v3:
	Added "Reviewed-by" tag.
---
 drivers/cxl/pci.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
index 1cb1494c28fe..2323169b6e5f 100644
--- a/drivers/cxl/pci.c
+++ b/drivers/cxl/pci.c
@@ -541,9 +541,9 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
 		return 0;
 	}
 
-	/* BIOS has CXL error control */
-	if (!host_bridge->native_cxl_error)
-		return -ENXIO;
+	/* BIOS has PCIe AER error control */
+	if (!host_bridge->native_aer)
+		return 0;
 
 	rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
 	if (rc)
-- 
2.17.1

[PATCH v3 2/3] PCI/AER: Export pcie_aer_is_native()

[Smita Koralahalli]

Export and move the declaration of pcie_aer_is_native() to a common header
file to be reused by cxl/pci module.

Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Robert Richter <rrichter@amd.com>
---
v2:
	Fixed $SUBJECT
v3:
	Added "Reviewd-by" tag.
---
 drivers/pci/pcie/aer.c     | 1 +
 drivers/pci/pcie/portdrv.h | 2 --
 include/linux/aer.h        | 2 ++
 3 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
index f6c24ded134c..87d90dbda023 100644
--- a/drivers/pci/pcie/aer.c
+++ b/drivers/pci/pcie/aer.c
@@ -229,6 +229,7 @@ int pcie_aer_is_native(struct pci_dev *dev)
 
 	return pcie_ports_native || host->native_aer;
 }
+EXPORT_SYMBOL_GPL(pcie_aer_is_native);
 
 int pci_enable_pcie_error_reporting(struct pci_dev *dev)
 {
diff --git a/drivers/pci/pcie/portdrv.h b/drivers/pci/pcie/portdrv.h
index 58a2b1a1cae4..1f3803bde7ee 100644
--- a/drivers/pci/pcie/portdrv.h
+++ b/drivers/pci/pcie/portdrv.h
@@ -29,10 +29,8 @@ extern bool pcie_ports_dpc_native;
 
 #ifdef CONFIG_PCIEAER
 int pcie_aer_init(void);
-int pcie_aer_is_native(struct pci_dev *dev);
 #else
 static inline int pcie_aer_init(void) { return 0; }
-static inline int pcie_aer_is_native(struct pci_dev *dev) { return 0; }
 #endif
 
 #ifdef CONFIG_HOTPLUG_PCI_PCIE
diff --git a/include/linux/aer.h b/include/linux/aer.h
index 3a3ab05e13fd..94ce49a5f8d5 100644
--- a/include/linux/aer.h
+++ b/include/linux/aer.h
@@ -45,6 +45,7 @@ struct aer_capability_regs {
 int pci_enable_pcie_error_reporting(struct pci_dev *dev);
 int pci_disable_pcie_error_reporting(struct pci_dev *dev);
 int pci_aer_clear_nonfatal_status(struct pci_dev *dev);
+int pcie_aer_is_native(struct pci_dev *dev);
 #else
 static inline int pci_enable_pcie_error_reporting(struct pci_dev *dev)
 {
@@ -58,6 +59,7 @@ static inline int pci_aer_clear_nonfatal_status(struct pci_dev *dev)
 {
 	return -EINVAL;
 }
+static inline int pcie_aer_is_native(struct pci_dev *dev) { return 0; }
 #endif
 
 void cper_print_aer(struct pci_dev *dev, int aer_severity,
-- 
2.17.1

[PATCH v3 3/3] cxl/pci: Replace host_bridge->native_aer with pcie_aer_is_native()

[Smita Koralahalli]

Use pcie_aer_is_native() to determine the native AER ownership as the
usage of host_bride->native_aer does not cover command line override of
AER ownership.

Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Robert Richter <rrichter@amd.com>
---
v2:
	Replaced pcie_aer_is_native() at a later stage for automated
	backports.
v3:
	Added more context to commit message.
	Added "Reviewed-by" tag.
---
 drivers/cxl/pci.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
index 2323169b6e5f..44a21ab7add5 100644
--- a/drivers/cxl/pci.c
+++ b/drivers/cxl/pci.c
@@ -529,7 +529,6 @@ static int cxl_pci_setup_regs(struct pci_dev *pdev, enum cxl_regloc_type type,
 
 static int cxl_pci_ras_unmask(struct pci_dev *pdev)
 {
-	struct pci_host_bridge *host_bridge = pci_find_host_bridge(pdev->bus);
 	struct cxl_dev_state *cxlds = pci_get_drvdata(pdev);
 	void __iomem *addr;
 	u32 orig_val, val, mask;
@@ -542,7 +541,7 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
 	}
 
 	/* BIOS has PCIe AER error control */
-	if (!host_bridge->native_aer)
+	if (!pcie_aer_is_native(pdev))
 		return 0;
 
 	rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
-- 
2.17.1

Re: [PATCH v3 1/3] cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS registers

[Jonathan Cameron]

On Thu, 3 Aug 2023 23:01:27 +0000
Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:

> According to Section 9.17.2, Table 9-26 of CXL Specification [1], owner
> of AER should also own CXL Protocol Error Management as there is no
> explicit control of CXL Protocol error. And the CXL RAS Cap registers
> reported on Protocol errors should check for AER _OSC rather than CXL
> Memory Error Reporting Control _OSC.
> 
> The CXL Memory Error Reporting Control _OSC specifically highlights
> handling Memory Error Logging and Signaling Enhancements. These kinds of
> errors are reported through a device's mailbox and can be managed
> independently from CXL Protocol Errors.
> 
> This change fixes handling and reporting CXL Protocol Errors and RAS
> registers natively with native AER and FW-First CXL Memory Error Reporting
> Control.
> 
> [1] Compute Express Link (CXL) Specification, Revision 3.1, Aug 1 2022.
> 
> Fixes: 248529edc86f ("cxl: add RAS status unmasking for CXL")
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Reviewed-by: Robert Richter <rrichter@amd.com>

I'd be tempted to add a comment on why this returns 0 rather than an
error.  I think that makes sense but it isn't immediately obvious from
the local context.

Otherwise LGTM

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>


> ---
> v2:
> 	Added fixes tag.
> 	Included what the patch fixes in commit message.
> v3:
> 	Added "Reviewed-by" tag.
> ---
>  drivers/cxl/pci.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
> index 1cb1494c28fe..2323169b6e5f 100644
> --- a/drivers/cxl/pci.c
> +++ b/drivers/cxl/pci.c
> @@ -541,9 +541,9 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>  		return 0;
>  	}
>  
> -	/* BIOS has CXL error control */
> -	if (!host_bridge->native_cxl_error)
> -		return -ENXIO;
> +	/* BIOS has PCIe AER error control */
> +	if (!host_bridge->native_aer)
> +		return 0;
>  
>  	rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
>  	if (rc)

Re: [PATCH v3 2/3] PCI/AER: Export pcie_aer_is_native()

[Jonathan Cameron]

On Thu, 3 Aug 2023 23:01:28 +0000
Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:

> Export and move the declaration of pcie_aer_is_native() to a common header
> file to be reused by cxl/pci module.
> 
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
> Reviewed-by: Robert Richter <rrichter@amd.com>

