Re: [PATCH v2 2/2] PCI/DPC: Add Error Disconnect Recover (EDR) support

[Bjorn Helgaas]

On Tue, Mar 19, 2019 at 01:47:29PM -0700, sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx wrote:
> From: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
> 
> As per PCI firmware specification v3.2 ECN
> (https://members.pcisig.com/wg/PCI-SIG/document/12614), when firmware
> owns Downstream Port Containment (DPC), its expected to use the "Error
> Disconnect Recover" (EDR) notification to alert OSPM of a DPC event and
> if OS supports EDR, its expected to handle the software state invalidation
> and port recovery in OS and let firmware know the recovery status via _OST
> ACPI call.
> 
> Since EDR is a hybrid service, DPC service shall be enabled in OS even
> if AER is not natively enabled in OS.
> 
> Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
> ---
>  drivers/pci/pcie/Kconfig        |  10 +
>  drivers/pci/pcie/dpc.c          | 326 ++++++++++++++++++++++++++++++--

I'll be looking for Keith's reviewed-by for this eventually.

It looks like this could be split into some smaller patches:

  - Add dpc_dev.native_dpc (hopefully we can find a less confusing name)
  - Convert dpc_handler() to dpc_handler() + dpc_process_error()
  - Add EDR support

>  drivers/pci/pcie/portdrv_core.c |   9 +-
>  3 files changed, 329 insertions(+), 16 deletions(-)
> 
> diff --git a/drivers/pci/pcie/Kconfig b/drivers/pci/pcie/Kconfig
> index 5cbdbca904ac..55f65d1cbc9e 100644
> --- a/drivers/pci/pcie/Kconfig
> +++ b/drivers/pci/pcie/Kconfig
> @@ -142,3 +142,13 @@ config PCIE_PTM
>  
>  	  This is only useful if you have devices that support PTM, but it
>  	  is safe to enable even if you don't.
> +
> +config PCIE_EDR
> +	bool "PCI Express Error Disconnect Recover support"
> +	default n
> +	depends on PCIE_DPC && ACPI
> +	help
> +	  This options adds Error Disconnect Recover support as specified
> +	  in PCI firmware specification v3.2 Downstream Port Containment
> +	  Related Enhancements ECN. Enable this if you want to support hybrid
> +	  DPC model which uses both firmware and OS to implement DPC.
> diff --git a/drivers/pci/pcie/dpc.c b/drivers/pci/pcie/dpc.c
> index 7b77754a82de..bdf4ca8a0acb 100644
> --- a/drivers/pci/pcie/dpc.c
> +++ b/drivers/pci/pcie/dpc.c
> @@ -11,6 +11,7 @@
>  #include <linux/interrupt.h>
>  #include <linux/init.h>
>  #include <linux/pci.h>
> +#include <linux/acpi.h>
>  
>  #include "portdrv.h"
>  #include "../pci.h"
> @@ -20,8 +21,23 @@ struct dpc_dev {
>  	u16			cap_pos;
>  	bool			rp_extensions;
>  	u8			rp_log_size;
> +	bool			native_dpc;

This is going to be way too confusing with a "native_dpc" in both the
struct pci_host_bridge and the struct dpc_dev.

