[PATCH 0/3] pci/iommu: Quirk non-compliant PCIe-to-PCI bridges

[PATCH 0/3] pci/iommu: Quirk non-compliant PCIe-to-PCI bridges

[Alex Williamson]

This series tries to address:

https://bugzilla.kernel.org/show_bug.cgi?id=44881

Where pci_find_upstream_pcie_bridge() gets lost trying to find the
upstream PCIe-to-PCI bridge for a device because the bridge doesn't
expose a PCIe capability.  To do this, we add a pci_is_pcie_bridge
function which includes a quirk to look to the next upstream device
as a sanity check.  We can then replace pci_find_upstream_pcie_bridge
with a function that's a bit more generic and less tied to intel-iommu
eccentricities.

This works on my VT-d system, but I don't have any of the offending
bridges.  Appreciate testing from those who do.  I'll link this series
to the bz above for more testing.  Thanks,

Alex

---

Alex Williamson (3):
      pci: Add PCI walk function and PCIe bridge test
      intel-iommu: Convert to pci_walk_up_to_first_match
      pci: Remove pci_find_pcie_upstream_bridge


 drivers/iommu/intel-iommu.c         |   77 ++++++++++++++++++++++-------------
 drivers/iommu/intel_irq_remapping.c |   13 ++++--
 drivers/pci/search.c                |   74 ++++++++++++++++++++++------------
 include/linux/pci.h                 |   26 +++++++++---
 4 files changed, 124 insertions(+), 66 deletions(-)

[PATCH 1/3] pci: Add PCI walk function and PCIe bridge test

[Alex Williamson]

These will replace pci_find_upstream_pcie_bridge, which is difficult
to use and rather specific to intel-iommu usage.  A quirked
pci_is_pcie_bridge function is provided to work around non-compliant
PCIe-to-PCI bridges such as those found in
https://bugzilla.kernel.org/show_bug.cgi?id=44881

Signed-off-by: Alex Williamson <alex.williamson-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 drivers/pci/search.c |   57 ++++++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/pci.h  |   23 ++++++++++++++++++++
 2 files changed, 80 insertions(+)

diff --git a/drivers/pci/search.c b/drivers/pci/search.c
index d0627fa..0357f74 100644
--- a/drivers/pci/search.c
+++ b/drivers/pci/search.c
@@ -17,6 +17,63 @@
 DECLARE_RWSEM(pci_bus_sem);
 EXPORT_SYMBOL_GPL(pci_bus_sem);
 
+/* Test for PCIe bridges. */
+bool pci_is_pcie_bridge(struct pci_dev *pdev)
+{
+	if (!pci_is_bridge(pdev))
+		return false;
+
+	if (pci_is_pcie(pdev))
+		return true;
+
+#ifdef CONFIG_PCI_QUIRKS
+	/*
+	 * If we're not on the root bus, look one device upstream of the
+	 * current device.  If that device is PCIe and is not a PCIe-to-PCI
+	 * bridge, then the current device is effectively PCIe as it must
+	 * be the PCIe-to-PCI bridge.  This handles several bridges that
+	 * violate the PCIe spec by not exposing a PCIe capability:
+	 * https://bugzilla.kernel.org/show_bug.cgi?id=44881
+	 */
+	if (!pci_is_root_bus(pdev->bus)) {
+		struct pci_dev *parent = pdev->bus->self;
+
+		if (pci_is_pcie(parent) &&
+		    pci_pcie_type(parent) != PCI_EXP_TYPE_PCI_BRIDGE)
+
+			return true;
+	}
+#endif
+	return false;
+}
+
+/*
+ * Walk upstream from the given pdev for the first device returning
+ * true for the provided match function.  If no match is found, return
+ * NULL.  *last records the previous step in the walk.
+ */
+struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
+					   bool (*match)(struct pci_dev *),
+					   struct pci_dev **last)
+{
+	*last = NULL;
+
+	if (match(pdev))
+		return pdev;
+
+	*last = pdev;
+
+	while (!pci_is_root_bus(pdev->bus)) {
+		*last = pdev;
+		pdev = pdev->bus->self;
+
+		if (match(pdev))
+			return pdev;
+	}
+
+	return NULL;
+}
+
 /*
  * find the upstream PCIe-to-PCI bridge of a PCI device
  * if the device is PCIE, return NULL
diff --git a/include/linux/pci.h b/include/linux/pci.h
index bd8ec30..e87423a 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1855,6 +1855,29 @@ static inline struct eeh_dev *pci_dev_to_eeh_dev(struct pci_dev *pdev)
 #endif
 
 /**
+ * pci_walk_up_to_first_match - Generic upstream search function
+ * @pdev: starting PCI device to search
+ * @match: match function to call on each device (true = match)
+ * @last: last device examined prior to returned device
+ *
+ * Walk upstream from the given device, calling match() at each device.
+ * Returns the first device matching match().  If the root bus is reached
+ * without finding a match, return NULL.  last returns the N-1 step in
+ * the search path.
+ */
+struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
+					   bool (*match)(struct pci_dev *),
+					   struct pci_dev **last);
+
+/**
+ * pci_is_pcie_bridge - Match a PCIe bridge device
+ * @pdev: device to test
+ *
+ * Return true if the given device is a PCIe bridge, false otherwise.
+ */
+bool pci_is_pcie_bridge(struct pci_dev *pdev);
+
+/**
  * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge of a device
  * @pdev: the PCI device
  *

[PATCH 2/3] intel-iommu: Convert to pci_walk_up_to_first_match

[Alex Williamson]

pci_find_upstream_pcie_bridge() attempts to return the PCIe-to-PCI
bridge upstream from the provided device.  It currently returns NULL
if 1) the provided device is PCIe (ie. it does not have such an
upstream bridge), 2) the device is on the root bus (a subtle case of
its claim to return the parent device when not connected to a PCIe
bridge), or 3) when it gets lost walking the bus and finds itself on
a PCIe device that is not a PCIe-to-PCI bridge.  This latter case not
only generates a WARN, but treats the provided device the same as if
it were PCIe.

