It's not very enlightening to see

  pci 0000:07:00.0: [Firmware Bug]: VPD access disabled

in the dmesg log because there's no clue about what the firmware bug is.
Expand the message to explain why we're disabling VPD.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Remove the "PCI Express Hot Plug Controller Driver" version message.  I
don't think it contains any useful information.  Remove unused #defines
and move the author information to a comment.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Remove the "PCI Hot Plug PCI Core" version message.  I don't think it
contains any useful information.  Remove unused #defines and move the
author information to a comment.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Remove the "service driver %s loaded" and unloaded messages.  All service
drivers already log something in their probe functions, where they can log
more useful details.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Add a log message when we enable AER on a Root Port and the hierarchy below
it.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

All other AER-related log messages use the PCI device, e.g.,
"pci 0000:00:1c.0", not the PCIe service device, e.g.,
"aer 0000:00:1c.0:pcie02".

Change the probe error messages to match the rest and include a little
context.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Remove the unused DRIVER_VERSION, DRIVER_AUTHOR, and DRIVER_DESC macros.
The author information is already included in a comment above.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

We already log a "Signaling PME" whenever the PME service driver claims a
Root Port.  In fact, we also log the same message for every device in the
hierarchy below the Root Port.

Log the "Signaling PME" once (only for the Root Port, since we can
trivially find out which devices are below the Root Port), and include the
IRQ number in the message to help connect the dots with /proc/interrupts.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Since we register pcie_pme_driver only for PCI_EXP_TYPE_ROOT_PORT, the PME
driver never claims Root Complex Event Collectors.

Remove unused code related to Root Complex Event Collectors.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Move PCI configuration space size macros (PCI_CFG_SPACE_SIZE and
PCI_CFG_SPACE_EXP_SIZE) from drivers/pci/pci.h to
include/uapi/linux/pci_regs.h so they can be used by more drivers and
eliminate duplicate definitions.

[bhelgaas: Expand comment to include PCI-X details]
Signed-off-by: NWang Sheng-Hui <shhuiw@foxmail.com>
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Body of an "if" statement wasn't indented.  Add a tab.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Reorder struct hisi_pcie to put generic fields first.  No functional change
intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Only interfaces used from outside the driver, e.g., those called by the
DesignWare core, need to accept pointers to the generic struct pcie_port.
Internal interfaces can accept pointers to the device-specific struct,
which makes them more straightforward.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Include the PCIE_HIP06_CTRL_OFF block base in the PCIE_SYS_STATE4 register
address so reads of PCIE_SYS_STATE4 don't have to mention both.  No
functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Reorder struct dra7xx_pcie to put generic fields first.  No functional
change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The xilinx-nwl driver never uses the platform drvdata pointer, so don't
bother setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Use a local "struct device *dev" for brevity and consistency with other
drivers.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

xilinx_pcie_assign_msi() doesn't use the struct xilinx_pcie_port pointer
passed to it, so remove the argument completely.  No functional change
intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The xilinx driver never uses the platform drvdata pointer, so don't
bother setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Use a local "struct device *dev" for brevity and consistency with other
drivers.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Add device-specific register accessors for consistency across host drivers.
No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Pass the struct xgene_pcie_port pointer, not addresses, to setup functions.
This enables future simplifications.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The xgene driver never uses the platform drvdata pointer, so don't
bother setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The tegra driver never uses the platform drvdata pointer, so don't
bother setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Use a local "struct device *dev" for brevity and consistency with other
drivers.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The tegra_pcie_phy_disable() path called pads_writel() with arguments in
the wrong order.  Swap them to be the "value, offset" order expected by
pads_writel().

Fixes: 6fe7c187 ("PCI: tegra: Support per-lane PHYs")
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NThierry Reding <treding@nvidia.com>
CC: stable@vger.kernel.org	# v4.7+

The rockchip driver never uses the platform drvdata pointer, so don't
bother setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NShawn Lin <shawn.lin@rock-chips.com>

Use a local "struct device *dev" for brevity and consistency with other
drivers.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NSimon Horman <horms+renesas@verge.net.au>

The DRV_NAME macro is only used once, so there's no real advantage to
having the macro at all.  Remove it and use the "rcar-pcie" name directly
in the struct platform_driver.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NSimon Horman <horms+renesas@verge.net.au>

rcar_pcie_get_resources() doesn't use the platform_device pointer passed to
it, so remove it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NSimon Horman <horms+renesas@verge.net.au>

The rcar driver never uses the platform drvdata pointer, so don't bother
setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NSimon Horman <horms+renesas@verge.net.au>

Reorder struct qcom_pcie to put generic fields first.  No functional change
intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Remove the struct qcom_pcie.dev member, which is a duplicate of the generic
pp.dev member.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Remove the struct qcom_pcie.dbi member, which is a duplicate of the generic
pp.dbi_base member.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

The qcom driver never uses the platform drvdata pointer, so don't bother
setting it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Use the existing "np" pointer instead of looking up dev->of_node again.  No
functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Use a local "struct device *dev" for brevity and consistency with other
drivers.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Reorder struct ls_pcie to put generic fields first.  No functional change
intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

ls_add_pcie_port() doesn't use the platform_device pointer passed to it, so
remove it.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

Do the basic pcie_port setup in the probe function for consistency with
other drivers.  No functional change intended.
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>