Merge tag 'pci-v5.10-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci

Pull PCI updates from Bjorn Helgaas:
 "Enumeration:
   - Print IRQ number used by PCIe Link Bandwidth Notification (Dongdong
     Liu)
   - Add schedule point in pci_read_config() to reduce max latency
     (Jiang Biao)
   - Add Kconfig options for MPS/MRRS strategy (Jim Quinlan)

  Resource management:
   - Fix pci_iounmap() memory leak when !CONFIG_GENERIC_IOMAP (Lorenzo
     Pieralisi)

  PCIe native device hotplug:
   - Reduce noisiness on hot removal (Lukas Wunner)

  Power management:
   - Revert "PCI/PM: Apply D2 delay as milliseconds, not microseconds"
     that was done on the basis of spec typo (Bjorn Helgaas)
   - Rename pci_dev.d3_delay to d3hot_delay to remove D3hot/D3cold
     ambiguity (Krzysztof Wilczyński)
   - Remove unused pcibios_pm_ops (Vaibhav Gupta)

  IOMMU:
   - Enable Translation Blocking for external devices to harden against
     DMA attacks (Rajat Jain)

  Error handling:
   - Add an ACPI APEI notifier chain for vendor CPER records to enable
     device-specific error handling (Shiju Jose)

  ASPM:
   - Remove struct aspm_register_info to simplify code (Saheed O.
     Bolarinwa)

  Amlogic Meson PCIe controller driver:
   - Build as module by default (Kevin Hilman)

  Ampere Altra PCIe controller driver:
   - Add MCFG quirk to work around non-standard ECAM implementation
     (Tuan Phan)

  Broadcom iProc PCIe controller driver:
   - Set affinity mask on MSI interrupts (Mark Tomlinson)

  Broadcom STB PCIe controller driver:
   - Make PCIE_BRCMSTB depend on ARCH_BRCMSTB (Jim Quinlan)
   - Add DT bindings for more Brcmstb chips (Jim Quinlan)
   - Add bcm7278 register info (Jim Quinlan)
   - Add bcm7278 PERST# support (Jim Quinlan)
   - Add suspend and resume pm_ops (Jim Quinlan)
   - Add control of rescal reset (Jim Quinlan)
   - Set additional internal memory DMA viewport sizes (Jim Quinlan)
   - Accommodate MSI for older chips (Jim Quinlan)
   - Set bus max burst size by chip type (Jim Quinlan)
   - Add support for bcm7211, bcm7216, bcm7445, bcm7278 (Jim Quinlan)

  Freescale i.MX6 PCIe controller driver:
   - Use dev_err_probe() to reduce redundant messages (Anson Huang)

  Freescale Layerscape PCIe controller driver:
   - Enforce 4K DMA buffer alignment in endpoint test (Hou Zhiqiang)
   - Add DT compatible strings for ls1088a, ls2088a (Xiaowei Bao)
   - Add endpoint support for ls1088a, ls2088a (Xiaowei Bao)
   - Add endpoint test support for lS1088a (Xiaowei Bao)
   - Add MSI-X support for ls1088a (Xiaowei Bao)

  HiSilicon HIP PCIe controller driver:
   - Handle HIP-specific errors via ACPI APEI (Yicong Yang)

  HiSilicon Kirin PCIe controller driver:
   - Return -EPROBE_DEFER if the GPIO isn't ready (Bean Huo)

  Intel VMD host bridge driver:
   - Factor out physical offset, bus offset, IRQ domain, IRQ allocation
     (Jon Derrick)
   - Use generic PCI PM correctly (Jon Derrick)

  Marvell Aardvark PCIe controller driver:
   - Fix compilation on s390 (Pali Rohár)
   - Implement driver 'remove' function and allow to build it as module
     (Pali Rohár)
   - Move PCIe reset card code to advk_pcie_train_link() (Pali Rohár)
   - Convert mvebu a3700 internal SMCC firmware return codes to errno
     (Pali Rohár)
   - Fix initialization with old Marvell's Arm Trusted Firmware (Pali
     Rohár)