We can replace 1) and 2) with explicit tests in the IOMMU code as this
is the code that knows the visibility of devices at the root bus.
That leaves the bus walk case 3), where the NULL return is actually an
error.

We hit this error because some PCIe-to-PCI bridge vendors ignore the
PCIe spec requirement that all PCIe devices must include a PCIe
capability, which is all that pci_is_pcie() actually tests.  When
quirks are enabled, pci_is_pcie_bridge() adds one more test by looking
at the next upstream device to test whether it is PCIe with a type
code of a PCIe-to-PCI/PCI-X bridge.  If it does not have this type
code, then by deduction, the current bridge must be a PCIe-to-PCI
bridge and thus be a PCIe bridge.

Therefore, pci_find_upstream_pcie_bridge() is replaced by an
explicit call to (pci_is_pcie() || pci_is_root_bus()) plus
pci_walk_up_to_first_match() with the pci_is_pcie_bridge() match
function.  This either returns the first PCIe bridge, which must
be a PCIe-to-PCI bridge or NULL if not found.  In the NULL case, we
know the original device is not on the root bus (already tested for
that), so we can safely walk from the provided device to the last
device and fall into the legacy PCI bridge code.

Signed-off-by: Alex Williamson <alex.williamson-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 drivers/iommu/intel-iommu.c         |   77 ++++++++++++++++++++++-------------
 drivers/iommu/intel_irq_remapping.c |   13 ++++--
 2 files changed, 57 insertions(+), 33 deletions(-)

diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index 0099667..277c18f 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -1688,9 +1688,13 @@ domain_context_mapping(struct dmar_domain *domain, struct pci_dev *pdev,
 		return ret;
 
 	/* dependent device mapping */
-	tmp = pci_find_upstream_pcie_bridge(pdev);
-	if (!tmp)
+	if (pci_is_pcie(pdev) || pci_is_root_bus(pdev->bus))
 		return 0;
+
+	tmp = pci_walk_up_to_first_match(pdev, pci_is_pcie_bridge, &parent);
+	if (!tmp)
+		tmp = parent;
+
 	/* Secondary interface's bus number and devfn 0 */
 	parent = pdev->bus->self;
 	while (parent != tmp) {
@@ -1702,7 +1706,7 @@ domain_context_mapping(struct dmar_domain *domain, struct pci_dev *pdev,
 			return ret;
 		parent = parent->bus->self;
 	}
-	if (pci_is_pcie(tmp)) /* this is a PCIe-to-PCI bridge */
+	if (pci_is_pcie_bridge(tmp)) /* this is a PCIe-to-PCI bridge */
 		return domain_context_mapping_one(domain,
 					pci_domain_nr(tmp->subordinate),
 					tmp->subordinate->number, 0,
@@ -1729,10 +1733,15 @@ static int domain_context_mapped(struct pci_dev *pdev)
 	ret = device_context_mapped(iommu, pdev->bus->number, pdev->devfn);
 	if (!ret)
 		return ret;
+
 	/* dependent device mapping */
-	tmp = pci_find_upstream_pcie_bridge(pdev);
+	if (pci_is_pcie(pdev) || pci_is_root_bus(pdev->bus))
+		return 0;
+
+	tmp = pci_walk_up_to_first_match(pdev, pci_is_pcie_bridge, &parent);
 	if (!tmp)
-		return ret;
+		tmp = parent;
+
 	/* Secondary interface's bus number and devfn 0 */
 	parent = pdev->bus->self;
 	while (parent != tmp) {
@@ -1742,7 +1751,7 @@ static int domain_context_mapped(struct pci_dev *pdev)
 			return ret;
 		parent = parent->bus->self;
 	}
-	if (pci_is_pcie(tmp))
+	if (pci_is_pcie_bridge(tmp))
 		return device_context_mapped(iommu, tmp->subordinate->number,
 					     0);
 	else
@@ -1973,7 +1982,7 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw)
 	struct intel_iommu *iommu;
 	struct dmar_drhd_unit *drhd;
 	struct device_domain_info *info, *tmp;
-	struct pci_dev *dev_tmp;
+	struct pci_dev *dev_tmp = NULL;
 	unsigned long flags;
 	int bus = 0, devfn = 0;
 	int segment;
@@ -1985,9 +1994,15 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw)
 
 	segment = pci_domain_nr(pdev->bus);
 
-	dev_tmp = pci_find_upstream_pcie_bridge(pdev);
-	if (dev_tmp) {
-		if (pci_is_pcie(dev_tmp)) {
+	if (!pci_is_pcie(pdev) && !pci_is_root_bus(pdev->bus)) {
+		struct pci_dev *last;
+
+		dev_tmp = pci_walk_up_to_first_match(pdev, pci_is_pcie_bridge,
+						     &last);
+		if (!dev_tmp)
+			dev_tmp = last;
+
+		if (pci_is_pcie_bridge(dev_tmp)) {
 			bus = dev_tmp->subordinate->number;
 			devfn = 0;
 		} else {
@@ -3738,26 +3753,24 @@ static void iommu_detach_dependent_devices(struct intel_iommu *iommu,
 {
 	struct pci_dev *tmp, *parent;
 
-	if (!iommu || !pdev)
+	if (!iommu || !pdev || pci_is_pcie(pdev) || pci_is_root_bus(pdev->bus))
 		return;
 
