From a6ed1f4e94e1883742ac79f09f9754ffacfae456 Mon Sep 17 00:00:00 2001
From: Yijing Wang <wangyijing@huawei.com>
Date: Fri, 19 Jun 2015 15:57:44 +0800
Subject: [PATCH 01/20] PCI: Use "slot" and "pci_slot" for struct hotplug_slot
 and struct pci_slot

Now in pci_hotplug_core.c, we randomly name a struct hotplug_slot and a
struct pci_slot.  It's easy to confuse them, so let us use "slot" for a
struct hotplug_slot and "pci_slot" for a struct pci_slot.

No functional change.

Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/hotplug/pci_hotplug_core.c | 122 ++++++++++++-------------
 1 file changed, 61 insertions(+), 61 deletions(-)

diff --git a/drivers/pci/hotplug/pci_hotplug_core.c b/drivers/pci/hotplug/pci_hotplug_core.c
index 56d8486dc167..d1fab97d6b01 100644
--- a/drivers/pci/hotplug/pci_hotplug_core.c
+++ b/drivers/pci/hotplug/pci_hotplug_core.c
@@ -83,12 +83,12 @@ GET_STATUS(attention_status, u8)
 GET_STATUS(latch_status, u8)
 GET_STATUS(adapter_status, u8)
 
-static ssize_t power_read_file(struct pci_slot *slot, char *buf)
+static ssize_t power_read_file(struct pci_slot *pci_slot, char *buf)
 {
 	int retval;
 	u8 value;
 
-	retval = get_power_status(slot->hotplug, &value);
+	retval = get_power_status(pci_slot->hotplug, &value);
 	if (retval)
 		return retval;
 
@@ -140,22 +140,22 @@ static struct pci_slot_attribute hotplug_slot_attr_power = {
 	.store = power_write_file
 };
 
-static ssize_t attention_read_file(struct pci_slot *slot, char *buf)
+static ssize_t attention_read_file(struct pci_slot *pci_slot, char *buf)
 {
 	int retval;
 	u8 value;
 
-	retval = get_attention_status(slot->hotplug, &value);
+	retval = get_attention_status(pci_slot->hotplug, &value);
 	if (retval)
 		return retval;
 
 	return sprintf(buf, "%d\n", value);
 }
 
-static ssize_t attention_write_file(struct pci_slot *slot, const char *buf,
+static ssize_t attention_write_file(struct pci_slot *pci_slot, const char *buf,
 				    size_t count)
 {
-	struct hotplug_slot_ops *ops = slot->hotplug->ops;
+	struct hotplug_slot_ops *ops = pci_slot->hotplug->ops;
 	unsigned long lattention;
 	u8 attention;
 	int retval = 0;
@@ -169,7 +169,7 @@ static ssize_t attention_write_file(struct pci_slot *slot, const char *buf,
 		goto exit;
 	}
 	if (ops->set_attention_status)
-		retval = ops->set_attention_status(slot->hotplug, attention);
+		retval = ops->set_attention_status(pci_slot->hotplug, attention);
 	module_put(ops->owner);
 
 exit:
@@ -184,12 +184,12 @@ static struct pci_slot_attribute hotplug_slot_attr_attention = {
 	.store = attention_write_file
 };
 
-static ssize_t latch_read_file(struct pci_slot *slot, char *buf)
+static ssize_t latch_read_file(struct pci_slot *pci_slot, char *buf)
 {
 	int retval;
 	u8 value;
 
-	retval = get_latch_status(slot->hotplug, &value);
+	retval = get_latch_status(pci_slot->hotplug, &value);
 	if (retval)
 		return retval;
 
@@ -201,12 +201,12 @@ static struct pci_slot_attribute hotplug_slot_attr_latch = {
 	.show = latch_read_file,
 };
 
-static ssize_t presence_read_file(struct pci_slot *slot, char *buf)
+static ssize_t presence_read_file(struct pci_slot *pci_slot, char *buf)
 {
 	int retval;
 	u8 value;
 
-	retval = get_adapter_status(slot->hotplug, &value);
+	retval = get_adapter_status(pci_slot->hotplug, &value);
 	if (retval)
 		return retval;
 
@@ -307,43 +307,43 @@ static bool has_test_file(struct pci_slot *pci_slot)
 	return false;
 }
 
-static int fs_add_slot(struct pci_slot *slot)
+static int fs_add_slot(struct pci_slot *pci_slot)
 {
 	int retval = 0;
 
 	/* Create symbolic link to the hotplug driver module */
-	pci_hp_create_module_link(slot);
+	pci_hp_create_module_link(pci_slot);
 
-	if (has_power_file(slot)) {
-		retval = sysfs_create_file(&slot->kobj,
+	if (has_power_file(pci_slot)) {
+		retval = sysfs_create_file(&pci_slot->kobj,
 					   &hotplug_slot_attr_power.attr);
 		if (retval)
 			goto exit_power;
 	}
 
-	if (has_attention_file(slot)) {
-		retval = sysfs_create_file(&slot->kobj,
+	if (has_attention_file(pci_slot)) {
+		retval = sysfs_create_file(&pci_slot->kobj,
 					   &hotplug_slot_attr_attention.attr);
 		if (retval)
 			goto exit_attention;
 	}
 
-	if (has_latch_file(slot)) {
-		retval = sysfs_create_file(&slot->kobj,
+	if (has_latch_file(pci_slot)) {
+		retval = sysfs_create_file(&pci_slot->kobj,
 					   &hotplug_slot_attr_latch.attr);
 		if (retval)
 			goto exit_latch;
 	}
 
-	if (has_adapter_file(slot)) {
-		retval = sysfs_create_file(&slot->kobj,
+	if (has_adapter_file(pci_slot)) {
+		retval = sysfs_create_file(&pci_slot->kobj,
 					   &hotplug_slot_attr_presence.attr);
 		if (retval)
 			goto exit_adapter;
 	}
 
-	if (has_test_file(slot)) {
-		retval = sysfs_create_file(&slot->kobj,
+	if (has_test_file(pci_slot)) {
+		retval = sysfs_create_file(&pci_slot->kobj,
 					   &hotplug_slot_attr_test.attr);
 		if (retval)
 			goto exit_test;
@@ -352,45 +352,45 @@ static int fs_add_slot(struct pci_slot *slot)
 	goto exit;
 
 exit_test:
-	if (has_adapter_file(slot))
-		sysfs_remove_file(&slot->kobj,
+	if (has_adapter_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj,
 				  &hotplug_slot_attr_presence.attr);
 exit_adapter:
-	if (has_latch_file(slot))
-		sysfs_remove_file(&slot->kobj, &hotplug_slot_attr_latch.attr);
+	if (has_latch_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj, &hotplug_slot_attr_latch.attr);
 exit_latch:
-	if (has_attention_file(slot))
-		sysfs_remove_file(&slot->kobj,
+	if (has_attention_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj,
 				  &hotplug_slot_attr_attention.attr);
 exit_attention:
-	if (has_power_file(slot))
-		sysfs_remove_file(&slot->kobj, &hotplug_slot_attr_power.attr);
+	if (has_power_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj, &hotplug_slot_attr_power.attr);
 exit_power:
-	pci_hp_remove_module_link(slot);
+	pci_hp_remove_module_link(pci_slot);
 exit:
 	return retval;
 }
 
-static void fs_remove_slot(struct pci_slot *slot)
+static void fs_remove_slot(struct pci_slot *pci_slot)
 {
-	if (has_power_file(slot))
-		sysfs_remove_file(&slot->kobj, &hotplug_slot_attr_power.attr);
+	if (has_power_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj, &hotplug_slot_attr_power.attr);
 
-	if (has_attention_file(slot))
-		sysfs_remove_file(&slot->kobj,
+	if (has_attention_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj,
 				  &hotplug_slot_attr_attention.attr);
 
-	if (has_latch_file(slot))
-		sysfs_remove_file(&slot->kobj, &hotplug_slot_attr_latch.attr);
+	if (has_latch_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj, &hotplug_slot_attr_latch.attr);
 
-	if (has_adapter_file(slot))
-		sysfs_remove_file(&slot->kobj,
+	if (has_adapter_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj,
 				  &hotplug_slot_attr_presence.attr);
 
-	if (has_test_file(slot))
-		sysfs_remove_file(&slot->kobj, &hotplug_slot_attr_test.attr);
+	if (has_test_file(pci_slot))
+		sysfs_remove_file(&pci_slot->kobj, &hotplug_slot_attr_test.attr);
 
-	pci_hp_remove_module_link(slot);
+	pci_hp_remove_module_link(pci_slot);
 }
 
 static struct hotplug_slot *get_slot_from_name(const char *name)
@@ -467,37 +467,37 @@ EXPORT_SYMBOL_GPL(__pci_hp_register);
 
 /**
  * pci_hp_deregister - deregister a hotplug_slot with the PCI hotplug subsystem
- * @hotplug: pointer to the &struct hotplug_slot to deregister
+ * @slot: pointer to the &struct hotplug_slot to deregister
  *
  * The @slot must have been registered with the pci hotplug subsystem
  * previously with a call to pci_hp_register().
  *
  * Returns 0 if successful, anything else for an error.
  */
-int pci_hp_deregister(struct hotplug_slot *hotplug)
+int pci_hp_deregister(struct hotplug_slot *slot)
 {
 	struct hotplug_slot *temp;
-	struct pci_slot *slot;
+	struct pci_slot *pci_slot;
 
-	if (!hotplug)
+	if (!slot)
 		return -ENODEV;
 
 	mutex_lock(&pci_hp_mutex);
-	temp = get_slot_from_name(hotplug_slot_name(hotplug));
-	if (temp != hotplug) {
+	temp = get_slot_from_name(hotplug_slot_name(slot));
+	if (temp != slot) {
 		mutex_unlock(&pci_hp_mutex);
 		return -ENODEV;
 	}
 
-	list_del(&hotplug->slot_list);
+	list_del(&slot->slot_list);
 
-	slot = hotplug->pci_slot;
-	fs_remove_slot(slot);
-	dbg("Removed slot %s from the list\n", hotplug_slot_name(hotplug));
+	pci_slot = slot->pci_slot;
+	fs_remove_slot(pci_slot);
+	dbg("Removed slot %s from the list\n", hotplug_slot_name(slot));
 
-	hotplug->release(hotplug);
-	slot->hotplug = NULL;
-	pci_destroy_slot(slot);
+	slot->release(slot);
+	pci_slot->hotplug = NULL;
+	pci_destroy_slot(pci_slot);
 	mutex_unlock(&pci_hp_mutex);
 
 	return 0;
@@ -506,7 +506,7 @@ EXPORT_SYMBOL_GPL(pci_hp_deregister);
 
 /**
  * pci_hp_change_slot_info - changes the slot's information structure in the core
- * @hotplug: pointer to the slot whose info has changed
+ * @slot: pointer to the slot whose info has changed
  * @info: pointer to the info copy into the slot's info structure
  *
  * @slot must have been registered with the pci
@@ -514,13 +514,13 @@ EXPORT_SYMBOL_GPL(pci_hp_deregister);
  *
  * Returns 0 if successful, anything else for an error.
  */
-int pci_hp_change_slot_info(struct hotplug_slot *hotplug,
+int pci_hp_change_slot_info(struct hotplug_slot *slot,
 			    struct hotplug_slot_info *info)
 {
-	if (!hotplug || !info)
+	if (!slot || !info)
 		return -ENODEV;
 
-	memcpy(hotplug->info, info, sizeof(struct hotplug_slot_info));
+	memcpy(slot->info, info, sizeof(struct hotplug_slot_info));
 
 	return 0;
 }

From ac10836b681289f7e430e52b106a209bbdcaa75e Mon Sep 17 00:00:00 2001
From: Yijing Wang <wangyijing@huawei.com>
Date: Fri, 19 Jun 2015 15:57:45 +0800
Subject: [PATCH 02/20] PCI: pciehp: Simplify pcie_poll_cmd()

Move first slot status read into while to simplify code.

Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/hotplug/pciehp_hpc.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c
index 2913f7e68a10..daf54bee720d 100644
--- a/drivers/pci/hotplug/pciehp_hpc.c
+++ b/drivers/pci/hotplug/pciehp_hpc.c
@@ -109,21 +109,17 @@ static int pcie_poll_cmd(struct controller *ctrl, int timeout)
 	struct pci_dev *pdev = ctrl_dev(ctrl);
 	u16 slot_status;
 
-	pcie_capability_read_word(pdev, PCI_EXP_SLTSTA, &slot_status);
-	if (slot_status & PCI_EXP_SLTSTA_CC) {
-		pcie_capability_write_word(pdev, PCI_EXP_SLTSTA,
-					   PCI_EXP_SLTSTA_CC);
-		return 1;
-	}
-	while (timeout > 0) {
-		msleep(10);
-		timeout -= 10;
+	while (true) {
 		pcie_capability_read_word(pdev, PCI_EXP_SLTSTA, &slot_status);
 		if (slot_status & PCI_EXP_SLTSTA_CC) {
 			pcie_capability_write_word(pdev, PCI_EXP_SLTSTA,
 						   PCI_EXP_SLTSTA_CC);
 			return 1;
 		}
+		if (timeout < 0)
+			break;
+		msleep(10);
+		timeout -= 10;
 	}
 	return 0;	/* timeout */
 }

From fb0cc3aa552642631c3a4e83deae2b5c1a1ef4fa Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Mon, 20 Jul 2015 09:10:36 -0500
Subject: [PATCH 03/20] iommu/vt-d: Cache PCI ATS state and Invalidate Queue
 Depth

We check the ATS state (enabled/disabled) and fetch the PCI ATS Invalidate
Queue Depth in performance-sensitive paths.  It's easy to cache these,
which removes dependencies on PCI.

Remember the ATS enabled state.  When enabling, read the queue depth once
and cache it in the device_domain_info struct.  This is similar to what
amd_iommu.c does.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Acked-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/iommu/intel-iommu.c | 28 ++++++++++++++++++----------
 1 file changed, 18 insertions(+), 10 deletions(-)

diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index a98a7b27aca1..c22a5490eea7 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -408,6 +408,10 @@ struct device_domain_info {
 	struct list_head global; /* link to global list */
 	u8 bus;			/* PCI bus number */
 	u8 devfn;		/* PCI devfn number */
+	struct {
+		u8 enabled:1;
+		u8 qdep;
+	} ats;			/* ATS state */
 	struct device *dev; /* it's NULL for PCIe-to-PCI bridge */
 	struct intel_iommu *iommu; /* IOMMU used by this device */
 	struct dmar_domain *domain; /* pointer to domain */
@@ -1391,19 +1395,26 @@ iommu_support_dev_iotlb (struct dmar_domain *domain, struct intel_iommu *iommu,
 
 static void iommu_enable_dev_iotlb(struct device_domain_info *info)
 {
+	struct pci_dev *pdev;
+
 	if (!info || !dev_is_pci(info->dev))
 		return;
 
-	pci_enable_ats(to_pci_dev(info->dev), VTD_PAGE_SHIFT);
+	pdev = to_pci_dev(info->dev);
+	if (pci_enable_ats(pdev, VTD_PAGE_SHIFT))
+		return;
+
+	info->ats.enabled = 1;
+	info->ats.qdep = pci_ats_queue_depth(pdev);
 }
 
 static void iommu_disable_dev_iotlb(struct device_domain_info *info)
 {
-	if (!info->dev || !dev_is_pci(info->dev) ||
-	    !pci_ats_enabled(to_pci_dev(info->dev)))
+	if (!info->ats.enabled)
 		return;
 
 	pci_disable_ats(to_pci_dev(info->dev));
+	info->ats.enabled = 0;
 }
 
 static void iommu_flush_dev_iotlb(struct dmar_domain *domain,
@@ -1415,16 +1426,11 @@ static void iommu_flush_dev_iotlb(struct dmar_domain *domain,
 
 	spin_lock_irqsave(&device_domain_lock, flags);
 	list_for_each_entry(info, &domain->devices, link) {
-		struct pci_dev *pdev;
-		if (!info->dev || !dev_is_pci(info->dev))
-			continue;
-
-		pdev = to_pci_dev(info->dev);
-		if (!pci_ats_enabled(pdev))
+		if (!info->ats.enabled)
 			continue;
 
 		sid = info->bus << 8 | info->devfn;
-		qdep = pci_ats_queue_depth(pdev);
+		qdep = info->ats.qdep;
 		qi_flush_dev_iotlb(info->iommu, sid, qdep, addr, mask);
 	}
 	spin_unlock_irqrestore(&device_domain_lock, flags);
@@ -2272,6 +2278,8 @@ static struct dmar_domain *dmar_insert_dev_info(struct intel_iommu *iommu,
 
 	info->bus = bus;
 	info->devfn = devfn;
+	info->ats.enabled = 0;
+	info->ats.qdep = 0;
 	info->dev = dev;
 	info->domain = domain;
 	info->iommu = iommu;

From 67546762978f523749eac157903e0b01c18e083a Mon Sep 17 00:00:00 2001
From: Yijing Wang <wangyijing@huawei.com>
Date: Fri, 17 Jul 2015 17:16:31 +0800
Subject: [PATCH 04/20] PCI: Protect pci_bus->slots with pci_slot_mutex, not
 pci_bus_sem

Rajat Jain reported a deadlock when PCIe hot-add and AER recovery happen at
the same time:

thread 1:

  pciehp_enable_slot
    pciehp_configure_device
      pci_bus_add_devices
        pci_bus_add_device
          device_attach
            device_lock(dev)                       # acquire device lock
            ...
            pciehp_probe
              init_slot
                pci_hp_register
                  pci_create_slot
                    down_write(pci_bus_sem)        # deadlock here

thread 2:

  aer_isr_one_error
    aer_process_err_device
      do_recovery
        broadcast_error_message(..., report_error_detected)
          pci_walk_bus(..., cb=report_error_detected, ...)
            down_read(&pci_bus_sem)                # acquire pci_bus_sem
            report_error_detected(dev)             # cb()
              device_lock(dev)                     # deadlock here

Previously, the bus->devices and bus->slots list were protected by
pci_bus_sem.  In pci_create_slot(), we held it for writing so we could
add to the bus->slots list.

Add a new local pci_slot_mutex to protect bus->slots.  Hold pci_bus_sem for
reading while searching the bus->devices list.

[bhelgaas: changelog]
Link: http://lkml.kernel.org/r/CAA93t1qpPqbih+UB0McA_d_+2rVaNkXsinAUxYzK9+JXSS+L-g@mail.gmail.com
Reported-by: Rajat Jain <rajatja@google.com>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/slot.c  | 18 +++++++++++-------
 include/linux/pci.h |  3 ++-
 2 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/drivers/pci/slot.c b/drivers/pci/slot.c
index 396c200b9ddb..4bd3fce93fa4 100644
--- a/drivers/pci/slot.c
+++ b/drivers/pci/slot.c
@@ -14,6 +14,7 @@
 
 struct kset *pci_slots_kset;
 EXPORT_SYMBOL_GPL(pci_slots_kset);
+static DEFINE_MUTEX(pci_slot_mutex);
 
 static ssize_t pci_slot_attr_show(struct kobject *kobj,
 					struct attribute *attr, char *buf)
@@ -106,9 +107,11 @@ static void pci_slot_release(struct kobject *kobj)
 	dev_dbg(&slot->bus->dev, "dev %02x, released physical slot %s\n",
 		slot->number, pci_slot_name(slot));
 
+	down_read(&pci_bus_sem);
 	list_for_each_entry(dev, &slot->bus->devices, bus_list)
 		if (PCI_SLOT(dev->devfn) == slot->number)
 			dev->slot = NULL;
+	up_read(&pci_bus_sem);
 
 	list_del(&slot->list);
 
@@ -194,9 +197,8 @@ static int rename_slot(struct pci_slot *slot, const char *name)
 static struct pci_slot *get_slot(struct pci_bus *parent, int slot_nr)
 {
 	struct pci_slot *slot;
-	/*
-	 * We already hold pci_bus_sem so don't worry
-	 */
+
+	/* We already hold pci_slot_mutex */
 	list_for_each_entry(slot, &parent->slots, list)
 		if (slot->number == slot_nr) {
 			kobject_get(&slot->kobj);
@@ -253,7 +255,7 @@ struct pci_slot *pci_create_slot(struct pci_bus *parent, int slot_nr,
 	int err = 0;
 	char *slot_name = NULL;
 
-	down_write(&pci_bus_sem);
+	mutex_lock(&pci_slot_mutex);
 
 	if (slot_nr == -1)
 		goto placeholder;
@@ -301,16 +303,18 @@ placeholder:
 	INIT_LIST_HEAD(&slot->list);
 	list_add(&slot->list, &parent->slots);
 