> +	pci_ers_result_t	error_state;
> +#ifdef CONFIG_ACPI
> +	struct acpi_device	*adev;
> +#endif
>  };
>  
> +#ifdef CONFIG_ACPI
> +
> +#define EDR_PORT_ENABLE_DSM     0x0C
> +#define EDR_PORT_LOCATE_DSM     0x0D
> +
> +static const guid_t pci_acpi_dsm_guid =
> +		GUID_INIT(0xe5c937d0, 0x3553, 0x4d7a,
> +			  0x91, 0x17, 0xea, 0x4d, 0x19, 0xc3, 0x43, 0x4d);
> +#endif
> +
>  static const char * const rp_pio_error_string[] = {
>  	"Configuration Request received UR Completion",	 /* Bit Position 0  */
>  	"Configuration Request received CA Completion",	 /* Bit Position 1  */
> @@ -67,6 +83,9 @@ void pci_save_dpc_state(struct pci_dev *dev)
>  	if (!dpc)
>  		return;
>  
> +	if (!dpc->native_dpc)
> +		return;
> +
>  	save_state = pci_find_saved_ext_cap(dev, PCI_EXT_CAP_ID_DPC);
>  	if (!save_state)
>  		return;
> @@ -88,6 +107,9 @@ void pci_restore_dpc_state(struct pci_dev *dev)
>  	if (!dpc)
>  		return;
>  
> +	if (!dpc->native_dpc)
> +		return;
> +
>  	save_state = pci_find_saved_ext_cap(dev, PCI_EXT_CAP_ID_DPC);
>  	if (!save_state)
>  		return;
> @@ -224,10 +246,9 @@ static int dpc_get_aer_uncorrect_severity(struct pci_dev *dev,
>  	return 1;
>  }
>  
> -static irqreturn_t dpc_handler(int irq, void *context)
> +static void dpc_process_error(struct dpc_dev *dpc)
>  {
>  	struct aer_err_info info;
> -	struct dpc_dev *dpc = context;
>  	struct pci_dev *pdev = dpc->dev->port;
>  	struct device *dev = &dpc->dev->device;
>  	u16 cap = dpc->cap_pos, status, source, reason, ext_reason;
> @@ -261,6 +282,13 @@ static irqreturn_t dpc_handler(int irq, void *context)
>  
>  	/* We configure DPC so it only triggers on ERR_FATAL */
>  	pcie_do_recovery(pdev, pci_channel_io_frozen, PCIE_PORT_SERVICE_DPC);
> +}
> +
> +static irqreturn_t dpc_handler(int irq, void *context)
> +{
> +	struct dpc_dev *dpc = context;
> +
> +	dpc_process_error(dpc);
>  
>  	return IRQ_HANDLED;
>  }
> @@ -283,6 +311,230 @@ static irqreturn_t dpc_irq(int irq, void *context)
>  	return IRQ_HANDLED;
>  }
>  
> +void dpc_error_resume(struct pci_dev *dev)

Looks like this should be static?

> +{
> +	struct dpc_dev *dpc;
> +
> +	dpc = to_dpc_dev(dev);
> +	if (!dpc)
> +		return;

I don't think it's possible for dpc to be NULL, is it?  Remove the
test if not.

> +	dpc->error_state = PCI_ERS_RESULT_RECOVERED;
> +}
> +
> +#ifdef CONFIG_ACPI
> +
> +/*
> + * _DSM wrapper function to enable/disable DPC port.
> + * @dpc   : DPC device structure
> + * @enable: status of DPC port (0 or 1).
> + *
> + * returns 0 on success or errno on failure.
> + */
> +static int acpi_enable_dpc_port(struct dpc_dev *dpc, bool enable)
> +{
> +	union acpi_object *obj;
> +	int status;
> +	union acpi_object argv4;
> +
> +	/* Check whether EDR_PORT_ENABLE_DSM is supported in firmware */
> +	status = acpi_check_dsm(dpc->adev->handle, &pci_acpi_dsm_guid, 1,
> +				1 << EDR_PORT_ENABLE_DSM);

I don't think this acpi_check_dsm() is necessary, is it?  I expect
acpi_evaluate_dsm() to fail nicely if there's no _DSM or it doesn't
support the function.

> +	if (!status)
> +		return -ENOTSUPP;
> +
> +	argv4.type = ACPI_TYPE_INTEGER;
> +	argv4.integer.value = enable;
> +
> +	obj = acpi_evaluate_dsm(dpc->adev->handle, &pci_acpi_dsm_guid, 1,
> +				EDR_PORT_ENABLE_DSM, &argv4);
> +	if (!obj)
> +		return -EIO;
> +
> +	if (obj->type == ACPI_TYPE_INTEGER && obj->integer.value == enable)
> +		status = 0;
> +	else
> +		status = -EIO;
> +
> +	ACPI_FREE(obj);
> +
> +	return status;
> +}
> +
> +/*
> + * _DSM wrapper function to locate DPC port.
> + * @dpc   : DPC device structure
> + *
> + * returns pci_dev or NULL.
> + */
> +static struct pci_dev *acpi_locate_dpc_port(struct dpc_dev *dpc)
> +{
> +	union acpi_object *obj;
> +	int status;
> +	u16 port;
> +
> +	/* Check whether EDR_PORT_LOCATE_DSM is supported in firmware */
> +	status = acpi_check_dsm(dpc->adev->handle, &pci_acpi_dsm_guid, 1,
> +				1 << EDR_PORT_LOCATE_DSM);

Unnecessary?

> +	if (!status)
> +		return dpc->dev->port;

If you *do* need the acpi_check_dsm(), I would have expected to return
NULL in this error case?

> +
> +

Extra blank line here.