  Microsoft Hyper-V host bridge driver:
   - Fix hibernation in case interrupts are not re-created (Dexuan Cui)

  NVIDIA Tegra PCIe controller driver:
   - Stop checking return value of debugfs_create() functions (Greg
     Kroah-Hartman)
   - Convert to use DEFINE_SEQ_ATTRIBUTE macro (Liu Shixin)

  Qualcomm PCIe controller driver:
   - Reset PCIe to work around Qsdk U-Boot issue (Ansuel Smith)

  Renesas R-Car PCIe controller driver:
   - Add DT documentation for r8a774a1, r8a774b1, r8a774e1 endpoints
     (Lad Prabhakar)
   - Add RZ/G2M, RZ/G2N, RZ/G2H IDs to endpoint test (Lad Prabhakar)
   - Add DT support for r8a7742 (Lad Prabhakar)

  Socionext UniPhier Pro5 controller driver:
   - Add DT descriptions of iATU register (host and endpoint) (Kunihiko
     Hayashi)

  Synopsys DesignWare PCIe controller driver:
   - Add link up check in dw_child_pcie_ops.map_bus() (racy, but seems
     unavoidable) (Hou Zhiqiang)
   - Fix endpoint Header Type check so multi-function devices work (Hou
     Zhiqiang)
   - Skip PCIE_MSI_INTR0* programming if MSI is disabled (Jisheng Zhang)
   - Stop leaking MSI page in suspend/resume (Jisheng Zhang)
   - Add common iATU register support instead of keystone-specific code
     (Kunihiko Hayashi)
   - Major config space access and other cleanups in dwc core and
     drivers that use it (al, exynos, histb, imx6, intel-gw, keystone,
     kirin, meson, qcom, tegra) (Rob Herring)
   - Add multiple PFs support for endpoint (Xiaowei Bao)
   - Add MSI-X doorbell mode in endpoint mode (Xiaowei Bao)

  Miscellaneous:
   - Use fallthrough pseudo-keyword (Gustavo A. R. Silva)
   - Fix "0 used as NULL pointer" warnings (Gustavo Pimentel)
   - Fix "cast truncates bits from constant value" warnings (Gustavo
     Pimentel)
   - Remove redundant zeroing for sg_init_table() (Julia Lawall)
   - Use scnprintf(), not snprintf(), in sysfs "show" functions
     (Krzysztof Wilczyński)
   - Remove unused assignments (Krzysztof Wilczyński)
   - Fix "0 used as NULL pointer" warning (Krzysztof Wilczyński)
   - Simplify bool comparisons (Krzysztof Wilczyński)
   - Use for_each_child_of_node() and for_each_node_by_name() (Qinglang
     Miao)
   - Simplify return expressions (Qinglang Miao)"

* tag 'pci-v5.10-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (147 commits)
  PCI: vmd: Update VMD PM to correctly use generic PCI PM
  PCI: vmd: Create IRQ allocation helper
  PCI: vmd: Create IRQ Domain configuration helper
  PCI: vmd: Create bus offset configuration helper
  PCI: vmd: Create physical offset helper
  PCI: v3-semi: Remove unneeded break
  PCI: dwc: Add link up check in dw_child_pcie_ops.map_bus()
  PCI/ASPM: Remove struct pcie_link_state.l1ss
  PCI/ASPM: Remove struct aspm_register_info.l1ss_cap
  PCI/ASPM: Pass L1SS Capabilities value, not struct aspm_register_info
  PCI/ASPM: Remove struct aspm_register_info.l1ss_ctl1
  PCI/ASPM: Remove struct aspm_register_info.l1ss_ctl2 (unused)
  PCI/ASPM: Remove struct aspm_register_info.l1ss_cap_ptr
  PCI/ASPM: Remove struct aspm_register_info.latency_encoding
  PCI/ASPM: Remove struct aspm_register_info.enabled
  PCI/ASPM: Remove struct aspm_register_info.support
  PCI/ASPM: Use 'parent' and 'child' for readability
  PCI/ASPM: Move LTR path check to where it's used
  PCI/ASPM: Move pci_clear_and_set_dword() earlier
  PCI: dwc: Fix MSI page leakage in suspend/resume
  ...