 	/* dependent device detach */
-	tmp = pci_find_upstream_pcie_bridge(pdev);
+	tmp = pci_walk_up_to_first_match(pdev, pci_is_pcie_bridge, &parent);
+	if (!tmp)
+		tmp = parent;
+
 	/* Secondary interface's bus number and devfn 0 */
-	if (tmp) {
-		parent = pdev->bus->self;
-		while (parent != tmp) {
-			iommu_detach_dev(iommu, parent->bus->number,
-					 parent->devfn);
-			parent = parent->bus->self;
-		}
-		if (pci_is_pcie(tmp)) /* this is a PCIe-to-PCI bridge */
-			iommu_detach_dev(iommu,
-				tmp->subordinate->number, 0);
-		else /* this is a legacy PCI bridge */
-			iommu_detach_dev(iommu, tmp->bus->number,
-					 tmp->devfn);
+	parent = pdev->bus->self;
+	while (parent != tmp) {
+		iommu_detach_dev(iommu, parent->bus->number, parent->devfn);
+		parent = parent->bus->self;
 	}
+	if (pci_is_pcie_bridge(tmp)) /* this is a PCIe-to-PCI bridge */
+		iommu_detach_dev(iommu, tmp->subordinate->number, 0);
+	else /* this is a legacy PCI bridge */
+		iommu_detach_dev(iommu, tmp->bus->number, tmp->devfn);
 }
 
 static void domain_remove_one_dev_info(struct dmar_domain *domain,
@@ -4148,7 +4161,7 @@ static void swap_pci_ref(struct pci_dev **from, struct pci_dev *to)
 static int intel_iommu_add_device(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
-	struct pci_dev *bridge, *dma_pdev = NULL;
+	struct pci_dev *dma_pdev = NULL;
 	struct iommu_group *group;
 	int ret;
 
@@ -4156,9 +4169,15 @@ static int intel_iommu_add_device(struct device *dev)
 			     pdev->bus->number, pdev->devfn))
 		return -ENODEV;
 
-	bridge = pci_find_upstream_pcie_bridge(pdev);
-	if (bridge) {
-		if (pci_is_pcie(bridge))
+	if (!pci_is_pcie(pdev) && !pci_is_root_bus(pdev->bus)) {
+		struct pci_dev *bridge, *last;
+
+		bridge = pci_walk_up_to_first_match(pdev, pci_is_pcie_bridge,
+						    &last);
+		if (!bridge)
+			bridge = last;
+
+		if (pci_is_pcie_bridge(bridge))
 			dma_pdev = pci_get_domain_bus_and_slot(
 						pci_domain_nr(pdev->bus),
 						bridge->subordinate->number, 0);
diff --git a/drivers/iommu/intel_irq_remapping.c b/drivers/iommu/intel_irq_remapping.c
index f3b8f23..873da76 100644
--- a/drivers/iommu/intel_irq_remapping.c
+++ b/drivers/iommu/intel_irq_remapping.c
@@ -373,7 +373,6 @@ static int set_hpet_sid(struct irte *irte, u8 id)
 
 static int set_msi_sid(struct irte *irte, struct pci_dev *dev)
 {
-	struct pci_dev *bridge;
 
 	if (!irte || !dev)
 		return -1;
@@ -385,9 +384,15 @@ static int set_msi_sid(struct irte *irte, struct pci_dev *dev)
 		return 0;
 	}
 
-	bridge = pci_find_upstream_pcie_bridge(dev);
-	if (bridge) {
-		if (pci_is_pcie(bridge))/* this is a PCIe-to-PCI/PCIX bridge */
+	if (!pci_is_pcie(dev) && !pci_is_root_bus(dev->bus)) {
+		struct pci_dev *bridge, *last;
+
+		bridge = pci_walk_up_to_first_match(dev,
+						    pci_is_pcie_bridge, &last);
+		if (!bridge)
+			bridge = last;
+
+		if (pci_is_pcie_bridge(bridge)) /* PCIe-to-PCI/PCIX bridge */
 			set_irte_sid(irte, SVT_VERIFY_BUS, SQ_ALL_16,
 				(bridge->bus->number << 8) | dev->bus->number);
 		else /* this is a legacy PCI bridge */

[PATCH 3/3] pci: Remove pci_find_pcie_upstream_bridge

[Alex Williamson]

This function no longer has any users.

Signed-off-by: Alex williamson <alex.williamson-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 drivers/pci/search.c |   35 -----------------------------------
 include/linux/pci.h  |   11 -----------
 2 files changed, 46 deletions(-)

diff --git a/drivers/pci/search.c b/drivers/pci/search.c
index 0357f74..bdbff05 100644
--- a/drivers/pci/search.c
+++ b/drivers/pci/search.c
@@ -74,41 +74,6 @@ struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
 	return NULL;
 }
 
-/*
- * find the upstream PCIe-to-PCI bridge of a PCI device
- * if the device is PCIE, return NULL
- * if the device isn't connected to a PCIe bridge (that is its parent is a
- * legacy PCI bridge and the bridge is directly connected to bus 0), return its
- * parent
- */
-struct pci_dev *
-pci_find_upstream_pcie_bridge(struct pci_dev *pdev)
-{
-	struct pci_dev *tmp = NULL;
-
-	if (pci_is_pcie(pdev))
-		return NULL;
-	while (1) {
-		if (pci_is_root_bus(pdev->bus))
-			break;
-		pdev = pdev->bus->self;
-		/* a p2p bridge */
-		if (!pci_is_pcie(pdev)) {
-			tmp = pdev;
-			continue;
-		}
-		/* PCI device should connect to a PCIe bridge */
-		if (pci_pcie_type(pdev) != PCI_EXP_TYPE_PCI_BRIDGE) {
-			/* Busted hardware? */
-			WARN_ON_ONCE(1);
-			return NULL;
-		}
-		return pdev;
-	}
-
-	return tmp;
-}
-
 static struct pci_bus *pci_do_find_bus(struct pci_bus *bus, unsigned char busnr)
 {
 	struct pci_bus* child;
diff --git a/include/linux/pci.h b/include/linux/pci.h
index e87423a..9d9dcf5 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1877,15 +1877,4 @@ struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
  */
 bool pci_is_pcie_bridge(struct pci_dev *pdev);
 
-/**
- * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge of a device
- * @pdev: the PCI device
- *
- * if the device is PCIE, return NULL
- * if the device isn't connected to a PCIe bridge (that is its parent is a
- * legacy PCI bridge and the bridge is directly connected to bus 0), return its
- * parent
- */
-struct pci_dev *pci_find_upstream_pcie_bridge(struct pci_dev *pdev);
-
 #endif /* LINUX_PCI_H */

RE: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test

[Sethi Varun-B16395]

Would these functions be used outside drivers/iommu? We recently added pci.h under drivers/iommu, maybe we can add a new file for these functions as well.

