Merge tag 'pci-v4.9-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci

Pull PCI updates from Bjorn Helgaas:
 "Summary of PCI changes for the v4.9 merge window:

  Enumeration:
   - microblaze: Add multidomain support for procfs (Bharat Kumar Gogada)

  Resource management:
   - Ignore requested alignment for PROBE_ONLY and fixed resources (Yongji Xie)
   - Ignore requested alignment for VF BARs (Yongji Xie)

  PCI device hotplug:
   - Make core explicitly non-modular (Paul Gortmaker)

  PCIe native device hotplug:
   - Rename pcie_isr() locals for clarity (Bjorn Helgaas)
   - Return IRQ_NONE when we can't read interrupt status (Bjorn Helgaas)
   - Remove unnecessary guard (Bjorn Helgaas)
   - Clean up dmesg "Slot(%s)" messages (Bjorn Helgaas)
   - Remove useless pciehp_get_latch_status() calls (Bjorn Helgaas)
   - Clear attention LED on device add (Keith Busch)
   - Allow exclusive userspace control of indicators (Keith Busch)
   - Process all hotplug events before looking for new ones (Mayurkumar Patel)
   - Don't re-read Slot Status when queuing hotplug event (Mayurkumar Patel)
   - Don't re-read Slot Status when handling surprise event (Mayurkumar Patel)
   - Make explicitly non-modular (Paul Gortmaker)

  Power management:
   - Afford direct-complete to devices with non-standard PM (Lukas Wunner)
   - Query platform firmware for device power state (Lukas Wunner)
   - Recognize D3cold in pci_update_current_state() (Lukas Wunner)
   - Avoid unnecessary resume after direct-complete (Lukas Wunner)
   - Make explicitly non-modular (Paul Gortmaker)

  Virtualization:
   - Mark Atheros AR9580 to avoid bus reset (Maik Broemme)
   - Check for pci_setup_device() failure in pci_iov_add_virtfn() (Po Liu)

  MSI:
   - Enable PCI_MSI_IRQ_DOMAIN support for ARC (Joao Pinto)

  AER:
   - Remove aerdriver.nosourceid kernel parameter (Bjorn Helgaas)
   - Remove aerdriver.forceload kernel parameter (Bjorn Helgaas)
   - Fix aer_probe() kernel-doc comment (Cao jin)
   - Add bus flag to skip source ID matching (Jon Derrick)
   - Avoid memory allocation in interrupt handling path (Jon Derrick)
   - Cache capability position (Keith Busch)
   - Make explicitly non-modular (Paul Gortmaker)
   - Remove duplicate AER severity translation (Tyler Baicar)
   - Send correct severity to calculate AER severity (Tyler Baicar)

  Precision Time Measurement:
   - Add Precision Time Measurement (PTM) support (Jonathan Yong)
   - Add PTM clock granularity information (Bjorn Helgaas)
   - Add pci_enable_ptm() for drivers to enable PTM on endpoints (Bjorn Helgaas)

  Generic host bridge driver:
   - Fix pci_remap_iospace() failure path (Lorenzo Pieralisi)
   - Make explicitly non-modular (Paul Gortmaker)

  Altera host bridge driver:
   - Remove redundant platform_get_resource() return value check (Bjorn Helgaas)
   - Poll for link training status after retraining the link (Ley Foon Tan)
   - Rework config accessors for use without a struct pci_bus (Ley Foon Tan)
   - Move retrain from fixup to altera_pcie_host_init() (Ley Foon Tan)
   - Make MSI explicitly non-modular (Paul Gortmaker)
   - Make explicitly non-modular (Paul Gortmaker)
   - Relax device number checking to allow SR-IOV (Po Liu)

  ARM Versatile host bridge driver:
   - Fix pci_remap_iospace() failure path (Lorenzo Pieralisi)

  Axis ARTPEC-6 host bridge driver:
   - Drop __init from artpec6_add_pcie_port() (Niklas Cassel)

  Freescale i.MX6 host bridge driver:
   - Make explicitly non-modular (Paul Gortmaker)