1 files changed, 181 insertions, 125 deletions

diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c
index aa1b12bac9a1..f375c21ceeb1 100644
--- a/drivers/pci/controller/vmd.c
+++ b/drivers/pci/controller/vmd.c
@@ -298,6 +298,33 @@ static struct msi_domain_info vmd_msi_domain_info = {
 	.chip		= &vmd_msi_controller,
 };
 
+static int vmd_create_irq_domain(struct vmd_dev *vmd)
+{
+	struct fwnode_handle *fn;
+
+	fn = irq_domain_alloc_named_id_fwnode("VMD-MSI", vmd->sysdata.domain);
+	if (!fn)
+		return -ENODEV;
+
+	vmd->irq_domain = pci_msi_create_irq_domain(fn, &vmd_msi_domain_info, NULL);
+	if (!vmd->irq_domain) {
+		irq_domain_free_fwnode(fn);
+		return -ENODEV;
+	}
+
+	return 0;
+}
+
+static void vmd_remove_irq_domain(struct vmd_dev *vmd)
+{
+	if (vmd->irq_domain) {
+		struct fwnode_handle *fn = vmd->irq_domain->fwnode;
+
+		irq_domain_remove(vmd->irq_domain);
+		irq_domain_free_fwnode(fn);
+	}
+}
+
 static char __iomem *vmd_cfg_addr(struct vmd_dev *vmd, struct pci_bus *bus,
 				  unsigned int devfn, int reg, int len)
 {
@@ -417,97 +444,175 @@ static int vmd_find_free_domain(void)
 	return domain + 1;
 }
 
-static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
+static int vmd_get_phys_offsets(struct vmd_dev *vmd, bool native_hint,
+				resource_size_t *offset1,
+				resource_size_t *offset2)
 {
-	struct pci_sysdata *sd = &vmd->sysdata;
-	struct fwnode_handle *fn;
-	struct resource *res;
-	u32 upper_bits;
-	unsigned long flags;
-	LIST_HEAD(resources);
-	resource_size_t offset[2] = {0};
-	resource_size_t membar2_offset = 0x2000;
-	struct pci_bus *child;
+	struct pci_dev *dev = vmd->dev;
+	u64 phys1, phys2;
 
-	/*
-	 * Shadow registers may exist in certain VMD device ids which allow
-	 * guests to correctly assign host physical addresses to the root ports
-	 * and child devices. These registers will either return the host value
-	 * or 0, depending on an enable bit in the VMD device.
-	 */
-	if (features & VMD_FEAT_HAS_MEMBAR_SHADOW) {
+	if (native_hint) {
 		u32 vmlock;
 		int ret;
 
-		membar2_offset = MB2_SHADOW_OFFSET + MB2_SHADOW_SIZE;
-		ret = pci_read_config_dword(vmd->dev, PCI_REG_VMLOCK, &vmlock);
+		ret = pci_read_config_dword(dev, PCI_REG_VMLOCK, &vmlock);
 		if (ret || vmlock == ~0)
 			return -ENODEV;
 
 		if (MB2_SHADOW_EN(vmlock)) {
 			void __iomem *membar2;
 
-			membar2 = pci_iomap(vmd->dev, VMD_MEMBAR2, 0);
+			membar2 = pci_iomap(dev, VMD_MEMBAR2, 0);
 			if (!membar2)
 				return -ENOMEM;
-			offset[0] = vmd->dev->resource[VMD_MEMBAR1].start -
-					(readq(membar2 + MB2_SHADOW_OFFSET) &
-					 PCI_BASE_ADDRESS_MEM_MASK);
-			offset[1] = vmd->dev->resource[VMD_MEMBAR2].start -
-					(readq(membar2 + MB2_SHADOW_OFFSET + 8) &
-					 PCI_BASE_ADDRESS_MEM_MASK);
-			pci_iounmap(vmd->dev, membar2);
-		}
-	}
-
-	if (features & VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP) {
-		int pos = pci_find_capability(vmd->dev, PCI_CAP_ID_VNDR);
+			phys1 = readq(membar2 + MB2_SHADOW_OFFSET);
+			phys2 = readq(membar2 + MB2_SHADOW_OFFSET + 8);
+			pci_iounmap(dev, membar2);
+		} else
+			return 0;
+	} else {
+		/* Hypervisor-Emulated Vendor-Specific Capability */
+		int pos = pci_find_capability(dev, PCI_CAP_ID_VNDR);
 		u32 reg, regu;
 