-Varun
> -----Original Message-----
> From: iommu-bounces@lists.linux-foundation.org [mailto:iommu-
> bounces@lists.linux-foundation.org] On Behalf Of Alex Williamson
> Sent: Saturday, May 11, 2013 2:49 AM
> To: bhelgaas@google.com; stephen@networkplumber.org
> Cc: linux-pci@vger.kernel.org; iommu@lists.linux-foundation.org;
> dwmw2@infradead.org
> Subject: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test
> 
> These will replace pci_find_upstream_pcie_bridge, which is difficult to
> use and rather specific to intel-iommu usage.  A quirked
> pci_is_pcie_bridge function is provided to work around non-compliant
> PCIe-to-PCI bridges such as those found in
> https://bugzilla.kernel.org/show_bug.cgi?id=44881
> 
> Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> ---
>  drivers/pci/search.c |   57
> ++++++++++++++++++++++++++++++++++++++++++++++++++
>  include/linux/pci.h  |   23 ++++++++++++++++++++
>  2 files changed, 80 insertions(+)
> 
> diff --git a/drivers/pci/search.c b/drivers/pci/search.c index
> d0627fa..0357f74 100644
> --- a/drivers/pci/search.c
> +++ b/drivers/pci/search.c
> @@ -17,6 +17,63 @@
>  DECLARE_RWSEM(pci_bus_sem);
>  EXPORT_SYMBOL_GPL(pci_bus_sem);
> 
> +/* Test for PCIe bridges. */
> +bool pci_is_pcie_bridge(struct pci_dev *pdev) {
> +	if (!pci_is_bridge(pdev))
> +		return false;
> +
> +	if (pci_is_pcie(pdev))
> +		return true;
> +
> +#ifdef CONFIG_PCI_QUIRKS
> +	/*
> +	 * If we're not on the root bus, look one device upstream of the
> +	 * current device.  If that device is PCIe and is not a PCIe-to-PCI
> +	 * bridge, then the current device is effectively PCIe as it must
> +	 * be the PCIe-to-PCI bridge.  This handles several bridges that
> +	 * violate the PCIe spec by not exposing a PCIe capability:
> +	 * https://bugzilla.kernel.org/show_bug.cgi?id=44881
> +	 */
> +	if (!pci_is_root_bus(pdev->bus)) {
> +		struct pci_dev *parent = pdev->bus->self;
> +
> +		if (pci_is_pcie(parent) &&
> +		    pci_pcie_type(parent) != PCI_EXP_TYPE_PCI_BRIDGE)
> +
> +			return true;
> +	}
> +#endif
> +	return false;
> +}
> +
> +/*
> + * Walk upstream from the given pdev for the first device returning
> + * true for the provided match function.  If no match is found, return
> + * NULL.  *last records the previous step in the walk.
> + */
> +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> +					   bool (*match)(struct pci_dev *),
> +					   struct pci_dev **last)
> +{
> +	*last = NULL;
> +
> +	if (match(pdev))
> +		return pdev;
> +
> +	*last = pdev;
> +
> +	while (!pci_is_root_bus(pdev->bus)) {
> +		*last = pdev;
> +		pdev = pdev->bus->self;
> +
> +		if (match(pdev))
> +			return pdev;
> +	}
> +
> +	return NULL;
> +}
> +
>  /*
>   * find the upstream PCIe-to-PCI bridge of a PCI device
>   * if the device is PCIE, return NULL
> diff --git a/include/linux/pci.h b/include/linux/pci.h index
> bd8ec30..e87423a 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -1855,6 +1855,29 @@ static inline struct eeh_dev
> *pci_dev_to_eeh_dev(struct pci_dev *pdev)  #endif
> 
>  /**
> + * pci_walk_up_to_first_match - Generic upstream search function
> + * @pdev: starting PCI device to search
> + * @match: match function to call on each device (true = match)
> + * @last: last device examined prior to returned device
> + *
> + * Walk upstream from the given device, calling match() at each device.
> + * Returns the first device matching match().  If the root bus is
> +reached
> + * without finding a match, return NULL.  last returns the N-1 step in
> + * the search path.
> + */
> +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> +					   bool (*match)(struct pci_dev *),
> +					   struct pci_dev **last);
> +
> +/**
> + * pci_is_pcie_bridge - Match a PCIe bridge device
> + * @pdev: device to test
> + *
> + * Return true if the given device is a PCIe bridge, false otherwise.
> + */
> +bool pci_is_pcie_bridge(struct pci_dev *pdev);
> +
> +/**
>   * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge of a
> device
>   * @pdev: the PCI device
>   *
> 
> _______________________________________________
> iommu mailing list
> iommu@lists.linux-foundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/iommu

Re: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test

[Alex Williamson]

On Mon, 2013-05-13 at 13:51 +0000, Sethi Varun-B16395 wrote:
> Would these functions be used outside drivers/iommu? We recently added pci.h under drivers/iommu, maybe we can add a new file for these functions as well.

