- Add the cpu_cache_invalidate_memregion() API for cache flushing in
response to physical memory reconfiguration, or memory-side data
invalidation from operations like secure erase or memory-device unlock.
- Add a facility for the kernel to warn about collisions between kernel
and userspace access to PCI configuration registers
- Add support for Restricted CXL Host (RCH) topologies (formerly CXL 1.1)
- Add handling and reporting of CXL errors reported via the PCIe AER
mechanism
- Add support for CXL Persistent Memory Security commands
- Add support for the "XOR" algorithm for CXL host bridge interleave
- Rework / simplify CXL to NVDIMM interactions
- Miscellaneous cleanups and fixes
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQSbo+XnGs+rwLz9XGXfioYZHlFsZwUCY5UpyAAKCRDfioYZHlFs
Z0ttAP4uxCjIibKsFVyexpSgI4vaZqQ9yt9NesmPwonc0XookwD+PlwP6Xc0d0Ox
t0gJ6+pwdh11NRzhcNE1pAaPcJZU4gs=
=HAQk
-----END PGP SIGNATURE-----
Merge tag 'cxl-for-6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl
Pull cxl updates from Dan Williams:
"Compute Express Link (CXL) updates for 6.2.
While it may seem backwards, the CXL update this time around includes
some focus on CXL 1.x enabling where the work to date had been with
CXL 2.0 (VH topologies) in mind.
First generation CXL can mostly be supported via BIOS, similar to DDR,
however it became clear there are use cases for OS native CXL error
handling and some CXL 3.0 endpoint features can be deployed on CXL 1.x
hosts (Restricted CXL Host (RCH) topologies). So, this update brings
RCH topologies into the Linux CXL device model.
In support of the ongoing CXL 2.0+ enabling two new core kernel
facilities are added.
One is the ability for the kernel to flag collisions between userspace
access to PCI configuration registers and kernel accesses. This is
brought on by the PCIe Data-Object-Exchange (DOE) facility, a hardware
mailbox over config-cycles.
The other is a cpu_cache_invalidate_memregion() API that maps to
wbinvd_on_all_cpus() on x86. To prevent abuse it is disabled in guest
VMs and architectures that do not support it yet. The CXL paths that
need it, dynamic memory region creation and security commands (erase /
unlock), are disabled when it is not present.
As for the CXL 2.0+ this cycle the subsystem gains support Persistent
Memory Security commands, error handling in response to PCIe AER
notifications, and support for the "XOR" host bridge interleave
algorithm.
Summary:
- Add the cpu_cache_invalidate_memregion() API for cache flushing in
response to physical memory reconfiguration, or memory-side data
invalidation from operations like secure erase or memory-device
unlock.
- Add a facility for the kernel to warn about collisions between
kernel and userspace access to PCI configuration registers
- Add support for Restricted CXL Host (RCH) topologies (formerly CXL
1.1)
- Add handling and reporting of CXL errors reported via the PCIe AER
mechanism
- Add support for CXL Persistent Memory Security commands
- Add support for the "XOR" algorithm for CXL host bridge interleave
- Rework / simplify CXL to NVDIMM interactions
- Miscellaneous cleanups and fixes"
* tag 'cxl-for-6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl: (71 commits)
cxl/region: Fix memdev reuse check
cxl/pci: Remove endian confusion
cxl/pci: Add some type-safety to the AER trace points
cxl/security: Drop security command ioctl uapi
cxl/mbox: Add variable output size validation for internal commands
cxl/mbox: Enable cxl_mbox_send_cmd() users to validate output size
cxl/security: Fix Get Security State output payload endian handling
cxl: update names for interleave ways conversion macros
cxl: update names for interleave granularity conversion macros
cxl/acpi: Warn about an invalid CHBCR in an existing CHBS entry
tools/testing/cxl: Require cache invalidation bypass
cxl/acpi: Fail decoder add if CXIMS for HBIG is missing
cxl/region: Fix spelling mistake "memergion" -> "memregion"
cxl/regs: Fix sparse warning
cxl/acpi: Set ACPI's CXL _OSC to indicate RCD mode support
tools/testing/cxl: Add an RCH topology
cxl/port: Add RCD endpoint port enumeration
cxl/mem: Move devm_cxl_add_endpoint() from cxl_core to cxl_mem
tools/testing/cxl: Add XOR Math support to cxl_test
cxl/acpi: Support CXL XOR Interleave Math (CXIMS)
...
- Core:
The bulk is the rework of the MSI subsystem to support per device MSI
interrupt domains. This solves conceptual problems of the current
PCI/MSI design which are in the way of providing support for PCI/MSI[-X]
and the upcoming PCI/IMS mechanism on the same device.
IMS (Interrupt Message Store] is a new specification which allows device
manufactures to provide implementation defined storage for MSI messages
contrary to the uniform and specification defined storage mechanisms for
PCI/MSI and PCI/MSI-X. IMS not only allows to overcome the size limitations
of the MSI-X table, but also gives the device manufacturer the freedom to
store the message in arbitrary places, even in host memory which is shared
with the device.
There have been several attempts to glue this into the current MSI code,
but after lengthy discussions it turned out that there is a fundamental
design problem in the current PCI/MSI-X implementation. This needs some
historical background.
When PCI/MSI[-X] support was added around 2003, interrupt management was
completely different from what we have today in the actively developed
architectures. Interrupt management was completely architecture specific
and while there were attempts to create common infrastructure the
commonalities were rudimentary and just providing shared data structures and
interfaces so that drivers could be written in an architecture agnostic
way.
The initial PCI/MSI[-X] support obviously plugged into this model which
resulted in some basic shared infrastructure in the PCI core code for
setting up MSI descriptors, which are a pure software construct for holding
data relevant for a particular MSI interrupt, but the actual association to
Linux interrupts was completely architecture specific. This model is still
supported today to keep museum architectures and notorious stranglers
alive.
In 2013 Intel tried to add support for hot-pluggable IO/APICs to the kernel,
which was creating yet another architecture specific mechanism and resulted
in an unholy mess on top of the existing horrors of x86 interrupt handling.
The x86 interrupt management code was already an incomprehensible maze of
indirections between the CPU vector management, interrupt remapping and the
actual IO/APIC and PCI/MSI[-X] implementation.
At roughly the same time ARM struggled with the ever growing SoC specific
extensions which were glued on top of the architected GIC interrupt
controller.
This resulted in a fundamental redesign of interrupt management and
provided the today prevailing concept of hierarchical interrupt
domains. This allowed to disentangle the interactions between x86 vector
domain and interrupt remapping and also allowed ARM to handle the zoo of
SoC specific interrupt components in a sane way.
The concept of hierarchical interrupt domains aims to encapsulate the
functionality of particular IP blocks which are involved in interrupt
delivery so that they become extensible and pluggable. The X86
encapsulation looks like this:
|--- device 1
[Vector]---[Remapping]---[PCI/MSI]--|...
|--- device N
where the remapping domain is an optional component and in case that it is
not available the PCI/MSI[-X] domains have the vector domain as their
parent. This reduced the required interaction between the domains pretty
much to the initialization phase where it is obviously required to
establish the proper parent relation ship in the components of the
hierarchy.
While in most cases the model is strictly representing the chain of IP
blocks and abstracting them so they can be plugged together to form a
hierarchy, the design stopped short on PCI/MSI[-X]. Looking at the hardware
it's clear that the actual PCI/MSI[-X] interrupt controller is not a global
entity, but strict a per PCI device entity.
Here we took a short cut on the hierarchical model and went for the easy
solution of providing "global" PCI/MSI domains which was possible because
the PCI/MSI[-X] handling is uniform across the devices. This also allowed
to keep the existing PCI/MSI[-X] infrastructure mostly unchanged which in
turn made it simple to keep the existing architecture specific management
alive.
A similar problem was created in the ARM world with support for IP block
specific message storage. Instead of going all the way to stack a IP block
specific domain on top of the generic MSI domain this ended in a construct
which provides a "global" platform MSI domain which allows overriding the
irq_write_msi_msg() callback per allocation.
In course of the lengthy discussions we identified other abuse of the MSI
infrastructure in wireless drivers, NTB etc. where support for
implementation specific message storage was just mindlessly glued into the
existing infrastructure. Some of this just works by chance on particular
platforms but will fail in hard to diagnose ways when the driver is used
on platforms where the underlying MSI interrupt management code does not
expect the creative abuse.
Another shortcoming of today's PCI/MSI-X support is the inability to
allocate or free individual vectors after the initial enablement of
MSI-X. This results in an works by chance implementation of VFIO (PCI
pass-through) where interrupts on the host side are not set up upfront to
avoid resource exhaustion. They are expanded at run-time when the guest
actually tries to use them. The way how this is implemented is that the
host disables MSI-X and then re-enables it with a larger number of
vectors again. That works by chance because most device drivers set up
all interrupts before the device actually will utilize them. But that's
not universally true because some drivers allocate a large enough number
of vectors but do not utilize them until it's actually required,
e.g. for acceleration support. But at that point other interrupts of the
device might be in active use and the MSI-X disable/enable dance can
just result in losing interrupts and therefore hard to diagnose subtle
problems.
Last but not least the "global" PCI/MSI-X domain approach prevents to
utilize PCI/MSI[-X] and PCI/IMS on the same device due to the fact that IMS
is not longer providing a uniform storage and configuration model.
The solution to this is to implement the missing step and switch from
global PCI/MSI domains to per device PCI/MSI domains. The resulting
hierarchy then looks like this:
|--- [PCI/MSI] device 1
[Vector]---[Remapping]---|...
|--- [PCI/MSI] device N
which in turn allows to provide support for multiple domains per device:
|--- [PCI/MSI] device 1
|--- [PCI/IMS] device 1
[Vector]---[Remapping]---|...
|--- [PCI/MSI] device N
|--- [PCI/IMS] device N
This work converts the MSI and PCI/MSI core and the x86 interrupt
domains to the new model, provides new interfaces for post-enable
allocation/free of MSI-X interrupts and the base framework for PCI/IMS.
PCI/IMS has been verified with the work in progress IDXD driver.
