b84106b4e2 ("PCI: Disable IO/MEM decoding for devices with non-compliant
BARs") disabled BAR sizing for BARs 0-5 of devices that don't comply with
the PCI spec. But it didn't do anything for expansion ROM BARs, so we
still try to size them, resulting in warnings like this on Broadwell-EP:
pci 0000:ff:12.0: BAR 6: failed to assign [mem size 0x00000001 pref]
Move the non-compliant BAR check from __pci_read_base() up to
pci_read_bases() so it applies to the expansion ROM BAR as well as
to BARs 0-5.
Note that direct callers of __pci_read_base(), like sriov_init(), will now
bypass this check. We haven't had reports of devices with broken SR-IOV
BARs yet.
[bhelgaas: changelog]
Fixes: b84106b4e2 ("PCI: Disable IO/MEM decoding for devices with non-compliant BARs")
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: stable@vger.kernel.org
CC: Thomas Gleixner <tglx@linutronix.de>
CC: Ingo Molnar <mingo@redhat.com>
CC: "H. Peter Anvin" <hpa@zytor.com>
CC: Andi Kleen <ak@linux.intel.com>
Resource flags are exposed to userspace via the sysfs "resource" file.
lspci reads the sysfs file to determine resource properties.
Add a "BAR Equivalent Indicator" flag so lspci can distinguish between
[virtual] and [enhanced] resources.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Sean O. Stalley <sean.stalley@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
No functional changes in this patch.
PCI I/O space mapping code does not depend on OF; therefore it can be moved
to PCI core code. This way we will be able to use it, e.g., in ACPI PCI
code.
Suggested-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Tomasz Nowicki <tn@semihalf.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: Arnd Bergmann <arnd@arndb.de>
CC: Liviu Dudau <Liviu.Dudau@arm.com>
Use functions provided by drivers/pci/ecam.h for mapping the config space
in drivers/pci/host/pci-host-common.c, and update its users to use 'struct
pci_config_window' and 'struct pci_ecam_ops'.
The changes are mostly to use 'struct pci_config_window' in place of
'struct gen_pci'. Some of the fields of gen_pci were only used temporarily
and can be eliminated by using local variables or function arguments, these
are not carried over to struct pci_config_window.
pci-thunder-ecam.c and pci-thunder-pem.c are the only users of the
pci_host_common_probe function and the gen_pci structure; these have been
updated to use the new API as well.
The patch does not introduce any functional changes other than a very minor
one: with the new code, on 64-bit platforms, we do just a single ioremap
for the whole config space.
Signed-off-by: Jayachandran C <jchandra@broadcom.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Add config option PCI_ECAM and file drivers/pci/ecam.c to provide generic
functions for accessing memory-mapped PCI config space.
The API is defined in drivers/pci/ecam.h and is written to replace the API
in drivers/pci/host/pci-host-common.h. The file defines a new 'struct
pci_config_window' to hold the information related to a PCI config area and
its mapping. This structure is expected to be used as sysdata for
controllers that have ECAM based mapping.
Helper functions are provided to setup the mapping, free the mapping and to
implement the map_bus method in 'struct pci_ops'
Signed-off-by: Jayachandran C <jchandra@broadcom.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
This renames pcibios_find_pci_bus() to pci_find_bus_by_node() to
avoid conflicts with those PCI subsystem weak function names, which
have prefix "pcibios". No logical changes introduced.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This renames pcibios_{add,remove}_pci_devices() to avoid conflicts
with names of the weak functions in PCI subsystem, which have the
prefix "pcibios". No logical changes introduced.
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Reviewed-By: Alistair Popple <alistair@popple.id.au>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
I'm trying to pass-through Broadcom BCM5720 NIC (Dell device 1f5b) on a
Dell R720 server. Everything works fine when the target VM has only one
CPU, but SMP guests reboot when the NIC driver accesses PCI config space
with hv_pcifront_read_config()/hv_pcifront_write_config(). The reboot
appears to be induced by the hypervisor and no crash is observed. Windows
event logs are not helpful at all ('Virtual machine ... has quit
unexpectedly'). The particular access point is always different and
putting debug between them (printk/mdelay/...) moves the issue further
away. The server model affects the issue as well: on Dell R420 I'm able to
pass-through BCM5720 NIC to SMP guests without issues.