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

> ---
> v2:
> 	Fixed $SUBJECT
> v3:
> 	Added "Reviewd-by" tag.
> ---
>  drivers/pci/pcie/aer.c     | 1 +
>  drivers/pci/pcie/portdrv.h | 2 --
>  include/linux/aer.h        | 2 ++
>  3 files changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
> index f6c24ded134c..87d90dbda023 100644
> --- a/drivers/pci/pcie/aer.c
> +++ b/drivers/pci/pcie/aer.c
> @@ -229,6 +229,7 @@ int pcie_aer_is_native(struct pci_dev *dev)
>  
>  	return pcie_ports_native || host->native_aer;
>  }
> +EXPORT_SYMBOL_GPL(pcie_aer_is_native);
>  
>  int pci_enable_pcie_error_reporting(struct pci_dev *dev)
>  {
> diff --git a/drivers/pci/pcie/portdrv.h b/drivers/pci/pcie/portdrv.h
> index 58a2b1a1cae4..1f3803bde7ee 100644
> --- a/drivers/pci/pcie/portdrv.h
> +++ b/drivers/pci/pcie/portdrv.h
> @@ -29,10 +29,8 @@ extern bool pcie_ports_dpc_native;
>  
>  #ifdef CONFIG_PCIEAER
>  int pcie_aer_init(void);
> -int pcie_aer_is_native(struct pci_dev *dev);
>  #else
>  static inline int pcie_aer_init(void) { return 0; }
> -static inline int pcie_aer_is_native(struct pci_dev *dev) { return 0; }
>  #endif
>  
>  #ifdef CONFIG_HOTPLUG_PCI_PCIE
> diff --git a/include/linux/aer.h b/include/linux/aer.h
> index 3a3ab05e13fd..94ce49a5f8d5 100644
> --- a/include/linux/aer.h
> +++ b/include/linux/aer.h
> @@ -45,6 +45,7 @@ struct aer_capability_regs {
>  int pci_enable_pcie_error_reporting(struct pci_dev *dev);
>  int pci_disable_pcie_error_reporting(struct pci_dev *dev);
>  int pci_aer_clear_nonfatal_status(struct pci_dev *dev);
> +int pcie_aer_is_native(struct pci_dev *dev);
>  #else
>  static inline int pci_enable_pcie_error_reporting(struct pci_dev *dev)
>  {
> @@ -58,6 +59,7 @@ static inline int pci_aer_clear_nonfatal_status(struct pci_dev *dev)
>  {
>  	return -EINVAL;
>  }
> +static inline int pcie_aer_is_native(struct pci_dev *dev) { return 0; }
>  #endif
>  
>  void cper_print_aer(struct pci_dev *dev, int aer_severity,

Re: [PATCH v3 3/3] cxl/pci: Replace host_bridge->native_aer with pcie_aer_is_native()

[Jonathan Cameron]

On Thu, 3 Aug 2023 23:01:29 +0000
Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:

> Use pcie_aer_is_native() to determine the native AER ownership as the
> usage of host_bride->native_aer does not cover command line override of
> AER ownership.
> 
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
> Reviewed-by: Robert Richter <rrichter@amd.com>
I hope there aren't systems getting this wrong but make sense to support
same override as for other cases.

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

> ---
> v2:
> 	Replaced pcie_aer_is_native() at a later stage for automated
> 	backports.
> v3:
> 	Added more context to commit message.
> 	Added "Reviewed-by" tag.
> ---
>  drivers/cxl/pci.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
> index 2323169b6e5f..44a21ab7add5 100644
> --- a/drivers/cxl/pci.c
> +++ b/drivers/cxl/pci.c
> @@ -529,7 +529,6 @@ static int cxl_pci_setup_regs(struct pci_dev *pdev, enum cxl_regloc_type type,
>  
>  static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>  {
> -	struct pci_host_bridge *host_bridge = pci_find_host_bridge(pdev->bus);
>  	struct cxl_dev_state *cxlds = pci_get_drvdata(pdev);
>  	void __iomem *addr;
>  	u32 orig_val, val, mask;
> @@ -542,7 +541,7 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>  	}
>  
>  	/* BIOS has PCIe AER error control */
> -	if (!host_bridge->native_aer)
> +	if (!pcie_aer_is_native(pdev))
>  		return 0;
>  
>  	rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);

Re: [PATCH v3 1/3] cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS registers

[Dave Jiang]

On 8/4/23 05:09, Jonathan Cameron wrote:
> On Thu, 3 Aug 2023 23:01:27 +0000
> Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:
> 
>> According to Section 9.17.2, Table 9-26 of CXL Specification [1], owner
>> of AER should also own CXL Protocol Error Management as there is no
>> explicit control of CXL Protocol error. And the CXL RAS Cap registers
>> reported on Protocol errors should check for AER _OSC rather than CXL
>> Memory Error Reporting Control _OSC.
>>
>> The CXL Memory Error Reporting Control _OSC specifically highlights
>> handling Memory Error Logging and Signaling Enhancements. These kinds of
>> errors are reported through a device's mailbox and can be managed
>> independently from CXL Protocol Errors.
>>
>> This change fixes handling and reporting CXL Protocol Errors and RAS
>> registers natively with native AER and FW-First CXL Memory Error Reporting
>> Control.
>>
>> [1] Compute Express Link (CXL) Specification, Revision 3.1, Aug 1 2022.
>>
>> Fixes: 248529edc86f ("cxl: add RAS status unmasking for CXL")
>> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
>> Reviewed-by: Robert Richter <rrichter@amd.com>
> 
> I'd be tempted to add a comment on why this returns 0 rather than an
> error.  I think that makes sense but it isn't immediately obvious from
> the local context.
> 
> Otherwise LGTM
> 
> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Echo Jonathan's comment.