There is work in progress to convert ARM over which will replace the
platform MSI train-wreck. The cleanup of VFIO, NTB and other creative
"solutions" are in the works as well.
- Drivers:
- Updates for the LoongArch interrupt chip drivers
- Support for MTK CIRQv2
- The usual small fixes and updates all over the place
-----BEGIN PGP SIGNATURE-----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=9erA
-----END PGP SIGNATURE-----
Merge tag 'irq-core-2022-12-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq updates from Thomas Gleixner:
"Updates for the interrupt core and driver subsystem:
The bulk is the rework of the MSI subsystem to support per device MSI
interrupt domains. This solves conceptual problems of the current
PCI/MSI design which are in the way of providing support for
PCI/MSI[-X] and the upcoming PCI/IMS mechanism on the same device.
IMS (Interrupt Message Store] is a new specification which allows
device manufactures to provide implementation defined storage for MSI
messages (as opposed to PCI/MSI and PCI/MSI-X that has a specified
message store which is uniform accross all devices). The PCI/MSI[-X]
uniformity allowed us to get away with "global" PCI/MSI domains.
IMS not only allows to overcome the size limitations of the MSI-X
table, but also gives the device manufacturer the freedom to store the
message in arbitrary places, even in host memory which is shared with
the device.
There have been several attempts to glue this into the current MSI
code, but after lengthy discussions it turned out that there is a
fundamental design problem in the current PCI/MSI-X implementation.
This needs some historical background.
When PCI/MSI[-X] support was added around 2003, interrupt management
was completely different from what we have today in the actively
developed architectures. Interrupt management was completely
architecture specific and while there were attempts to create common
infrastructure the commonalities were rudimentary and just providing
shared data structures and interfaces so that drivers could be written
in an architecture agnostic way.
The initial PCI/MSI[-X] support obviously plugged into this model
which resulted in some basic shared infrastructure in the PCI core
code for setting up MSI descriptors, which are a pure software
construct for holding data relevant for a particular MSI interrupt,
but the actual association to Linux interrupts was completely
architecture specific. This model is still supported today to keep
museum architectures and notorious stragglers alive.
In 2013 Intel tried to add support for hot-pluggable IO/APICs to the
kernel, which was creating yet another architecture specific mechanism
and resulted in an unholy mess on top of the existing horrors of x86
interrupt handling. The x86 interrupt management code was already an
incomprehensible maze of indirections between the CPU vector
management, interrupt remapping and the actual IO/APIC and PCI/MSI[-X]
implementation.
At roughly the same time ARM struggled with the ever growing SoC
specific extensions which were glued on top of the architected GIC
interrupt controller.
This resulted in a fundamental redesign of interrupt management and
provided the today prevailing concept of hierarchical interrupt
domains. This allowed to disentangle the interactions between x86
vector domain and interrupt remapping and also allowed ARM to handle
the zoo of SoC specific interrupt components in a sane way.
The concept of hierarchical interrupt domains aims to encapsulate the
functionality of particular IP blocks which are involved in interrupt
delivery so that they become extensible and pluggable. The X86
encapsulation looks like this:
|--- device 1
[Vector]---[Remapping]---[PCI/MSI]--|...
|--- device N
where the remapping domain is an optional component and in case that
it is not available the PCI/MSI[-X] domains have the vector domain as
their parent. This reduced the required interaction between the
domains pretty much to the initialization phase where it is obviously
required to establish the proper parent relation ship in the
components of the hierarchy.
While in most cases the model is strictly representing the chain of IP
blocks and abstracting them so they can be plugged together to form a
hierarchy, the design stopped short on PCI/MSI[-X]. Looking at the
hardware it's clear that the actual PCI/MSI[-X] interrupt controller
is not a global entity, but strict a per PCI device entity.
Here we took a short cut on the hierarchical model and went for the
easy solution of providing "global" PCI/MSI domains which was possible
because the PCI/MSI[-X] handling is uniform across the devices. This
also allowed to keep the existing PCI/MSI[-X] infrastructure mostly
unchanged which in turn made it simple to keep the existing
architecture specific management alive.
A similar problem was created in the ARM world with support for IP
block specific message storage. Instead of going all the way to stack
a IP block specific domain on top of the generic MSI domain this ended
in a construct which provides a "global" platform MSI domain which
allows overriding the irq_write_msi_msg() callback per allocation.
In course of the lengthy discussions we identified other abuse of the
MSI infrastructure in wireless drivers, NTB etc. where support for
implementation specific message storage was just mindlessly glued into
the existing infrastructure. Some of this just works by chance on
particular platforms but will fail in hard to diagnose ways when the
driver is used on platforms where the underlying MSI interrupt
management code does not expect the creative abuse.
Another shortcoming of today's PCI/MSI-X support is the inability to
allocate or free individual vectors after the initial enablement of
MSI-X. This results in an works by chance implementation of VFIO (PCI
pass-through) where interrupts on the host side are not set up upfront
to avoid resource exhaustion. They are expanded at run-time when the
guest actually tries to use them. The way how this is implemented is
that the host disables MSI-X and then re-enables it with a larger
number of vectors again. That works by chance because most device
drivers set up all interrupts before the device actually will utilize
them. But that's not universally true because some drivers allocate a
large enough number of vectors but do not utilize them until it's
actually required, e.g. for acceleration support. But at that point
other interrupts of the device might be in active use and the MSI-X
disable/enable dance can just result in losing interrupts and
therefore hard to diagnose subtle problems.
Last but not least the "global" PCI/MSI-X domain approach prevents to
utilize PCI/MSI[-X] and PCI/IMS on the same device due to the fact
that IMS is not longer providing a uniform storage and configuration
model.
The solution to this is to implement the missing step and switch from
global PCI/MSI domains to per device PCI/MSI domains. The resulting
hierarchy then looks like this:
|--- [PCI/MSI] device 1
[Vector]---[Remapping]---|...
|--- [PCI/MSI] device N
which in turn allows to provide support for multiple domains per
device:
|--- [PCI/MSI] device 1
|--- [PCI/IMS] device 1
[Vector]---[Remapping]---|...
|--- [PCI/MSI] device N
|--- [PCI/IMS] device N
This work converts the MSI and PCI/MSI core and the x86 interrupt
domains to the new model, provides new interfaces for post-enable
allocation/free of MSI-X interrupts and the base framework for
PCI/IMS. PCI/IMS has been verified with the work in progress IDXD
driver.
There is work in progress to convert ARM over which will replace the
platform MSI train-wreck. The cleanup of VFIO, NTB and other creative
"solutions" are in the works as well.
Drivers:
- Updates for the LoongArch interrupt chip drivers
- Support for MTK CIRQv2
- The usual small fixes and updates all over the place"
* tag 'irq-core-2022-12-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (134 commits)
irqchip/ti-sci-inta: Fix kernel doc
irqchip/gic-v2m: Mark a few functions __init
irqchip/gic-v2m: Include arm-gic-common.h
irqchip/irq-mvebu-icu: Fix works by chance pointer assignment
iommu/amd: Enable PCI/IMS
iommu/vt-d: Enable PCI/IMS
x86/apic/msi: Enable PCI/IMS
PCI/MSI: Provide pci_ims_alloc/free_irq()
PCI/MSI: Provide IMS (Interrupt Message Store) support
genirq/msi: Provide constants for PCI/IMS support
x86/apic/msi: Enable MSI_FLAG_PCI_MSIX_ALLOC_DYN
PCI/MSI: Provide post-enable dynamic allocation interfaces for MSI-X
PCI/MSI: Provide prepare_desc() MSI domain op
PCI/MSI: Split MSI-X descriptor setup
genirq/msi: Provide MSI_FLAG_MSIX_ALLOC_DYN
genirq/msi: Provide msi_domain_alloc_irq_at()
genirq/msi: Provide msi_domain_ops:: Prepare_desc()
genirq/msi: Provide msi_desc:: Msi_data
genirq/msi: Provide struct msi_map
x86/apic/msi: Remove arch_create_remap_msi_irq_domain()
...