While I'm obviously failing to reveal the essence of the issue I was able
to come up with a (possible) solution: if explicit barriers are added to
hv_pcifront_read_config()/hv_pcifront_write_config() the issue goes away.
The essential minimum is rmb() at the end on _hv_pcifront_read_config() and
wmb() at the end of _hv_pcifront_write_config() but I'm not confident it
will be sufficient for all hardware. I suggest the following barriers:
1) wmb()/mb() between choosing the function and writing to its space.
2) mb() before releasing the spinlock in both _hv_pcifront_read_config()/
_hv_pcifront_write_config() to ensure that consecutive reads/writes to
the space won't get re-ordered as drivers may count on that.
Config space access is not supposed to be performance-critical so these
explicit barriers should not cause any slowdown.
[bhelgaas: use Linux "barriers" terminology]
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Jake Oshins <jakeo@microsoft.com>
We cache the PCI_EXP_SLTCAP_HPC bit in pci_dev->is_hotplug_bridge on device
probe, so there's no need to read it again on allocation of port service
devices.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Add driver for the PCI Express Downstream Port Containment extended
capability. DPC is an optional capability to contain uncorrectable errors
below a port.
For more information on DPC, please see PCI Express Base Specification
Revision 4, section 7.31, or view the PCI-SIG DPC ECN here:
https://pcisig.com/sites/default/files/specification_documents/ECN_DPC_2012-02-09_finalized.pdf
When a DPC event is triggered, the hardware disables downstream links, so
the DPC driver schedules removal for all devices below this port. This may
happen concurrently with a PCIe hotplug driver if enabled. When all
downstream devices are removed and the link state transitions to disabled,
the DPC driver clears the DPC status and interrupt bits so the link may
retrain for a newly connected device.
[bhelgaas: clear (not set) DPC_CTL bits on remove, whitespace cleanup]
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Lukas Wunner <lukas@wunner.de>
Add the Downstream Port Containment (PCIE_PORT_SERVICE_DPC) portdrv service
type, available if the device has the DPC extended capability.
[bhelgaas: split to separate patch, changelog]
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The names of port service devices previously used one nibble to encode the
port type and another nibble to encode the service type. We're about to
add a fifth service type, so change device names to use one *byte* to
encode the service type.
For example, a hotplug port service on a downstream bridge was previously
called "pcie24" and is now called "pcie204". The "2" encodes the device
type (PCI_EXP_TYPE_DOWNSTREAM - 4), and the "4" (now "04") encodes the
service (PCIE_PORT_SERVICE_HP).
Based on Lukas Wunner's patch:
b688d6e487
[bhelgaas: split to separate patch, expand changelog]
Based-on-patch-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Currently dw_pcie_setup_rc() configures memory base and memory limit in the
type1 configuration header for the root complex. In doing so it uses the
CPU address (pp->mem_base) rather than the bus address (pp->mem_bus_addr).
This is wrong and it is useless since the configuration is overwritten
later on when pci_bus_assign_resources() is called.
Remove this configuration from dw_pcie_setup_rc().
Signed-off-by: Gabriele Paoloni <gabriele.paoloni@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Pratyush Anand <pratyush.anand@gmail.com>
Kernel hang is observed when pci-hyperv module is release with device
drivers still attached. E.g., when I do 'rmmod pci_hyperv' with BCM5720
device pass-through-ed (tg3 module) I see the following:
NMI watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [rmmod:2104]
...
Call Trace:
[<ffffffffa0641487>] tg3_read_mem+0x87/0x100 [tg3]
[<ffffffffa063f000>] ? 0xffffffffa063f000
[<ffffffffa0644375>] tg3_poll_fw+0x85/0x150 [tg3]
[<ffffffffa0649877>] tg3_chip_reset+0x357/0x8c0 [tg3]
[<ffffffffa064ca8b>] tg3_halt+0x3b/0x190 [tg3]
[<ffffffffa0657611>] tg3_stop+0x171/0x230 [tg3]
...