-		pci_read_config_dword(vmd->dev, pos + 4, &reg);
+		pci_read_config_dword(dev, pos + 4, &reg);
 
 		/* "SHDW" */
 		if (pos && reg == 0x53484457) {
-			pci_read_config_dword(vmd->dev, pos + 8, &reg);
-			pci_read_config_dword(vmd->dev, pos + 12, &regu);
-			offset[0] = vmd->dev->resource[VMD_MEMBAR1].start -
-					(((u64) regu << 32 | reg) &
-					 PCI_BASE_ADDRESS_MEM_MASK);
-
-			pci_read_config_dword(vmd->dev, pos + 16, &reg);
-			pci_read_config_dword(vmd->dev, pos + 20, &regu);
-			offset[1] = vmd->dev->resource[VMD_MEMBAR2].start -
-					(((u64) regu << 32 | reg) &
-					 PCI_BASE_ADDRESS_MEM_MASK);
+			pci_read_config_dword(dev, pos + 8, &reg);
+			pci_read_config_dword(dev, pos + 12, &regu);
+			phys1 = (u64) regu << 32 | reg;
+
+			pci_read_config_dword(dev, pos + 16, &reg);
+			pci_read_config_dword(dev, pos + 20, &regu);
+			phys2 = (u64) regu << 32 | reg;
+		} else
+			return 0;
+	}
+
+	*offset1 = dev->resource[VMD_MEMBAR1].start -
+			(phys1 & PCI_BASE_ADDRESS_MEM_MASK);
+	*offset2 = dev->resource[VMD_MEMBAR2].start -
+			(phys2 & PCI_BASE_ADDRESS_MEM_MASK);
+
+	return 0;
+}
+
+static int vmd_get_bus_number_start(struct vmd_dev *vmd)
+{
+	struct pci_dev *dev = vmd->dev;
+	u16 reg;
+
+	pci_read_config_word(dev, PCI_REG_VMCAP, &reg);
+	if (BUS_RESTRICT_CAP(reg)) {
+		pci_read_config_word(dev, PCI_REG_VMCONFIG, &reg);
+
+		switch (BUS_RESTRICT_CFG(reg)) {
+		case 0:
+			vmd->busn_start = 0;
+			break;
+		case 1:
+			vmd->busn_start = 128;
+			break;
+		case 2:
+			vmd->busn_start = 224;
+			break;
+		default:
+			pci_err(dev, "Unknown Bus Offset Setting (%d)\n",
+				BUS_RESTRICT_CFG(reg));
+			return -ENODEV;
 		}
 	}
 