  Intel VMD host bridge driver:
   - Add quirk for AER to ignore source ID (Jon Derrick)
   - Allocate IRQ lists with correct MSI-X count (Jon Derrick)
   - Convert to use pci_alloc_irq_vectors() API (Jon Derrick)
   - Eliminate vmd_vector member from list type (Jon Derrick)
   - Eliminate index member from IRQ list (Jon Derrick)
   - Synchronize with RCU freeing MSI IRQ descs (Keith Busch)
   - Request userspace control of PCIe hotplug indicators (Keith Busch)
   - Move VMD driver to drivers/pci/host (Keith Busch)

  Marvell Aardvark host bridge driver:
   - Fix pci_remap_iospace() failure path (Lorenzo Pieralisi)
   - Remove redundant dev_err call in advk_pcie_probe() (Wei Yongjun)

  Microsoft Hyper-V host bridge driver:
   - Use zero-length array in struct pci_packet (Dexuan Cui)
   - Use pci_function_description[0] in struct definitions (Dexuan Cui)
   - Remove the unused 'wrk' in struct hv_pcibus_device (Dexuan Cui)
   - Handle vmbus_sendpacket() failure in hv_compose_msi_msg() (Dexuan Cui)
   - Handle hv_pci_generic_compl() error case (Dexuan Cui)
   - Use list_move_tail() instead of list_del() + list_add_tail() (Wei Yongjun)

  NVIDIA Tegra host bridge driver:
   - Fix pci_remap_iospace() failure path (Lorenzo Pieralisi)
   - Remove redundant _data suffix (Thierry Reding)
   - Use of_device_get_match_data() (Thierry Reding)

  Qualcomm host bridge driver:
   - Make explicitly non-modular (Paul Gortmaker)

  Renesas R-Car host bridge driver:
   - Consolidate register space lookup and ioremap (Bjorn Helgaas)
   - Don't disable/unprepare clocks on prepare/enable failure (Geert Uytterhoeven)
   - Add multi-MSI support (Grigory Kletsko)
   - Fix pci_remap_iospace() failure path (Lorenzo Pieralisi)
   - Fix some checkpatch warnings (Sergei Shtylyov)
   - Try increasing PCIe link speed to 5 GT/s at boot (Sergei Shtylyov)

  Rockchip host bridge driver:
   - Add DT bindings for Rockchip PCIe controller (Shawn Lin)
   - Add Rockchip PCIe controller support (Shawn Lin)
   - Improve the deassert sequence of four reset pins (Shawn Lin)
   - Fix wrong transmitted FTS count (Shawn Lin)
   - Increase the Max Credit update interval (Rajat Jain)

  Samsung Exynos host bridge driver:
   - Make explicitly non-modular (Paul Gortmaker)

  ST Microelectronics SPEAr13xx host bridge driver:
   - Make explicitly non-modular (Paul Gortmaker)

  Synopsys DesignWare host bridge driver:
   - Return data directly from dw_pcie_readl_rc() (Bjorn Helgaas)
   - Exchange viewport of `MEMORYs' and `CFGs/IOs' (Dong Bo)
   - Check LTSSM training bit before deciding link is up (Jisheng Zhang)
   - Move link wait definitions to .c file (Joao Pinto)
   - Wait for iATU enable (Joao Pinto)
   - Add iATU Unroll feature (Joao Pinto)
   - Fix pci_remap_iospace() failure path (Lorenzo Pieralisi)
   - Make explicitly non-modular (Paul Gortmaker)
   - Relax device number checking to allow SR-IOV (Po Liu)
   - Keep viewport fixed for IO transaction if num_viewport > 2 (Pratyush Anand)
   - Remove redundant platform_get_resource() return value check (Wei Yongjun)

  TI DRA7xx host bridge driver:
   - Make explicitly non-modular (Paul Gortmaker)

  TI Keystone host bridge driver:
   - Propagate request_irq() failure (Wei Yongjun)