[<ffffffffa064c550>] tg3_remove_one+0x90/0x140 [tg3]
[<ffffffff813bee59>] pci_device_remove+0x39/0xc0
[<ffffffff814a3201>] __device_release_driver+0xa1/0x160
[<ffffffff814a32e3>] device_release_driver+0x23/0x30
[<ffffffff813b794a>] pci_stop_bus_device+0x8a/0xa0
[<ffffffff813b7ab6>] pci_stop_root_bus+0x36/0x60
[<ffffffffa02c3f38>] hv_pci_remove+0x238/0x260 [pci_hyperv]
The problem seems to be that we report local resources release before
stopping the bus and removing devices from it and device drivers may try to
perform some operations with these resources on shutdown. Move resources
release report after we do pci_stop_root_bus().
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Jake Oshins <jakeo@microsoft.com>
I.MX6+ has a dedicated bit for resetting PCIe core, which should be used
instead of a regular reset sequence since using the latter will hang the
SoC.
This commit is based on c34068d48273e24d392d9a49a38be807954420ed from
http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git
Tested-by: Gary Bisson <gary.bisson@boundarydevices.com>
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Use enumerated type instead of a boolean flag to specify the variant of
the PCIe IP block (6Q, 6SX, etc). This patch has zero functional impact,
however it makes the code easier to extend for the case of more than 2
possible variants of an IP block (of which there are).
[bhelgaas: rewrap comment, remove extra blank line]
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Linux 4.5 introduced a behavioral change in device probing during the
suspend process with commit 013c074f86 ("PM / sleep: prohibit devices
probing during suspend/hibernation"): It defers device probing during the
entire suspend process, starting from the prepare phase and ending with the
complete phase. A rule existed before that "we rely on subsystems not to
do any probing once a device is suspended" but it is enforced only now
(Alan Stern, https://lkml.org/lkml/2015/9/15/908).
This resulted in a WARN splat if a PCI device (e.g., Thunderbolt) is
plugged in while the system is asleep: Upon waking up, pciehp_resume()
discovers new devices in the resume phase and immediately tries to bind
them to a driver. Since probing is now deferred, device_attach() returns
-EPROBE_DEFER, which provoked a WARN in pci_bus_add_device().
Linux 4.6-rc1 aggravates the situation with commit ab1a187bba ("PCI:
Check device_attach() return value always"): If device_attach() returns a
negative value, pci_bus_add_device() now removes the sysfs and procfs
entries for the device and pci_bus_add_devices() subsequently locks up with
a BUG. Even with the BUG fixed we're still in trouble because the device
remains on the deferred probing list even though its sysfs and procfs
entries are gone and its children won't be added.
Fix by not interpreting -EPROBE_DEFER as failure. The device will be
probed eventually (through device_unblock_probing() in dpm_complete()) and
there is proper locking in place to avoid races (e.g., if devices are
unplugged again und thus deleted from the system before deferred probing
happens, I have tested this). Also, those functions which dereference
dev->driver (e.g. pci_pm_*()) do contain proper NULL pointer checks. So it
seems safe to ignore -EPROBE_DEFER.
Fixes: ab1a187bba ("PCI: Check device_attach() return value always")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Grygorii Strashko <grygorii.strashko@ti.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Previously when pci_bus_add_device() called device_attach() and it returned
a negative value, we emitted a WARN but carried on.
Commit ab1a187bba ("PCI: Check device_attach() return value always"),
introduced in Linux 4.6-rc1, changed this to unwind all steps preceding
device_attach() and to not set dev->is_added = 1.
The latter leads to a BUG if pci_bus_add_device() was called from
pci_bus_add_devices(). Fix by not recursing to a child bus if
device_attach() failed for the bridge leading to it.
This can be triggered by plugging in a PCI device (e.g. Thunderbolt) while
the system is asleep. The system locks up when woken because
device_attach() returns -EPROBE_DEFER.
Fixes: ab1a187bba ("PCI: Check device_attach() return value always")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The 32-bit addressing modes in the I/O and Prefetchable Memory registers
are required to be read-only. Since the underlying access method allows
them to be set, emulate their read-only nature and always set them.