+	return 0;
+}
+
+static irqreturn_t vmd_irq(int irq, void *data)
+{
+	struct vmd_irq_list *irqs = data;
+	struct vmd_irq *vmdirq;
+	int idx;
+
+	idx = srcu_read_lock(&irqs->srcu);
+	list_for_each_entry_rcu(vmdirq, &irqs->irq_list, node)
+		generic_handle_irq(vmdirq->virq);
+	srcu_read_unlock(&irqs->srcu, idx);
+
+	return IRQ_HANDLED;
+}
+
+static int vmd_alloc_irqs(struct vmd_dev *vmd)
+{
+	struct pci_dev *dev = vmd->dev;
+	int i, err;
+
+	vmd->msix_count = pci_msix_vec_count(dev);
+	if (vmd->msix_count < 0)
+		return -ENODEV;
+
+	vmd->msix_count = pci_alloc_irq_vectors(dev, 1, vmd->msix_count,
+						PCI_IRQ_MSIX);
+	if (vmd->msix_count < 0)
+		return vmd->msix_count;
+
+	vmd->irqs = devm_kcalloc(&dev->dev, vmd->msix_count, sizeof(*vmd->irqs),
+				 GFP_KERNEL);
+	if (!vmd->irqs)
+		return -ENOMEM;
+
+	for (i = 0; i < vmd->msix_count; i++) {
+		err = init_srcu_struct(&vmd->irqs[i].srcu);
+		if (err)
+			return err;
+
+		INIT_LIST_HEAD(&vmd->irqs[i].irq_list);
+		err = devm_request_irq(&dev->dev, pci_irq_vector(dev, i),
+				       vmd_irq, IRQF_NO_THREAD,
+				       "vmd", &vmd->irqs[i]);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
+{
+	struct pci_sysdata *sd = &vmd->sysdata;
+	struct resource *res;
+	u32 upper_bits;
+	unsigned long flags;
+	LIST_HEAD(resources);
+	resource_size_t offset[2] = {0};
+	resource_size_t membar2_offset = 0x2000;
+	struct pci_bus *child;
+	int ret;
+
+	/*
+	 * Shadow registers may exist in certain VMD device ids which allow
+	 * guests to correctly assign host physical addresses to the root ports
+	 * and child devices. These registers will either return the host value
+	 * or 0, depending on an enable bit in the VMD device.
+	 */
+	if (features & VMD_FEAT_HAS_MEMBAR_SHADOW) {
+		membar2_offset = MB2_SHADOW_OFFSET + MB2_SHADOW_SIZE;
+		ret = vmd_get_phys_offsets(vmd, true, &offset[0], &offset[1]);
+		if (ret)
+			return ret;
+	} else if (features & VMD_FEAT_HAS_MEMBAR_SHADOW_VSCAP) {
+		ret = vmd_get_phys_offsets(vmd, false, &offset[0], &offset[1]);
+		if (ret)
+			return ret;
+	}
+
 	/*
 	 * Certain VMD devices may have a root port configuration option which
 	 * limits the bus range to between 0-127, 128-255, or 224-255
 	 */
 	if (features & VMD_FEAT_HAS_BUS_RESTRICTIONS) {
-		u16 reg16;
-
-		pci_read_config_word(vmd->dev, PCI_REG_VMCAP, &reg16);
-		if (BUS_RESTRICT_CAP(reg16)) {
-			pci_read_config_word(vmd->dev, PCI_REG_VMCONFIG,
-					     &reg16);
-
-			switch (BUS_RESTRICT_CFG(reg16)) {
-			case 1:
-				vmd->busn_start = 128;
-				break;
-			case 2:
-				vmd->busn_start = 224;
-				break;
-			case 3:
-				pci_err(vmd->dev, "Unknown Bus Offset Setting\n");
-				return -ENODEV;
-			default:
-				break;
-			}
-		}
+		ret = vmd_get_bus_number_start(vmd);
+		if (ret)
+			return ret;
 	}
 
 	res = &vmd->dev->resource[VMD_CFGBAR];
@@ -568,17 +673,9 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
 
 	sd->node = pcibus_to_node(vmd->dev->bus);
 
-	fn = irq_domain_alloc_named_id_fwnode("VMD-MSI", vmd->sysdata.domain);
-	if (!fn)
-		return -ENODEV;
-
-	vmd->irq_domain = pci_msi_create_irq_domain(fn, &vmd_msi_domain_info,
-						    NULL);
-
-	if (!vmd->irq_domain) {
-		irq_domain_free_fwnode(fn);
-		return -ENODEV;
-	}
+	ret = vmd_create_irq_domain(vmd);
+	if (ret)
+		return ret;
 