  Xilinx AXI host bridge driver:
   - Keep both legacy and MSI interrupt domain references (Bharat Kumar Gogada)
   - Clear interrupt register for invalid interrupt (Bharat Kumar Gogada)
   - Clear correct MSI set bit (Bharat Kumar Gogada)
   - Dispose of MSI virtual IRQ (Bharat Kumar Gogada)
   - Make explicitly non-modular (Paul Gortmaker)
   - Relax device number checking to allow SR-IOV (Po Liu)

  Xilinx NWL host bridge driver:
   - Expand error logging (Bharat Kumar Gogada)
   - Enable all MSI interrupts using MSI mask (Bharat Kumar Gogada)
   - Make explicitly non-modular (Paul Gortmaker)

  Miscellaneous:
   - Drop CONFIG_KEXEC_CORE ifdeffery (Lukas Wunner)
   - portdrv: Make explicitly non-modular (Paul Gortmaker)
   - Make DPC explicitly non-modular (Paul Gortmaker)"

* tag 'pci-v4.9-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (105 commits)
  x86/PCI: VMD: Move VMD driver to drivers/pci/host
  PCI: rockchip: Fix wrong transmitted FTS count
  PCI: rockchip: Improve the deassert sequence of four reset pins
  PCI: rockchip: Increase the Max Credit update interval
  PCI: rcar: Try increasing PCIe link speed to 5 GT/s at boot
  PCI/AER: Fix aer_probe() kernel-doc comment
  PCI: Ignore requested alignment for VF BARs
  PCI: Ignore requested alignment for PROBE_ONLY and fixed resources
  PCI: Avoid unnecessary resume after direct-complete
  PCI: Recognize D3cold in pci_update_current_state()
  PCI: Query platform firmware for device power state
  PCI: Afford direct-complete to devices with non-standard PM
  PCI/AER: Cache capability position
  PCI/AER: Avoid memory allocation in interrupt handling path
  x86/PCI: VMD: Request userspace control of PCIe hotplug indicators
  PCI: pciehp: Allow exclusive userspace control of indicators
  ACPI / APEI: Send correct severity to calculate AER severity
  PCI/AER: Remove duplicate AER severity translation
  x86/PCI: VMD: Synchronize with RCU freeing MSI IRQ descs
  x86/PCI: VMD: Eliminate index member from IRQ list
  ...

1 files changed, 58 insertions, 17 deletions

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 415956c5c593..ba34907538f6 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -576,8 +576,9 @@ static const struct pci_platform_pm_ops *pci_platform_pm;
 
 int pci_set_platform_pm(const struct pci_platform_pm_ops *ops)
 {
-	if (!ops->is_manageable || !ops->set_state || !ops->choose_state ||
-	    !ops->sleep_wake || !ops->run_wake || !ops->need_resume)
+	if (!ops->is_manageable || !ops->set_state  || !ops->get_state ||
+	    !ops->choose_state  || !ops->sleep_wake || !ops->run_wake  ||
+	    !ops->need_resume)
 		return -EINVAL;
 	pci_platform_pm = ops;
 	return 0;
@@ -594,6 +595,11 @@ static inline int platform_pci_set_power_state(struct pci_dev *dev,
 	return pci_platform_pm ? pci_platform_pm->set_state(dev, t) : -ENOSYS;
 }
 
+static inline pci_power_t platform_pci_get_power_state(struct pci_dev *dev)
+{
+	return pci_platform_pm ? pci_platform_pm->get_state(dev) : PCI_UNKNOWN;
+}
+
 static inline pci_power_t platform_pci_choose_state(struct pci_dev *dev)
 {
 	return pci_platform_pm ?
@@ -725,26 +731,25 @@ static int pci_raw_set_power_state(struct pci_dev *dev, pci_power_t state)
 }
 