The intention is to make them generic enough for pci-core, unlike
pci_find_upstream_pcie_bridge, which was rather specific to iommu usage.
Thanks,

Alex

> > -----Original Message-----
> > From: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org [mailto:iommu-
> > bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org] On Behalf Of Alex Williamson
> > Sent: Saturday, May 11, 2013 2:49 AM
> > To: bhelgaas-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org; stephen-OTpzqLSitTUnbdJkjeBofR2eb7JE58TQ@public.gmane.org
> > Cc: linux-pci-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org;
> > dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org
> > Subject: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test
> > 
> > These will replace pci_find_upstream_pcie_bridge, which is difficult to
> > use and rather specific to intel-iommu usage.  A quirked
> > pci_is_pcie_bridge function is provided to work around non-compliant
> > PCIe-to-PCI bridges such as those found in
> > https://bugzilla.kernel.org/show_bug.cgi?id=44881
> > 
> > Signed-off-by: Alex Williamson <alex.williamson-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> > ---
> >  drivers/pci/search.c |   57
> > ++++++++++++++++++++++++++++++++++++++++++++++++++
> >  include/linux/pci.h  |   23 ++++++++++++++++++++
> >  2 files changed, 80 insertions(+)
> > 
> > diff --git a/drivers/pci/search.c b/drivers/pci/search.c index
> > d0627fa..0357f74 100644
> > --- a/drivers/pci/search.c
> > +++ b/drivers/pci/search.c
> > @@ -17,6 +17,63 @@
> >  DECLARE_RWSEM(pci_bus_sem);
> >  EXPORT_SYMBOL_GPL(pci_bus_sem);
> > 
> > +/* Test for PCIe bridges. */
> > +bool pci_is_pcie_bridge(struct pci_dev *pdev) {
> > +	if (!pci_is_bridge(pdev))
> > +		return false;
> > +
> > +	if (pci_is_pcie(pdev))
> > +		return true;
> > +
> > +#ifdef CONFIG_PCI_QUIRKS
> > +	/*
> > +	 * If we're not on the root bus, look one device upstream of the
> > +	 * current device.  If that device is PCIe and is not a PCIe-to-PCI
> > +	 * bridge, then the current device is effectively PCIe as it must
> > +	 * be the PCIe-to-PCI bridge.  This handles several bridges that
> > +	 * violate the PCIe spec by not exposing a PCIe capability:
> > +	 * https://bugzilla.kernel.org/show_bug.cgi?id=44881
> > +	 */
> > +	if (!pci_is_root_bus(pdev->bus)) {
> > +		struct pci_dev *parent = pdev->bus->self;
> > +
> > +		if (pci_is_pcie(parent) &&
> > +		    pci_pcie_type(parent) != PCI_EXP_TYPE_PCI_BRIDGE)
> > +
> > +			return true;
> > +	}
> > +#endif
> > +	return false;
> > +}
> > +
> > +/*
> > + * Walk upstream from the given pdev for the first device returning
> > + * true for the provided match function.  If no match is found, return
> > + * NULL.  *last records the previous step in the walk.
> > + */
> > +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> > +					   bool (*match)(struct pci_dev *),
> > +					   struct pci_dev **last)
> > +{
> > +	*last = NULL;
> > +
> > +	if (match(pdev))
> > +		return pdev;
> > +
> > +	*last = pdev;
> > +
> > +	while (!pci_is_root_bus(pdev->bus)) {
> > +		*last = pdev;
> > +		pdev = pdev->bus->self;
> > +
> > +		if (match(pdev))
> > +			return pdev;
> > +	}
> > +
> > +	return NULL;
> > +}
> > +
> >  /*
> >   * find the upstream PCIe-to-PCI bridge of a PCI device
> >   * if the device is PCIE, return NULL
> > diff --git a/include/linux/pci.h b/include/linux/pci.h index
> > bd8ec30..e87423a 100644
> > --- a/include/linux/pci.h
> > +++ b/include/linux/pci.h
> > @@ -1855,6 +1855,29 @@ static inline struct eeh_dev
> > *pci_dev_to_eeh_dev(struct pci_dev *pdev)  #endif
> > 
> >  /**
> > + * pci_walk_up_to_first_match - Generic upstream search function
> > + * @pdev: starting PCI device to search
> > + * @match: match function to call on each device (true = match)
> > + * @last: last device examined prior to returned device
> > + *
> > + * Walk upstream from the given device, calling match() at each device.
> > + * Returns the first device matching match().  If the root bus is
> > +reached
> > + * without finding a match, return NULL.  last returns the N-1 step in
> > + * the search path.
> > + */
> > +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> > +					   bool (*match)(struct pci_dev *),
> > +					   struct pci_dev **last);
> > +
> > +/**
> > + * pci_is_pcie_bridge - Match a PCIe bridge device
> > + * @pdev: device to test
> > + *
> > + * Return true if the given device is a PCIe bridge, false otherwise.
> > + */
> > +bool pci_is_pcie_bridge(struct pci_dev *pdev);
> > +
> > +/**
> >   * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge of a
> > device
> >   * @pdev: the PCI device
> >   *
> > 
> > _______________________________________________
> > iommu mailing list
> > iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
> > https://lists.linuxfoundation.org/mailman/listinfo/iommu
> 
> 

RE: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test

[Sethi Varun-B16395]

Any idea when would this patch set be integrated? Freescale PAMU driver depends on this patch. Currently I am using the pci_find_pcie_upstream_bridge routine in the device group creation routine.