Reviewed-by: Dave Jiang <dave.jiang@intel.com>

> 
> 
>> ---
>> v2:
>> 	Added fixes tag.
>> 	Included what the patch fixes in commit message.
>> v3:
>> 	Added "Reviewed-by" tag.
>> ---
>>   drivers/cxl/pci.c | 6 +++---
>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
>> index 1cb1494c28fe..2323169b6e5f 100644
>> --- a/drivers/cxl/pci.c
>> +++ b/drivers/cxl/pci.c
>> @@ -541,9 +541,9 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>>   		return 0;
>>   	}
>>   
>> -	/* BIOS has CXL error control */
>> -	if (!host_bridge->native_cxl_error)
>> -		return -ENXIO;
>> +	/* BIOS has PCIe AER error control */
>> +	if (!host_bridge->native_aer)
>> +		return 0;
>>   
>>   	rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
>>   	if (rc)
> 

Re: [PATCH v3 2/3] PCI/AER: Export pcie_aer_is_native()

[Dave Jiang]

On 8/3/23 16:01, Smita Koralahalli wrote:
> Export and move the declaration of pcie_aer_is_native() to a common header
> file to be reused by cxl/pci module.
> 
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
> Reviewed-by: Robert Richter <rrichter@amd.com>

Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> ---
> v2:
> 	Fixed $SUBJECT
> v3:
> 	Added "Reviewd-by" tag.
> ---
>   drivers/pci/pcie/aer.c     | 1 +
>   drivers/pci/pcie/portdrv.h | 2 --
>   include/linux/aer.h        | 2 ++
>   3 files changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
> index f6c24ded134c..87d90dbda023 100644
> --- a/drivers/pci/pcie/aer.c
> +++ b/drivers/pci/pcie/aer.c
> @@ -229,6 +229,7 @@ int pcie_aer_is_native(struct pci_dev *dev)
>   
>   	return pcie_ports_native || host->native_aer;
>   }
> +EXPORT_SYMBOL_GPL(pcie_aer_is_native);
>   
>   int pci_enable_pcie_error_reporting(struct pci_dev *dev)
>   {
> diff --git a/drivers/pci/pcie/portdrv.h b/drivers/pci/pcie/portdrv.h
> index 58a2b1a1cae4..1f3803bde7ee 100644
> --- a/drivers/pci/pcie/portdrv.h
> +++ b/drivers/pci/pcie/portdrv.h
> @@ -29,10 +29,8 @@ extern bool pcie_ports_dpc_native;
>   
>   #ifdef CONFIG_PCIEAER
>   int pcie_aer_init(void);
> -int pcie_aer_is_native(struct pci_dev *dev);
>   #else
>   static inline int pcie_aer_init(void) { return 0; }
> -static inline int pcie_aer_is_native(struct pci_dev *dev) { return 0; }
>   #endif
>   
>   #ifdef CONFIG_HOTPLUG_PCI_PCIE
> diff --git a/include/linux/aer.h b/include/linux/aer.h
> index 3a3ab05e13fd..94ce49a5f8d5 100644
> --- a/include/linux/aer.h
> +++ b/include/linux/aer.h
> @@ -45,6 +45,7 @@ struct aer_capability_regs {
>   int pci_enable_pcie_error_reporting(struct pci_dev *dev);
>   int pci_disable_pcie_error_reporting(struct pci_dev *dev);
>   int pci_aer_clear_nonfatal_status(struct pci_dev *dev);
> +int pcie_aer_is_native(struct pci_dev *dev);
>   #else
>   static inline int pci_enable_pcie_error_reporting(struct pci_dev *dev)
>   {
> @@ -58,6 +59,7 @@ static inline int pci_aer_clear_nonfatal_status(struct pci_dev *dev)
>   {
>   	return -EINVAL;
>   }
> +static inline int pcie_aer_is_native(struct pci_dev *dev) { return 0; }
>   #endif
>   
>   void cper_print_aer(struct pci_dev *dev, int aer_severity,