- Switch to using devm_gpiod_get_optional() so we can stop exporting
devm_gpiod_get_from_of_node() (Dmitry Torokhov)
* pci/ctrl/aardvark:
PCI: aardvark: Switch to using devm_gpiod_get_optional()
- Restore MSI remapping configuration during resume because the
configuration is cleared out by firmware when suspending (Nirmal Patel)
- Reset the hierarchy below VMD when probing the VMD; we attempted this
before, but with the wrong device, so it didn't work (Francisco Munoz)
* remotes/lorenzo/pci/vmd:
PCI: vmd: Fix secondary bus reset for Intel bridges
PCI: vmd: Disable MSI remapping after suspend
- Switch from devm_gpiod_get_from_of_node() to devm_fwnode_gpiod_get()
(Dmitry Torokhov)
* remotes/lorenzo/pci/tegra:
PCI: tegra: Switch to using devm_fwnode_gpiod_get
- Add sentinel to mt7621_pcie_quirks_match[] to prevent oops when parsing
the table (John Thomson)
* remotes/lorenzo/pci/mt7621:
PCI: mt7621: Add sentinel to quirks table
- Enable Multi-MSI (Jim Quinlan)
- Wait for 100ms after PERST# deassert for power and clocks to stabilize
(Jim Quinlan)
- Use readl_poll_timeout_atomic() instead of hand-rolled timeout loop (Jim
Quinlan)
- Drop needless "inline" annotations (Jim Quinlan)
- Set RCB_MPS mode bit so data for reads up to MPS are returned in a single
completion (Jim Quinlan)
* remotes/lorenzo/pci/brcmstb:
PCI: brcmstb: Set RCB_{MPS,64B}_MODE bits
PCI: brcmstb: Drop needless 'inline' annotations
PCI: brcmstb: Replace status loops with read_poll_timeout_atomic()
PCI: brcmstb: Wait for 100ms following PERST# deassert
PCI: brcmstb: Enable Multi-MSI
- Remove EfiMemoryMappedIO regions from the E820 map to allow PCI core to
allocate BARs from them. The only purpose of EfiMemoryMappedIO is to
tell the OS to map things needed by EFI runtime services, so it's often
used for PCI host bridge apertures. If we can't allocate from those
apertures, we can't hot-add devices (Bjorn Helgaas)
* pci/resource:
x86/PCI: Use pr_info() when possible
x86/PCI: Fix log message typo
x86/PCI: Tidy E820 removal messages
PCI: Skip allocate_resource() if too little space available
efi/x86: Remove EfiMemoryMappedIO from E820 map
- Squash portdrv_core.c and portdrv_pci.c into portdrv.c to make it easier
to find things (Bjorn Helgaas)
- Allow AER service only for Root Ports & RCECs so portdrv can successfully
bind to other devices that have AER but lack MSI (which they don't need
for AER), which allows power management for those devices (Bjorn Helgaas)
* pci/portdrv:
PCI/portdrv: Allow AER service only for Root Ports & RCECs
PCI/portdrv: Unexport pcie_port_service_register(), pcie_port_service_unregister()
PCI/portdrv: Move private things to portdrv.c
PCI/portdrv: Squash into portdrv.c
- Use METHOD_NAME__UID instead of plain string to make it easier to find
all uses (Yipeng Zou)
* pci/misc:
PCI/ACPI: Use METHOD_NAME__UID instead of plain string
- Enable pciehp by default if USB4 is enabled because USB4/Thunderbolt
tunneling depends on native PCIe hotplug (Albert Zhou)
- Make sure pciehp binds only to Downstream Ports, not Upstream Ports
(Rafael J. Wysocki)
- Remove unused get_mode1_ECC_cap callback in shpchp (Ian Cowan)
- Enable pciehp Command Completed Interrupt only if supported to reduce
confusion when looking at lspci output (Pali Rohár)
* pci/hotplug:
PCI: pciehp: Enable Command Completed Interrupt only if supported
PCI: shpchp: Remove unused get_mode1_ECC_cap callback
PCI: acpiphp: Avoid setting is_hotplug_bridge for PCIe Upstream Ports
PCI/portdrv: Set PCIE_PORT_SERVICE_HP for Root and Downstream Ports only
PCI: pciehp: Enable by default if USB4 enabled
- Only read/write PCIe Link 2 registers for devices with Links and PCIe
Capability version >= 2 (Maciej W. Rozycki)
- Revert a patch that cleared PCI_STATUS during enumeration because it
broke Linux guests on Apple's virtualization framework (Bjorn Helgaas)
- Assign PCI domain IDs using IDAs so IDs can be easily reused after
loading/unloading host bridge drivers (Pali Rohár)
- Fix pci_device_is_present(), which previously always returned "false" for
VFs because their vendor ID is always 0xfff (Michael S. Tsirkin)
- Check for alloc failure in pci_request_irq() (Zeng Heng)
* pci/enumeration:
PCI: Check for alloc failure in pci_request_irq()
PCI: Fix pci_device_is_present() for VFs by checking PF
PCI: Assign PCI domain IDs by ida_alloc()
Revert "PCI: Clear PCI_STATUS when setting up device"
PCI: Access Link 2 registers only for devices with Links
pci_bus_alloc_from_region() allocates MMIO space by iterating through all
the resources available on the bus. The available resource might be
reduced if the caller requires 32-bit space or we're avoiding BIOS or E820
areas.
Don't bother calling allocate_resource() if we need more space than is
available in this resource. This prevents some pointless and annoying
messages about avoided areas.
Link: https://lore.kernel.org/r/20221208190341.1560157-3-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Previously portdrv allowed the AER service for any device with an AER
capability (assuming Linux had control of AER) even though the AER service
driver only attaches to Root Port and RCECs.
Because get_port_device_capability() included AER for non-RP, non-RCEC
devices, we tried to initialize the AER IRQ even though these devices
don't generate AER interrupts.
Intel DG1 and DG2 discrete graphics cards contain a switch leading to a
GPU. The switch supports AER but not MSI, so initializing an AER IRQ
failed, and portdrv failed to claim the switch port at all. The GPU itself
could be suspended, but the switch could not be put in a low-power state
because it had no driver.
Don't allow the AER service on non-Root Port, non-Root Complex Event
Collector devices. This means we won't enable Bus Mastering if the device
doesn't require MSI, the AER service will not appear in sysfs, and the AER
service driver will not bind to the device.
Link: https://lore.kernel.org/r/20221207084105.84947-1-mika.westerberg@linux.intel.com
Link: https://lore.kernel.org/r/20221210002922.1749403-1-helgaas@kernel.org
Based-on-patch-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Switch the driver away from legacy gpio/of_gpio API to gpiod API, and
remove use of of_get_named_gpio_flags() which I want to make private to
gpiolib.
Link: https://lore.kernel.org/r/Y5EAft42YiT66mVj@google.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The No Command Completed Support bit in the Slot Capabilities register
indicates whether Command Completed Interrupt Enable is unsupported.
We already check whether No Command Completed Support bit is set in
pcie_wait_cmd(), and do not wait in this case.
Don't enable this Command Completed Interrupt at all if NCCS is set, so
that when users dump configuration space from userspace, the dump does not
confuse them by saying that Command Completed Interrupt is not supported,
but it is enabled.
Link: https://lore.kernel.org/r/20220927141926.8895-2-kabel@kernel.org
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
Switch the driver to the generic version of gpiod API (and away from
OF-specific variant), so that we can stop exporting
devm_gpiod_get_from_of_node().
Link: https://lore.kernel.org/r/Y3KMEZFv6dpxA+Gv@google.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Pali Rohár <pali@kernel.org>
Current driver is missing a sentinel in the struct soc_device_attribute
array, which causes an oops when assessed by the
soc_device_match(mt7621_pcie_quirks_match) call.
This was only exposed once the CONFIG_SOC_MT7621 mt7621 soc_dev_attr
was fixed to register the SOC as a device, in:
commit 7c18b64bba ("mips: ralink: mt7621: do not use kzalloc too early")
Fix it by adding the required sentinel.
Link: https://lore.kernel.org/lkml/26ebbed1-0fe9-4af9-8466-65f841d0b382@app.fastmail.com
Link: https://lore.kernel.org/r/20221205204645.301301-1-git@johnthomson.fastmail.com.au
Fixes: b483b4e4d3 ("staging: mt7621-pci: add quirks for 'E2' revision using 'soc_device_attribute'")
Signed-off-by: John Thomson <git@johnthomson.fastmail.com.au>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Sergio Paracuellos <sergio.paracuellos@gmail.com>
The reset was never applied in the current implementation because Intel
Bridges owned by VMD are parentless. Internally, pci_reset_bus() applies
a reset to the parent of the PCI device supplied as argument, but in this
case it failed because there wasn't a parent.
In more detail, this change allows the VMD driver to enumerate NVMe devices
in pass-through configurations when guest reboots are performed. There was
an attempted to fix this, but later we discovered that the code inside
pci_reset_bus() wasn’t triggering secondary bus resets. Therefore, we
updated the parameters passed to it, and now NVMe SSDs attached to VMD
bridges are properly enumerated in VT-d pass-through scenarios.
Link: https://lore.kernel.org/r/20221206001637.4744-1-francisco.munoz.ruiz@linux.intel.com
Fixes: 6aab562229 ("PCI: vmd: Clean up domain before enumeration")
Signed-off-by: Francisco Munoz <francisco.munoz.ruiz@linux.intel.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Nirmal Patel <nirmal.patel@linux.intel.com>
Reviewed-by: Jonathan Derrick <jonathan.derrick@linux.dev>
Single vector allocation which allocates the next free index in the IMS
space. The free function releases.
All allocated vectors are released also via pci_free_vectors() which is
also releasing MSI/MSI-X vectors.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.961711347@linutronix.de
IMS (Interrupt Message Store) is a new specification which allows
implementation specific storage of MSI messages contrary to the
strict standard specified MSI and MSI-X message stores.
This requires new device specific interrupt domains to handle the
implementation defined storage which can be an array in device memory or
host/guest memory which is shared with hardware queues.
Add a function to create IMS domains for PCI devices. IMS domains are using
the new per device domain mechanism and are configured by the device driver
via a template. IMS domains are created as secondary device domains so they
work side on side with MSI[-X] on the same device.
The IMS domains have a few constraints:
- The index space is managed by the core code.
Device memory based IMS provides a storage array with a fixed size
which obviously requires an index. But there is no association between
index and functionality so the core can randomly allocate an index in
the array.
System memory based IMS does not have the concept of an index as the
storage is somewhere in memory. In that case the index is purely
software based to keep track of the allocations.
- There is no requirement for consecutive index ranges
This is currently a limitation of the MSI core and can be implemented
if there is a justified use case by changing the internal storage from
xarray to maple_tree. For now it's single vector allocation.
- The interrupt chip must provide the following callbacks:
- irq_mask()
- irq_unmask()
- irq_write_msi_msg()
- The interrupt chip must provide the following optional callbacks
when the irq_mask(), irq_unmask() and irq_write_msi_msg() callbacks
cannot operate directly on hardware, e.g. in the case that the
interrupt message store is in queue memory:
- irq_bus_lock()
- irq_bus_unlock()
These callbacks are invoked from preemptible task context and are
allowed to sleep. In this case the mandatory callbacks above just
store the information. The irq_bus_unlock() callback is supposed to
make the change effective before returning.
- Interrupt affinity setting is handled by the underlying parent
interrupt domain and communicated to the IMS domain via
irq_write_msi_msg(). IMS domains cannot have a irq_set_affinity()
callback. That's a reasonable restriction similar to the PCI/MSI
device domain implementations.
The domain is automatically destroyed when the PCI device is removed.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.904316841@linutronix.de
MSI-X vectors can be allocated after the initial MSI-X enablement, but this
needs explicit support of the underlying interrupt domains.
Provide a function to query the ability and functions to allocate/free
individual vectors post-enable.
The allocation can either request a specific index in the MSI-X table or
with the index argument MSI_ANY_INDEX it allocates the next free vector.