Signed-off-by: David Daney <david.daney@cavium.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The R-Car PCIe driver requires the use of IRQ domains for its MSI code:
drivers/pci/host/pcie-rcar.c:635:9: error: implicit declaration of function 'irq_find_mapping' [-Werror=implicit-function-declaration]
drivers/pci/host/pcie-rcar.c:666:8: error: implicit declaration of function 'irq_create_mapping' [-Werror=implicit-function-declaration]
...
Add a Kconfig select to ensure that the feature is always enabled.
This is not consistent with what the other drivers do at the moment, but I
have another patch that changes them to do it like this one, which is more
logical.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
This patch modifies all the callers of vmbus_mmio_allocate()
to call vmbus_mmio_free() instead of release_mem_region().
Signed-off-by: Jake Oshins <jakeo@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The current XUSB pad controller bindings are insufficient to describe
PHY devices attached to USB controllers. New bindings have been created
to overcome these restrictions. As a side-effect each root port now is
assigned a set of PHY devices, one for each lane associated with the
root port. This has the benefit of allowing fine-grained control of the
power management for each lane.
Acked-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The Marvell Armada 7K/8K SoCs integrate a PCIe controller from Synopsys.
Add a new driver that provides the small glue needed to use the existing
Designware driver to make it work on Marvell Armada 7K/8K SoCs.
The MSI support will be enabled at a later point.
[bhelgaas: use dev_dbg(), dw_pcie_wait_for_link()]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
iomem_is_exclusive() requires a CPU physical address, but on some arches we
supplied a PCI bus address instead.
On most arches, pci_resource_to_user(res) returns "res->start", which is a
CPU physical address. But on microblaze, mips, powerpc, and sparc, it
returns the PCI bus address corresponding to "res->start".
The result is that pci_mmap_resource() may fail when it shouldn't (if the
bus address happens to match an existing resource), or it may succeed when
it should fail (if the resource is exclusive but the bus address doesn't
match it).
Call iomem_is_exclusive() with "res->start", which is always a CPU physical
address, not the result of pci_resource_to_user().
Fixes: e8de1481fd ("resource: allow MMIO exclusivity for device drivers")
Suggested-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: Arjan van de Ven <arjan@linux.intel.com>
Fix the misuse of goto statement in ks_pcie_get_irq_controller_info() as
simple return is more appropriate for this function. While at it add an
error log for absence of interrupt controller node.
[bhelgaas: drop "ret" altogether since we always know the return value]
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: Rob Herring <robh+dt@kernel.org>
CC: Pawel Moll <pawel.moll@arm.com>
CC: Mark Rutland <mark.rutland@arm.com>
CC: Ian Campbell <ijc+devicetree@hellion.org.uk>
CC: Kumar Gala <galak@codeaurora.org>
It appears that Gen2 is a misnomer for the R-Car PCIE driver
which also supports Gen 1 and Gen 3 SoCs. Accordingly, drop Gen 2
from the help text and Kconfig symbol.
Also, re-arange the Kconfig symbol name to use PCIE as the prefix.
This appears to be in keeping with other PCIE Kconfig symbols.
Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Geert Uytterhoeven <geert+renesas@glider.be>
Freescale has stated [1] that the LVDS clock source of the IMX6 does not
pass the PCI Gen2 clock jitter test, therefore unless an external Gen2
compliant external clock source is present and supplied back to the IMX6
PCIe core via LVDS CLK1/CLK2 you can not claim Gen2 compliance.
Add a DT property to specify Gen1 vs Gen2 and check this before allowing a
Gen2 link.
We default to Gen1 if the property is not present because at this time
there are no IMX6 boards in mainline that 'input' a clock on LVDS
CLK1/CLK2.
In order to be Gen2 compliant on IMX6 you need to:
- Have a Gen2 compliant external clock generator and route that clock back
to either LVDS CLK1 or LVDS CLK2 as an input (see IMX6SX-SabreSD
reference design).
- Specify this clock in the PCIe node in the DT (i.e.,
IMX6QDL_CLK_LVDS1_IN or IMX6QDL_CLK_LVDS2_IN instead of
IMX6QDL_CLK_LVDS1_GATE which configures it as a CLK output).
