Re: [PATCH v15 8/9] PCI/DPC: Unify and plumb error handling into DPC

[poza@codeaurora.org]

On 2018-05-10 18:52, Bjorn Helgaas wrote:
> On Thu, May 03, 2018 at 01:03:57AM -0400, Oza Pawandeep wrote:
>> Current DPC driver does not do recovery, e.g. calling end-point's 
>> driver's
>> callbacks, which sanitize the sw.
>> 
>> DPC driver implements link_reset callback, and calls 
>> pci_do_recovery().
>> 
>> Signed-off-by: Oza Pawandeep <poza@codeaurora.org>
>> 
>> diff --git a/drivers/pci/pcie/dpc.c b/drivers/pci/pcie/dpc.c
>> index 80ec384..aed7c9f 100644
>> --- a/drivers/pci/pcie/dpc.c
>> +++ b/drivers/pci/pcie/dpc.c
>> @@ -73,29 +73,21 @@ static void dpc_wait_link_inactive(struct dpc_dev 
>> *dpc)
>>  	pcie_wait_for_link(pdev, false);
>>  }
>> 
>> -static void dpc_work(struct work_struct *work)
>> +static pci_ers_result_t dpc_reset_link(struct pci_dev *pdev)
>>  {
>> -	struct dpc_dev *dpc = container_of(work, struct dpc_dev, work);
>> -	struct pci_dev *dev, *temp, *pdev = dpc->dev->port;
>> -	struct pci_bus *parent = pdev->subordinate;
>> -	u16 cap = dpc->cap_pos, ctl;
>> -
>> -	pci_lock_rescan_remove();
>> -	list_for_each_entry_safe_reverse(dev, temp, &parent->devices,
>> -					 bus_list) {
>> -		pci_dev_get(dev);
>> -		pci_dev_set_disconnected(dev, NULL);
>> -		if (pci_has_subordinate(dev))
>> -			pci_walk_bus(dev->subordinate,
>> -				     pci_dev_set_disconnected, NULL);
>> -		pci_stop_and_remove_bus_device(dev);
>> -		pci_dev_put(dev);
>> -	}
>> -	pci_unlock_rescan_remove();
> 
> I think it would be good to have a comment here about why this 
> "reset_link"
> function doesn't actually reset the link, e.g.,
> 
>   /*
>    * DPC disables the Link automatically in hardware, so it has
>    * already been reset by the time we get here.
>    */
> 
>> +	struct dpc_dev *dpc;
>> +	struct pcie_device *pciedev;
>> +	struct device *devdpc;
>> +	u16 cap, ctl;
>> +
>> +	devdpc = pcie_port_find_device(pdev, PCIE_PORT_SERVICE_DPC);
>> +	pciedev = to_pcie_device(devdpc);
>> +	dpc = get_service_data(pciedev);
>> +	cap = dpc->cap_pos;
> 
> And maybe one about waiting until the link is inactive, then clearing 
> DPC
> Trigger Status to allow the port to leave DPC.
> 
>>  	dpc_wait_link_inactive(dpc);
>>  	if (dpc->rp_extensions && dpc_wait_rp_inactive(dpc))
>> -		return;
>> +		return PCI_ERS_RESULT_DISCONNECT;
>>  	if (dpc->rp_extensions && dpc->rp_pio_status) {
>>  		pci_write_config_dword(pdev, cap + PCI_EXP_DPC_RP_PIO_STATUS,
>>  				       dpc->rp_pio_status);
>> @@ -108,6 +100,17 @@ static void dpc_work(struct work_struct *work)
>>  	pci_read_config_word(pdev, cap + PCI_EXP_DPC_CTL, &ctl);
>>  	pci_write_config_word(pdev, cap + PCI_EXP_DPC_CTL,
>>  			      ctl | PCI_EXP_DPC_CTL_INT_EN);
>> +
>> +	return PCI_ERS_RESULT_RECOVERED;
>> +}
>> +
>> +static void dpc_work(struct work_struct *work)
>> +{
>> +	struct dpc_dev *dpc = container_of(work, struct dpc_dev, work);
>> +	struct pci_dev *pdev = dpc->dev->port;
>> +
>> +	/* From DPC point of view error is always FATAL. */
>> +	pcie_do_recovery(pdev, DPC_FATAL);
>>  }
>> 
>>  static void dpc_process_rp_pio_error(struct dpc_dev *dpc)
>> @@ -288,6 +291,7 @@ static struct pcie_port_service_driver dpcdriver = 
>> {
>>  	.service	= PCIE_PORT_SERVICE_DPC,
>>  	.probe		= dpc_probe,
>>  	.remove		= dpc_remove,
>> +	.reset_link	= dpc_reset_link,
>>  };
>> 
>>  static int __init dpc_service_init(void)
>> diff --git a/drivers/pci/pcie/err.c b/drivers/pci/pcie/err.c
>> index 877785d..526aba8 100644
>> --- a/drivers/pci/pcie/err.c
>> +++ b/drivers/pci/pcie/err.c
>> @@ -181,11 +181,12 @@ static pci_ers_result_t 
>> default_reset_link(struct pci_dev *dev)
>>  	return PCI_ERS_RESULT_RECOVERED;
>>  }
>> 
>> -static pci_ers_result_t reset_link(struct pci_dev *dev)
>> +static pci_ers_result_t reset_link(struct pci_dev *dev, int severity)
>>  {
>>  	struct pci_dev *udev;
>>  	pci_ers_result_t status;
>>  	struct pcie_port_service_driver *driver;
>> +	u32 service;
>> 
>>  	if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE) {
>>  		/* Reset this port for all subordinates */
>> @@ -196,7 +197,12 @@ static pci_ers_result_t reset_link(struct pci_dev 
>> *dev)
>>  	}
>> 
>>  	/* Use the aer driver of the component firstly */
>> -	driver = pcie_port_find_service(udev, PCIE_PORT_SERVICE_AER);
>> +	if (severity == DPC_FATAL)
>> +		service = PCIE_PORT_SERVICE_DPC;
>> +	else
>> +		service = PCIE_PORT_SERVICE_AER;
>> +
>> +	driver = pcie_port_find_service(udev, service);
> 
> This is where I was wondering about passing in "service" directly 
> instead
> of "severity".