> +	obj = acpi_evaluate_dsm(dpc->adev->handle, &pci_acpi_dsm_guid, 1,
> +				EDR_PORT_LOCATE_DSM, NULL);
> +	if (!obj)
> +		return NULL;
> +
> +	if (obj->type == ACPI_TYPE_INTEGER) {
> +		/*
> +		 * Firmware returns DPC port BDF details in following format:
> +		 *	15:8 = bus
> +		 *	7:3 = device
> +		 *	2:0 = function
> +		 */
> +		port = obj->integer.value;
> +		ACPI_FREE(obj);
> +	} else {
> +		ACPI_FREE(obj);
> +		return NULL;
> +	}
> +
> +	return pci_get_domain_bus_and_slot(0, PCI_BUS_NUM(port), port & 0xff);
> +}
> +
> +/*
> + * _OST wrapper function to let firmware know the status of EDR event.
> + * @dpc   : DPC device structure.
> + * @status: Status of EDR event.
> + *

Spurious blank line in the comment.

> + */
> +static int acpi_send_edr_status(struct dpc_dev *dpc,  u16 status)
> +{
> +	u32 ost_status;
> +	struct pci_dev *pdev = dpc->dev->port;
> +
> +	dev_dbg(&pdev->dev, "Sending EDR status :%x\n", status);
> +
> +	ost_status =  PCI_DEVID(pdev->bus->number, pdev->devfn);
> +	ost_status = (ost_status << 16) | status;
> +
> +	if (!acpi_has_method(dpc->adev->handle, "_OST"))
> +		return -ENOTSUPP;

This acpi_has_method() check is superfluous, isn't it?  I would expect
acpi_evaluate_ost() to fail gracefully if there's no _OST.

I do see several other instances of code of the form:

  if (!acpi_has_method(handle, "XXXX"))
    return false;

  status = acpi_evaluate_*(handle, "XXXX", ...);

But I think that's a pointless pattern.  I think we could just try to
evaluate the method directly, since the evaluation will fail if the
method doesn't exist:

  status = acpi_evaluate_*(handle, "XXXX", ...);

> +	status = acpi_evaluate_ost(dpc->adev->handle,
> +				   ACPI_NOTIFY_DISCONNECT_RECOVER,
> +				   ost_status, NULL);
> +	if (ACPI_FAILURE(status))
> +		return -EINVAL;
> +
> +	return 0;
> +}
> +
> +/*
> + * Helper function used for disconnecting the child devices when EDR event is
> + * received from firmware.
> + */
> +static void dpc_disconnect_devices(struct pci_dev *dev)
> +{
> +	struct pci_dev *udev;
> +	struct pci_bus *parent;
> +	struct pci_dev *pdev, *temp;
> +
> +	dev_dbg(&dev->dev, "Disconnecting the child devices\n");
> +
> +	if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE)
> +		udev = dev;
> +	else
> +		udev = dev->bus->self;
> +
> +	parent = udev->subordinate;
> +	pci_walk_bus(parent, pci_dev_set_disconnected, NULL);
> +
> +	pci_lock_rescan_remove();
> +	pci_dev_get(dev);
> +	list_for_each_entry_safe_reverse(pdev, temp, &parent->devices,
> +					 bus_list) {
> +		pci_stop_and_remove_bus_device(pdev);
> +	}
> +	pci_dev_put(dev);
> +	pci_unlock_rescan_remove();
> +}
> +
> +static void edr_handle_event(acpi_handle handle, u32 event, void *data)
> +{
> +	struct dpc_dev *dpc = (struct dpc_dev *) data;
> +	struct pci_dev *pdev;
> +	u16 status, cap;
> +
> +	if (event != ACPI_NOTIFY_DISCONNECT_RECOVER)
> +		return;
> +
> +	if (!data) {
> +		pr_err("Invalid EDR event\n");

In what instance can "data" be NULL?  I think *never*, unless ACPI
screwed up and lost the context you supplied to
acpi_install_notify_handler().  In that case, we should do something
more significant than print a message and just return.  I think we
should just omit this test, and if ACPI screwed up, we'll take a null
pointer dereference oops, we'll see exactly where, and we'll have a
chance to fix the problem.

> +		return;
> +	}
> +
> +	dev_dbg(&dpc->dev->port->dev, "Valid EDR event received\n");

Set "pdev" at declaration and use it in these printks.

> +
> +	/*
> +	 * Check if _DSM(0xD) is available, and if present locate the
> +	 * port which issued EDR event.
> +	 */
> +	pdev = acpi_locate_dpc_port(dpc);

Can this be done at probe-time instead of event-time?