The return value is a struct msi_map which on success contains both index
and the Linux interrupt number. In case of failure index is negative and
the Linux interrupt number is 0.
The allocation function is for a single MSI-X index at a time as that's
sufficient for the most urgent use case VFIO to get rid of the 'disable
MSI-X, reallocate, enable-MSI-X' cycle which is prone to lost interrupts
and redirections to the legacy and obviously unhandled INTx.
As single index allocation is also sufficient for the use cases Jason
Gunthorpe pointed out: Allocation of a MSI-X or IMS vector for a network
queue. See Link below.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/all/20211126232735.547996838@linutronix.de
Link: https://lore.kernel.org/r/20221124232326.731233614@linutronix.de
The setup of MSI descriptors for PCI/MSI-X interrupts depends partially on
the MSI index for which the descriptor is initialized.
Dynamic MSI-X vector allocation post MSI-X enablement allows to allocate
vectors at a given index or at any free index in the available table
range. The latter requires that the descriptor is initialized after the
MSI core has chosen an index.
Implement the prepare_desc() op in the PCI/MSI-X specific msi_domain_ops
which is invoked before the core interrupt descriptor and the associated
Linux interrupt number is allocated.
That callback is also provided for the upcoming PCI/IMS implementations so
the implementation specific interrupt domain can do their domain specific
initialization of the MSI descriptors.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.673658806@linutronix.de
The upcoming mechanism to allocate MSI-X vectors after enabling MSI-X needs
to share some of the MSI-X descriptor setup.
The regular descriptor setup on enable has the following code flow:
1) Allocate descriptor
2) Setup descriptor with PCI specific data
3) Insert descriptor
4) Allocate interrupts which in turn scans the inserted
descriptors
This cannot be easily changed because the PCI/MSI code needs to handle the
legacy architecture specific allocation model and the irq domain model
where quite some domains have the assumption that the above flow is how it
works.
Ideally the code flow should look like this:
1) Invoke allocation at the MSI core
2) MSI core allocates descriptor
3) MSI core calls back into the irq domain which fills in
the domain specific parts
This could be done for underlying parent MSI domains which support
post-enable allocation/free but that would create significantly different
code pathes for MSI/MSI-X enable.
Though for dynamic allocation which wants to share the allocation code with
the upcoming PCI/IMS support it's the right thing to do.
Split the MSI-X descriptor setup into the preallocation part which just sets
the index and fills in the horrible hack of virtual IRQs and the real PCI
specific MSI-X setup part which solely depends on the index in the
descriptor. This allows to provide a common dynamic allocation interface at
the MSI core level for both PCI/MSI-X and PCI/IMS.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.616292598@linutronix.de
The check for special MSI domains like VMD which prevents the interrupt
remapping code to overwrite device::msi::domain is not longer required and
has been replaced by an x86 specific version which is aware of MSI parent
domains.
Remove it.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232326.093093200@linutronix.de
Provide a template and the necessary callbacks to create PCI/MSI and
PCI/MSI-X domains.
The domains are created when MSI or MSI-X is enabled. The domain's lifetime
is either the device lifetime or in case that e.g. MSI-X was tried first
and failed, then the MSI-X domain is removed and a MSI domain is created as
both are mutually exclusive and reside in the default domain ID slot of the
per device domain pointer array.
Also expand pci_msi_domain_supports() to handle feature checks correctly
even in the case that the per device domain was not yet created by checking
the features supported by the MSI parent.
Add the necessary setup calls into the MSI and MSI-X enable code path.
These setup calls are backwards compatible. They return success when there
is no parent domain found, which means the existing global domains or the
legacy allocation path keep just working.
Co-developed-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232325.975388241@linutronix.de
The upcoming per device MSI domains will create different domains for MSI
and MSI-X. Split the write message function into MSI and MSI-X helpers so
they can be used by those new domain functions seperately.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124232325.857982142@linutronix.de
Switch to the new domain id aware interfaces to phase out the previous
ones. No functional change.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124230314.455168748@linutronix.de
This reflects the functionality better. No functional change.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20221124230314.103554618@linutronix.de
Use bullet-list RST syntax for kernel-doc parameters' flags and interrupt
mode descriptions. Otherwise Sphinx produces "Unexpected identation" errors
and warnings.
Fixes: 5c0997dc33 ("PCI/MSI: Move pci_alloc_irq_vectors() to api.c")
Fixes: 017239c8db ("PCI/MSI: Move pci_irq_vector() to api.c")
Fixes: be37b8428b ("PCI/MSI: Move pci_irq_get_affinity() to api.c")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Suggested-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Ahmed S. Darwish <darwi@linutronix.de>
Link: https://lore.kernel.org/r/20221203100511.222136-1-bagasdotme@gmail.com
Some new devices such as CXL devices may want to record additional error
information on a corrected error. Add a callback to allow the PCI device
driver to do additional logging such as providing additional stats for user
space RAS monitoring.
For CXL device, this is actually a need due to CXL needing to write to the
CXL RAS capability structure correctable error status register in order to
clear the unmasked correctable errors. See CXL spec rev3.0 8.2.4.16.
Suggested-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/166984619233.2804404.3966368388544312674.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Resolve conflicts in drivers/vfio/vfio_main.c by using the iommfd version.
The rc fix was done a different way when iommufd patches reworked this
code.
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The function hv_set_affinity was removed in commit 831c1ae7 ("PCI: hv:
Make the code arch neutral by adding arch specific interfaces").
Signed-off-by: Olaf Hering <olaf@aepfle.de>
Link: https://lore.kernel.org/r/20221107171831.25283-1-olaf@aepfle.de
Signed-off-by: Wei Liu <wei.liu@kernel.org>
Use epf_db[i] dereference instead of readl() because epf_db is
in memory allocated by dma_alloc_coherent(), not I/O.
Remove useless/duplicated readl() in the process.
Link: https://lore.kernel.org/r/20221102141014.1025893-7-Frank.Li@nxp.com
Signed-off-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
NTB spad entry item size is sizeof(u32), replace hardcoded 4 with it.
Link: https://lore.kernel.org/r/20221102141014.1025893-6-Frank.Li@nxp.com
Signed-off-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
epf_db_phy member in struct epf_ntb is not used, remove it.
Link: https://lore.kernel.org/r/20221102141014.1025893-5-Frank.Li@nxp.com
Signed-off-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
Replace pci_epc_mem_free_addr() with pci_epf_free_space() in the
error handle path to match pci_epf_alloc_space().
Link: https://lore.kernel.org/r/20221102141014.1025893-4-Frank.Li@nxp.com
Fixes: e35f56bb03 ("PCI: endpoint: Support NTB transfer between RC and EP")
Signed-off-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Cleanup warning found by scripts/kernel-doc.
Consolidate terms:
- host, host1 to HOST
- vhost, vHost, Vhost, VHOST2 to VHOST
Link: https://lore.kernel.org/r/20221102141014.1025893-2-Frank.Li@nxp.com
Signed-off-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Manivannan Sadhasivam <mani@kernel.org>
Baikal-T1 SoC is equipped with DWC PCIe v4.60a host controller. It can be
trained to work up to Gen.3 speed over up to x4 lanes. The host controller
is attached to the DW PCIe 3.0 PCS via the PIPE-4 interface, which in its
turn is connected to the DWC 10G PHY. The whole system is supposed to be
fed up with four clock sources: DBI peripheral clock, AXI application
clocks and external PHY/core reference clock generating the 100MHz signal.
In addition to that the platform provide a way to reset each part of the
controller: sticky/non-sticky bits, host controller core, PIPE interface,
PCS/PHY and Hot/Power reset signal. The driver also provides a way to
handle the GPIO-based PERST# signal.
Note due to the Baikal-T1 MMIO peculiarity we have to implement the DBI
interface accessors which make sure the IO operations are dword-aligned.
Link: https://lore.kernel.org/r/20221113191301.5526-21-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Currently almost each platform driver uses its own resets and clocks
naming in order to get the corresponding descriptors. It makes the code
harder to maintain and comprehend especially seeing the DWC PCIe core main
resets and clocks signals set hasn't changed much for about at least one
major IP-core release. So in order to organize things around these signals
we suggest to create a generic interface for them in accordance with the
naming introduced in the DWC PCIe IP-core reference manual:
Application clocks:
- "dbi" - data bus interface clock (on some DWC PCIe platforms it's
referred as "pclk", "pcie", "sys", "ahb", "cfg", "iface",
"gio", "reg", "pcie_apb_sys");
- "mstr" - AXI-bus master interface clock (some DWC PCIe glue drivers
refer to this clock as "port", "bus", "pcie_bus",
"bus_master/master_bus/axi_m", "pcie_aclk");
- "slv" - AXI-bus slave interface clock (also called as "port", "bus",
"pcie_bus", "bus_slave/slave_bus/axi_s", "pcie_aclk",
"pcie_inbound_axi").
Core clocks:
- "pipe" - core-PCS PIPE interface clock coming from external PHY (it's
normally named by the platform drivers as just "pipe");
- "core" - primary clock of the controller (none of the platform drivers
declare such a clock but in accordance with the ref. manual
the devices may have it separately specified);
- "aux" - auxiliary PMC domain clock (it is named by some platforms as
"pcie_aux" and just "aux");
- "ref" - Generic reference clock (it is a generic clock source, which
can be used as a signal source for multiple interfaces, some
platforms call it as "ref", "general", "pcie_phy",
"pcie_phy_ref").
Application resets:
- "dbi" - Data-bus interface reset (it's CSR interface clock and is
normally called as "apb" though technically it's not APB but
DWC PCIe-specific interface);
- "mstr" - AXI-bus master reset (some platforms call it as "port", "apps",
"bus", "axi_m");
- "slv" - ABI-bus slave reset (some platforms call it as "port", "apps",
"bus", "axi_s").