Re: [PATCH v3 3/3] cxl/pci: Replace host_bridge->native_aer with pcie_aer_is_native()

[Dave Jiang]

On 8/3/23 16:01, Smita Koralahalli wrote:
> Use pcie_aer_is_native() to determine the native AER ownership as the
> usage of host_bride->native_aer does not cover command line override of
> AER ownership.
> 
> Signed-off-by: Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com>
> Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
> Reviewed-by: Robert Richter <rrichter@amd.com>

Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> ---
> v2:
> 	Replaced pcie_aer_is_native() at a later stage for automated
> 	backports.
> v3:
> 	Added more context to commit message.
> 	Added "Reviewed-by" tag.
> ---
>   drivers/cxl/pci.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
> index 2323169b6e5f..44a21ab7add5 100644
> --- a/drivers/cxl/pci.c
> +++ b/drivers/cxl/pci.c
> @@ -529,7 +529,6 @@ static int cxl_pci_setup_regs(struct pci_dev *pdev, enum cxl_regloc_type type,
>   
>   static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>   {
> -	struct pci_host_bridge *host_bridge = pci_find_host_bridge(pdev->bus);
>   	struct cxl_dev_state *cxlds = pci_get_drvdata(pdev);
>   	void __iomem *addr;
>   	u32 orig_val, val, mask;
> @@ -542,7 +541,7 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>   	}
>   
>   	/* BIOS has PCIe AER error control */
> -	if (!host_bridge->native_aer)
> +	if (!pcie_aer_is_native(pdev))
>   		return 0;
>   
>   	rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);

Re: [PATCH v3 1/3] cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS registers

[Smita Koralahalli]

On 8/16/2023 11:06 AM, Dave Jiang wrote:
> 
> 
> On 8/4/23 05:09, Jonathan Cameron wrote:
>> On Thu, 3 Aug 2023 23:01:27 +0000
>> Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:
>>
>>> According to Section 9.17.2, Table 9-26 of CXL Specification [1], owner
>>> of AER should also own CXL Protocol Error Management as there is no
>>> explicit control of CXL Protocol error. And the CXL RAS Cap registers
>>> reported on Protocol errors should check for AER _OSC rather than CXL
>>> Memory Error Reporting Control _OSC.
>>>
>>> The CXL Memory Error Reporting Control _OSC specifically highlights
>>> handling Memory Error Logging and Signaling Enhancements. These kinds of
>>> errors are reported through a device's mailbox and can be managed
>>> independently from CXL Protocol Errors.
>>>
>>> This change fixes handling and reporting CXL Protocol Errors and RAS
>>> registers natively with native AER and FW-First CXL Memory Error 
>>> Reporting
>>> Control.
>>>
>>> [1] Compute Express Link (CXL) Specification, Revision 3.1, Aug 1 2022.
>>>
>>> Fixes: 248529edc86f ("cxl: add RAS status unmasking for CXL")
>>> Signed-off-by: Smita Koralahalli 
>>> <Smita.KoralahalliChannabasappa@amd.com>
>>> Reviewed-by: Robert Richter <rrichter@amd.com>
>>
>> I'd be tempted to add a comment on why this returns 0 rather than an
>> error.  I think that makes sense but it isn't immediately obvious from
>> the local context.
>>
>> Otherwise LGTM
>>
>> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
> 
> Echo Jonathan's comment.
> 
> Reviewed-by: Dave Jiang <dave.jiang@intel.com>

Yes, and Dan is probably against returning error code.
https://lore.kernel.org/all/64d1b3e78629f_5ea6e2944@dwillia2-xfh.jf.intel.com.notmuch/

But I think returning zero is required as we don't want to interfere 
with cxl device access when operating in native cxl memory error 
reporting. Returning error code will basically fail cxl_pci_probe() and 
thus fail to create a cxl device node.

I was thinking a single line comment as: "Return zero to not block the 
communication with the cxl device when in native memory error reporting 
mode".

Agree? Or anything more that needs to be added?

