- Always set DesignWare "TLP Digest" bit so generic code can enable ECRC
  via the AER Capability (Vidya Sagar)

- Drop support for config space in DT 'ranges' (Rob Herring)

- Increase width of outbound iATU size to u64 (Shradha Todi)

- Add upper limit address for outbound iATU (Shradha Todi)

- Allow dwc-based drivers that don't override any default ops (Jisheng
  Zhang)

- Drop unnecessary dw_pcie_ops from the al driver (Jisheng Zhang)

* pci/dwc:
  PCI: al: Remove useless dw_pcie_ops
  PCI: dwc: Don't assume the ops in dw_pcie always exist
  PCI: dwc: Add upper limit address for outbound iATU
  PCI: dwc: Change size to u64 for EP outbound iATU
  PCI: dwc: Drop support for config space in 'ranges'
  PCI: dwc: Work around ECRC configuration issue

- Retrain Link to work around Gen2 training defect (Nadeem Athani)

* remotes/lorenzo/pci/cadence:
  PCI: cadence: Retrain Link to work around Gen2 training defect

- Add support for BCM4908 with external PERST# signal controller (Rafał
  Miłecki)

* remotes/lorenzo/pci/brcmstb:
  PCI: brcmstb: support BCM4908 with external PERST# signal controller
  dt-bindings: PCI: brcmstb: add BCM4908 binding

- Align checking of syscall user config accessor return codes (Heiner
  Kallweit)

- Fix "ordering" comment typos (Bjorn Helgaas)

- Fix 'ARM/TEXAS INSTRUMENT KEYSTONE CLOCKSOURCE' capitalization in
  MAINTAINERS (Bjorn Helgaas)

- Add Silicom Denmark vendor ID (Martin Hundebøll)

- Apply CONFIG_PCI_DEBUG to entire drivers/pci hierarchy (Junhao He)

- Remove WARN_ON(in_interrupt()) (Sebastian Andrzej Siewior)

* pci/misc:
  PCI: Remove WARN_ON(in_interrupt())
  PCI: Apply CONFIG_PCI_DEBUG to entire drivers/pci hierarchy
  PCI: Add Silicom Denmark vendor ID
  MAINTAINERS: Fix 'ARM/TEXAS INSTRUMENT KEYSTONE CLOCKSOURCE' capitalization
  Fix "ordering" comment typos
  PCI: Align checking of syscall user config accessors

- Fix merge botch in cdns_pcie_host_map_dma_ranges() (Krzysztof Wilczyński)

* pci/host-probe-refactor:
  PCI: cadence: Fix DMA range mapping early return error

- Decline requests to resize BARs if platform requires us to preserve
  resource assignments (Ard Biesheuvel)

* pci/resource:
  PCI: Decline to resize resources if boot config must be preserved

- Remove bandwidth notification for now to avoid reporting spam (Bjorn
  Helgaas)

* pci/link:
  PCI/LINK: Remove bandwidth notification

- Remove unused acpiphp_callback typedef (Chen Lin)

* pci/hotplug:
  PCI: acpiphp: Remove unused acpiphp_callback typedef

- Clear AER status of the reporting device (Keith Busch)

- Clear AER status from Root Port when resetting Downstream Port (Keith
  Busch)

- Retain status from error notification (Keith Busch)

- Log the type of Port that was reset for error handling (Keith Busch)

- Report reset for frozen channel (Keith Busch)

* pci/error:
  PCI/portdrv: Report reset for frozen channel
  PCI/AER: Specify the type of Port that was reset
  PCI/ERR: Retain status from error notification
  PCI/AER: Clear AER status from Root Port when resetting Downstream Port
  PCI/ERR: Clear status of the reporting device

- Unexport acpi_pci_osc_control_set() (Bjorn Helgaas)

- Remove unnecessary osc_lock mutex (Bjorn Helgaas)

- Clarify _OSC failure message (Bjorn Helgaas)

- Fix pci-bridge-emul array overruns and improve safety (Russell King)

- Fix pci_register_io_range() memory leak (Geert Uytterhoeven)

* pci/enumeration:
  PCI: Fix pci_register_io_range() memory leak
  PCI: pci-bridge-emul: Fix array overruns, improve safety
  PCI/ACPI: Clarify message about _OSC failure
  PCI/ACPI: Remove unnecessary osc_lock
  PCI/ACPI: Make acpi_pci_osc_control_set() static

We have removed the assumption that dw_pcie_ops always exists in the dwc
core driver, so we can remove the useless dw_pcie_ops now.

