linux-stable/drivers/pci
Rafael J. Wysocki 95c80bc695 PCI/PME: Fix hotplug/sysfs remove deadlock in pcie_pme_remove()
Dongdong reported a deadlock triggered by a hotplug event during a sysfs
"remove" operation:

  pciehp 0000:00:0c.0:pcie004: Slot(0-1): Link Up
  # echo 1 > 0000:00:0c.0/remove

  PME and hotplug share an MSI/MSI-X vector.  The sysfs "remove" side is:

    remove_store
       pci_stop_and_remove_bus_device_locked
	 pci_lock_rescan_remove
	 pci_stop_and_remove_bus_device
	   ...
	   pcie_pme_remove
	     pcie_pme_suspend
	       synchronize_irq        # wait for hotplug IRQ handler
	 pci_unlock_rescan_remove

  The hotplug side is:

    pciehp_ist
       pciehp_handle_presence_or_link_change
	 pciehp_configure_device
	   pci_lock_rescan_remove     # wait for pci_unlock_rescan_remove()

  INFO: task bash:10913 blocked for more than 120 seconds.

  # ps -ax |grep D
   PID TTY      STAT   TIME COMMAND
  10913 ttyAMA0  Ds+    0:00 -bash
  14022 ?        D      0:00 [irq/745-pciehp]

  # cat /proc/14022/stack
  __switch_to+0x94/0xd8
  pci_lock_rescan_remove+0x20/0x28
  pciehp_configure_device+0x30/0x140
  pciehp_handle_presence_or_link_change+0x324/0x458
  pciehp_ist+0x1dc/0x1e0

  # cat /proc/10913/stack
  __switch_to+0x94/0xd8
  synchronize_irq+0x8c/0xc0
  pcie_pme_suspend+0xa4/0x118
  pcie_pme_remove+0x20/0x40
  pcie_port_remove_service+0x3c/0x58
  ...
  pcie_port_device_remove+0x2c/0x48
  pcie_portdrv_remove+0x68/0x78
  pci_device_remove+0x48/0x120
  ...
  pci_stop_bus_device+0x84/0xc0
  pci_stop_and_remove_bus_device_locked+0x24/0x40
  remove_store+0xa4/0xb8
  dev_attr_store+0x44/0x60
  sysfs_kf_write+0x58/0x80

It is incorrect to call pcie_pme_suspend() from pcie_pme_remove() for two
reasons.

First, pcie_pme_suspend() calls synchronize_irq(), which will wait for the
native hotplug interrupt handler as well as for the PME one, because they
share one IRQ (as per the spec).  That may deadlock if hotplug is signaled
while pcie_pme_remove() is running and the latter calls
pci_lock_rescan_remove() before the former.

Second, if pcie_pme_suspend() figures out that wakeup needs to be enabled
for the port, it will return without disabling the interrupt as expected by
pcie_pme_remove() which was overlooked by commit c7b5a4e6e8 ("PCI / PM:
Fix native PME handling during system suspend/resume").

To fix that, rework pcie_pme_remove() to disable the PME interrupt, clear
its status and prevent the PME worker function from re-enabling it before
calling free_irq() on it, which should be sufficient.