Thanks,
Smita
> 
>>
>>
>>> ---
>>> v2:
>>>     Added fixes tag.
>>>     Included what the patch fixes in commit message.
>>> v3:
>>>     Added "Reviewed-by" tag.
>>> ---
>>>   drivers/cxl/pci.c | 6 +++---
>>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
>>> index 1cb1494c28fe..2323169b6e5f 100644
>>> --- a/drivers/cxl/pci.c
>>> +++ b/drivers/cxl/pci.c
>>> @@ -541,9 +541,9 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>>>           return 0;
>>>       }
>>> -    /* BIOS has CXL error control */
>>> -    if (!host_bridge->native_cxl_error)
>>> -        return -ENXIO;
>>> +    /* BIOS has PCIe AER error control */
>>> +    if (!host_bridge->native_aer)
>>> +        return 0;
>>>       rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
>>>       if (rc)
>>

Re: [PATCH v3 1/3] cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS registers

[Smita Koralahalli]

On 8/16/2023 2:33 PM, Smita Koralahalli wrote:
> On 8/16/2023 11:06 AM, Dave Jiang wrote:
>>
>>
>> On 8/4/23 05:09, Jonathan Cameron wrote:
>>> On Thu, 3 Aug 2023 23:01:27 +0000
>>> Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:
>>>
>>>> According to Section 9.17.2, Table 9-26 of CXL Specification [1], owner
>>>> of AER should also own CXL Protocol Error Management as there is no
>>>> explicit control of CXL Protocol error. And the CXL RAS Cap registers
>>>> reported on Protocol errors should check for AER _OSC rather than CXL
>>>> Memory Error Reporting Control _OSC.
>>>>
>>>> The CXL Memory Error Reporting Control _OSC specifically highlights
>>>> handling Memory Error Logging and Signaling Enhancements. These 
>>>> kinds of
>>>> errors are reported through a device's mailbox and can be managed
>>>> independently from CXL Protocol Errors.
>>>>
>>>> This change fixes handling and reporting CXL Protocol Errors and RAS
>>>> registers natively with native AER and FW-First CXL Memory Error 
>>>> Reporting
>>>> Control.
>>>>
>>>> [1] Compute Express Link (CXL) Specification, Revision 3.1, Aug 1 2022.
>>>>
>>>> Fixes: 248529edc86f ("cxl: add RAS status unmasking for CXL")
>>>> Signed-off-by: Smita Koralahalli 
>>>> <Smita.KoralahalliChannabasappa@amd.com>
>>>> Reviewed-by: Robert Richter <rrichter@amd.com>
>>>
>>> I'd be tempted to add a comment on why this returns 0 rather than an
>>> error.  I think that makes sense but it isn't immediately obvious from
>>> the local context.
>>>
>>> Otherwise LGTM
>>>
>>> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
>>
>> Echo Jonathan's comment.
>>
>> Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> 
> Yes, and Dan is probably against returning error code.

Against returning zero. My bad sorry!
> https://lore.kernel.org/all/64d1b3e78629f_5ea6e2944@dwillia2-xfh.jf.intel.com.notmuch/ 
> 
> 
> But I think returning zero is required as we don't want to interfere 
> with cxl device access when operating in native cxl memory error 
> reporting. Returning error code will basically fail cxl_pci_probe() and 
> thus fail to create a cxl device node.
> 
> I was thinking a single line comment as: "Return zero to not block the 
> communication with the cxl device when in native memory error reporting 
> mode".
> 
> Agree? Or anything more that needs to be added?
> 
> Thanks,
> Smita
>>
>>>
>>>
>>>> ---
>>>> v2:
>>>>     Added fixes tag.
>>>>     Included what the patch fixes in commit message.
>>>> v3:
>>>>     Added "Reviewed-by" tag.
>>>> ---
>>>>   drivers/cxl/pci.c | 6 +++---
>>>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
>>>> index 1cb1494c28fe..2323169b6e5f 100644
>>>> --- a/drivers/cxl/pci.c
>>>> +++ b/drivers/cxl/pci.c
>>>> @@ -541,9 +541,9 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>>>>           return 0;
>>>>       }
>>>> -    /* BIOS has CXL error control */
>>>> -    if (!host_bridge->native_cxl_error)
>>>> -        return -ENXIO;
>>>> +    /* BIOS has PCIe AER error control */
>>>> +    if (!host_bridge->native_aer)
>>>> +        return 0;
>>>>       rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
>>>>       if (rc)
>>>
> 

