Re: [PATCH v4 15/15] PCI/AER: Enable internal errors for CXL Upstream and Downstream Switch Ports

["Bowman, Terry" <terry.bowman@amd.com>]

On 12/12/2024 4:44 AM, Alejandro Lucero Palau wrote:
> On 12/12/24 09:44, Alejandro Lucero Palau wrote:
>> On 12/11/24 23:40, Terry Bowman wrote:
>>> The AER service driver enables PCIe Uncorrectable Internal Errors 
>>> (UIE) and
>>> Correctable Internal errors (CIE) for CXL Root Ports and CXL RCEC's. The
>>> UIE and CIE are used in reporting CXL Protocol Errors. The same UIE/CIE
>>> enablement is needed for CXL PCIe Upstream and Downstream Ports 
>>> inorder to
>>> notify the associated Root Port and OS.[1]
>>>
>>> Export the AER service driver's pci_aer_unmask_internal_errors() 
>>> function
>>> to CXL namespace.
>>>
>>> Remove the function's dependency on the CONFIG_PCIEAER_CXL kernel config
>>> because it is now an exported function.
>>>
>>> Call pci_aer_unmask_internal_errors() during RAS initialization in:
>>> cxl_uport_init_ras_reporting() and cxl_dport_init_ras_reporting().
>>>
>>> [1] PCIe Base Spec r6.2-1.0, 6.2.3.2.2 Masking Individual Errors
>>>
>>> Signed-off-by: Terry Bowman <terry.bowman@amd.com>
>>> ---
>>>   drivers/cxl/core/pci.c | 2 ++
>>>   drivers/pci/pcie/aer.c | 5 +++--
>>>   include/linux/aer.h    | 1 +
>>>   3 files changed, 6 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/cxl/core/pci.c b/drivers/cxl/core/pci.c
>>> index 9734a4c55b29..740ac5d8809f 100644
>>> --- a/drivers/cxl/core/pci.c
>>> +++ b/drivers/cxl/core/pci.c
>>> @@ -886,6 +886,7 @@ void cxl_uport_init_ras_reporting(struct cxl_port 
>>> *port)
>>>         cxl_assign_port_error_handlers(pdev);
>>>       devm_add_action_or_reset(port->uport_dev, 
>>> cxl_clear_port_error_handlers, pdev);
>>> +    pci_aer_unmask_internal_errors(pdev);
>>>   }
>>>   EXPORT_SYMBOL_NS_GPL(cxl_uport_init_ras_reporting, CXL);
>>>   @@ -920,6 +921,7 @@ void cxl_dport_init_ras_reporting(struct 
>>> cxl_dport *dport)
>>>         cxl_assign_port_error_handlers(pdev);
>>>       devm_add_action_or_reset(dport_dev, 
>>> cxl_clear_port_error_handlers, pdev);
>>> +    pci_aer_unmask_internal_errors(pdev);
>>>   }
>>>   EXPORT_SYMBOL_NS_GPL(cxl_dport_init_ras_reporting, CXL);
>>>   diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
>>> index 861521872318..0fa1b1ed48c9 100644
>>> --- a/drivers/pci/pcie/aer.c
>>> +++ b/drivers/pci/pcie/aer.c
>>> @@ -949,7 +949,6 @@ static bool is_internal_error(struct aer_err_info 
>>> *info)
>>>       return info->status & PCI_ERR_UNC_INTN;
>>>   }
>>>   -#ifdef CONFIG_PCIEAER_CXL
>>
>> This ifdef move puzzles me. I would expect to use it when the next 
>> function is invoked instead of moving it here.
>>
>> It seems weird to have such a config but code using those related 
>> functions not aware of it.
>>
>>
>>>   /**
>>>    * pci_aer_unmask_internal_errors - unmask internal errors
>>>    * @dev: pointer to the pcie_dev data structure
>>> @@ -960,7 +959,7 @@ static bool is_internal_error(struct aer_err_info 
>>> *info)
>>>    * Note: AER must be enabled and supported by the device which must be
>>>    * checked in advance, e.g. with pcie_aer_is_native().
>>>    */
>>> -static void pci_aer_unmask_internal_errors(struct pci_dev *dev)
>>> +void pci_aer_unmask_internal_errors(struct pci_dev *dev)
>>>   {
>>>       int aer = dev->aer_cap;
>>>       u32 mask;
>>> @@ -973,7 +972,9 @@ static void pci_aer_unmask_internal_errors(struct 
>>> pci_dev *dev)
>>>       mask &= ~PCI_ERR_COR_INTERNAL;
>>>       pci_write_config_dword(dev, aer + PCI_ERR_COR_MASK, mask);
>>>   }
>>> +EXPORT_SYMBOL_NS_GPL(pci_aer_unmask_internal_errors, CXL);
>
> Forgot to mention all these exports are changing in 6.13 with the second 
> macro param being now an string, so just
>
> EXPORT_SYMBOL_NS_GPL(pci_aer_unmask_internal_errors, "CXL");
>
>
> Not affected in the codebase linked to this patchset, but I hope it 
> helps you when getting weird errors with a newer kernel.

Thanks for the heads-up?

- Terry

>
>>>   +#ifdef CONFIG_PCIEAER_CXL
>>>   static bool is_cxl_mem_dev(struct pci_dev *dev)
>>>   {
>>>       /*
>>> diff --git a/include/linux/aer.h b/include/linux/aer.h
>>> index 4b97f38f3fcf..093293f9f12b 100644
>>> --- a/include/linux/aer.h
>>> +++ b/include/linux/aer.h
>>> @@ -55,5 +55,6 @@ void pci_print_aer(struct pci_dev *dev, int 
>>> aer_severity,
>>>   int cper_severity_to_aer(int cper_severity);
>>>   void aer_recover_queue(int domain, unsigned int bus, unsigned int 
>>> devfn,
>>>                  int severity, struct aer_capability_regs *aer_regs);
>>> +void pci_aer_unmask_internal_errors(struct pci_dev *dev);
>>>   #endif //_AER_H_