 /**
- * pci_update_current_state - Read PCI power state of given device from its
- *                            PCI PM registers and cache it
+ * pci_update_current_state - Read power state of given device and cache it
  * @dev: PCI device to handle.
  * @state: State to cache in case the device doesn't have the PM capability
+ *
+ * The power state is read from the PMCSR register, which however is
+ * inaccessible in D3cold.  The platform firmware is therefore queried first
+ * to detect accessibility of the register.  In case the platform firmware
+ * reports an incorrect state or the device isn't power manageable by the
+ * platform at all, we try to detect D3cold by testing accessibility of the
+ * vendor ID in config space.
  */
 void pci_update_current_state(struct pci_dev *dev, pci_power_t state)
 {
-	if (dev->pm_cap) {
+	if (platform_pci_get_power_state(dev) == PCI_D3cold ||
+	    !pci_device_is_present(dev)) {
+		dev->current_state = PCI_D3cold;
+	} else if (dev->pm_cap) {
 		u16 pmcsr;
 
-		/*
-		 * Configuration space is not accessible for device in
-		 * D3cold, so just keep or set D3cold for safety
-		 */
-		if (dev->current_state == PCI_D3cold)
-			return;
-		if (state == PCI_D3cold) {
-			dev->current_state = PCI_D3cold;
-			return;
-		}
 		pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
 		dev->current_state = (pmcsr & PCI_PM_CTRL_STATE_MASK);
 	} else {
@@ -1983,9 +1988,22 @@ static pci_power_t pci_target_state(struct pci_dev *dev)
 		default:
 			target_state = state;
 		}
-	} else if (!dev->pm_cap) {
+
+		return target_state;
+	}
+
+	if (!dev->pm_cap)
 		target_state = PCI_D0;
-	} else if (device_may_wakeup(&dev->dev)) {
+
+	/*
+	 * If the device is in D3cold even though it's not power-manageable by
+	 * the platform, it may have been powered down by non-standard means.
+	 * Best to let it slumber.
+	 */
+	if (dev->current_state == PCI_D3cold)
+		target_state = PCI_D3cold;
+
+	if (device_may_wakeup(&dev->dev)) {
 		/*
 		 * Find the deepest state from which the device can generate
 		 * wake-up events, make it the target state and enable device
@@ -4983,6 +5001,13 @@ static resource_size_t pci_specified_resource_alignment(struct pci_dev *dev)
 
 	spin_lock(&resource_alignment_lock);
 	p = resource_alignment_param;
+	if (!*p)
+		goto out;
+	if (pci_has_flag(PCI_PROBE_ONLY)) {
+		pr_info_once("PCI: Ignoring requested alignments (PCI_PROBE_ONLY)\n");
+		goto out;
+	}
+
 	while (*p) {
 		count = 0;
 		if (sscanf(p, "%d%n", &align_order, &count) == 1 &&
@@ -5047,6 +5072,7 @@ static resource_size_t pci_specified_resource_alignment(struct pci_dev *dev)
 		}
 		p++;
 	}
+out:
 	spin_unlock(&resource_alignment_lock);
 	return align;
 }
@@ -5065,6 +5091,15 @@ void pci_reassigndev_resource_alignment(struct pci_dev *dev)
 	resource_size_t align, size;
 	u16 command;
 
+	/*
+	 * VF BARs are read-only zero according to SR-IOV spec r1.1, sec
+	 * 3.4.1.11.  Their resources are allocated from the space
+	 * described by the VF BARx register in the PF's SR-IOV capability.
+	 * We can't influence their alignment here.
+	 */
+	if (dev->is_virtfn)
+		return;
+
 	/* check if specified PCI is target device to reassign */
 	align = pci_specified_resource_alignment(dev);
 	if (!align)
@@ -5087,6 +5122,12 @@ void pci_reassigndev_resource_alignment(struct pci_dev *dev)
 		r = &dev->resource[i];
 		if (!(r->flags & IORESOURCE_MEM))
 			continue;
+		if (r->flags & IORESOURCE_PCI_FIXED) {
+			dev_info(&dev->dev, "Ignoring requested alignment for BAR%d: %pR\n",
+				i, r);
+			continue;
+		}
+
 		size = resource_size(r);
 		if (size < align) {
 			size = align;