Link: https://lore.kernel.org/r/20210128144324.2fa8577c@xhacker.debianSigned-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Jonathan Chocron <jonnyc@amazon.com>

Some dwc-based device drivers, especially host-only drivers, may work well
with the default read_dbi/write_dbi/link_up implementations in
pcie-designware.c, so remove the assumption that every driver implements
them to simplify those drivers.

Link: https://lore.kernel.org/r/20210128144258.10329aa4@xhacker.debianSigned-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

The size parameter is unsigned long type which can accept size > 4GB. In
that case, the upper limit address must be programmed. Add support to
program the upper limit address and set INCREASE_REGION_SIZE in case size >
4GB.

Link: https://lore.kernel.org/r/1612250918-19610-1-git-send-email-shradha.t@samsung.comSigned-off-by: Shradha Todi <shradha.t@samsung.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Pankaj Dubey <pankaj.dubey@samsung.com>
Reviewed-by: Rob Herring <robh@kernel.org>

Since outbound iATU permits size to be greater than 4GB for which the
support is also available, allow EP function to send u64 size instead of
truncating to u32.

Link: https://lore.kernel.org/r/1609929900-19082-1-git-send-email-shradha.t@samsung.comSigned-off-by: Shradha Todi <shradha.t@samsung.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Pankaj Dubey <pankaj.dubey@samsung.com>

Since commit a0fd361d ("PCI: dwc: Move "dbi", "dbi2", and
"addr_space" resource setup into common code"), the code
setting dbi_base when the config space is defined in 'ranges' property
instead of 'reg' is dead code as dbi_base is never NULL.

Rather than fix this, let's just drop the code. Using ranges has been
deprecated since 2014. The only platforms using this were exynos5440,
i.MX6 and Spear13xx. Exynos5440 is dead and has been removed. i.MX6 and
Spear13xx had PCIe support added just before this was deprecated and
were fixed within a kernel release or 2.

Link: https://lore.kernel.org/r/20201215194149.86831-1-robh@kernel.orgReported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

DesignWare core has a TLP digest (TD) override bit in one of the control
registers of ATU. This bit also needs to be programmed for proper ECRC
functionality. This is currently identified as an issue with DesignWare
IP version 4.90a.

[bhelgaas: fix typos/grammar errors]
Link: https://lore.kernel.org/r/20201230165723.673-1-vidyas@nvidia.comSigned-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>

The PCI error recovery always resets the link for a frozen state, so the
port driver should return that a reset is required for its result. This
will get the .slot_reset() callback invoked, which is necessary to
restore the port's config space. Without this, the driver had been
relying on downstream drivers to return this status.

Link: https://lore.kernel.org/r/20210104230300.1277180-6-kbusch@kernel.orgTested-by: Hedi Berriche <hedi.berriche@hpe.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Hedi Berriche <hedi.berriche@hpe.com>

The AER driver may be called upon to reset either a Downstream or a Root
Port. Check which type it is to properly identify it when logging that
the reset occurred.

Link: https://lore.kernel.org/r/20210104230300.1277180-5-kbusch@kernel.orgTested-by: Hedi Berriche <hedi.berriche@hpe.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Hedi Berriche <hedi.berriche@hpe.com>

Overwriting the frozen detected status with the result of the link reset
loses the NEED_RESET result that drivers are depending on for error
handling to report the .slot_reset() callback. Retain this status so
that subsequent error handling has the correct flow.

Link: https://lore.kernel.org/r/20210104230300.1277180-4-kbusch@kernel.orgReported-by: Hinko Kocevar <hinko.kocevar@ess.eu>
Tested-by: Hedi Berriche <hedi.berriche@hpe.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Sean V Kelley <sean.v.kelley@intel.com>
Acked-by: Hedi Berriche <hedi.berriche@hpe.com>

The pci_dev parameter given to aer_root_reset() may be a Downstream Port
rather than the Root Port. Get the Root Port from the provided device in
order to clear the root's AER status.

Link: https://lore.kernel.org/r/20210104230300.1277180-3-kbusch@kernel.orgTested-by: Hedi Berriche <hedi.berriche@hpe.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Sean V Kelley <sean.v.kelley@intel.com>
Acked-by: Hedi Berriche <hedi.berriche@hpe.com>

Error handling operates on the first Downstream Port above the detected
error, but the error may have been reported by a downstream device.
Clear the AER status of the device that reported the error rather than
the first Downstream Port.