+	down_read(&pci_bus_sem);
 	list_for_each_entry(dev, &parent->devices, bus_list)
 		if (PCI_SLOT(dev->devfn) == slot_nr)
 			dev->slot = slot;
+	up_read(&pci_bus_sem);
 
 	dev_dbg(&parent->dev, "dev %02x, created physical slot %s\n",
 		slot_nr, pci_slot_name(slot));
 
 out:
 	kfree(slot_name);
-	up_write(&pci_bus_sem);
+	mutex_unlock(&pci_slot_mutex);
 	return slot;
 err:
 	kfree(slot);
@@ -332,9 +336,9 @@ void pci_destroy_slot(struct pci_slot *slot)
 	dev_dbg(&slot->bus->dev, "dev %02x, dec refcount to %d\n",
 		slot->number, atomic_read(&slot->kobj.kref.refcount) - 1);
 
-	down_write(&pci_bus_sem);
+	mutex_lock(&pci_slot_mutex);
 	kobject_put(&slot->kobj);
-	up_write(&pci_bus_sem);
+	mutex_unlock(&pci_slot_mutex);
 }
 EXPORT_SYMBOL_GPL(pci_destroy_slot);
 
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 8a0321a8fb59..aaee493174e2 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -446,7 +446,8 @@ struct pci_bus {
 	struct list_head children;	/* list of child buses */
 	struct list_head devices;	/* list of devices on this bus */
 	struct pci_dev	*self;		/* bridge device as seen by parent */
-	struct list_head slots;		/* list of slots on this bus */
+	struct list_head slots;		/* list of slots on this bus;
+					   protected by pci_slot_mutex */
 	struct resource *resource[PCI_BRIDGE_RESOURCE_NUM];
 	struct list_head resources;	/* address space routed to this bus */
 	struct resource busn_res;	/* bus numbers routed to this bus */

From 017ffe64e8b8c8db0f50433a71da41c6a4e12710 Mon Sep 17 00:00:00 2001
From: Yijing Wang <wangyijing@huawei.com>
Date: Fri, 17 Jul 2015 17:16:32 +0800
Subject: [PATCH 05/20] PCI: Hold pci_slot_mutex while searching bus->slots
 list

Previously, pci_setup_device() and similar functions searched the
pci_bus->slots list without any locking.  It was possible for another
thread to update the list while we searched it.

Add pci_dev_assign_slot() to search the list while holding pci_slot_mutex.

[bhelgaas: changelog, fold in CONFIG_SYSFS fix]
Tested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 arch/powerpc/kernel/pci_of_scan.c |  6 +-----
 arch/sparc/kernel/pci.c           |  6 +-----
 drivers/pci/probe.c               |  6 +-----
 drivers/pci/slot.c                | 11 +++++++++++
 include/linux/pci.h               |  5 +++++
 5 files changed, 19 insertions(+), 15 deletions(-)

diff --git a/arch/powerpc/kernel/pci_of_scan.c b/arch/powerpc/kernel/pci_of_scan.c
index 42e02a2d570b..5e2debfc6ce5 100644
--- a/arch/powerpc/kernel/pci_of_scan.c
+++ b/arch/powerpc/kernel/pci_of_scan.c
@@ -126,7 +126,6 @@ struct pci_dev *of_create_pci_dev(struct device_node *node,
 {
 	struct pci_dev *dev;
 	const char *type;
-	struct pci_slot *slot;
 
 	dev = pci_alloc_dev(bus);
 	if (!dev)
@@ -145,10 +144,7 @@ struct pci_dev *of_create_pci_dev(struct device_node *node,
 	dev->needs_freset = 0;		/* pcie fundamental reset required */
 	set_pcie_port_type(dev);
 
-	list_for_each_entry(slot, &dev->bus->slots, list)
-		if (PCI_SLOT(dev->devfn) == slot->number)
-			dev->slot = slot;
-
+	pci_dev_assign_slot(dev);
 	dev->vendor = get_int_prop(node, "vendor-id", 0xffff);
 	dev->device = get_int_prop(node, "device-id", 0xffff);
 	dev->subsystem_vendor = get_int_prop(node, "subsystem-vendor-id", 0);
diff --git a/arch/sparc/kernel/pci.c b/arch/sparc/kernel/pci.c
index c928bc64b4ba..3a0e1a986bfe 100644
--- a/arch/sparc/kernel/pci.c
+++ b/arch/sparc/kernel/pci.c
@@ -249,7 +249,6 @@ static struct pci_dev *of_create_pci_dev(struct pci_pbm_info *pbm,
 					 struct pci_bus *bus, int devfn)
 {
 	struct dev_archdata *sd;
-	struct pci_slot *slot;
 	struct platform_device *op;
 	struct pci_dev *dev;
 	const char *type;
@@ -290,10 +289,7 @@ static struct pci_dev *of_create_pci_dev(struct pci_pbm_info *pbm,
 	dev->multifunction = 0;		/* maybe a lie? */
 	set_pcie_port_type(dev);
 
-	list_for_each_entry(slot, &dev->bus->slots, list)
-		if (PCI_SLOT(dev->devfn) == slot->number)
-			dev->slot = slot;
-
+	pci_dev_assign_slot(dev);
 	dev->vendor = of_getintprop_default(node, "vendor-id", 0xffff);
 	dev->device = of_getintprop_default(node, "device-id", 0xffff);
 	dev->subsystem_vendor =
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c
index cefd636681b6..2a9ce16cb374 100644
--- a/drivers/pci/probe.c
+++ b/drivers/pci/probe.c
@@ -1133,7 +1133,6 @@ int pci_setup_device(struct pci_dev *dev)
 {
 	u32 class;
 	u8 hdr_type;
-	struct pci_slot *slot;
 	int pos = 0;
 	struct pci_bus_region region;
 	struct resource *res;
@@ -1149,10 +1148,7 @@ int pci_setup_device(struct pci_dev *dev)
 	dev->error_state = pci_channel_io_normal;
 	set_pcie_port_type(dev);
 
-	list_for_each_entry(slot, &dev->bus->slots, list)
-		if (PCI_SLOT(dev->devfn) == slot->number)
-			dev->slot = slot;
-
+	pci_dev_assign_slot(dev);
 	/* Assume 32-bit PCI; let 64-bit PCI cards (which are far rarer)
 	   set this higher, assuming the system even supports it.  */
 	dev->dma_mask = 0xffffffff;
diff --git a/drivers/pci/slot.c b/drivers/pci/slot.c
index 4bd3fce93fa4..429d34c348b9 100644
--- a/drivers/pci/slot.c
+++ b/drivers/pci/slot.c
@@ -194,6 +194,17 @@ static int rename_slot(struct pci_slot *slot, const char *name)
 	return result;
 }
 
+void pci_dev_assign_slot(struct pci_dev *dev)
+{
+	struct pci_slot *slot;
+
+	mutex_lock(&pci_slot_mutex);
+	list_for_each_entry(slot, &dev->bus->slots, list)
+		if (PCI_SLOT(dev->devfn) == slot->number)
+			dev->slot = slot;
+	mutex_unlock(&pci_slot_mutex);
+}
+
 static struct pci_slot *get_slot(struct pci_bus *parent, int slot_nr)
 {
 	struct pci_slot *slot;
diff --git a/include/linux/pci.h b/include/linux/pci.h
index aaee493174e2..b3ba7fef2916 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -798,6 +798,11 @@ struct pci_slot *pci_create_slot(struct pci_bus *parent, int slot_nr,
 				 const char *name,
 				 struct hotplug_slot *hotplug);
 void pci_destroy_slot(struct pci_slot *slot);
+#ifdef CONFIG_SYSFS
+void pci_dev_assign_slot(struct pci_dev *dev);
+#else
+static inline void pci_dev_assign_slot(struct pci_dev *dev) { }
+#endif
 int pci_scan_slot(struct pci_bus *bus, int devfn);
 struct pci_dev *pci_scan_single_device(struct pci_bus *bus, int devfn);
 void pci_device_add(struct pci_dev *dev, struct pci_bus *bus);

From 1469d17dd341458267a08c8d8bb517cb986ec56d Mon Sep 17 00:00:00 2001
From: Jarod Wilson <jarod@redhat.com>
Date: Tue, 21 Jul 2015 12:25:30 -0400
Subject: [PATCH 06/20] PCI: pciehp: Handle invalid data when reading from
 non-existent devices

It's platform-dependent, but an MMIO read to a non-existent PCI device
generally returns data with all bits set.  This happens when the host
bridge or Root Complex times out waiting for a response from the device and
fabricates return data to complete the CPU's read.

One example, reported in the bugzilla below, involved this hierarchy:

  pci 0000:00:1c.0: PCI bridge to [bus 02-3a] Root Port
  pci 0000:02:00.0: PCI bridge to [bus 03-0a] Upstream Port
  pci 0000:03:03.0: PCI bridge to [bus 05-07] Downstream Port
  pci 0000:05:00.0: PCI bridge to [bus 06-07] Thunderbolt Upstream Port
  pci 0000:06:00.0: PCI bridge to [bus 07]    Thunderbolt Downstream Port
  pci 0000:07:00.0: BCM57762 NIC

Unplugging the Thunderbolt switch and the NIC below it resulted in this:

  pciehp 0000:03:03.0: Surprise Removal
  tg3 0000:07:00.0: tg3_abort_hw timed out, TX_MODE_ENABLE will not clear MAC_TX_MODE=ffffffff
  pciehp 0000:06:00.0: unloading service driver pciehp
  pciehp 0000:06:00.0: pcie_isr: intr_loc 11f
  pciehp 0000:06:00.0: Switch interrupt received
  pciehp 0000:06:00.0: Latch open on Slot
  pciehp 0000:06:00.0: Attention button interrupt received
  pciehp 0000:06:00.0: Button pressed on Slot
  pciehp 0000:06:00.0: Presence/Notify input change
  pciehp 0000:06:00.0: Card present on Slot
  pciehp 0000:06:00.0: Power fault interrupt received
  pciehp 0000:06:00.0: Data Link Layer State change
  pciehp 0000:06:00.0: Link Up event

The pciehp driver correctly noticed that the Thunderbolt switch (05:00.0
and 06:00.0) and NIC (07:00.0) had been removed, and it called their driver
remove methods.

Since the NIC was already gone, tg3 received 0xffffffff when it tried to
read from the device.  The resulting timeout is a tg3 issue and not of
interest here.