> +	if (!pdev) {
> +		dev_err(&dpc->dev->port->dev, "No valid port found\n");
> +		return;
> +	}
> +
> +	/*
> +	 * Get DPC priv data for given pdev
> +	 */
> +	dpc = to_dpc_dev(pdev);
> +	dpc->error_state = PCI_ERS_RESULT_DISCONNECT;
> +	pdev = dpc->dev->port;

It's quite confusing to reassign pdev here.  The comment about "Get
DPC priv data for given pdev" is really superfluous since that much is
obvious from the code.  What's *not* obvious is whether this "pdev =
dpc->dev->port" changes anything and if so, why.

> +	cap = dpc->cap_pos;
> +
> +	/*
> +	 * Check if the port supports DPC:
> +	 *
> +	 * if port does not support DPC, then let firmware handle
> +	 * the error recovery and OS is responsible for cleaning
> +	 * up the child devices.
> +	 *
> +	 * if port supports DPC, then fall back to default error
> +	 * recovery.

Capitalize first letter of sentences.

> +	 *
> +	 */
> +	if (cap) {
> +		/* Check if there is a valid DPC trigger */
> +		pci_read_config_word(pdev, cap + PCI_EXP_DPC_STATUS, &status);
> +		if (!(status & PCI_EXP_DPC_STATUS_TRIGGER)) {
> +			dev_err(&pdev->dev, "Invalid DPC trigger\n");

Include "status" value in the printk.

> +			return;
> +		}
> +		dpc_process_error(dpc);
> +	}
> +
> +	if (dpc->error_state == PCI_ERS_RESULT_RECOVERED) {
> +		/*
> +		 * Recovery is successful, so send
> +		 * _OST(0xF, BDF << 16 | 0x80, "") to firmware.
> +		 */
> +		status = 0x80;
> +	} else {
> +		/*
> +		 * Recovery is not successful, so disconnect child devices
> +		 * and send _OST(0xF, BDF << 16 | 0x81, "") to firmware.
> +		 */
> +		dpc_disconnect_devices(pdev);
> +		status = 0x81;
> +	}
> +
> +	acpi_send_edr_status(dpc, status);
> +}
> +
> +#endif
> +
>  #define FLAG(x, y) (((x) & (y)) ? '+' : '-')
>  static int dpc_probe(struct pcie_device *dev)
>  {
> @@ -291,9 +543,10 @@ static int dpc_probe(struct pcie_device *dev)
>  	struct device *device = &dev->device;
>  	int status;
>  	u16 ctl, cap;
> -
> -	if (pcie_aer_get_firmware_first(pdev))
> -		return -ENOTSUPP;
> +#ifdef CONFIG_ACPI
> +	struct acpi_device *adev = ACPI_COMPANION(&pdev->dev);
> +	acpi_status astatus;
> +#endif
>  
>  	dpc = devm_kzalloc(device, sizeof(*dpc), GFP_KERNEL);
>  	if (!dpc)
> @@ -302,16 +555,53 @@ static int dpc_probe(struct pcie_device *dev)
>  	dpc->cap_pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_DPC);
>  	dpc->dev = dev;
>  	set_service_data(dev, dpc);
> +	dpc->error_state = PCI_ERS_RESULT_NONE;
> +
> +	if (!pcie_aer_get_firmware_first(pdev))
> +		if (pci_aer_available() && dpc->cap_pos)
> +			dpc->native_dpc = 1;
>  
> -	status = devm_request_threaded_irq(device, dev->irq, dpc_irq,
> -					   dpc_handler, IRQF_SHARED,
> -					   "pcie-dpc", dpc);
> -	if (status) {
> -		dev_warn(device, "request IRQ%d failed: %d\n", dev->irq,
> -			 status);
> -		return status;
> +	/*
> +	 * If native support is not enabled and ACPI is not
> +	 * enabled then return error.
> +	 */
> +	if (!dpc->native_dpc && !IS_ENABLED(CONFIG_APCI))
> +		return -ENODEV;
> +
> +	if (dpc->native_dpc) {
> +		status = devm_request_threaded_irq(device, dev->irq, dpc_irq,
> +						   dpc_handler, IRQF_SHARED,
> +						   "pcie-dpc", dpc);
> +		if (status) {
> +			dev_warn(device, "request IRQ%d failed: %d\n", dev->irq,
> +				 status);
> +			return status;
> +		}
>  	}
>  
> +#ifdef CONFIG_ACPI
> +	if (!dpc->native_dpc) {
> +		if (!adev) {
> +			dev_err(device, "No valid acpi device found\n");

s/acpi/ACPI/ (in comments, printk, other English text)

> +			return -ENODEV;
> +		}
> +
> +		dpc->adev = adev;
> +
> +		/* Register ACPI notifier for EDR event */

"Register handler for System events, one of which is the EDR event"?