I will take care of all your comments made on all the patches.
but this one I do not understand.

passing service directly instead of severity ? (I do not think you meant 
to write severity there)
perhaps do you mean following ?

if (severity == DPC_FATAL)
          pcie_port_find_service(udev, PCIE_PORT_SERVICE_DPC);
else
          pcie_port_find_service(udev, PCIE_PORT_SERVICE_AER);



> 
>>  	if (driver && driver->reset_link) {
>>  		status = driver->reset_link(udev);
>> @@ -302,7 +308,7 @@ static pci_ers_result_t do_fatal_recovery(struct 
>> pci_dev *dev, int severity)
>>  		pci_dev_put(pdev);
>>  	}
>> 
>> -	result = reset_link(udev);
>> +	result = reset_link(udev, severity);
>>  	if (result == PCI_ERS_RESULT_RECOVERED)
>>  		if (pcie_wait_for_link(udev, true))
>>  			pci_rescan_bus(udev->bus);
>> @@ -326,7 +332,8 @@ void pcie_do_recovery(struct pci_dev *dev, int 
>> severity)
>>  	pci_ers_result_t status;
>>  	enum pci_channel_state state;
>> 
>> -	if (severity == AER_FATAL) {
>> +	if ((severity == AER_FATAL) ||
>> +	   (severity == DPC_FATAL)) {
>>  		status = do_fatal_recovery(dev, severity);
>>  		if (status != PCI_ERS_RESULT_RECOVERED)
>>  			goto failed;
>> diff --git a/include/linux/aer.h b/include/linux/aer.h
>> index 8f87bbe..0c506fe 100644
>> --- a/include/linux/aer.h
>> +++ b/include/linux/aer.h
>> @@ -14,6 +14,7 @@
>>  #define AER_NONFATAL			0
>>  #define AER_FATAL			1
>>  #define AER_CORRECTABLE			2
>> +#define DPC_FATAL			4
>> 
>>  struct pci_dev;
>> 
>> --
>> 2.7.4
>>