Core resets:
- "non-sticky" - non-sticky CSR flags reset;
- "sticky" - sticky CSR flags reset;
- "pipe" - PIPE-interface (Core-PCS) logic reset (some platforms
call it just "pipe");
- "core" - controller primary reset (resets everything except PMC
module, some platforms refer to this signal as "soft",
"pci");
- "phy" - PCS/PHY block reset (strictly speaking it is normally
connected to the input of an external block, but the
reference manual says it must be available for the PMC
working correctly, some existing platforms call it
"pciephy", "phy", "link");
- "hot" - PMC hot reset signal (also called as "sleep");
- "pwr" - cold reset signal (can be referred as "pwr", "turnoff").
Bus reset:
- "perst" - PCIe standard signal used to reset the PCIe peripheral
devices.
As you can see each platform uses it's own naming for basically the same
set of the signals. In the framework of this commit we suggest to add a
set of the clocks and reset signals resources, corresponding names and
identifiers for each denoted entity. At current stage the platforms will
be able to use the provided infrastructure to automatically request all
these resources and manipulate with them in the Host/EP init callbacks.
Alas it isn't that easy to create a common cold/hot reset procedure due to
too many platform-specifics in the procedure, like the external flags
exposure and the delays requirement.
Link: https://lore.kernel.org/r/20221113191301.5526-20-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Since the iATU CSR region is now retrieved in the DW PCIe resources getter
there is no much benefits in the iATU detection procedures splitting up.
Therefore let's join the iATU unroll/viewport detection procedure with the
rest of the iATU parameters detection code. The resultant method will be
as coherent as before, while the redundant functions will be eliminated
thus producing more readable code.
Link: https://lore.kernel.org/r/20221113191301.5526-19-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Currently the DW PCIe Root Port and Endpoint CSR spaces are retrieved in
the separate parts of the DW PCIe core driver. It doesn't really make
sense since the both controller types have identical set of the core CSR
regions: DBI, DBI CS2 and iATU/eDMA. Thus we can simplify the DW PCIe Host
and EP initialization methods by moving the platform-specific registers
space getting and mapping into a common method. It gets to be even more
justified seeing the CSRs base address pointers are preserved in the
common DW PCIe descriptor. Note all the OF-based common DW PCIe settings
initialization will be moved to the new method too in order to have a
single function for all the generic platform properties handling in single
place.
A nice side-effect of this change is that the pcie-designware-host.c and
pcie-designware-ep.c drivers are cleaned up from all the direct dw_pcie
storage modification, which makes the DW PCIe core, Root Port and Endpoint
modules more coherent.
Link: https://lore.kernel.org/r/20221113191301.5526-18-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Since in addition to the already available iATU unrolled mapping we are
about to add a few more DW PCIe platform-specific capabilities (CDM-check
and generic clocks/resets resources) let's add a generic interface to set
and get the flags indicating their availability. The new interface shall
improve maintainability of the platform-specific code.
Link: https://lore.kernel.org/r/20221113191301.5526-17-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
In accordance with the generic PCIe Root Port DT-bindings the "dma-ranges"
property has the same format as the "ranges" property. The only difference
is in their semantics. The "dma-ranges" property describes the PCIe-to-CPU
memory mapping in opposite to the CPU-to-PCIe mapping of the "ranges"
property. Even though the DW PCIe controllers are normally equipped with
the internal Address Translation Unit which inbound and outbound tables
can be used to implement both properties semantics, it was surprising for
me to discover that the host-related part of the DW PCIe driver currently
supports the "ranges" property only while the "dma-ranges" windows are
just ignored. Having the "dma-ranges" supported in the driver would be
very handy for the platforms, that don't tolerate the 1:1 CPU-PCIe memory
mapping and require a customized PCIe memory layout. So let's fix that by
introducing the "dma-ranges" property support.
First of all we suggest to rename the dw_pcie_prog_inbound_atu() method to
dw_pcie_prog_ep_inbound_atu() and create a new version of the
dw_pcie_prog_inbound_atu() function. Thus we'll have two methods for the
RC and EP controllers respectively in the same way as it has been
developed for the outbound ATU setup methods.
Secondly aside with the memory window index and type the new
dw_pcie_prog_inbound_atu() function will accept CPU address, PCIe address
and size as its arguments. These parameters define the PCIe and CPU memory
ranges which will be used to setup the respective inbound ATU mapping. The
passed parameters need to be verified against the ATU ranges constraints
in the same way as it is done for the outbound ranges.
Finally the DMA-ranges detected for the PCIe controller need to be
converted to the inbound ATU entries during the host controller
initialization procedure. It will be done in the framework of the
dw_pcie_iatu_setup() method. Note before setting the inbound ranges up we
need to disable all the inbound ATU entries in order to prevent unexpected
PCIe TLPs translations defined by some third party software like
bootloaders.
Link: https://lore.kernel.org/r/20221113191301.5526-16-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
It is reported that on some systems pciehp binds to an Upstream Port and
attempts to operate it which causes devices below the Port to disappear
from the bus.
This happens because acpiphp sets dev->is_hotplug_bridge for that Port
(after receiving a Device Check notification on it from the platform
firmware via ACPI) during the enumeration of PCI devices.
get_port_device_capability() sees that dev->is_hotplug_bridge is set and
adds PCIE_PORT_SERVICE_HP to Port services, which allows pciehp to bind to
the Port in question.
Even though this particular problem can be addressed by making the
portdrv_core checks more robust, it also causes power management to work
differently on the affected systems which generally is not desirable (PCIe
Ports with dev->is_hotplug_bridge set have to pass additional tests to be
allowed to go into the D3hot/cold power states which affects runtime PM of
devices below these Ports).
For this reason, amend check_hotplug_bridge() with a PCIe type check to
prevent it from setting dev->is_hotplug_bridge for Upstream Ports.
Suggested-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/2262230.ElGaqSPkdT@kreacher
Reported-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Tested-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
It is reported that on some systems pciehp binds to an Upstream Port and
attempts to operate it which causes devices below the Port to disappear
from the bus.
This happens because acpiphp sets dev->is_hotplug_bridge for that Port
(after receiving a Device Check notification on it from the platform
firmware via ACPI) during the enumeration of PCI devices.
get_port_device_capability() sees that dev->is_hotplug_bridge is set and
adds PCIE_PORT_SERVICE_HP to Port services (which allows pciehp to bind to
the Port in question) without consulting the PCIe type, which should be
either Root Port or Downstream Port for the hotplug capability to be
present.
Per PCIe r6.0, sec 7.5.3.2, the Slot Implemented bit is only valid for
Downstream Ports (including Root Ports), and PCIe hotplug depends on the
Slot Capabilities / Control / Status registers.
Make get_port_device_capability() more robust by adding a PCIe type check
to it before adding PCIE_PORT_SERVICE_HP to Port services which helps to
avoid the problem.
[bhelgaas: add spec citation]
Suggested-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/4786090.31r3eYUQgx@kreacher
Reported-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
When kvasprintf() fails to allocate memory, it returns a NULL pointer.
Return error from pci_request_irq() so we don't dereference it.
[bhelgaas: commit log]
Fixes: 704e8953d3 ("PCI/irq: Add pci_request_irq() and pci_free_irq() helpers")
Link: https://lore.kernel.org/r/20221121020029.3759444-1-zengheng4@huawei.com
Signed-off-by: Zeng Heng <zengheng4@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
This is a simple mechanical transformation done by:
@@
expression E;
@@
- prandom_u32_max
+ get_random_u32_below
(E)
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Darrick J. Wong <djwong@kernel.org> # for xfs
Reviewed-by: SeongJae Park <sj@kernel.org> # for damon
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> # for infiniband
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> # for arm
Acked-by: Ulf Hansson <ulf.hansson@linaro.org> # for mmc
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
We have stubs for most OF interfaces even when CONFIG_OF is not set, so we
allow building of most controller drivers in that case for compile testing.
When CONFIG_OF is not set, "of_match_ptr(<match_table>)" compiles to NULL,
which leaves <match_table> unused, resulting in errors like this:
$ make W=1
drivers/pci/controller/pci-xgene.c:636:34: error: ‘xgene_pcie_match_table’ defined but not used [-Werror=unused-const-variable=]
Drop of_match_ptr() to avoid the unused variable warning.
See also 1dff012f63 ("PCI: Drop of_match_ptr() to avoid unused
variables").
Link: https://lore.kernel.org/r/20221025191339.667614-2-helgaas@kernel.org
Link: https://lore.kernel.org/r/20221116205100.1136224-2-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Now that the PCI/MSI core code does early checking for multi-MSI support
X86_IRQ_ALLOC_CONTIGUOUS_VECTORS is not required anymore.
Remove the flag and rely on MSI_FLAG_MULTI_PCI_MSI.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20221111122015.865042356@linutronix.de
All these sanity checks are now done _before_ any allocation work
happens. No point in doing it twice.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.749446904@linutronix.de
With interrupt domains the sanity check for MSI-X vector validation can be
done _before_ any allocation happens. The sanity check only applies to the
allocation functions which have an 'entries' array argument. The entries
array is filled by the caller with the requested MSI-X indices. Some drivers
have gaps in the index space which is not supported on all architectures.
The PCI/MSI irq domain has a 'feature' bit to enforce this validation late
during the allocation phase.
Just do it right away before doing any other work along with the other
sanity checks on that array.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.691357406@linutronix.de
Similar to PCI multi-MSI reject MSI-X enablement when a irq domain is
attached to the device which does not support MSI-X.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.631728309@linutronix.de
When hierarchical MSI interrupt domains are enabled then there is no point
to do tons of work and detect the missing support for multi-MSI late in the
allocation path.
Just query the domain feature flags right away. The query function is going
to be used for other purposes later and has a mode argument which influences
the result:
ALLOW_LEGACY returns true when:
- there is no irq domain attached (legacy support)
- there is a irq domain attached which has the feature flag set
DENY_LEGACY returns only true when:
- there is a irq domain attached which has the feature flag set
This allows to use the function universally without ifdeffery in the
calling code.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.574339988@linutronix.de
There is no point in doing the same sanity checks over and over in a loop
during MSI-X enablement. Put them in front of the loop and return early
when they fail.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.516946468@linutronix.de
There is no way to navigate msi.c without banging the head against the wall
every now and then because MSI and MSI-X specific functions are
intermingled and the code flow is completely non-obvious.