[1] https://community.freescale.com/message/453209
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
CC: Fabio Estevam <fabio.estevam@freescale.com>
CC: Zhu Richard <Richard.Zhu@freescale.com>
CC: Akshay Bhat <akshay.bhat@timesys.com>
CC: Rob Herring <robh+dt@kernel.org>
CC: Shawn Guo <shawnguo@kernel.org>
Currently the reset-gpio DT property which controls the PCI bus device
reset signal defaults to active-low reset sequence (L=reset state,
H=operation state) plus the code in reset function isn't GPIO polarity
aware - it doesn't matter if the defined reset-gpio is active-low or
active-high, it will always result into active-low reset sequence.
I've tried to fix it properly and change the reset-gpio reset sequence to
be polarity-aware, but this patch has been accepted and then reverted as it
has introduced few backward incompatible issues:
1. Some DTBs, for example, imx6qdl-sabresd, don't define reset-gpio
polarity correctly:
reset-gpio = <&gpio7 12 0>;
which means that it's defined as active-high, but in reality it's
active-low; thus it wouldn't work without a DTS fix.
2. The logic in the reset function is inverted:
gpio_set_value_cansleep(imx6_pcie->reset_gpio, 0)
msleep(100);
gpio_set_value_cansleep(imx6_pcie->reset_gpio, 1);
so even if some of the i.MX6 boards had reset-gpio polarity defined
correctly in their DTSes, they would stop working.
As we can't break old DTBs, we can't fix them, so we need to introduce this
new DT reset-gpio-active-high boolean property so we can support boards
with active-high reset sequence.
This active-high reset sequence is for example needed on Apalis SoMs, where
GPIO1_IO28, used to PCIe reset is not connected directly to PERST# PCIe
signal, but it's ORed with RESETBMCU coming off the PMIC, and thus is
inverted, active-high.
Tested-by: Tim Harvey <tharvey@gateworks.com> # Gateworks Ventana boards (which have active-low PERST#)
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Acked-by: Rob Herring <robh@kernel.org>
Add initial PCIe support for the imx6 SoC derivate imx6sx. PCI MSI support
is untested as the necessary suspend/resume quirk is not included in this
patch.
This patch is heavily based on patches by Richard Zhu.
[bhelgaas: factor out refclk enable, fix adjacent typos in imx6q-pcie.txt]
Signed-off-by: Christoph Fritz <chf.fritz@googlemail.com>
Acked-by: Richard Zhu <Richard.Zhu@freescale.com>
Acked-by: Lucas Stach <l.stach@pengutronix.de>
Factor out ref clock enable to make it cleaner to add imx6sx support. No
functional change intended.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Christoph Fritz <chf.fritz@googlemail.com>
Intel Sunrise Point root ports implement ACS but use dwords for the
capability and control registers, putting the control register at the wrong
offset.
Use quirks to enable and test ACS for these devices, which match the
standard functions modulo the broken control register offset.
Note that lspci assumes devices implement ACS per spec, so it shows invalid
ACS data for these devices.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The original thought was that if a device implemented ACS, then surely
we want to use that... well, it turns out that devices can make an ACS
capability so broken that we still need to fall back to quirks.
Reverse the order of ACS enabling to give quirks first shot at it.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
All of the i40e (XL710/X710) 10/20/40GbE NICs lack support for indicating
INTx is asserted via the interrupt bit in the PCI status register. The
DisINTx bit in the command register is functional, causing these devices to
be incorrectly detected as supporting INTx masking. Quirk them to properly
indicate no INTx masking support.
Device IDs copied from i40e_devids.h.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: John Ronciak <john.ronciak@intel.com>
CC: Jesse Brandeburg <jesse.brandeburg@intel.com>
After 104daa71b3 ("PCI: Determine actual VPD size on first access"), the
PCI core computes the valid VPD size by parsing the VPD starting at offset
0x0. We don't attempt to read past that valid size because that causes
some devices to crash.
However, some devices do have data past that valid size. For example,
Chelsio adapters contain two VPD structures, and the driver needs both of
them.
Add pci_set_vpd_size(). If a driver knows it is safe to read past the end
of the VPD data structure at offset 0, it can use pci_set_vpd_size() to
allow access to as much data as it needs.