-Varun

> -----Original Message-----
> From: Alex Williamson [mailto:alex.williamson@redhat.com]
> Sent: Monday, May 13, 2013 8:19 PM
> To: Sethi Varun-B16395
> Cc: bhelgaas@google.com; stephen@networkplumber.org; linux-
> pci@vger.kernel.org; iommu@lists.linux-foundation.org;
> dwmw2@infradead.org
> Subject: Re: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test
> 
> On Mon, 2013-05-13 at 13:51 +0000, Sethi Varun-B16395 wrote:
> > Would these functions be used outside drivers/iommu? We recently added
> pci.h under drivers/iommu, maybe we can add a new file for these
> functions as well.
> 
> The intention is to make them generic enough for pci-core, unlike
> pci_find_upstream_pcie_bridge, which was rather specific to iommu usage.
> Thanks,
> 
> Alex
> 
> > > -----Original Message-----
> > > From: iommu-bounces@lists.linux-foundation.org [mailto:iommu-
> > > bounces@lists.linux-foundation.org] On Behalf Of Alex Williamson
> > > Sent: Saturday, May 11, 2013 2:49 AM
> > > To: bhelgaas@google.com; stephen@networkplumber.org
> > > Cc: linux-pci@vger.kernel.org; iommu@lists.linux-foundation.org;
> > > dwmw2@infradead.org
> > > Subject: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test
> > >
> > > These will replace pci_find_upstream_pcie_bridge, which is difficult
> > > to use and rather specific to intel-iommu usage.  A quirked
> > > pci_is_pcie_bridge function is provided to work around non-compliant
> > > PCIe-to-PCI bridges such as those found in
> > > https://bugzilla.kernel.org/show_bug.cgi?id=44881
> > >
> > > Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> > > ---
> > >  drivers/pci/search.c |   57
> > > ++++++++++++++++++++++++++++++++++++++++++++++++++
> > >  include/linux/pci.h  |   23 ++++++++++++++++++++
> > >  2 files changed, 80 insertions(+)
> > >
> > > diff --git a/drivers/pci/search.c b/drivers/pci/search.c index
> > > d0627fa..0357f74 100644
> > > --- a/drivers/pci/search.c
> > > +++ b/drivers/pci/search.c
> > > @@ -17,6 +17,63 @@
> > >  DECLARE_RWSEM(pci_bus_sem);
> > >  EXPORT_SYMBOL_GPL(pci_bus_sem);
> > >
> > > +/* Test for PCIe bridges. */
> > > +bool pci_is_pcie_bridge(struct pci_dev *pdev) {
> > > +	if (!pci_is_bridge(pdev))
> > > +		return false;
> > > +
> > > +	if (pci_is_pcie(pdev))
> > > +		return true;
> > > +
> > > +#ifdef CONFIG_PCI_QUIRKS
> > > +	/*
> > > +	 * If we're not on the root bus, look one device upstream of the
> > > +	 * current device.  If that device is PCIe and is not a PCIe-to-PCI
> > > +	 * bridge, then the current device is effectively PCIe as it must
> > > +	 * be the PCIe-to-PCI bridge.  This handles several bridges that
> > > +	 * violate the PCIe spec by not exposing a PCIe capability:
> > > +	 * https://bugzilla.kernel.org/show_bug.cgi?id=44881
> > > +	 */
> > > +	if (!pci_is_root_bus(pdev->bus)) {
> > > +		struct pci_dev *parent = pdev->bus->self;
> > > +
> > > +		if (pci_is_pcie(parent) &&
> > > +		    pci_pcie_type(parent) != PCI_EXP_TYPE_PCI_BRIDGE)
> > > +
> > > +			return true;
> > > +	}
> > > +#endif
> > > +	return false;
> > > +}
> > > +
> > > +/*
> > > + * Walk upstream from the given pdev for the first device returning
> > > + * true for the provided match function.  If no match is found,
> > > +return
> > > + * NULL.  *last records the previous step in the walk.
> > > + */
> > > +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> > > +					   bool (*match)(struct pci_dev *),
> > > +					   struct pci_dev **last)
> > > +{
> > > +	*last = NULL;
> > > +
> > > +	if (match(pdev))
> > > +		return pdev;
> > > +
> > > +	*last = pdev;
> > > +
> > > +	while (!pci_is_root_bus(pdev->bus)) {
> > > +		*last = pdev;
> > > +		pdev = pdev->bus->self;
> > > +
> > > +		if (match(pdev))
> > > +			return pdev;
> > > +	}
> > > +
> > > +	return NULL;
> > > +}
> > > +
> > >  /*
> > >   * find the upstream PCIe-to-PCI bridge of a PCI device
> > >   * if the device is PCIE, return NULL diff --git
> > > a/include/linux/pci.h b/include/linux/pci.h index bd8ec30..e87423a
> > > 100644
> > > --- a/include/linux/pci.h
> > > +++ b/include/linux/pci.h
> > > @@ -1855,6 +1855,29 @@ static inline struct eeh_dev
> > > *pci_dev_to_eeh_dev(struct pci_dev *pdev)  #endif
> > >
> > >  /**
> > > + * pci_walk_up_to_first_match - Generic upstream search function
> > > + * @pdev: starting PCI device to search
> > > + * @match: match function to call on each device (true = match)
> > > + * @last: last device examined prior to returned device
> > > + *
> > > + * Walk upstream from the given device, calling match() at each
> device.
> > > + * Returns the first device matching match().  If the root bus is
> > > +reached
> > > + * without finding a match, return NULL.  last returns the N-1 step
> > > +in
> > > + * the search path.
> > > + */
> > > +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> > > +					   bool (*match)(struct pci_dev *),
> > > +					   struct pci_dev **last);
> > > +
> > > +/**
> > > + * pci_is_pcie_bridge - Match a PCIe bridge device
> > > + * @pdev: device to test
> > > + *
> > > + * Return true if the given device is a PCIe bridge, false
> otherwise.
> > > + */
> > > +bool pci_is_pcie_bridge(struct pci_dev *pdev);
> > > +
> > > +/**
> > >   * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge
> > > of a device
> > >   * @pdev: the PCI device
> > >   *
> > >
> > > _______________________________________________
> > > iommu mailing list
> > > iommu@lists.linux-foundation.org
> > > https://lists.linuxfoundation.org/mailman/listinfo/iommu
> >
> >
> 
> 
> 

Re: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test

[Alex Williamson]

On Wed, 2013-05-22 at 14:34 +0000, Sethi Varun-B16395 wrote:
> Any idea when would this patch set be integrated? Freescale PAMU
> driver depends on this patch. Currently I am using the
> pci_find_pcie_upstream_bridge routine in the device group creation
> routine.