Link: https://lore.kernel.org/r/20210104230300.1277180-2-kbusch@kernel.orgTested-by: Hedi Berriche <hedi.berriche@hpe.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Sean V Kelley <sean.v.kelley@intel.com>
Acked-by: Hedi Berriche <hedi.berriche@hpe.com>

Remove the 'acpiphp_callback' typedef as it is not used.

Link: https://lore.kernel.org/r/1613443120-4279-1-git-send-email-chen45464546@163.comSigned-off-by: Chen Lin <chen.lin5@zte.com.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Krzysztof Wilczyński <kw@linux.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Function cdns_pcie_host_map_dma_ranges() iterates over a PCIe host bridge
DMA ranges using the resource_list_for_each_entry() iterator, returning an
error if cdns_pcie_host_bar_config() fails.

49e427e6 ("Merge branch 'pci/host-probe-refactor'") botched a merge so
it *always* returned after the first DMA range, even if no error occurred.

Fix the error checking so we return early only when an error occurs.

[bhelgaas: commit log]
Fixes: 49e427e6 ("Merge branch 'pci/host-probe-refactor'")
Link: https://lore.kernel.org/r/20210216205935.3112661-1-kw@linux.comSigned-off-by: Krzysztof Wilczyński <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

Kmemleak reports:

  unreferenced object 0xc328de40 (size 64):
    comm "kworker/1:1", pid 21, jiffies 4294938212 (age 1484.670s)
    hex dump (first 32 bytes):
      00 00 00 00 00 00 00 00 e0 d8 fc eb 00 00 00 00  ................
      00 00 10 fe 00 00 00 00 00 00 00 00 00 00 00 00  ................

  backtrace:
    [<ad758d10>] pci_register_io_range+0x3c/0x80
    [<2c7f139e>] of_pci_range_to_resource+0x48/0xc0
    [<f079ecc8>] devm_of_pci_get_host_bridge_resources.constprop.0+0x2ac/0x3ac
    [<e999753b>] devm_of_pci_bridge_init+0x60/0x1b8
    [<a895b229>] devm_pci_alloc_host_bridge+0x54/0x64
    [<e451ddb0>] rcar_pcie_probe+0x2c/0x644

In case a PCI host driver's probe is deferred, the same I/O range may be
allocated again, and be ignored, causing a memory leak.

Fix this by (a) letting logic_pio_register_range() return -EEXIST if the
passed range already exists, so pci_register_io_range() will free it, and
by (b) making pci_register_io_range() not consider -EEXIST an error
condition.

Link: https://lore.kernel.org/r/20210202100332.829047-1-geert+renesas@glider.beSigned-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

We allow up to PCI_EXP_SLTSTA2 registers to be accessed, but the
pcie_cap_regs_behavior[] array only covers up to PCI_EXP_RTSTA.  Expand
this array to avoid walking off the end of it.

Do the same for pci_regs_behavior for consistency[], and add a
BUILD_BUG_ON() to also check the bridge->conf structure size.

Fixes: 23a5fba4 ("PCI: Introduce PCI bridge emulated config space common logic")
Link: https://lore.kernel.org/r/E1l6z9W-0006Re-MQ@rmk-PC.armlinux.org.ukSigned-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Pali Rohár <pali@kernel.org>

WARN_ON(in_interrupt()) is used for historic reasons to ensure proper usage
of down_read() and predates might_sleep() and lockdep.

down_read() has might_sleep() which also catches users from preemption
disabled regions while in_interrupt() does not.

Remove WARN_ON(in_interrupt()) because there are now better debugging
facilities.

Link: https://lore.kernel.org/r/20210208194400.384003-1-bigeasy@linutronix.deSigned-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

Cadence controller will not initiate autonomous speed change if strapped
as Gen2. The Retrain Link bit is set as quirk to enable this speed change.

Link: https://lore.kernel.org/r/20210209144622.26683-3-nadeem@cadence.comSigned-off-by: Nadeem Athani <nadeem@cadence.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>

CONFIG_PCI_DEBUG=y adds -DDEBUG to CFLAGS, which enables things like
pr_debug() and dev_dbg() (and hence pci_dbg()).  Previously we added
-DDEBUG for files in drivers/pci/, but not files in subdirectories of
drivers/pci/.

Add -DDEBUG to CFLAGS for all files below drivers/pci/ so CONFIG_PCI_DEBUG
applies to the entire hierarchy.