Similarly, since the 06:00.0 Thunderbolt switch was already gone,
pcie_isr() received 0xffff when it tried to read PCI_EXP_SLTSTA, and pciehp
thought that was valid status showing that many events had happened: the
latch had been opened, the attention button had been pressed, a card was
now present, and the link was now up.  These are all wrong, of course, but
pciehp went on to try to power up and enumerate devices below the
non-existent bridge:

  pciehp 0000:06:00.0: PCI slot - powering on due to button press
  pciehp 0000:06:00.0: Surprise Insertion
  pci 0000:07:00.0 id reading try 50 times with interval 20 ms to get ffffffff

[bhelgaas: changelog, also check in pcie_poll_cmd() & pcie_do_write_cmd()]
Link: https://bugzilla.kernel.org/show_bug.cgi?id=99841
Suggested-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/hotplug/pciehp_hpc.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c
index daf54bee720d..8f3d3cff3581 100644
--- a/drivers/pci/hotplug/pciehp_hpc.c
+++ b/drivers/pci/hotplug/pciehp_hpc.c
@@ -111,6 +111,12 @@ static int pcie_poll_cmd(struct controller *ctrl, int timeout)
 
 	while (true) {
 		pcie_capability_read_word(pdev, PCI_EXP_SLTSTA, &slot_status);
+		if (slot_status == (u16) ~0) {
+			ctrl_info(ctrl, "%s: no response from device\n",
+				  __func__);
+			return 0;
+		}
+
 		if (slot_status & PCI_EXP_SLTSTA_CC) {
 			pcie_capability_write_word(pdev, PCI_EXP_SLTSTA,
 						   PCI_EXP_SLTSTA_CC);
@@ -186,6 +192,11 @@ static void pcie_do_write_cmd(struct controller *ctrl, u16 cmd,
 	pcie_wait_cmd(ctrl);
 
 	pcie_capability_read_word(pdev, PCI_EXP_SLTCTL, &slot_ctrl);
+	if (slot_ctrl == (u16) ~0) {
+		ctrl_info(ctrl, "%s: no response from device\n", __func__);
+		goto out;
+	}
+
 	slot_ctrl &= ~mask;
 	slot_ctrl |= (cmd & mask);
 	ctrl->cmd_busy = 1;
@@ -201,6 +212,7 @@ static void pcie_do_write_cmd(struct controller *ctrl, u16 cmd,
 	if (wait)
 		pcie_wait_cmd(ctrl);
 
+out:
 	mutex_unlock(&ctrl->ctrl_lock);
 }
 
@@ -542,6 +554,11 @@ static irqreturn_t pcie_isr(int irq, void *dev_id)
 	intr_loc = 0;
 	do {
 		pcie_capability_read_word(pdev, PCI_EXP_SLTSTA, &detected);
+		if (detected == (u16) ~0) {
+			ctrl_info(ctrl, "%s: no response from device\n",
+				  __func__);
+			return IRQ_HANDLED;
+		}
 
 		detected &= (PCI_EXP_SLTSTA_ABP | PCI_EXP_SLTSTA_PFD |
 			     PCI_EXP_SLTSTA_MRLSC | PCI_EXP_SLTSTA_PDC |

From 58fa2405bd44805cb6166603100b0183ce26a0c8 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Wed, 1 Jul 2015 17:01:03 -0500
Subject: [PATCH 07/20] PCI: pciehp: Remove unused interrupt events

The list of interrupt events (INT_BUTTON_IGNORE, INT_PRESENCE_ON, etc.) was
copied from other hotplug drivers, but pciehp doesn't use them all.

Remove the interrupt events that aren't used by pciehp.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/hotplug/pciehp.h | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/drivers/pci/hotplug/pciehp.h b/drivers/pci/hotplug/pciehp.h
index 57cd1327346f..aca84e1b81be 100644
--- a/drivers/pci/hotplug/pciehp.h
+++ b/drivers/pci/hotplug/pciehp.h
@@ -101,18 +101,14 @@ struct controller {
 	unsigned int power_fault_detected;
 };
 
-#define INT_BUTTON_IGNORE		0
 #define INT_PRESENCE_ON			1
 #define INT_PRESENCE_OFF		2
 #define INT_SWITCH_CLOSE		3
 #define INT_SWITCH_OPEN			4
 #define INT_POWER_FAULT			5
-#define INT_POWER_FAULT_CLEAR		6
-#define INT_BUTTON_PRESS		7
-#define INT_BUTTON_RELEASE		8
-#define INT_BUTTON_CANCEL		9
-#define INT_LINK_UP			10
-#define INT_LINK_DOWN			11
+#define INT_BUTTON_PRESS		6
+#define INT_LINK_UP			7
+#define INT_LINK_DOWN			8
 
 #define STATIC_STATE			0
 #define BLINKINGON_STATE		1

From 2db0f71f56795f6b04d386b15c3b32b570558ebd Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Wed, 1 Jul 2015 17:17:49 -0500
Subject: [PATCH 08/20] PCI: pciehp: Remove ignored MRL sensor interrupt events

We queued interrupt events for the MRL being opened or closed, but the code
in interrupt_event_handler() that handles these events ignored them.

Stop enabling MRL interrupts and remove the ignored events.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 drivers/pci/hotplug/pciehp.h     | 10 ++++------
 drivers/pci/hotplug/pciehp_hpc.c | 17 +++--------------
 2 files changed, 7 insertions(+), 20 deletions(-)

diff --git a/drivers/pci/hotplug/pciehp.h b/drivers/pci/hotplug/pciehp.h
index aca84e1b81be..62d6fe6c3714 100644
--- a/drivers/pci/hotplug/pciehp.h
+++ b/drivers/pci/hotplug/pciehp.h
@@ -103,12 +103,10 @@ struct controller {
 
 #define INT_PRESENCE_ON			1
 #define INT_PRESENCE_OFF		2
-#define INT_SWITCH_CLOSE		3
-#define INT_SWITCH_OPEN			4
-#define INT_POWER_FAULT			5
-#define INT_BUTTON_PRESS		6
-#define INT_LINK_UP			7
-#define INT_LINK_DOWN			8
+#define INT_POWER_FAULT			3
+#define INT_BUTTON_PRESS		4
+#define INT_LINK_UP			5
+#define INT_LINK_DOWN			6
 
 #define STATIC_STATE			0
 #define BLINKINGON_STATE		1
diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c
index 8f3d3cff3581..5c24e938042f 100644
--- a/drivers/pci/hotplug/pciehp_hpc.c
+++ b/drivers/pci/hotplug/pciehp_hpc.c
@@ -543,7 +543,7 @@ static irqreturn_t pcie_isr(int irq, void *dev_id)
 	struct pci_dev *dev;
 	struct slot *slot = ctrl->slot;
 	u16 detected, intr_loc;
-	u8 open, present;
+	u8 present;
 	bool link;
 
 	/*
@@ -561,7 +561,7 @@ static irqreturn_t pcie_isr(int irq, void *dev_id)
 		}
 
 		detected &= (PCI_EXP_SLTSTA_ABP | PCI_EXP_SLTSTA_PFD |
-			     PCI_EXP_SLTSTA_MRLSC | PCI_EXP_SLTSTA_PDC |
+			     PCI_EXP_SLTSTA_PDC |
 			     PCI_EXP_SLTSTA_CC | PCI_EXP_SLTSTA_DLLSC);
 		detected &= ~intr_loc;
 		intr_loc |= detected;
@@ -594,15 +594,6 @@ static irqreturn_t pcie_isr(int irq, void *dev_id)
 	if (!(intr_loc & ~PCI_EXP_SLTSTA_CC))
 		return IRQ_HANDLED;
 
-	/* Check MRL Sensor Changed */
-	if (intr_loc & PCI_EXP_SLTSTA_MRLSC) {
-		pciehp_get_latch_status(slot, &open);
-		ctrl_info(ctrl, "Latch %s on Slot(%s)\n",
-			  open ? "open" : "close", slot_name(slot));
-		pciehp_queue_interrupt_event(slot, open ? INT_SWITCH_OPEN :
-					     INT_SWITCH_CLOSE);
-	}
-
 	/* Check Attention Button Pressed */
 	if (intr_loc & PCI_EXP_SLTSTA_ABP) {
 		ctrl_info(ctrl, "Button pressed on Slot(%s)\n",
@@ -662,13 +653,11 @@ void pcie_enable_notification(struct controller *ctrl)
 		cmd |= PCI_EXP_SLTCTL_ABPE;
 	else
 		cmd |= PCI_EXP_SLTCTL_PDCE;
-	if (MRL_SENS(ctrl))
-		cmd |= PCI_EXP_SLTCTL_MRLSCE;
 	if (!pciehp_poll_mode)
 		cmd |= PCI_EXP_SLTCTL_HPIE | PCI_EXP_SLTCTL_CCIE;
 
 	mask = (PCI_EXP_SLTCTL_PDCE | PCI_EXP_SLTCTL_ABPE |
-		PCI_EXP_SLTCTL_MRLSCE | PCI_EXP_SLTCTL_PFDE |
+		PCI_EXP_SLTCTL_PFDE |
 		PCI_EXP_SLTCTL_HPIE | PCI_EXP_SLTCTL_CCIE |
 		PCI_EXP_SLTCTL_DLLSCE);
 

From 3e14675335a15e1a91573837a15ff3bbb50bc14f Mon Sep 17 00:00:00 2001
From: Rob Herring <robh@kernel.org>
Date: Mon, 27 Jul 2015 15:55:21 -0500
Subject: [PATCH 09/20] PCI: Kill off set_irq_flags() usage
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

set_irq_flags is ARM-specific with custom flags which have genirq
equivalents.  Convert drivers to use the genirq interfaces directly, so we
can kill off set_irq_flags.  The translation of flags is as follows:

  IRQF_VALID -> !IRQ_NOREQUEST
  IRQF_PROBE -> !IRQ_NOPROBE
  IRQF_NOAUTOEN -> IRQ_NOAUTOEN

For IRQs managed by an irqdomain, the irqdomain core code handles clearing
and setting IRQ_NOREQUEST already, so there is no need to do this in .map()
functions, and we can simply remove the set_irq_flags calls.  Some users
also modify IRQ_NOPROBE, and this has been maintained although it is not
clear that is really needed.  There appears to be a great deal of blind
copy and paste of this code.

Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Jingoo Han <jingoohan1@gmail.com>
CC: Kishon Vijay Abraham I <kishon@ti.com>
CC: Murali Karicheri <m-karicheri2@ti.com>
CC: Thierry Reding <thierry.reding@gmail.com>
CC: Stephen Warren <swarren@wwwdotorg.org>
CC: Alexandre Courbot <gnurou@gmail.com>
CC: Jingoo Han <jingoohan1@gmail.com>
CC: Pratyush Anand <pratyush.anand@gmail.com>
CC: Simon Horman <horms@verge.net.au>
CC: Michal Simek <michal.simek@xilinx.com>
CC: "Sören Brinkmann" <soren.brinkmann@xilinx.com>
---
 drivers/pci/host/pci-dra7xx.c      | 1 -
 drivers/pci/host/pci-keystone-dw.c | 2 --
 drivers/pci/host/pci-tegra.c       | 1 -
 drivers/pci/host/pci-xgene-msi.c   | 1 -
 drivers/pci/host/pcie-designware.c | 1 -
 drivers/pci/host/pcie-rcar.c       | 1 -
 drivers/pci/host/pcie-xilinx.c     | 2 --
 7 files changed, 9 deletions(-)

diff --git a/drivers/pci/host/pci-dra7xx.c b/drivers/pci/host/pci-dra7xx.c
index 80db09e47800..1a0d124b199a 100644
--- a/drivers/pci/host/pci-dra7xx.c
+++ b/drivers/pci/host/pci-dra7xx.c
@@ -155,7 +155,6 @@ static int dra7xx_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
 {
 	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }
diff --git a/drivers/pci/host/pci-keystone-dw.c b/drivers/pci/host/pci-keystone-dw.c
index f34892e0edb4..f1d0749ebbf0 100644
--- a/drivers/pci/host/pci-keystone-dw.c
+++ b/drivers/pci/host/pci-keystone-dw.c
@@ -196,7 +196,6 @@ static int ks_dw_pcie_msi_map(struct irq_domain *domain, unsigned int irq,
 	irq_set_chip_and_handler(irq, &ks_dw_pcie_msi_irq_chip,
 				 handle_level_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }
@@ -277,7 +276,6 @@ static int ks_dw_pcie_init_legacy_irq_map(struct irq_domain *d,
 	irq_set_chip_and_handler(irq, &ks_dw_pcie_legacy_irq_chip,
 				 handle_level_irq);
 	irq_set_chip_data(irq, d->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }
diff --git a/drivers/pci/host/pci-tegra.c b/drivers/pci/host/pci-tegra.c
index 10c05718dbfd..81df0c1fe063 100644
--- a/drivers/pci/host/pci-tegra.c
+++ b/drivers/pci/host/pci-tegra.c
@@ -1248,7 +1248,6 @@ static int tegra_msi_map(struct irq_domain *domain, unsigned int irq,
 {
 	irq_set_chip_and_handler(irq, &tegra_msi_irq_chip, handle_simple_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	tegra_cpuidle_pcie_irqs_in_use();
 
diff --git a/drivers/pci/host/pci-xgene-msi.c b/drivers/pci/host/pci-xgene-msi.c
index 2d31d4d6fd08..8e559d181be1 100644
--- a/drivers/pci/host/pci-xgene-msi.c
+++ b/drivers/pci/host/pci-xgene-msi.c
@@ -223,7 +223,6 @@ static int xgene_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
 	irq_domain_set_info(domain, virq, msi_irq,
 			    &xgene_msi_bottom_irq_chip, domain->host_data,
 			    handle_simple_irq, NULL, NULL);
-	set_irq_flags(virq, IRQF_VALID);
 
 	return 0;
 }
diff --git a/drivers/pci/host/pcie-designware.c b/drivers/pci/host/pcie-designware.c
index 69486be7181e..5c6b562e5781 100644
--- a/drivers/pci/host/pcie-designware.c
+++ b/drivers/pci/host/pcie-designware.c
@@ -350,7 +350,6 @@ static int dw_pcie_msi_map(struct irq_domain *domain, unsigned int irq,
 {
 	irq_set_chip_and_handler(irq, &dw_msi_irq_chip, handle_simple_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }
diff --git a/drivers/pci/host/pcie-rcar.c b/drivers/pci/host/pcie-rcar.c
index c086210f2ffd..7678fe0820d7 100644
--- a/drivers/pci/host/pcie-rcar.c
+++ b/drivers/pci/host/pcie-rcar.c
@@ -664,7 +664,6 @@ static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
 {
 	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }
diff --git a/drivers/pci/host/pcie-xilinx.c b/drivers/pci/host/pcie-xilinx.c
index f1a06a091ccb..7201511972ad 100644
--- a/drivers/pci/host/pcie-xilinx.c
+++ b/drivers/pci/host/pcie-xilinx.c
@@ -338,7 +338,6 @@ static int xilinx_pcie_msi_map(struct irq_domain *domain, unsigned int irq,
 {
 	irq_set_chip_and_handler(irq, &xilinx_msi_irq_chip, handle_simple_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }
@@ -377,7 +376,6 @@ static int xilinx_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
 {
 	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
 	irq_set_chip_data(irq, domain->host_data);
-	set_irq_flags(irq, IRQF_VALID);
 
 	return 0;
 }

From 95e16587ccd0f330e349360de048ebd4432d4b2c Mon Sep 17 00:00:00 2001
From: Alex Williamson <alex.williamson@redhat.com>
Date: Mon, 10 Aug 2015 12:32:04 -0600
Subject: [PATCH 10/20] PCI: Add ACS quirks for Intel I219-LM/V

The Intel 100-series chipset now includes the integrated Ethernet as part
of a multifunction package.  The Ethernet function does not include native
ACS support, but Intel confirms that the device is not capable of peer-to-
peer within the package.  We can therefore quirk it to expose the
isolation.

Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: John Ronciak <john.ronciak@gmail.com>
---
 drivers/pci/quirks.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
index e9fd0e90fa3b..a9148483d184 100644
--- a/drivers/pci/quirks.c
+++ b/drivers/pci/quirks.c
@@ -3848,6 +3848,9 @@ static const struct pci_dev_acs_enabled {
 	{ PCI_VENDOR_ID_INTEL, 0x105F, pci_quirk_mf_endpoint_acs },
 	{ PCI_VENDOR_ID_INTEL, 0x1060, pci_quirk_mf_endpoint_acs },
 	{ PCI_VENDOR_ID_INTEL, 0x10D9, pci_quirk_mf_endpoint_acs },
+	/* I219 */
+	{ PCI_VENDOR_ID_INTEL, 0x15b7, pci_quirk_mf_endpoint_acs },
+	{ PCI_VENDOR_ID_INTEL, 0x15b8, pci_quirk_mf_endpoint_acs },
 	/* Intel PCH root ports */
 	{ PCI_VENDOR_ID_INTEL, PCI_ANY_ID, pci_quirk_intel_pch_acs },
 	{ 0x19a2, 0x710, pci_quirk_mf_endpoint_acs }, /* Emulex BE3-R */

From edc90fee916b4f0d14af9c6b5c08666747488ef8 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:05:46 -0500
Subject: [PATCH 11/20] PCI: Allocate ATS struct during enumeration

Previously, we allocated pci_ats structures when an IOMMU driver called
pci_enable_ats().  An SR-IOV VF shares the STU setting with its PF, so when
enabling ATS on the VF, we allocated a pci_ats struct for the PF if it
didn't already have one.  We held the sriov->lock to serialize threads
concurrently enabling ATS on several VFS so only one would allocate the PF
pci_ats.

Gregor reported a deadlock here:

  pci_enable_sriov
    sriov_enable
      virtfn_add
        mutex_lock(dev->sriov->lock)      # acquire sriov->lock
        pci_device_add
          device_add
            BUS_NOTIFY_ADD_DEVICE notifier chain
            iommu_bus_notifier
              amd_iommu_add_device        # iommu_ops.add_device
                init_iommu_group
                  iommu_group_get_for_dev
                    iommu_group_add_device
                      __iommu_attach_device
                        amd_iommu_attach_device  # iommu_ops.attach_device
                          attach_device
                            pci_enable_ats
                              mutex_lock(dev->sriov->lock) # deadlock

There's no reason to delay allocating the pci_ats struct, and if we
allocate it for each device at enumeration-time, there's no need for
locking in pci_enable_ats().

Allocate pci_ats struct during enumeration, when we initialize other
capabilities.

Note that this implementation requires ATS to be enabled on the PF first,
before on any of the VFs because the PF controls the STU for all the VFs.

Link: http://permalink.gmane.org/gmane.linux.kernel.iommu/9433
Reported-by: Gregor Dick <gdick@solarflare.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c       | 102 +++++++++++++++++-----------------------
 drivers/pci/probe.c     |   3 ++
 drivers/pci/remove.c    |   1 +
 include/linux/pci-ats.h |   2 +-
 include/linux/pci.h     |   9 ++++
 5 files changed, 58 insertions(+), 59 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index a8099d4d0c9d..2026f5388796 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -17,7 +17,7 @@
 
 #include "pci.h"
 
-static int ats_alloc_one(struct pci_dev *dev, int ps)
+static void ats_alloc_one(struct pci_dev *dev)
 {
 	int pos;
 	u16 cap;
@@ -25,20 +25,19 @@ static int ats_alloc_one(struct pci_dev *dev, int ps)
 
 	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ATS);
 	if (!pos)
-		return -ENODEV;
+		return;
 
 	ats = kzalloc(sizeof(*ats), GFP_KERNEL);
-	if (!ats)
-		return -ENOMEM;
+	if (!ats) {
+		dev_warn(&dev->dev, "can't allocate space for ATS state\n");
+		return;
+	}
 
 	ats->pos = pos;
-	ats->stu = ps;
 	pci_read_config_word(dev, pos + PCI_ATS_CAP, &cap);
 	ats->qdep = PCI_ATS_CAP_QDEP(cap) ? PCI_ATS_CAP_QDEP(cap) :
 					    PCI_ATS_MAX_QDEP;
 	dev->ats = ats;
-
-	return 0;
 }
 
 static void ats_free_one(struct pci_dev *dev)
@@ -47,6 +46,16 @@ static void ats_free_one(struct pci_dev *dev)
 	dev->ats = NULL;
 }
 
+void pci_ats_init(struct pci_dev *dev)
+{
+	ats_alloc_one(dev);
+}
+
+void pci_ats_free(struct pci_dev *dev)
+{
+	ats_free_one(dev);
+}
+
 /**
  * pci_enable_ats - enable the ATS capability
  * @dev: the PCI device
@@ -56,43 +65,35 @@ static void ats_free_one(struct pci_dev *dev)
  */
 int pci_enable_ats(struct pci_dev *dev, int ps)
 {
-	int rc;
 	u16 ctrl;
 