Reorder everthing so common helpers, MSI and MSI-X specific functions are
grouped together.
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.459089736@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_msi_enabled() and add kernel-doc for the function.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.331584998@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_msi_enabled() and make its kernel-doc comprehensive.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.271447896@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_irq_get_affinity() and let its kernel-doc match rest of the
file.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.214792769@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_disable_msix() and make its kernel-doc comprehensive.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.156785224@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_msix_vec_count() and make its kernel-doc comprehensive.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.099461602@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_free_irq_vectors() and make its kernel-doc comprehensive.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122015.042870570@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_irq_vector() and let its kernel-doc match the rest of the file.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.984490384@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_alloc_irq_vectors_affinity() and let its kernel-doc reference
pci_alloc_irq_vectors() documentation added in parent commit.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.927531290@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Make pci_alloc_irq_vectors() a real function instead of wrapper and add
proper kernel doc to it.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.870888193@linutronix.de
To disentangle the maze in msi.c, all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_enable_msix_range() and make its kernel-doc comprehensive.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.813792885@linutronix.de
To disentangle the maze in msi.c all exported device-driver MSI APIs are
now to be grouped in one file, api.c.
Move pci_enable_msi() and make its kernel-doc comprehensive.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.755178149@linutronix.de
msi.c is a maze of randomly sorted functions which makes the code
unreadable. As a first step split the driver visible API and the internal
implementation which also allows proper API documentation via one file.
Create drivers/pci/msi/api.c to group all exported device-driver PCI/MSI
APIs in one C file.
Begin by moving pci_disable_msi() there and add kernel-doc for the function
as appropriate.
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.696798036@linutronix.de
The upcoming support for per device MSI interrupt domains needs to share
some of the inline helpers with the MSI implementation.
Move them to the header file.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.640052354@linutronix.de
Follow the style of <linux/pci.h>
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.582175082@linutronix.de
Adjust to reality and remove another layer of pointless Kconfig
indirection. CONFIG_GENERIC_MSI_IRQ is good enough to serve
all purposes.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20221111122014.524842979@linutronix.de
What a zoo:
PCI_MSI
select GENERIC_MSI_IRQ
PCI_MSI_IRQ_DOMAIN
def_bool y
depends on PCI_MSI
select GENERIC_MSI_IRQ_DOMAIN
Ergo PCI_MSI enables PCI_MSI_IRQ_DOMAIN which in turn selects
GENERIC_MSI_IRQ_DOMAIN. So all the dependencies on PCI_MSI_IRQ_DOMAIN are
just an indirection to PCI_MSI.
Match the reality and just admit that PCI_MSI requires
GENERIC_MSI_IRQ_DOMAIN.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.467556921@linutronix.de
Let the core do the freeing of descriptors and just keep it around for the
legacy case.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.409654736@linutronix.de
Set the bus token in the msi_domain_info structure and let the core code
handle the update.
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122014.352437595@linutronix.de
PCI/MSI and PCI/MSI-X are mutually exclusive, but the MSI-X enable code
lacks a check for already enabled MSI.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ashok Raj <ashok.raj@intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20221111122013.653556720@linutronix.de
Per PCIe r6.0, sec 6.30.1, a data object Length of 0x0 indicates 2^18
DWORDs (256K DW or 1MB) being transferred. Adjust the value of data object
length for this case on both sending side and receiving side.
Don't bother checking whether Length is greater than SZ_1M because all
values of the 18-bit Length field are valid, and it is impossible to
represent anything larger than SZ_1M:
0x00000 256K DW (1M bytes)
0x00001 1 DW (4 bytes)
...
0x3ffff 256K-1 DW (1M - 4 bytes)
[bhelgaas: commit log]
Link: https://lore.kernel.org/r/20221116015637.3299664-1-ming4.li@intel.com
Fixes: 9d24322e88 ("PCI/DOE: Add DOE mailbox support functions")
Signed-off-by: Li Ming <ming4.li@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
Cc: stable@vger.kernel.org # v6.0+
PCI config space access from user space has traditionally been
unrestricted with writes being an understood risk for device operation.
Unfortunately, device breakage or odd behavior from config writes lacks
indicators that can leave driver writers confused when evaluating
failures. This is especially true with the new PCIe Data Object
Exchange (DOE) mailbox protocol where backdoor shenanigans from user
space through things such as vendor defined protocols may affect device
operation without complete breakage.
A prior proposal restricted read and writes completely.[1] Greg and
Bjorn pointed out that proposal is flawed for a couple of reasons.
First, lspci should always be allowed and should not interfere with any
device operation. Second, setpci is a valuable tool that is sometimes
necessary and it should not be completely restricted.[2] Finally
methods exist for full lock of device access if required.
Even though access should not be restricted it would be nice for driver
writers to be able to flag critical parts of the config space such that
interference from user space can be detected.
Introduce pci_request_config_region_exclusive() to mark exclusive config
regions. Such regions trigger a warning and kernel taint if accessed
via user space.
Create pci_warn_once() to restrict the user from spamming the log.
[1] https://lore.kernel.org/all/161663543465.1867664.5674061943008380442.stgit@dwillia2-desk3.amr.corp.intel.com/
[2] https://lore.kernel.org/all/YF8NGeGv9vYcMfTV@kroah.com/
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Suggested-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20220926215711.2893286-2-ira.weiny@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
This patch switches the driver away from legacy gpio/of_gpio API to
gpiod API, and removes use of of_get_named_gpio_flags() which I want to
make private to gpiolib.
Link: https://lore.kernel.org/r/20220906204301.3736813-1-dmitry.torokhov@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Jeffrey added Multi-MSI support to the pci-hyperv driver by the 4 patches:
08e61e861a ("PCI: hv: Fix multi-MSI to allow more than one MSI vector")
455880dfe2 ("PCI: hv: Fix hv_arch_irq_unmask() for multi-MSI")
b4b77778ec ("PCI: hv: Reuse existing IRTE allocation in compose_msi_msg()")
a2bad844a6 ("PCI: hv: Fix interrupt mapping for multi-MSI")
It turns out that the third patch (b4b77778ec) causes a performance
regression because all the interrupts now happen on 1 physical CPU (or two
pCPUs, if one pCPU doesn't have enough vectors). When a guest has many PCI
devices, it may suffer from soft lockups if the workload is heavy, e.g.,
see https://lwn.net/ml/linux-kernel/20220804025104.15673-1-decui@microsoft.com/
Commit b4b77778ec itself is good. The real issue is that the hypercall in
hv_irq_unmask() -> hv_arch_irq_unmask() ->
hv_do_hypercall(HVCALL_RETARGET_INTERRUPT...) only changes the target
virtual CPU rather than physical CPU; with b4b77778ec, the pCPU is
determined only once in hv_compose_msi_msg() where only vCPU0 is specified;
consequently the hypervisor only uses 1 target pCPU for all the interrupts.
Note: before b4b77778ec, the pCPU is determined twice, and when the pCPU
is determined the second time, the vCPU in the effective affinity mask is
used (i.e., it isn't always vCPU0), so the hypervisor chooses different
pCPU for each interrupt.
The hypercall will be fixed in future to update the pCPU as well, but
that will take quite a while, so let's restore the old behavior in
hv_compose_msi_msg(), i.e., don't reuse the existing IRTE allocation for
single-MSI and MSI-X; for multi-MSI, we choose the vCPU in a round-robin
manner for each PCI device, so the interrupts of different devices can
happen on different pCPUs, though the interrupts of each device happen on
some single pCPU.
The hypercall fix may not be backported to all old versions of Hyper-V, so
we want to have this guest side change forever (or at least till we're sure
the old affected versions of Hyper-V are no longer supported).
Fixes: b4b77778ec ("PCI: hv: Reuse existing IRTE allocation in compose_msi_msg()")
Co-developed-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Signed-off-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Co-developed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/20221104222953.11356-1-decui@microsoft.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
pci_device_is_present() previously didn't work for VFs because it reads the
Vendor and Device ID, which are 0xffff for VFs, which looks like they
aren't present. Check the PF instead.
Wei Gong reported that if virtio I/O is in progress when the driver is
unbound or "0" is written to /sys/.../sriov_numvfs, the virtio I/O
operation hangs, which may result in output like this:
task:bash state:D stack: 0 pid: 1773 ppid: 1241 flags:0x00004002
Call Trace:
schedule+0x4f/0xc0
blk_mq_freeze_queue_wait+0x69/0xa0
blk_mq_freeze_queue+0x1b/0x20
blk_cleanup_queue+0x3d/0xd0
virtblk_remove+0x3c/0xb0 [virtio_blk]
virtio_dev_remove+0x4b/0x80
...
device_unregister+0x1b/0x60
unregister_virtio_device+0x18/0x30
virtio_pci_remove+0x41/0x80
pci_device_remove+0x3e/0xb0
This happened because pci_device_is_present(VF) returned "false" in
virtio_pci_remove(), so it called virtio_break_device(). The broken vq
meant that vring_interrupt() skipped the vq.callback() that would have
completed the virtio I/O operation via virtblk_done().
[bhelgaas: commit log, simplify to always use pci_physfn(), add stable tag]
Link: https://lore.kernel.org/r/20221026060912.173250-1-mst@redhat.com
Reported-by: Wei Gong <gongwei833x@gmail.com>
Tested-by: Wei Gong <gongwei833x@gmail.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org
When the PHY is the reference clock provider then it must be initialized
and powered on before the reset on the client is deasserted, otherwise
the link will never come up. The order was changed in cf236e0c0d.
Restore the correct order to make the driver work again on boards where
the PHY provides the reference clock. This also changes the order for
boards where the Soc is the PHY reference clock divider, but this
shouldn't do any harm.
Link: https://lore.kernel.org/r/20221101095714.440001-1-s.hauer@pengutronix.de
Fixes: cf236e0c0d ("PCI: imx6: Do not hide PHY driver callbacks and refine the error handling")
Tested-by: Richard Zhu <hongxing.zhu@nxp.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Change to follow the Kconfig style guide. This patch fixes to use tab
rather than space to indent, while help text is indented an additional
two spaces.