Fixes: c7b5a4e6e8 ("PCI / PM: Fix native PME handling during system suspend/resume")
Link: https://lore.kernel.org/linux-pci/c7697e7c-e1af-13e4-8491-0a3996e6ab5d@huawei.com
Reported-by: Dongdong Liu <liudongdong3@huawei.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
[bhelgaas: add URL and deadlock details from Dongdong]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
2019-02-28 13:56:27 -06:00
..
controller pci-v4.21-changes 2019-01-05 17:57:34 -08:00
endpoint PCI: consolidate PCI config entry in drivers/pci 2018-11-23 11:45:34 +09:00
hotplug PCI: pciehp: Disable Data Link Layer State Changed event on suspend 2019-02-15 14:18:00 -06:00
pcie PCI/PME: Fix hotplug/sysfs remove deadlock in pcie_pme_remove() 2019-02-28 13:56:27 -06:00
switch switchtec: Add MRPC DMA mode support 2018-12-13 09:03:32 -06:00
access.c PCI: Uninline PCI bus accessors for better ftracing 2018-10-04 16:37:37 -05:00
ats.c PCI: Enable PASID only if entire path supports End-End TLP prefixes 2018-06-30 14:10:03 -05:00
bus.c PCI: Fix is_added/is_busmaster race condition 2018-07-31 11:27:54 -05:00
ecam.c PCI: Add SPDX GPL-2.0 to replace GPL v2 boilerplate 2018-01-28 15:48:29 -06:00
host-bridge.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
iov.c PCI/IOV: Add flag so platforms can skip VF scanning 2019-01-01 19:04:37 -06:00
irq.c PCI: Use IRQF_ONESHOT if pci_request_irq() called with no handler 2018-07-31 10:43:43 -05:00
Kconfig pci-v4.21-changes 2019-01-05 17:57:34 -08:00
Makefile PCI/ACPI: Allow ACPI to be built without CONFIG_PCI set 2018-12-20 10:19:49 +01:00
mmap.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
msi.c genirq/core: Introduce struct irq_affinity_desc 2018-12-19 11:32:08 +01:00
of.c iommu/of: make of_pci_map_rid() available for other devices too 2018-09-25 09:47:52 +02:00
p2pdma.c pci-v4.21-changes 2019-01-05 17:57:34 -08:00
pci-acpi.c PCI / ACPI: Identify untrusted PCI devices 2018-12-05 12:01:55 +03:00
pci-bridge-emul.c PCI: Introduce PCI bridge emulated config space common logic 2018-10-18 17:50:19 +01:00
pci-bridge-emul.h PCI: Introduce PCI bridge emulated config space common logic 2018-10-18 17:50:19 +01:00
pci-driver.c pci-v4.21-changes 2019-01-05 17:57:34 -08:00
pci-label.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
pci-mid.c x86/cpu: Sanitize FAM6_ATOM naming 2018-10-02 10:14:32 +02:00
pci-pf-stub.c PCI/IOV: Add pci-pf-stub driver for PFs that only enable VFs 2018-04-24 16:47:16 -05:00
pci-stub.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
pci-sysfs.c Merge branch 'pci/hotplug' 2018-08-15 14:58:52 -05:00
pci.c PCI: Blacklist power management of Gigabyte X299 DESIGNARE EX PCIe ports 2019-02-11 08:44:33 -06:00
pci.h PCI: Add missing include to drivers/pci.h 2018-12-06 14:55:41 -06:00
probe.c PCI / ACPI: Identify untrusted PCI devices 2018-12-05 12:01:55 +03:00
proc.c Remove 'type' argument from access_ok() function 2019-01-03 18:57:57 -08:00
quirks.c PCI: Override Synopsys USB 3.x HAPS device class 2018-12-17 16:25:16 -06:00
remove.c PCI/ASPM: Fix link_state teardown on device removal 2018-09-17 16:32:23 -05:00
rom.c PCI: Make pci_get_rom_size() static 2018-06-29 21:17:26 -05:00
search.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
setup-bus.c PCI: Equalize hotplug memory and io for occupied and empty slots 2018-10-02 16:04:40 -05:00
setup-irq.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
setup-res.c PCI: Remove messages about reassigning resources 2018-04-11 08:46:50 -05:00
slot.c PCI/ERR: Use slot reset if available 2018-09-21 12:18:10 -05:00
syscall.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00
vc.c Merge branch 'pci/spdx' into next 2018-02-01 11:40:07 -06:00
vpd.c PCI/VPD: Check for VPD access completion before checking for timeout 2018-08-14 16:04:46 -05:00
xen-pcifront.c PCI: Tidy comments 2018-03-19 14:20:43 -05:00