[bhelgaas: changelog, split patches, rename to pci_set_vpd_size() and
return int (not ssize_t)]
Fixes: 104daa71b3 ("PCI: Determine actual VPD size on first access")
Tested-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Casey Leedom <leedom@chelsio.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
If ibm_get_table_from_acpi() fails then "table" isn't initialized. Check
for failure so we don't reference "table" unless it's been initialized.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The PCI config access checked the file capabilities correctly, but used
the itnernal security capability check rather than the helper function
that is actually meant for that.
The security_capable() has unusual return values and is not meant to be
used elsewhere (the only other use is in the capability checking
functions that we actually intend people to use, and this odd PCI usage
really stood out when looking around the capability code.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Keystone PCI hardware generates error interrupts at RC using a platform IRQ
instead of a standard MSI or legacy IRQ. Add a simple error handler that
logs the fatal interrupt status to the console.
[bhelgaas: s/node/dev->of_node/, tidy comments, return irqreturn_t directly]
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rob Herring <robh@kernel.org>
CC: Pawel Moll <pawel.moll@arm.com>
CC: Mark Rutland <mark.rutland@arm.com>
CC: Ian Campbell <ijc+devicetree@hellion.org.uk>
CC: Kumar Gala <galak@codeaurora.org>
The MIC x200 NTB forwards DMA transactions upstream using multiple alien
RIDs. These RIDs have to be added as aliases to the DMA device to allow
buffer access when the IOMMU is enabled.
Signed-off-by: Jacek Lawrynowicz <jacek.lawrynowicz@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Acked-by: David Woodhouse <David.Woodhouse@intel.com>
Solve IOMMU support issues with PCIe non-transparent bridges that use
Requester ID look-up tables (RID-LUT), e.g., the PEX8733.
The NTB connects devices in two independent PCI domains. Devices separated
by the NTB are not able to discover each other. A PCI packet being
forwared from one domain to another has to have its RID modified so it
appears on correct bus and completions are forwarded back to the original
domain through the NTB. The RID is translated using a preprogrammed table
(LUT) and the PCI packet propagates upstream away from the NTB. If the
destination system has IOMMU enabled, the packet will be discarded because
the new RID is unknown to the IOMMU. Adding a DMA alias for the new RID
allows IOMMU to properly recognize the packet.
Each device behind the NTB has a unique RID assigned in the RID-LUT. The
current DMA alias implementation supports only a single alias, so it's not
possible to support mutiple devices behind the NTB when IOMMU is enabled.
Enable all possible aliases on a given bus (256) that are stored in a
bitset. Alias devfn is directly translated to a bit number. The bitset is
not allocated for devices that have no need for DMA aliases.
More details can be found in the following article:
http://www.plxtech.com/files/pdf/technical/expresslane/RTC_Enabling%20MulitHostSystemDesigns.pdf
Signed-off-by: Jacek Lawrynowicz <jacek.lawrynowicz@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Acked-by: David Woodhouse <David.Woodhouse@intel.com>
Acked-by: Joerg Roedel <jroedel@suse.de>
One of the quirks that adds DMA aliases logs an informational message in
dmesg. Move that to pci_add_dma_alias() so all users log the message
consistently. No functional change intended (except extra message).
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Add a pci_add_dma_alias() interface to encapsulate the details of adding an
alias. No functional change intended.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Now that pcie_port_acpi_setup() always returns 0, make it and its callers
void functions and stop checking the return values.
[bhelgaas: changelog]
Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Host bridges we discover via ACPI, i.e., PNP0A03 and PNP0A08 devices, may
have an _OSC method by which the OS can ask the platform for control of
PCIe features like native hotplug, power management events, AER, etc.
Previously, if we found a bridge without an ACPI device, we assumed we did
not have permission to use any of these PCIe features. That seems
unreasonably restrictive.
If we find no ACPI device, assume we can take control of all PCIe features.
The Intel Volume Management Device (VMD) is one such bridge with no ACPI
device. Prior to this change, users had to boot with "pcie_ports=native"
to get hotplug and other services to work below the VMD Root Port.