 	BUG_ON(dev->ats && dev->ats->is_enabled);
 
+	if (!dev->ats)
+		return -EINVAL;
+
 	if (ps < PCI_ATS_MIN_STU)
 		return -EINVAL;
 
-	if (dev->is_physfn || dev->is_virtfn) {
-		struct pci_dev *pdev = dev->is_physfn ? dev : dev->physfn;
-
-		mutex_lock(&pdev->sriov->lock);
-		if (pdev->ats)
-			rc = pdev->ats->stu == ps ? 0 : -EINVAL;
-		else
-			rc = ats_alloc_one(pdev, ps);
-
-		if (!rc)
-			pdev->ats->ref_cnt++;
-		mutex_unlock(&pdev->sriov->lock);
-		if (rc)
-			return rc;
-	}
-
-	if (!dev->is_physfn) {
-		rc = ats_alloc_one(dev, ps);
-		if (rc)
-			return rc;
-	}
-
+	/*
+	 * Note that enabling ATS on a VF fails unless it's already enabled
+	 * with the same STU on the PF.
+	 */
 	ctrl = PCI_ATS_CTRL_ENABLE;
-	if (!dev->is_virtfn)
-		ctrl |= PCI_ATS_CTRL_STU(ps - PCI_ATS_MIN_STU);
+	if (dev->is_virtfn) {
+		struct pci_dev *pdev = dev->physfn;
+
+		if (pdev->ats->stu != ps)
+			return -EINVAL;
+
+		atomic_inc(&pdev->ats->ref_cnt);  /* count enabled VFs */
+	} else {
+		dev->ats->stu = ps;
+		ctrl |= PCI_ATS_CTRL_STU(dev->ats->stu - PCI_ATS_MIN_STU);
+	}
 	pci_write_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, ctrl);
 
 	dev->ats->is_enabled = 1;
-
 	return 0;
 }
 EXPORT_SYMBOL_GPL(pci_enable_ats);
@@ -107,24 +108,20 @@ void pci_disable_ats(struct pci_dev *dev)
 
 	BUG_ON(!dev->ats || !dev->ats->is_enabled);
 
+	if (atomic_read(&dev->ats->ref_cnt))
+		return;		/* VFs still enabled */
+
+	if (dev->is_virtfn) {
+		struct pci_dev *pdev = dev->physfn;
+
+		atomic_dec(&pdev->ats->ref_cnt);
+	}
+
 	pci_read_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, &ctrl);
 	ctrl &= ~PCI_ATS_CTRL_ENABLE;
 	pci_write_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, ctrl);
 
 	dev->ats->is_enabled = 0;
-
-	if (dev->is_physfn || dev->is_virtfn) {
-		struct pci_dev *pdev = dev->is_physfn ? dev : dev->physfn;
-
-		mutex_lock(&pdev->sriov->lock);
-		pdev->ats->ref_cnt--;
-		if (!pdev->ats->ref_cnt)
-			ats_free_one(pdev);
-		mutex_unlock(&pdev->sriov->lock);
-	}
-
-	if (!dev->is_physfn)
-		ats_free_one(dev);
 }
 EXPORT_SYMBOL_GPL(pci_disable_ats);
 
@@ -140,7 +137,6 @@ void pci_restore_ats_state(struct pci_dev *dev)
 	ctrl = PCI_ATS_CTRL_ENABLE;
 	if (!dev->is_virtfn)
 		ctrl |= PCI_ATS_CTRL_STU(dev->ats->stu - PCI_ATS_MIN_STU);
-
 	pci_write_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, ctrl);
 }
 EXPORT_SYMBOL_GPL(pci_restore_ats_state);
@@ -159,23 +155,13 @@ EXPORT_SYMBOL_GPL(pci_restore_ats_state);
  */
 int pci_ats_queue_depth(struct pci_dev *dev)
 {
-	int pos;
-	u16 cap;
-
 	if (dev->is_virtfn)
 		return 0;
 
 	if (dev->ats)
 		return dev->ats->qdep;
 
-	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ATS);
-	if (!pos)
-		return -ENODEV;
-
-	pci_read_config_word(dev, pos + PCI_ATS_CAP, &cap);
-
-	return PCI_ATS_CAP_QDEP(cap) ? PCI_ATS_CAP_QDEP(cap) :
-				       PCI_ATS_MAX_QDEP;
+	return -ENODEV;
 }
 EXPORT_SYMBOL_GPL(pci_ats_queue_depth);
 
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c
index cefd636681b6..c206398ca67e 100644
--- a/drivers/pci/probe.c
+++ b/drivers/pci/probe.c
@@ -1540,6 +1540,9 @@ static void pci_init_capabilities(struct pci_dev *dev)
 	/* Single Root I/O Virtualization */
 	pci_iov_init(dev);
 
+	/* Address Translation Services */
+	pci_ats_init(dev);
+
 	/* Enable ACS P2P upstream forwarding */
 	pci_enable_acs(dev);
 }
diff --git a/drivers/pci/remove.c b/drivers/pci/remove.c
index 8a280e9c2ad1..27617b862ca8 100644
--- a/drivers/pci/remove.c
+++ b/drivers/pci/remove.c
@@ -26,6 +26,7 @@ static void pci_stop_dev(struct pci_dev *dev)
 		dev->is_added = 0;
 	}
 
+	pci_ats_free(dev);
 	if (dev->bus->self)
 		pcie_aspm_exit_link_state(dev);
 }
diff --git a/include/linux/pci-ats.h b/include/linux/pci-ats.h
index 72031785fe1d..e2dcc2ff3d0e 100644
--- a/include/linux/pci-ats.h
+++ b/include/linux/pci-ats.h
@@ -8,7 +8,7 @@ struct pci_ats {
 	int pos;        /* capability position */
 	int stu;        /* Smallest Translation Unit */
 	int qdep;       /* Invalidate Queue Depth */
-	int ref_cnt;    /* Physical Function reference count */
+	atomic_t ref_cnt; /* number of VFs with ATS enabled */
 	unsigned int is_enabled:1;      /* Enable bit is set */
 };
 
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 8a0321a8fb59..1817819ba57b 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1294,6 +1294,15 @@ int  ht_create_irq(struct pci_dev *dev, int idx);
 void ht_destroy_irq(unsigned int irq);
 #endif /* CONFIG_HT_IRQ */
 
+#ifdef CONFIG_PCI_ATS
+/* Address Translation Service */
+void pci_ats_init(struct pci_dev *dev);
+void pci_ats_free(struct pci_dev *dev);
+#else
+static inline void pci_ats_init(struct pci_dev *dev) { }
+static inline void pci_ats_free(struct pci_dev *dev) { }
+#endif
+
 void pci_cfg_access_lock(struct pci_dev *dev);
 bool pci_cfg_access_trylock(struct pci_dev *dev);
 void pci_cfg_access_unlock(struct pci_dev *dev);

From d544d75ac96aa1b0a8a378826626a0fbd8ce4380 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:15:19 -0500
Subject: [PATCH 12/20] PCI: Embed ATS info directly into struct pci_dev

The pci_ats struct is small and will get smaller, so I don't think it's
worth allocating it separately from the pci_dev struct.

Embed the ATS fields directly into struct pci_dev.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c       | 61 ++++++++++++++---------------------------
 drivers/pci/remove.c    |  1 -
 include/linux/pci-ats.h | 10 +------
 include/linux/pci.h     |  8 ++++--
 4 files changed, 27 insertions(+), 53 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index 2026f5388796..690ae6e6786c 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -21,29 +21,15 @@ static void ats_alloc_one(struct pci_dev *dev)
 {
 	int pos;
 	u16 cap;
-	struct pci_ats *ats;
 
 	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ATS);
 	if (!pos)
 		return;
 
-	ats = kzalloc(sizeof(*ats), GFP_KERNEL);
-	if (!ats) {
-		dev_warn(&dev->dev, "can't allocate space for ATS state\n");
-		return;
-	}
-
-	ats->pos = pos;
-	pci_read_config_word(dev, pos + PCI_ATS_CAP, &cap);
-	ats->qdep = PCI_ATS_CAP_QDEP(cap) ? PCI_ATS_CAP_QDEP(cap) :
+	dev->ats_cap = pos;
+	pci_read_config_word(dev, dev->ats_cap + PCI_ATS_CAP, &cap);
+	dev->ats_qdep = PCI_ATS_CAP_QDEP(cap) ? PCI_ATS_CAP_QDEP(cap) :
 					    PCI_ATS_MAX_QDEP;
-	dev->ats = ats;
-}
-
-static void ats_free_one(struct pci_dev *dev)
-{
-	kfree(dev->ats);
-	dev->ats = NULL;
 }
 
 void pci_ats_init(struct pci_dev *dev)
@@ -51,11 +37,6 @@ void pci_ats_init(struct pci_dev *dev)
 	ats_alloc_one(dev);
 }
 
-void pci_ats_free(struct pci_dev *dev)
-{
-	ats_free_one(dev);
-}
-
 /**
  * pci_enable_ats - enable the ATS capability
  * @dev: the PCI device
@@ -67,9 +48,9 @@ int pci_enable_ats(struct pci_dev *dev, int ps)
 {
 	u16 ctrl;
 
-	BUG_ON(dev->ats && dev->ats->is_enabled);
+	BUG_ON(dev->ats_cap && dev->ats_enabled);
 
-	if (!dev->ats)
+	if (!dev->ats_cap)
 		return -EINVAL;
 
 	if (ps < PCI_ATS_MIN_STU)
@@ -83,17 +64,17 @@ int pci_enable_ats(struct pci_dev *dev, int ps)
 	if (dev->is_virtfn) {
 		struct pci_dev *pdev = dev->physfn;
 
-		if (pdev->ats->stu != ps)
+		if (pdev->ats_stu != ps)
 			return -EINVAL;
 
-		atomic_inc(&pdev->ats->ref_cnt);  /* count enabled VFs */
+		atomic_inc(&pdev->ats_ref_cnt);  /* count enabled VFs */
 	} else {
-		dev->ats->stu = ps;
-		ctrl |= PCI_ATS_CTRL_STU(dev->ats->stu - PCI_ATS_MIN_STU);
+		dev->ats_stu = ps;
+		ctrl |= PCI_ATS_CTRL_STU(dev->ats_stu - PCI_ATS_MIN_STU);
 	}
-	pci_write_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, ctrl);
+	pci_write_config_word(dev, dev->ats_cap + PCI_ATS_CTRL, ctrl);
 