 	/*
 	 * Override the irq domain bus token so the domain can be distinguished
@@ -594,13 +691,13 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
 				       &vmd_ops, sd, &resources);
 	if (!vmd->bus) {
 		pci_free_resource_list(&resources);
-		irq_domain_remove(vmd->irq_domain);
-		irq_domain_free_fwnode(fn);
+		vmd_remove_irq_domain(vmd);
 		return -ENODEV;
 	}
 
 	vmd_attach_resources(vmd);
-	dev_set_msi_domain(&vmd->bus->dev, vmd->irq_domain);
+	if (vmd->irq_domain)
+		dev_set_msi_domain(&vmd->bus->dev, vmd->irq_domain);
 
 	pci_scan_child_bus(vmd->bus);
 	pci_assign_unassigned_bus_resources(vmd->bus);
@@ -620,24 +717,10 @@ static int vmd_enable_domain(struct vmd_dev *vmd, unsigned long features)
 	return 0;
 }
 
-static irqreturn_t vmd_irq(int irq, void *data)
-{
-	struct vmd_irq_list *irqs = data;
-	struct vmd_irq *vmdirq;
-	int idx;
-
-	idx = srcu_read_lock(&irqs->srcu);
-	list_for_each_entry_rcu(vmdirq, &irqs->irq_list, node)
-		generic_handle_irq(vmdirq->virq);
-	srcu_read_unlock(&irqs->srcu, idx);
-
-	return IRQ_HANDLED;
-}
-
 static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)
 {
 	struct vmd_dev *vmd;
-	int i, err;
+	int err;
 
 	if (resource_size(&dev->resource[VMD_CFGBAR]) < (1 << 20))
 		return -ENOMEM;
@@ -660,32 +743,9 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)
 	    dma_set_mask_and_coherent(&dev->dev, DMA_BIT_MASK(32)))
 		return -ENODEV;
 
-	vmd->msix_count = pci_msix_vec_count(dev);
-	if (vmd->msix_count < 0)
-		return -ENODEV;
-
-	vmd->msix_count = pci_alloc_irq_vectors(dev, 1, vmd->msix_count,
-					PCI_IRQ_MSIX);
-	if (vmd->msix_count < 0)
-		return vmd->msix_count;
-
-	vmd->irqs = devm_kcalloc(&dev->dev, vmd->msix_count, sizeof(*vmd->irqs),
-				 GFP_KERNEL);
-	if (!vmd->irqs)
-		return -ENOMEM;
-
-	for (i = 0; i < vmd->msix_count; i++) {
-		err = init_srcu_struct(&vmd->irqs[i].srcu);
-		if (err)
-			return err;
-
-		INIT_LIST_HEAD(&vmd->irqs[i].irq_list);
-		err = devm_request_irq(&dev->dev, pci_irq_vector(dev, i),
-				       vmd_irq, IRQF_NO_THREAD,
-				       "vmd", &vmd->irqs[i]);
-		if (err)
-			return err;
-	}
+	err = vmd_alloc_irqs(vmd);
+	if (err)
+		return err;
 
 	spin_lock_init(&vmd->cfg_lock);
 	pci_set_drvdata(dev, vmd);
@@ -709,15 +769,13 @@ static void vmd_cleanup_srcu(struct vmd_dev *vmd)
 static void vmd_remove(struct pci_dev *dev)
 {
 	struct vmd_dev *vmd = pci_get_drvdata(dev);
-	struct fwnode_handle *fn = vmd->irq_domain->fwnode;
 
 	sysfs_remove_link(&vmd->dev->dev.kobj, "domain");
 	pci_stop_root_bus(vmd->bus);
 	pci_remove_root_bus(vmd->bus);
 	vmd_cleanup_srcu(vmd);
 	vmd_detach_resources(vmd);
-	irq_domain_remove(vmd->irq_domain);
-	irq_domain_free_fwnode(fn);
+	vmd_remove_irq_domain(vmd);
 }
 
 #ifdef CONFIG_PM_SLEEP
@@ -730,7 +788,6 @@ static int vmd_suspend(struct device *dev)
 	for (i = 0; i < vmd->msix_count; i++)
 		devm_free_irq(dev, pci_irq_vector(pdev, i), &vmd->irqs[i]);
 
-	pci_save_state(pdev);
 	return 0;
 }
 
@@ -748,7 +805,6 @@ static int vmd_resume(struct device *dev)
 			return err;
 	}
 
-	pci_restore_state(pdev);
 	return 0;
 }
 #endif