[bhelgaas: changelog]
Suggested-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Add support for the 1st gen Light Ridge controller, which is built into
these systems:
iMac12,1 2011 21.5"
iMac12,2 2011 27"
Macmini5,1 2011 i5 2.3 GHz
Macmini5,2 2011 i5 2.5 GHz
Macmini5,3 2011 i7 2.0 GHz
MacBookPro8,1 2011 13"
MacBookPro8,2 2011 15"
MacBookPro8,3 2011 17"
MacBookPro9,1 2012 15"
MacBookPro9,2 2012 13"
Light Ridge (CV82524) was the very first copper Thunderbolt controller,
introduced 2010 alongside its fiber-optic cousin Light Peak (CVL2510).
Consequently the chip suffers from some teething troubles:
- MSI is broken for hotplug signaling on the downstream bridges: The chip
just never sends an interrupt. It requests 32 MSIs for each of its six
bridges and the pcieport driver only allocates one per bridge. However
I've verified that even if 32 MSIs are allocated there's no interrupt
on hotplug. The only option is thus to disable MSI, which is also what
OS X does. Apparently all Thunderbolt chips up to revision 1 of Cactus
Ridge 4C are plagued by this issue so quirk those as well.
- The chip supports a maximum hop_count of 32, unlike its successors
which support only 12. Fixup ring_interrupt_active() to cope with
values >= 32.
- Another peculiarity is that the chip supports a maximum of 13 ports
whereas its successors support 12. However the additional port (#5)
seems to be unusable as reading its TB_CFG_PORT config space results in
TB_CFG_ERROR_INVALID_CONFIG_SPACE. Add a quirk to mark the port
disabled on the root switch, assuming that's necessary on all Macs
using this chip.
Tested-by: Lukas Wunner <lukas@wunner.de> [MacBookPro9,1]
Tested-by: William Brown <william@blackhats.net.au> [MacBookPro8,2]
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Intel Gen 1 and 2 chips use the same ID for NHI, bridges and switch. Gen 3
chips and onward use a distinct ID for the NHI.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Use the SET_NOIRQ_SYSTEM_SLEEP_PM_OPS helper macro for mvebu_pcie_pm_ops.
The macro also sets up freeze_noirq, thaw_noirq and poweroff_noirq,
restore_noirq accordingly.
Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The mvebu_pcie_pm_ops structure is never modified, so declare it as const.
Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
dw_pcie_host_init() looks up host bridge resources, ioremaps them, creates
IRQ domains, and enumerates devices below the bridge. dw_pcie_setup_rc()
programs the Root Complex registers. The Root Complex may lose power
during suspend-to-RAM, and when we resume, we want to redo the latter but
not the former.
Move some Root Complex programming from dw_pcie_host_init() to
dw_pcie_setup_rc() where it belongs. DesignWare-based drivers can call
dw_pcie_setup_rc() in their resume paths.
[Niklas Cassel <niklas.cassel@axis.com>: This change moves outbound ATU
programming, which uses pp->mem_base, to dw_pcie_setup_rc(). Apply the
dra7xx pp->mem_base update before calling dw_pcie_setup_rc().]
[bhelgaas: changelog, fold in dra7xx fix from Niklas]
Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Pratyush Anand <pratyush.anand@gmail.com>
Commit 5c5fb40de8 ("PCI: imx6: Add support for active-low reset GPIO")
cause regressions on some boards like MX6 Gateworks Ventana, for example.
The reason for the breakage is that this commit sets the GPIO polarity in
the wrong logic level.
Also, the commit log is wrong because active-low reset GPIO is what the
driver used to support since the beginning.
So keep the old behavior that ignores the GPIO polarity specified in the
device tree and treat the PCI reset GPIO as active-low.
Reported-by: Krzysztof Hałasa <khalasa@piap.pl>
Tested-by: Tim Harvey <tharvey@gateworks.com> # Gateworks Ventana
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Tim Harvey <tharvey@gateworks.com>
Acked-by: Lucas Stach <l.stach@pengutronix.de>
CC: stable@vger.kernel.org # v4.5+
Fix spelling of "initalization".
[bhelgaas: also fix pci/pci.c]
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Prarit Bhargava