> +		astatus = acpi_install_notify_handler(adev->handle,
> +						      ACPI_SYSTEM_NOTIFY,
> +						      edr_handle_event,
> +						      dpc);
> +
> +		if (ACPI_FAILURE(astatus)) {
> +			dev_err(device, "Install notifier failed\n");

Mention "ACPI notify handler" here.

> +			return -EBUSY;
> +		}
> +
> +		acpi_enable_dpc_port(dpc, true);
> +	}
> +#endif
>  	pci_read_config_word(pdev, dpc->cap_pos + PCI_EXP_DPC_CAP, &cap);
>  	pci_read_config_word(pdev, dpc->cap_pos + PCI_EXP_DPC_CTL, &ctl);
>  
> @@ -325,8 +615,12 @@ static int dpc_probe(struct pcie_device *dev)
>  		}
>  	}
>  
> -	ctl = (ctl & 0xfff4) | PCI_EXP_DPC_CTL_EN_FATAL | PCI_EXP_DPC_CTL_INT_EN;
> -	pci_write_config_word(pdev, dpc->cap_pos + PCI_EXP_DPC_CTL, ctl);
> +	if (dpc->native_dpc) {
> +		ctl = (ctl & 0xfff4) | PCI_EXP_DPC_CTL_EN_FATAL |
> +			PCI_EXP_DPC_CTL_INT_EN;
> +		pci_write_config_word(pdev, dpc->cap_pos + PCI_EXP_DPC_CTL,
> +				      ctl);
> +	}
>  
>  	dev_info(device, "DPC error containment capabilities: Int Msg #%d, RPExt%c PoisonedTLP%c SwTrigger%c RP PIO Log %d, DL_ActiveErr%c\n",
>  		cap & PCI_EXP_DPC_IRQ, FLAG(cap, PCI_EXP_DPC_CAP_RP_EXT),
> @@ -344,6 +638,9 @@ static void dpc_remove(struct pcie_device *dev)
>  	struct pci_dev *pdev = dev->port;
>  	u16 ctl;
>  
> +	if (!dpc->native_dpc)
> +		return;
> +
>  	pci_read_config_word(pdev, dpc->cap_pos + PCI_EXP_DPC_CTL, &ctl);
>  	ctl &= ~(PCI_EXP_DPC_CTL_EN_FATAL | PCI_EXP_DPC_CTL_INT_EN);
>  	pci_write_config_word(pdev, dpc->cap_pos + PCI_EXP_DPC_CTL, ctl);
> @@ -356,6 +653,7 @@ static struct pcie_port_service_driver dpcdriver = {
>  	.probe		= dpc_probe,
>  	.remove		= dpc_remove,
>  	.reset_link	= dpc_reset_link,
> +	.error_resume   = dpc_error_resume,
>  };
>  
>  int __init pcie_dpc_init(void)
> diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c
> index 865f12f4b314..050cbb7a5083 100644
> --- a/drivers/pci/pcie/portdrv_core.c
> +++ b/drivers/pci/pcie/portdrv_core.c
> @@ -249,9 +249,14 @@ static int get_port_device_capability(struct pci_dev *dev)
>  		pcie_pme_interrupt_enable(dev, false);
>  	}
>  
> +	/*
> +	 * If EDR support is enabled in OS, then even if AER is not handled in
> +	 * OS, DPC service can be enabled.
> +	 */
>  	if (pci_find_ext_capability(dev, PCI_EXT_CAP_ID_DPC) &&
> -	    pci_aer_available() && services & PCIE_PORT_SERVICE_AER &&
> -	    (pcie_ports_native || host->native_dpc))
> +	    ((IS_ENABLED(CONFIG_PCIE_EDR) && !host->native_dpc) ||
> +	    (pci_aer_available() && services & PCIE_PORT_SERVICE_AER &&
> +	    (pcie_ports_native || host->native_dpc))))

Holy cow, I think I'll have to schedule an hour sometime to figure
out what's going on here :)

>  		services |= PCIE_PORT_SERVICE_DPC;
>  
>  	if (pci_pcie_type(dev) == PCI_EXP_TYPE_DOWNSTREAM || -- 2.20.1
>