Link: https://lore.kernel.org/r/20220815025006.48167-1-mie@igel.co.jp
Fixes: e35f56bb03 ("PCI: endpoint: Support NTB transfer between RC and EP")
Signed-off-by: Shunsuke Mie <mie@igel.co.jp>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
On Qualcomm platforms like SC8280XP and SA8540P, interconnect bandwidth
must be requested before enabling interconnect clocks.
Add basic support for managing an optional "pcie-mem" interconnect path
by setting a low constraint before enabling clocks and updating it after
the link is up.
Note that it is not possible for a controller driver to set anything but
a maximum peak bandwidth as expected average bandwidth will vary with
use case and actual use (and power policy?). This very much remains an
unresolved problem with the interconnect framework.
Also note that no constraint is set for the SC8280XP/SA8540P "cpu-pcie"
path for now as it is not clear what an appropriate constraint would be
(and the system does not crash when left unspecified).
Link: https://lore.kernel.org/r/20221102090705.23634-3-johan+linaro@kernel.org
Fixes: 70574511f3 ("PCI: qcom: Add support for SC8280XP")
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Brian Masney <bmasney@redhat.com>
Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>
Acked-by: Georgi Djakov <djakov@kernel.org>
MSI remapping is disabled by VMD driver for Intel's Icelake and
newer systems in order to improve performance by setting
VMCONFIG_MSI_REMAP. By design VMCONFIG_MSI_REMAP register is cleared
by firmware during boot. The same register gets cleared when system
is put in S3 power state. VMD driver needs to set this register again
in order to avoid interrupt issues with devices behind VMD if MSI
remapping was disabled before.
Link: https://lore.kernel.org/r/20221109142652.450998-1-nirmal.patel@linux.intel.com
Fixes: ee81ee84f8 ("PCI: vmd: Disable MSI-X remapping when possible")
Signed-off-by: Nirmal Patel <nirmal.patel@linux.intel.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Francisco Munoz <francisco.munoz.ruiz@linux.intel.com>
Set RCB_MPS mode bit so that data for PCIe read requests up to the size of
the Maximum Payload Size (MPS) are returned in one completion, and data for
PCIe read requests greater than the MPS are split at the specified Read
Completion Boundary setting.
Set RCB_64B so that the Read Compeletion Boundary is 64B.
Link: https://lore.kernel.org/r/20221011184211.18128-6-jim2101024@gmail.com
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
A number of inline functions are called rarely and/or are not
time-critical. Take out the "inline" and let the compiler do its work.
Link: https://lore.kernel.org/r/20221011184211.18128-5-jim2101024@gmail.com
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
It would be nice to replace the PCIe link-up loop as well but
there are too many uses of this that do not poll (and the
read_poll_timeout uses "timeout==0" to loop forever).
Link: https://lore.kernel.org/r/20221011184211.18128-4-jim2101024@gmail.com
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Be prudent and give some time for power and clocks to become stable. As
described in the PCIe CEM specification sections 2.2 and 2.2.1; as well as
PCIe r5.0, 6.6.1.
Link: https://lore.kernel.org/r/20221011184211.18128-3-jim2101024@gmail.com
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
We always wanted to enable Multi-MSI but didn't have a test device until
recently. In addition, there are some devices out there that will ask for
multiple MSI but refuse to work if they are only granted one.
Link: https://lore.kernel.org/r/20221011184211.18128-2-jim2101024@gmail.com
Signed-off-by: Jim Quinlan <jim2101024@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Many host controller drivers #include <linux/of_irq.h> even though they
don't need it. Remove the unnecessary #includes.
Link: https://lore.kernel.org/r/20221031153954.1163623-6-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Roy Zang <roy.zang@nxp.com>
pci-xgene-msi.c uses irq_domain_add_linear() and related interfaces, so it
needs <linux/irqdomain.h> but doesn't include it directly; it relies on the
fact that <linux/of_irq.h> includes it.
But pci-xgene-msi.c *doesn't* need <linux/of_irq.h> itself. Include
<linux/irqdomain.h> directly to remove this implicit dependency so a future
patch can drop <linux/of_irq.h>.
Link: https://lore.kernel.org/r/20221031153954.1163623-5-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
pci-mvebu.c uses irq_domain_add_linear() and related interfaces but relies
on <linux/irqdomain.h> but doesn't include it directly; it relies on the
fact that <linux/of_irq.h> includes it.
Include <linux/irqdomain.h> directly to remove this implicit dependency.
Link: https://lore.kernel.org/r/20221031153954.1163623-4-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
pcie-microchip-host.c uses irq_domain_add_linear() and related interfaces,
so it needs <linux/irqdomain.h> but doesn't include it directly; it relies
on the fact that <linux/of_irq.h> includes it.
But pcie-microchip-host.c *doesn't* need <linux/of_irq.h> itself. Include
<linux/irqdomain.h> directly to remove this implicit dependency so a future
patch can drop <linux/of_irq.h>.
Link: https://lore.kernel.org/r/20221031153954.1163623-3-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
pcie-altera-msi.c uses irq_domain_add_linear() and related interfaces, so
it needs <linux/irqdomain.h> but doesn't include it directly; it relies on
the fact that <linux/of_irq.h> includes it.
But pcie-altera-msi.c *doesn't* need <linux/of_irq.h> itself. Include
<linux/irqdomain.h> directly to remove this implicit dependency so a future
patch can drop <linux/of_irq.h>.
Link: https://lore.kernel.org/r/20221031153954.1163623-2-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Some of the platforms (like Tegra194 and Tegra234) have open slots and
not having an endpoint connected to the slot is not an error.
So, changing the macro from dev_err to dev_info to log the event.
Link: https://lore.kernel.org/r/20220913101237.4337-1-vidyas@nvidia.com
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Acked-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
When pci_create_attr() fails, pci_remove_resource_files() is called which
will iterate over the res_attr[_wc] arrays and frees every non NULL entry.
To avoid a double free here set the array entry only after it's clear we
successfully initialized it.
Fixes: b562ec8f74 ("PCI: Don't leak memory if sysfs_create_bin_file() fails")
Link: https://lore.kernel.org/r/20221007070735.GX986@pengutronix.de/
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org
Create a sysfs bin attribute called "allocate" under the existing
"p2pmem" group. The only allowable operation on this file is the mmap()
call.
When mmap() is called on this attribute, the kernel allocates a chunk of
memory from the genalloc and inserts the pages into the VMA. The
dev_pagemap .page_free callback will indicate when these pages are no
longer used and they will be put back into the genalloc.
On device unbind, remove the sysfs file before the memremap_pages are
cleaned up. This ensures unmap_mapping_range() is called on the files
inode and no new mappings can be created.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Link: https://lore.kernel.org/r/20221021174116.7200-9-logang@deltatee.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Replace assignment of PCI domain IDs from atomic_inc_return() to
ida_alloc().
Use two IDAs, one for static domain allocations (those which are defined in
device tree) and second for dynamic allocations (all other).
During removal of root bus / host bridge, also release the domain ID. The
released ID can be reused again, for example when dynamically loading and
unloading native PCI host bridge drivers.
This change also allows to mix static device tree assignment and dynamic by
kernel as all static allocations are reserved in dynamic pool.
[bhelgaas: set "err" if "bus->domain_nr < 0"]
Link: https://lore.kernel.org/r/20220714184130.5436-1-pali@kernel.org
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
PCIe r2.0, sec 7.8 added Link Capabilities/Status/Control 2 registers to
the PCIe Capability with Capability Version 2.
Previously we assumed these registers were implemented for all PCIe
Capabilities of version 2 or greater, but in fact they are only
implemented for devices with Links.
Update pcie_capability_reg_implemented() to check whether the device has
a Link.
[bhelgaas: commit log, squash export]
Link: https://lore.kernel.org/r/alpine.DEB.2.21.2209100057070.2275@angie.orcam.me.uk
Link: https://lore.kernel.org/r/alpine.DEB.2.21.2209100057300.2275@angie.orcam.me.uk
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The local variable 'vector' must be u32 rather than u8: see the
struct hv_msi_desc3.
'vector_count' should be u16 rather than u8: see struct hv_msi_desc,
hv_msi_desc2 and hv_msi_desc3.
Fixes: a2bad844a6 ("PCI: hv: Fix interrupt mapping for multi-MSI")
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Cc: Jeffrey Hugo <quic_jhugo@quicinc.com>
Cc: Carl Vanderlip <quic_carlv@quicinc.com>
Reviewed-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Link: https://lore.kernel.org/r/20221027205256.17678-1-decui@microsoft.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
The Requester ID/Process Address Space ID (PASID) combination
identifies an address space distinct from the PCI bus address space,
e.g., an address space defined by an IOMMU.
But the PCIe fabric routes Memory Requests based on the TLP address,
ignoring any PASID (PCIe r6.0, sec 2.2.10.4), so a TLP with PASID that
SHOULD go upstream to the IOMMU may instead be routed as a P2P
Request if its address falls in a bridge window.
To ensure that all Memory Requests with PASID are routed upstream,
only enable PASID if ACS P2P Request Redirect and Upstream Forwarding
are enabled for the path leading to the device.
Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Suggested-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Tested-by: Tony Zhu <tony.zhu@intel.com>
Link: https://lore.kernel.org/r/20221031005917.45690-5-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The pci_epf_test_notifier function should be installed also if only
core_init_notifier is enabled. Fix the current logic.
Link: https://lore.kernel.org/r/20220825090101.20474-1-hayashi.kunihiko@socionext.com
Fixes: 5e50ee27d4 ("PCI: pci-epf-test: Add support to defer core initialization")
Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Om Prakash Singh <omp@nvidia.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
[devm_]gpiod_get_from_of_node in drivers usage should be limited
so that gpiolib can be cleaned up; let's switch to the generic device
property API.
It may even help with handling secondary fwnodes when gpiolib is taught
to handle gpios described by swnodes.