Re: [PATCH v3 1/3] cxl/pci: Fix appropriate checking for _OSC while handling CXL RAS registers

[Dave Jiang]

On 8/16/23 14:33, Smita Koralahalli wrote:
> On 8/16/2023 11:06 AM, Dave Jiang wrote:
>>
>>
>> On 8/4/23 05:09, Jonathan Cameron wrote:
>>> On Thu, 3 Aug 2023 23:01:27 +0000
>>> Smita Koralahalli <Smita.KoralahalliChannabasappa@amd.com> wrote:
>>>
>>>> According to Section 9.17.2, Table 9-26 of CXL Specification [1], owner
>>>> of AER should also own CXL Protocol Error Management as there is no
>>>> explicit control of CXL Protocol error. And the CXL RAS Cap registers
>>>> reported on Protocol errors should check for AER _OSC rather than CXL
>>>> Memory Error Reporting Control _OSC.
>>>>
>>>> The CXL Memory Error Reporting Control _OSC specifically highlights
>>>> handling Memory Error Logging and Signaling Enhancements. These 
>>>> kinds of
>>>> errors are reported through a device's mailbox and can be managed
>>>> independently from CXL Protocol Errors.
>>>>
>>>> This change fixes handling and reporting CXL Protocol Errors and RAS
>>>> registers natively with native AER and FW-First CXL Memory Error 
>>>> Reporting
>>>> Control.
>>>>
>>>> [1] Compute Express Link (CXL) Specification, Revision 3.1, Aug 1 2022.
>>>>
>>>> Fixes: 248529edc86f ("cxl: add RAS status unmasking for CXL")
>>>> Signed-off-by: Smita Koralahalli 
>>>> <Smita.KoralahalliChannabasappa@amd.com>
>>>> Reviewed-by: Robert Richter <rrichter@amd.com>
>>>
>>> I'd be tempted to add a comment on why this returns 0 rather than an
>>> error.  I think that makes sense but it isn't immediately obvious from
>>> the local context.
>>>
>>> Otherwise LGTM
>>>
>>> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
>>
>> Echo Jonathan's comment.
>>
>> Reviewed-by: Dave Jiang <dave.jiang@intel.com>
> 
> Yes, and Dan is probably against returning error code.
> https://lore.kernel.org/all/64d1b3e78629f_5ea6e2944@dwillia2-xfh.jf.intel.com.notmuch/
> 
> But I think returning zero is required as we don't want to interfere 
> with cxl device access when operating in native cxl memory error 
> reporting. Returning error code will basically fail cxl_pci_probe() and 
> thus fail to create a cxl device node.
> 
> I was thinking a single line comment as: "Return zero to not block the 
> communication with the cxl device when in native memory error reporting 
> mode".

Looks reasonable to me. Unless Dan feels it needs to fail the probe.

> 
> Agree? Or anything more that needs to be added?
> 
> Thanks,
> Smita
>>
>>>
>>>
>>>> ---
>>>> v2:
>>>>     Added fixes tag.
>>>>     Included what the patch fixes in commit message.
>>>> v3:
>>>>     Added "Reviewed-by" tag.
>>>> ---
>>>>   drivers/cxl/pci.c | 6 +++---
>>>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
>>>> index 1cb1494c28fe..2323169b6e5f 100644
>>>> --- a/drivers/cxl/pci.c
>>>> +++ b/drivers/cxl/pci.c
>>>> @@ -541,9 +541,9 @@ static int cxl_pci_ras_unmask(struct pci_dev *pdev)
>>>>           return 0;
>>>>       }
>>>> -    /* BIOS has CXL error control */
>>>> -    if (!host_bridge->native_cxl_error)
>>>> -        return -ENXIO;
>>>> +    /* BIOS has PCIe AER error control */
>>>> +    if (!host_bridge->native_aer)
>>>> +        return 0;
>>>>       rc = pcie_capability_read_word(pdev, PCI_EXP_DEVCTL, &cap);
>>>>       if (rc)
>>>
>