I don't know, Bjorn?  The bugzilla referenced in patch 0/3 has several
reports from users that this series solves the problem for them.
Thanks,

Alex

> > -----Original Message-----
> > From: Alex Williamson [mailto:alex.williamson-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org]
> > Sent: Monday, May 13, 2013 8:19 PM
> > To: Sethi Varun-B16395
> > Cc: bhelgaas-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org; stephen-OTpzqLSitTUnbdJkjeBofR2eb7JE58TQ@public.gmane.org; linux-
> > pci-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org;
> > dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org
> > Subject: Re: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test
> > 
> > On Mon, 2013-05-13 at 13:51 +0000, Sethi Varun-B16395 wrote:
> > > Would these functions be used outside drivers/iommu? We recently added
> > pci.h under drivers/iommu, maybe we can add a new file for these
> > functions as well.
> > 
> > The intention is to make them generic enough for pci-core, unlike
> > pci_find_upstream_pcie_bridge, which was rather specific to iommu usage.
> > Thanks,
> > 
> > Alex
> > 
> > > > -----Original Message-----
> > > > From: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org [mailto:iommu-
> > > > bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org] On Behalf Of Alex Williamson
> > > > Sent: Saturday, May 11, 2013 2:49 AM
> > > > To: bhelgaas-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org; stephen-OTpzqLSitTUnbdJkjeBofR2eb7JE58TQ@public.gmane.org
> > > > Cc: linux-pci-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org;
> > > > dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org
> > > > Subject: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test
> > > >
> > > > These will replace pci_find_upstream_pcie_bridge, which is difficult
> > > > to use and rather specific to intel-iommu usage.  A quirked
> > > > pci_is_pcie_bridge function is provided to work around non-compliant
> > > > PCIe-to-PCI bridges such as those found in
> > > > https://bugzilla.kernel.org/show_bug.cgi?id=44881
> > > >
> > > > Signed-off-by: Alex Williamson <alex.williamson-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> > > > ---
> > > >  drivers/pci/search.c |   57
> > > > ++++++++++++++++++++++++++++++++++++++++++++++++++
> > > >  include/linux/pci.h  |   23 ++++++++++++++++++++
> > > >  2 files changed, 80 insertions(+)
> > > >
> > > > diff --git a/drivers/pci/search.c b/drivers/pci/search.c index
> > > > d0627fa..0357f74 100644
> > > > --- a/drivers/pci/search.c
> > > > +++ b/drivers/pci/search.c
> > > > @@ -17,6 +17,63 @@
> > > >  DECLARE_RWSEM(pci_bus_sem);
> > > >  EXPORT_SYMBOL_GPL(pci_bus_sem);
> > > >
> > > > +/* Test for PCIe bridges. */
> > > > +bool pci_is_pcie_bridge(struct pci_dev *pdev) {
> > > > +	if (!pci_is_bridge(pdev))
> > > > +		return false;
> > > > +
> > > > +	if (pci_is_pcie(pdev))
> > > > +		return true;
> > > > +
> > > > +#ifdef CONFIG_PCI_QUIRKS
> > > > +	/*
> > > > +	 * If we're not on the root bus, look one device upstream of the
> > > > +	 * current device.  If that device is PCIe and is not a PCIe-to-PCI
> > > > +	 * bridge, then the current device is effectively PCIe as it must
> > > > +	 * be the PCIe-to-PCI bridge.  This handles several bridges that
> > > > +	 * violate the PCIe spec by not exposing a PCIe capability:
> > > > +	 * https://bugzilla.kernel.org/show_bug.cgi?id=44881
> > > > +	 */
> > > > +	if (!pci_is_root_bus(pdev->bus)) {
> > > > +		struct pci_dev *parent = pdev->bus->self;
> > > > +
> > > > +		if (pci_is_pcie(parent) &&
> > > > +		    pci_pcie_type(parent) != PCI_EXP_TYPE_PCI_BRIDGE)
> > > > +
> > > > +			return true;
> > > > +	}
> > > > +#endif
> > > > +	return false;
> > > > +}
> > > > +
> > > > +/*
> > > > + * Walk upstream from the given pdev for the first device returning
> > > > + * true for the provided match function.  If no match is found,
> > > > +return
> > > > + * NULL.  *last records the previous step in the walk.
> > > > + */
> > > > +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> > > > +					   bool (*match)(struct pci_dev *),
> > > > +					   struct pci_dev **last)
> > > > +{
> > > > +	*last = NULL;
> > > > +
> > > > +	if (match(pdev))
> > > > +		return pdev;
> > > > +
> > > > +	*last = pdev;
> > > > +
> > > > +	while (!pci_is_root_bus(pdev->bus)) {
> > > > +		*last = pdev;
> > > > +		pdev = pdev->bus->self;
> > > > +
> > > > +		if (match(pdev))
> > > > +			return pdev;
> > > > +	}
> > > > +
> > > > +	return NULL;
> > > > +}
> > > > +
> > > >  /*
> > > >   * find the upstream PCIe-to-PCI bridge of a PCI device
> > > >   * if the device is PCIE, return NULL diff --git
> > > > a/include/linux/pci.h b/include/linux/pci.h index bd8ec30..e87423a
> > > > 100644
> > > > --- a/include/linux/pci.h
> > > > +++ b/include/linux/pci.h
> > > > @@ -1855,6 +1855,29 @@ static inline struct eeh_dev
> > > > *pci_dev_to_eeh_dev(struct pci_dev *pdev)  #endif
> > > >
> > > >  /**
> > > > + * pci_walk_up_to_first_match - Generic upstream search function
> > > > + * @pdev: starting PCI device to search
> > > > + * @match: match function to call on each device (true = match)
> > > > + * @last: last device examined prior to returned device
> > > > + *
> > > > + * Walk upstream from the given device, calling match() at each
> > device.
> > > > + * Returns the first device matching match().  If the root bus is
> > > > +reached
> > > > + * without finding a match, return NULL.  last returns the N-1 step
> > > > +in
> > > > + * the search path.
> > > > + */
> > > > +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> > > > +					   bool (*match)(struct pci_dev *),
> > > > +					   struct pci_dev **last);
> > > > +
> > > > +/**
> > > > + * pci_is_pcie_bridge - Match a PCIe bridge device
> > > > + * @pdev: device to test
> > > > + *
> > > > + * Return true if the given device is a PCIe bridge, false
> > otherwise.
> > > > + */
> > > > +bool pci_is_pcie_bridge(struct pci_dev *pdev);
> > > > +
> > > > +/**
> > > >   * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge
> > > > of a device
> > > >   * @pdev: the PCI device
> > > >   *
> > > >
> > > > _______________________________________________
> > > > iommu mailing list
> > > > iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
> > > > https://lists.linuxfoundation.org/mailman/listinfo/iommu
> > >
> > >
> > 
> > 
> > 
> 