-	dev->ats->is_enabled = 1;
+	dev->ats_enabled = 1;
 	return 0;
 }
 EXPORT_SYMBOL_GPL(pci_enable_ats);
@@ -106,22 +87,22 @@ void pci_disable_ats(struct pci_dev *dev)
 {
 	u16 ctrl;
 
-	BUG_ON(!dev->ats || !dev->ats->is_enabled);
+	BUG_ON(!dev->ats_cap || !dev->ats_enabled);
 
-	if (atomic_read(&dev->ats->ref_cnt))
+	if (atomic_read(&dev->ats_ref_cnt))
 		return;		/* VFs still enabled */
 
 	if (dev->is_virtfn) {
 		struct pci_dev *pdev = dev->physfn;
 
-		atomic_dec(&pdev->ats->ref_cnt);
+		atomic_dec(&pdev->ats_ref_cnt);
 	}
 
-	pci_read_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, &ctrl);
+	pci_read_config_word(dev, dev->ats_cap + PCI_ATS_CTRL, &ctrl);
 	ctrl &= ~PCI_ATS_CTRL_ENABLE;
-	pci_write_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, ctrl);
+	pci_write_config_word(dev, dev->ats_cap + PCI_ATS_CTRL, ctrl);
 
-	dev->ats->is_enabled = 0;
+	dev->ats_enabled = 0;
 }
 EXPORT_SYMBOL_GPL(pci_disable_ats);
 
@@ -136,8 +117,8 @@ void pci_restore_ats_state(struct pci_dev *dev)
 
 	ctrl = PCI_ATS_CTRL_ENABLE;
 	if (!dev->is_virtfn)
-		ctrl |= PCI_ATS_CTRL_STU(dev->ats->stu - PCI_ATS_MIN_STU);
-	pci_write_config_word(dev, dev->ats->pos + PCI_ATS_CTRL, ctrl);
+		ctrl |= PCI_ATS_CTRL_STU(dev->ats_stu - PCI_ATS_MIN_STU);
+	pci_write_config_word(dev, dev->ats_cap + PCI_ATS_CTRL, ctrl);
 }
 EXPORT_SYMBOL_GPL(pci_restore_ats_state);
 
@@ -158,8 +139,8 @@ int pci_ats_queue_depth(struct pci_dev *dev)
 	if (dev->is_virtfn)
 		return 0;
 
-	if (dev->ats)
-		return dev->ats->qdep;
+	if (dev->ats_cap)
+		return dev->ats_qdep;
 
 	return -ENODEV;
 }
diff --git a/drivers/pci/remove.c b/drivers/pci/remove.c
index 27617b862ca8..8a280e9c2ad1 100644
--- a/drivers/pci/remove.c
+++ b/drivers/pci/remove.c
@@ -26,7 +26,6 @@ static void pci_stop_dev(struct pci_dev *dev)
 		dev->is_added = 0;
 	}
 
-	pci_ats_free(dev);
 	if (dev->bus->self)
 		pcie_aspm_exit_link_state(dev);
 }
diff --git a/include/linux/pci-ats.h b/include/linux/pci-ats.h
index e2dcc2ff3d0e..5d81d47b0a95 100644
--- a/include/linux/pci-ats.h
+++ b/include/linux/pci-ats.h
@@ -4,14 +4,6 @@
 #include <linux/pci.h>
 
 /* Address Translation Service */
-struct pci_ats {
-	int pos;        /* capability position */
-	int stu;        /* Smallest Translation Unit */
-	int qdep;       /* Invalidate Queue Depth */
-	atomic_t ref_cnt; /* number of VFs with ATS enabled */
-	unsigned int is_enabled:1;      /* Enable bit is set */
-};
-
 #ifdef CONFIG_PCI_ATS
 
 int pci_enable_ats(struct pci_dev *dev, int ps);
@@ -26,7 +18,7 @@ int pci_ats_queue_depth(struct pci_dev *dev);
  */
 static inline int pci_ats_enabled(struct pci_dev *dev)
 {
-	return dev->ats && dev->ats->is_enabled;
+	return dev->ats_cap && dev->ats_enabled;
 }
 
 #else /* CONFIG_PCI_ATS */
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 1817819ba57b..8bc16b5e4747 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -343,6 +343,7 @@ struct pci_dev {
 	unsigned int	msi_enabled:1;
 	unsigned int	msix_enabled:1;
 	unsigned int	ari_enabled:1;	/* ARI forwarding */
+	unsigned int	ats_enabled:1;	/* Address Translation Service */
 	unsigned int	is_managed:1;
 	unsigned int    needs_freset:1; /* Dev requires fundamental reset */
 	unsigned int	state_saved:1;
@@ -375,7 +376,10 @@ struct pci_dev {
 		struct pci_sriov *sriov;	/* SR-IOV capability related */
 		struct pci_dev *physfn;	/* the PF this VF is associated with */
 	};
-	struct pci_ats	*ats;	/* Address Translation Service */
+	int		ats_cap;	/* ATS Capability offset */
+	int		ats_stu;	/* ATS Smallest Translation Unit */
+	int		ats_qdep;	/* ATS Invalidate Queue Depth */
+	atomic_t	ats_ref_cnt;	/* number of VFs with ATS enabled */
 #endif
 	phys_addr_t rom; /* Physical address of ROM if it's not from the BAR */
 	size_t romlen; /* Length of ROM if it's not from the BAR */
@@ -1297,10 +1301,8 @@ void ht_destroy_irq(unsigned int irq);
 #ifdef CONFIG_PCI_ATS
 /* Address Translation Service */
 void pci_ats_init(struct pci_dev *dev);
-void pci_ats_free(struct pci_dev *dev);
 #else
 static inline void pci_ats_init(struct pci_dev *dev) { }
-static inline void pci_ats_free(struct pci_dev *dev) { }
 #endif
 
 void pci_cfg_access_lock(struct pci_dev *dev);

From 67930995d7fb8ae7d2078822b563010b289ace2e Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:27:34 -0500
Subject: [PATCH 13/20] PCI: Reduce size of ATS structure elements

The extended capabilities list is linked with 12-bit pointers, and the ATS
Smallest Translation Unit and Invalidate Queue Depth fields are both 5
bits.

Use u16 and u8 to hold the extended capability address and the stu and qdep
values.  No functional change.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 include/linux/pci.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/include/linux/pci.h b/include/linux/pci.h
index 8bc16b5e4747..238b77e8ca41 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -376,9 +376,9 @@ struct pci_dev {
 		struct pci_sriov *sriov;	/* SR-IOV capability related */
 		struct pci_dev *physfn;	/* the PF this VF is associated with */
 	};
-	int		ats_cap;	/* ATS Capability offset */
-	int		ats_stu;	/* ATS Smallest Translation Unit */
-	int		ats_qdep;	/* ATS Invalidate Queue Depth */
+	u16		ats_cap;	/* ATS Capability offset */
+	u8		ats_stu;	/* ATS Smallest Translation Unit */
+	u8		ats_qdep;	/* ATS Invalidate Queue Depth */
 	atomic_t	ats_ref_cnt;	/* number of VFs with ATS enabled */
 #endif
 	phys_addr_t rom; /* Physical address of ROM if it's not from the BAR */

From 3c765399524308ab36777a443ce77e19810a97d7 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:30:26 -0500
Subject: [PATCH 14/20] PCI: Rationalize pci_ats_queue_depth() error checking

We previously returned -ENODEV for devices that don't support ATS (except
that we always returned 0 for VFs, whether or not they support ATS).

For consistency, always return -EINVAL (not -ENODEV) if the device doesn't
support ATS.  Return zero for VFs that support ATS.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index 690ae6e6786c..9a98b3a4f983 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -136,13 +136,13 @@ EXPORT_SYMBOL_GPL(pci_restore_ats_state);
  */
 int pci_ats_queue_depth(struct pci_dev *dev)
 {
+	if (!dev->ats_cap)
+		return -EINVAL;
+
 	if (dev->is_virtfn)
 		return 0;
 
-	if (dev->ats_cap)
-		return dev->ats_qdep;
-
-	return -ENODEV;
+	return dev->ats_qdep;
 }
 EXPORT_SYMBOL_GPL(pci_ats_queue_depth);
 

From afdd596c42c4540308eae9f4c685d9bccc525cfc Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:35:18 -0500
Subject: [PATCH 15/20] PCI: Inline the ATS setup code into pci_ats_init()

The ATS setup code in ats_alloc_one() is only used by pci_ats_init(), so
inline it there.  No functional change.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index 9a98b3a4f983..95905f3c08f1 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -17,7 +17,7 @@
 
 #include "pci.h"
 
-static void ats_alloc_one(struct pci_dev *dev)
+void pci_ats_init(struct pci_dev *dev)
 {
 	int pos;
 	u16 cap;
@@ -32,11 +32,6 @@ static void ats_alloc_one(struct pci_dev *dev)
 					    PCI_ATS_MAX_QDEP;
 }
 
-void pci_ats_init(struct pci_dev *dev)
-{
-	ats_alloc_one(dev);
-}
-
 /**
  * pci_enable_ats - enable the ATS capability
  * @dev: the PCI device

From c39127dbaf6c267f18d1cac14e2b0ecbfaa52d63 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:38:13 -0500
Subject: [PATCH 16/20] PCI: Use pci_physfn() rather than looking up physfn by
 hand

Use the pci_physfn() helper rather than looking up physfn by hand.
No functional change.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index 95905f3c08f1..0b5b0ed7a436 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -42,6 +42,7 @@ void pci_ats_init(struct pci_dev *dev)
 int pci_enable_ats(struct pci_dev *dev, int ps)
 {
 	u16 ctrl;
+	struct pci_dev *pdev;
 
 	BUG_ON(dev->ats_cap && dev->ats_enabled);
 
@@ -57,8 +58,7 @@ int pci_enable_ats(struct pci_dev *dev, int ps)
 	 */
 	ctrl = PCI_ATS_CTRL_ENABLE;
 	if (dev->is_virtfn) {
-		struct pci_dev *pdev = dev->physfn;
-
+		pdev = pci_physfn(dev);
 		if (pdev->ats_stu != ps)
 			return -EINVAL;
 
@@ -80,6 +80,7 @@ EXPORT_SYMBOL_GPL(pci_enable_ats);
  */
 void pci_disable_ats(struct pci_dev *dev)
 {
+	struct pci_dev *pdev;
 	u16 ctrl;
 
 	BUG_ON(!dev->ats_cap || !dev->ats_enabled);
@@ -88,8 +89,7 @@ void pci_disable_ats(struct pci_dev *dev)
 		return;		/* VFs still enabled */
 
 	if (dev->is_virtfn) {
-		struct pci_dev *pdev = dev->physfn;
-
+		pdev = pci_physfn(dev);
 		atomic_dec(&pdev->ats_ref_cnt);
 	}
 

From a021f3019db7b40f0fe85b0a64c6be21e3665fbf Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:43:27 -0500
Subject: [PATCH 17/20] PCI: Clean up ATS error handling

There's no need to BUG() if we enable ATS when it's already enabled.  We
don't need to BUG() when disabling ATS on a device that doesn't support ATS
or if it's already disabled.  If ATS is enabled, certainly we found an ATS
capability in the past, so it should still be there now.