[bhelgaas: commit log]
Link: https://lore.kernel.org/r/1612438215-33105-1-git-send-email-yangyicong@hisilicon.comSigned-off-by: Junhao He <hejunhao2@hisilicon.com>
Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Krzysztof Wilczyński <kw@linux.com>

Update pci_ids.h with the vendor ID for Silicom Denmark. The define is
going to be referenced in driver(s) for FPGA accelerated smart NICs.

Link: https://lore.kernel.org/r/20210208150158.2877414-1-mhu@silicom.dkSigned-off-by: Martin Hundebøll <mhu@silicom.dk>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Reviewed-by: Krzysztof Wilczyński <kw@linux.com>

The PCIe Bandwidth Change Notification feature logs messages when the link
bandwidth changes.  Some users have reported that these messages occur
often enough to significantly reduce NVMe performance.  GPUs also seem to
generate these messages.

We don't know why the link bandwidth changes, but in the reported cases
there's no indication that it's caused by hardware failures.

Remove the bandwidth change notifications for now.  Hopefully we can add
this back when we have a better understanding of why this happens and how
we can make the messages useful instead of overwhelming.

Link: https://lore.kernel.org/r/20200115221008.GA191037@google.com/
Link: https://lore.kernel.org/r/155605909349.3575.13433421148215616375.stgit@gimli.home/
Link: https://bugzilla.kernel.org/show_bug.cgi?id=206197Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

Fix capitalization typo in 'ARM/TEXAS INSTRUMENT KEYSTONE CLOCKSOURCE'.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

Fix comment typos in "ordering".
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Kalle Valo <kvalo@codeaurora.org>
Acked-by: Vasily Gorbik <gor@linux.ibm.com>	# s390

After 34e32072 ("PCI: handle positive error codes"),
pci_user_read_config_*() and pci_user_write_config_*() return 0 or negative
errno values, not PCIBIOS_* values like PCIBIOS_SUCCESSFUL or
PCIBIOS_BAD_REGISTER_NUMBER.

Remove comparisons with PCIBIOS_SUCCESSFUL and check only for non-zero.  It
happens that PCIBIOS_SUCCESSFUL is zero, so this is not a functional
change, but it aligns this code with the user accessors.

[bhelgaas: commit log]
Fixes: 34e32072 ("PCI: handle positive error codes")
Link: https://lore.kernel.org/r/f1220314-e518-1e18-bf94-8e6f8c703758@gmail.comSigned-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>

This message:

  acpi PNP0A08:02: _OSC failed (AE_NOT_FOUND); disabling ASPM

is alarming and slightly misleading.  Per the PCI Firmware Spec, r3.2, sec
4.5.1, _OSC is required for PCIe hierarchies.  If _OSC is absent or fails,
the OS must not attempt to use any of the features defined for _OSC.  That
includes native hotplug, native PME, AER, and other things as well as ASPM.

Rephrase the message to:

  acpi PNP0A08:02: _OSC: platform retains control of PCIe features (AE_NOT_FOUND)

Previous discussion at https://lore.kernel.org/r/20200602223618.GA845676@bjorn-Precision-5520/Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Sinan Kaya <okaya@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Yicong Yang <yangyicong@hisilicon.com>

9778c14b ("ACPI/PCI: Fix possible race condition on _OSC evaluation")
added locking around _OSC calls to protect the acpi_osc_data_list that
stored the results.

63f10f0f ("PCI/ACPI: move _OSC code to pci_root.c") moved the results
from acpi_osc_data_list to the struct acpi_pci_root, where it no longer
needs locking, but did not remove the lock.

Remove the unnecessary locking around _OSC calls.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

acpi_pci_osc_control_set() is only called from pci_root.c, so stop
exporting it and make it static.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

BCM4908 uses external MISC block for controlling PERST# signal. Use it
as a reset controller.

Link: https://lore.kernel.org/r/20201210180421.7230-3-zajec5@gmail.comSigned-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>

BCM4908 is a SoC family with PCIe controller sharing design with the one
for STB. BCM4908 has different power management and memory controller so
few tweaks are required.

PERST# signal on BCM4908 is handled by an external MISC block so it
needs specifying a reset phandle.

Link: https://lore.kernel.org/r/20201210180421.7230-2-zajec5@gmail.comSigned-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>

The _DSM #5 method in the ACPI host bridge object tells us whether the OS
must preserve the resource assignments done by firmware. If this is the
case, we should not permit drivers to resize BARs on the fly. Make
pci_resize_resource() take this into account.

Link: https://lore.kernel.org/r/20210109095353.13417-1-ardb@kernel.orgSigned-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org	# v5.4+