Re: [PATCH 1/3] pci: Add PCI walk function and PCIe bridge test

[Bjorn Helgaas]

On Fri, May 10, 2013 at 3:18 PM, Alex Williamson
<alex.williamson@redhat.com> wrote:
> These will replace pci_find_upstream_pcie_bridge, which is difficult
> to use and rather specific to intel-iommu usage.  A quirked
> pci_is_pcie_bridge function is provided to work around non-compliant
> PCIe-to-PCI bridges such as those found in
> https://bugzilla.kernel.org/show_bug.cgi?id=44881
>
> Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> ---
>  drivers/pci/search.c |   57 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  include/linux/pci.h  |   23 ++++++++++++++++++++
>  2 files changed, 80 insertions(+)
>
> diff --git a/drivers/pci/search.c b/drivers/pci/search.c
> index d0627fa..0357f74 100644
> --- a/drivers/pci/search.c
> +++ b/drivers/pci/search.c
> @@ -17,6 +17,63 @@
>  DECLARE_RWSEM(pci_bus_sem);
>  EXPORT_SYMBOL_GPL(pci_bus_sem);
>
> +/* Test for PCIe bridges. */
> +bool pci_is_pcie_bridge(struct pci_dev *pdev)
> +{
> +       if (!pci_is_bridge(pdev))
> +               return false;
> +
> +       if (pci_is_pcie(pdev))
> +               return true;
> +
> +#ifdef CONFIG_PCI_QUIRKS
> +       /*
> +        * If we're not on the root bus, look one device upstream of the
> +        * current device.  If that device is PCIe and is not a PCIe-to-PCI
> +        * bridge, then the current device is effectively PCIe as it must
> +        * be the PCIe-to-PCI bridge.  This handles several bridges that
> +        * violate the PCIe spec by not exposing a PCIe capability:
> +        * https://bugzilla.kernel.org/show_bug.cgi?id=44881
> +        */
> +       if (!pci_is_root_bus(pdev->bus)) {
> +               struct pci_dev *parent = pdev->bus->self;
> +
> +               if (pci_is_pcie(parent) &&
> +                   pci_pcie_type(parent) != PCI_EXP_TYPE_PCI_BRIDGE)
> +
> +                       return true;
> +       }
> +#endif
> +       return false;
> +}

I like this strategy.  But I'd rather it not be a general-purpose PCI
interface, because if pci_is_pcie_bridge() is true, people will assume
they can perform PCIe operations on the device, and they can't.  The
only use for this is to figure out the source ID the IOMMU will see,
so I think this should just go in the IOMMU code.

> +/*
> + * Walk upstream from the given pdev for the first device returning
> + * true for the provided match function.  If no match is found, return
> + * NULL.  *last records the previous step in the walk.
> + */
> +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> +                                          bool (*match)(struct pci_dev *),
> +                                          struct pci_dev **last)
> +{
> +       *last = NULL;
> +
> +       if (match(pdev))
> +               return pdev;
> +
> +       *last = pdev;
> +
> +       while (!pci_is_root_bus(pdev->bus)) {
> +               *last = pdev;
> +               pdev = pdev->bus->self;
> +
> +               if (match(pdev))
> +                       return pdev;
> +       }
> +
> +       return NULL;
> +}

Same here.  I don't really see much potential for other uses of this,
so it seems like you might as well just put this in the IOMMU code and
make it call pci_is_pcie_bridge() directly.

The "source ID == upstream PCIe bridge" mapping is deeply ingrained in
your skull, but I think it would make the intent of the code clearer
if the function names mentioned the source ID somehow.  Otherwise new
readers like me have to come up with that association on our own.  But
since I'm proposing putting all this in the IOMMU code, it's totally
up to you :)

Bjorn

> +
>  /*
>   * find the upstream PCIe-to-PCI bridge of a PCI device
>   * if the device is PCIE, return NULL
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index bd8ec30..e87423a 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -1855,6 +1855,29 @@ static inline struct eeh_dev *pci_dev_to_eeh_dev(struct pci_dev *pdev)
>  #endif
>
>  /**
> + * pci_walk_up_to_first_match - Generic upstream search function
> + * @pdev: starting PCI device to search
> + * @match: match function to call on each device (true = match)
> + * @last: last device examined prior to returned device
> + *
> + * Walk upstream from the given device, calling match() at each device.
> + * Returns the first device matching match().  If the root bus is reached
> + * without finding a match, return NULL.  last returns the N-1 step in
> + * the search path.
> + */
> +struct pci_dev *pci_walk_up_to_first_match(struct pci_dev *pdev,
> +                                          bool (*match)(struct pci_dev *),
> +                                          struct pci_dev **last);
> +
> +/**
> + * pci_is_pcie_bridge - Match a PCIe bridge device
> + * @pdev: device to test
> + *
> + * Return true if the given device is a PCIe bridge, false otherwise.
> + */
> +bool pci_is_pcie_bridge(struct pci_dev *pdev);
> +
> +/**
>   * pci_find_upstream_pcie_bridge - find upstream PCIe-to-PCI bridge of a device
>   * @pdev: the PCI device
>   *
>