Clean up these error paths.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index 0b5b0ed7a436..9355f754c7c2 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -44,11 +44,12 @@ int pci_enable_ats(struct pci_dev *dev, int ps)
 	u16 ctrl;
 	struct pci_dev *pdev;
 
-	BUG_ON(dev->ats_cap && dev->ats_enabled);
-
 	if (!dev->ats_cap)
 		return -EINVAL;
 
+	if (WARN_ON(pci_ats_enabled(dev)))
+		return -EBUSY;
+
 	if (ps < PCI_ATS_MIN_STU)
 		return -EINVAL;
 
@@ -83,7 +84,8 @@ void pci_disable_ats(struct pci_dev *dev)
 	struct pci_dev *pdev;
 	u16 ctrl;
 
-	BUG_ON(!dev->ats_cap || !dev->ats_enabled);
+	if (WARN_ON(!pci_ats_enabled(dev)))
+		return;
 
 	if (atomic_read(&dev->ats_ref_cnt))
 		return;		/* VFs still enabled */
@@ -107,8 +109,6 @@ void pci_restore_ats_state(struct pci_dev *dev)
 
 	if (!pci_ats_enabled(dev))
 		return;
-	if (!pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ATS))
-		BUG();
 
 	ctrl = PCI_ATS_CTRL_ENABLE;
 	if (!dev->is_virtfn)

From ff9bee895c4d11a519a6b2c49451376025a6af4e Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Fri, 17 Jul 2015 15:55:48 -0500
Subject: [PATCH 18/20] PCI: Move ATS declarations to linux/pci.h so they're
 all together

Move ATS declarations to linux/pci.h so they're all in one place.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 include/linux/pci-ats.h | 41 -----------------------------------------
 include/linux/pci.h     | 10 +++++++++-
 2 files changed, 9 insertions(+), 42 deletions(-)

diff --git a/include/linux/pci-ats.h b/include/linux/pci-ats.h
index 5d81d47b0a95..57e0b8250947 100644
--- a/include/linux/pci-ats.h
+++ b/include/linux/pci-ats.h
@@ -3,47 +3,6 @@
 
 #include <linux/pci.h>
 
-/* Address Translation Service */
-#ifdef CONFIG_PCI_ATS
-
-int pci_enable_ats(struct pci_dev *dev, int ps);
-void pci_disable_ats(struct pci_dev *dev);
-int pci_ats_queue_depth(struct pci_dev *dev);
-
-/**
- * pci_ats_enabled - query the ATS status
- * @dev: the PCI device
- *
- * Returns 1 if ATS capability is enabled, or 0 if not.
- */
-static inline int pci_ats_enabled(struct pci_dev *dev)
-{
-	return dev->ats_cap && dev->ats_enabled;
-}
-
-#else /* CONFIG_PCI_ATS */
-
-static inline int pci_enable_ats(struct pci_dev *dev, int ps)
-{
-	return -ENODEV;
-}
-
-static inline void pci_disable_ats(struct pci_dev *dev)
-{
-}
-
-static inline int pci_ats_queue_depth(struct pci_dev *dev)
-{
-	return -ENODEV;
-}
-
-static inline int pci_ats_enabled(struct pci_dev *dev)
-{
-	return 0;
-}
-
-#endif /* CONFIG_PCI_ATS */
-
 #ifdef CONFIG_PCI_PRI
 
 int pci_enable_pri(struct pci_dev *pdev, u32 reqs);
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 238b77e8ca41..307f96a58e8b 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1301,8 +1301,16 @@ void ht_destroy_irq(unsigned int irq);
 #ifdef CONFIG_PCI_ATS
 /* Address Translation Service */
 void pci_ats_init(struct pci_dev *dev);
+int pci_enable_ats(struct pci_dev *dev, int ps);
+void pci_disable_ats(struct pci_dev *dev);
+int pci_ats_queue_depth(struct pci_dev *dev);
+static inline int pci_ats_enabled(struct pci_dev *dev) { return dev->ats_cap && dev->ats_enabled; }
 #else
-static inline void pci_ats_init(struct pci_dev *dev) { }
+static inline void pci_ats_init(struct pci_dev *d) { }
+static inline int pci_enable_ats(struct pci_dev *d, int ps) { return -ENODEV; }
+static inline void pci_disable_ats(struct pci_dev *d) { }
+static inline int pci_ats_queue_depth(struct pci_dev *d) { return -ENODEV; }
+static inline int pci_ats_enabled(struct pci_dev *d) { return 0; }
 #endif
 
 void pci_cfg_access_lock(struct pci_dev *dev);

From a71f938f3a9a7bc879296cd34ecae9effe5edf3f Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Mon, 20 Jul 2015 09:24:32 -0500
Subject: [PATCH 19/20] PCI: Stop caching ATS Invalidate Queue Depth

Stop caching the Invalidate Queue Depth in struct pci_dev.
pci_ats_queue_depth() is typically called only once per device, and it
returns a fixed value per-device, so callers who need the value frequently
can cache it themselves.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c   | 9 ++++-----
 include/linux/pci.h | 1 -
 2 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index 9355f754c7c2..ceda7dc556d4 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -20,16 +20,12 @@
 void pci_ats_init(struct pci_dev *dev)
 {
 	int pos;
-	u16 cap;
 
 	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ATS);
 	if (!pos)
 		return;
 
 	dev->ats_cap = pos;
-	pci_read_config_word(dev, dev->ats_cap + PCI_ATS_CAP, &cap);
-	dev->ats_qdep = PCI_ATS_CAP_QDEP(cap) ? PCI_ATS_CAP_QDEP(cap) :
-					    PCI_ATS_MAX_QDEP;
 }
 
 /**
@@ -131,13 +127,16 @@ EXPORT_SYMBOL_GPL(pci_restore_ats_state);
  */
 int pci_ats_queue_depth(struct pci_dev *dev)
 {
+	u16 cap;
+
 	if (!dev->ats_cap)
 		return -EINVAL;
 
 	if (dev->is_virtfn)
 		return 0;
 
-	return dev->ats_qdep;
+	pci_read_config_word(dev, dev->ats_cap + PCI_ATS_CAP, &cap);
+	return PCI_ATS_CAP_QDEP(cap) ? PCI_ATS_CAP_QDEP(cap) : PCI_ATS_MAX_QDEP;
 }
 EXPORT_SYMBOL_GPL(pci_ats_queue_depth);
 
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 307f96a58e8b..4b484fdfd66f 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -378,7 +378,6 @@ struct pci_dev {
 	};
 	u16		ats_cap;	/* ATS Capability offset */
 	u8		ats_stu;	/* ATS Smallest Translation Unit */
-	u8		ats_qdep;	/* ATS Invalidate Queue Depth */
 	atomic_t	ats_ref_cnt;	/* number of VFs with ATS enabled */
 #endif
 	phys_addr_t rom; /* Physical address of ROM if it's not from the BAR */

From f7ef1340bb501717372a39f4807d0ad519ebd432 Mon Sep 17 00:00:00 2001
From: Bjorn Helgaas <bhelgaas@google.com>
Date: Mon, 20 Jul 2015 09:23:37 -0500
Subject: [PATCH 20/20] PCI: Remove pci_ats_enabled()

Remove pci_ats_enabled().  There are no callers outside the ATS code
itself.  We don't need to check ats_cap, because if we don't find an ATS
capability, we'll never set ats_enabled.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/pci/ats.c   | 6 +++---
 include/linux/pci.h | 2 --
 2 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/pci/ats.c b/drivers/pci/ats.c
index ceda7dc556d4..eeb9fb2b47aa 100644
--- a/drivers/pci/ats.c
+++ b/drivers/pci/ats.c
@@ -43,7 +43,7 @@ int pci_enable_ats(struct pci_dev *dev, int ps)
 	if (!dev->ats_cap)
 		return -EINVAL;
 
-	if (WARN_ON(pci_ats_enabled(dev)))
+	if (WARN_ON(dev->ats_enabled))
 		return -EBUSY;
 
 	if (ps < PCI_ATS_MIN_STU)
@@ -80,7 +80,7 @@ void pci_disable_ats(struct pci_dev *dev)
 	struct pci_dev *pdev;
 	u16 ctrl;
 
-	if (WARN_ON(!pci_ats_enabled(dev)))
+	if (WARN_ON(!dev->ats_enabled))
 		return;
 
 	if (atomic_read(&dev->ats_ref_cnt))
@@ -103,7 +103,7 @@ void pci_restore_ats_state(struct pci_dev *dev)
 {
 	u16 ctrl;
 
-	if (!pci_ats_enabled(dev))
+	if (!dev->ats_enabled)
 		return;
 
 	ctrl = PCI_ATS_CTRL_ENABLE;
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 4b484fdfd66f..806da7634f91 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1303,13 +1303,11 @@ void pci_ats_init(struct pci_dev *dev);
 int pci_enable_ats(struct pci_dev *dev, int ps);
 void pci_disable_ats(struct pci_dev *dev);
 int pci_ats_queue_depth(struct pci_dev *dev);
-static inline int pci_ats_enabled(struct pci_dev *dev) { return dev->ats_cap && dev->ats_enabled; }
 #else
 static inline void pci_ats_init(struct pci_dev *d) { }
 static inline int pci_enable_ats(struct pci_dev *d, int ps) { return -ENODEV; }
 static inline void pci_disable_ats(struct pci_dev *d) { }
 static inline int pci_ats_queue_depth(struct pci_dev *d) { return -ENODEV; }
-static inline int pci_ats_enabled(struct pci_dev *d) { return 0; }
 #endif
 
 void pci_cfg_access_lock(struct pci_dev *dev);