Link: https://lore.kernel.org/r/20220903-gpiod_get_from_of_node-remove-v1-1-b29adfb27a6c@gmail.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
[lpieralisi@kernel.org: commit log]
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Since there is no release callback defined for the PCI EPC device,
the below warning is thrown by driver core when a PCI endpoint driver is
removed:
Device 'e65d0000.pcie-ep' does not have a release() function, it is broken and must be fixed. See Documentation/core-api/kobject.rst.
WARNING: CPU: 0 PID: 139 at drivers/base/core.c:2232 device_release+0x78/0x8c
Hence, add the release callback and also move the kfree(epc) from
pci_epc_destroy() so that the epc memory is freed when all references are
dropped.
Link: https://lore.kernel.org/r/20220623003817.298173-1-yoshihiro.shimoda.uh@renesas.com
Tested-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Dual mode DesignWare PCIe IP has PTM capability enabled (if supported) even
in the EP mode. The PCIe compliance for the EP mode expects PTM
capabilities (ROOT_CAPABLE, RES_CAPABLE, CLK_GRAN) be disabled.
Hence disable PTM for the EP mode.
Link: https://lore.kernel.org/r/20220919143340.4527-3-vidyas@nvidia.com
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Acked-by: Jingoo Han <jingoohan1@gmail.com>
commit aeaa0bfe89 ("PCI: dwc: Move N_FTS setup to common setup")
incorrectly uses pci->link_gen in deriving the index to the
n_fts[] array also introducing the issue of accessing beyond the
boundaries of array for greater than Gen-2 speeds. This change fixes
that issue.
Link: https://lore.kernel.org/r/20220926111923.22487-1-vidyas@nvidia.com
Fixes: aeaa0bfe89 ("PCI: dwc: Move N_FTS setup to common setup")
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Jingoo Han <jingoohan1@gmail.com>
1a1daf097e ("PCI/PM: Remove unused pci_driver.suspend_late() hook")
removed the legacy .suspend_late() hook, which was the only user of the
"state" parameter to pci_legacy_suspend_late(), but it neglected to remove
the parameter.
Remove the unused "state" parameter to pci_legacy_suspend_late().
Link: https://lore.kernel.org/r/20221025193502.669091-1-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
pcie_port_service_register() and pcie_port_service_unregister() are used
only by the pciehp, aer, dpc, and pme PCIe port service drivers, none of
which can be modules. Unexport pcie_port_service_register() and
pcie_port_service_unregister(). No functional change intended.
Link: https://lore.kernel.org/r/20221019204127.44463-4-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Previously several things used by portdrv_core.c and portdrv_pci.c were
shared by defining them in portdrv.h. Now that portdrv_core.c and
portdrv_pci.c have been squashed, move things that can be private into
portdrv.c. No functional change intended.
Link: https://lore.kernel.org/r/20221019204127.44463-3-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Squash portdrv_core.c and portdrv_pci.c into portdrv.c to make it easier to
find things. The whole thing is less than 1000 lines, and it's a pain to
bounce back and forth between two files.
Several portdrv_core.c functions were non-static because they were
referenced from portdrv_pci.c. Make them static since they're now all in
portdrv.c.
No functional change intended.
Link: https://lore.kernel.org/r/20221019204127.44463-2-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
This reverts commit 8bb7ff12a9.
Commit 8bb7ff12a9 ("PCI: tegra: Use PCI_CONF1_EXT_ADDRESS() macro")
updated the Tegra PCI driver to use the macro PCI_CONF1_EXT_ADDRESS()
instead of a local function in the Tegra PCI driver. This broke PCI for
some Tegra platforms because, when calculating the offset value, the mask
applied to the lower 8-bits changed from 0xff to 0xfc.
For now, fix this by reverting this commit.
Fixes: 8bb7ff12a9 ("PCI: tegra: Use PCI_CONF1_EXT_ADDRESS() macro")
Link: https://lore.kernel.org/r/20221017084006.11770-1-jonathanh@nvidia.com
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Call phy_set_mode_ext() to notify the PHY driver that the PHY is being
used in the EP mode.
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Jingoo Han <jingoohan1@gmail.com>
Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Acked-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Link: https://lore.kernel.org/r/20220927092207.161501-6-dmitry.baryshkov@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Call phy_set_mode_ext() to notify the PHY driver that the PHY is being
used in the RC mode.
Reviewed-by: Jingoo Han <jingoohan1@gmail.com>
Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Acked-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Link: https://lore.kernel.org/r/20220927092207.161501-5-dmitry.baryshkov@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
-----BEGIN PGP SIGNATURE-----
iQJIBAABCgAyFiEEgMe7l+5h9hnxdsnuWYigwDrT+vwFAmNKzlUUHGJoZWxnYWFz
QGdvb2dsZS5jb20ACgkQWYigwDrT+vxBbA//QYuw8iZHzshJntCyWlTRg742x2jI
2NrLojVG/ccSlqnl/KRultaPFdSlZb221UASt7CAEIGoP6+/SNepazEEJ4QJhMX6
inY/hYrVd2OHhCnO+OoNd7DKfbeJRkFVMHOGJdpbQ7jMwa99YSaN6xEMfALcZIrg
8CqYKDpdJW4avYyKlfXayY42d8gljWJwfLuK1xEnnIrdvXQEB3vj3yKxziadhBtA
EJbLWORVF+yQwq63NCPkLvCe0ZuV09a4q4IhlzWo+30FLMOsk2z4HBi1B2fwlawV
tnYCt0i6FNC6e3DJyWX/hkAiIbJUadnych+LcGq+/FY30tCBtNkNaQN3XsrvdvI4
5piacj1av21R7JHSyk64M4jdNlE7MYonQNsc8vp4yDtlAjQYkaj7BJdsIq3MCTei
ce6mxguVb2hrMr4fZXc8ka7XZkofJNv/XhfuQAorzKfEqA8cI4enICsPdoTFyZDV
pYYuHIJXY4rTm0wUD+TlDg/Cn6gyLu3WYVyEn6q/E53Twj1rr5ske7igDMDVYYly
qMg8CTPHYzHJ0wbzkf0CfND7SVAAE1BnIzr4Drfbrm+fi11HVhd0u/0XiWf2NthW
/ZFQUDBh0SwSHgUnO+ONqcR3arhIT4PEsEJadD4kuW4nIyPS7t1Nfu0V3XUZDa14
U6UeL6URnXBDV7U=
=gs3v
-----END PGP SIGNATURE-----
Merge tag 'pci-v6.1-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull pci fix from Bjorn Helgaas:
"Revert the attempt to distribute spare resources to unconfigured
hotplug bridges at boot time.
This fixed some dock hot-add scenarios, but Jonathan Cameron reported
that it broke a topology with a multi-function device where one
function was a Switch Upstream Port and the other was an Endpoint"
* tag 'pci-v6.1-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci:
Revert "PCI: Distribute available resources for root buses, too"
This reverts commit e96e27fc6f.
Jonathan reported that this commit broke this topology, where all the space
available on bus 02 was assigned to the 02:00.0 bridge window, leaving none
for the e1000 device at 02:00.1:
pci 0000:00:04.0: bridge window [mem 0x10200000-0x103fffff] to [bus 02-04]
pci 0000:02:00.0: bridge window [mem 0x10200000-0x103fffff] to [bus 03-04]
pci 0000:02:00.1: BAR 0: failed to assign [mem size 0x00020000]
e1000 0000:02:00.1: can't ioremap BAR 0: [??? 0x00000000 flags 0x0]
Link: https://lore.kernel.org/r/20221014124553.0000696f@huawei.com
Reported-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQRTLbB6QfY48x44uB6AXGG7T9hjvgUCY0ZjFAAKCRCAXGG7T9hj
vjEsAP4rFMnqc6AXy4Mpvv8cxBtEuQZbwEqgBrMJUvK1jZQrBQD/dOJK2GBCVcfD
2yaVlefFiJGTw5WUlbPeohUlTZ8pJwg=
=xsHV
-----END PGP SIGNATURE-----
Merge tag 'for-linus-6.1-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen updates from Juergen Gross:
- Some minor typo fixes
- A fix of the Xen pcifront driver for supporting the device model to
run in a Linux stub domain
- A cleanup of the pcifront driver
- A series to enable grant-based virtio with Xen on x86
- A cleanup of Xen PV guests to distinguish between safe and faulting
MSR accesses
- Two fixes of the Xen gntdev driver
- Two fixes of the new xen grant DMA driver
* tag 'for-linus-6.1-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: Kconfig: Fix spelling mistake "Maxmium" -> "Maximum"
xen/pv: support selecting safe/unsafe msr accesses
xen/pv: refactor msr access functions to support safe and unsafe accesses
xen/pv: fix vendor checks for pmu emulation
xen/pv: add fault recovery control to pmu msr accesses
xen/virtio: enable grant based virtio on x86
xen/virtio: use dom0 as default backend for CONFIG_XEN_VIRTIO_FORCE_GRANT
xen/virtio: restructure xen grant dma setup
xen/pcifront: move xenstore config scanning into sub-function
xen/gntdev: Accommodate VMA splitting
xen/gntdev: Prevent leaking grants
xen/virtio: Fix potential deadlock when accessing xen_grant_dma_devices
xen/virtio: Fix n_pages calculation in xen_grant_dma_map(unmap)_page()
xen/xenbus: Fix spelling mistake "hardward" -> "hardware"
xen-pcifront: Handle missed Connected state
-----BEGIN PGP SIGNATURE-----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=pOqY
-----END PGP SIGNATURE-----
Merge tag 'pci-v6.1-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull pci updates from Bjorn Helgaas:
"Resource management:
- Distribute spare resources to unconfigured hotplug bridges at
boot-time (not just when hot-adding such a bridge), which makes
hot-adding devices to docks work better.
- Revert to a BAR assignment inherited from firmware only when the
address is actually reachable via any upstream bridges, which fixes
some cases where firmware doesn't configure all devices.
- Add a sysfs interface to resize BARs so this can be done before
assigning devices to a VM through VFIO.
Power management:
- Disable Precision Time Management for all devices on suspend to
enable lower-power PM state. We previously did this just for Root
Ports, which isn't enough because downstream devices can still
generate PTM messages, which cause errors if it's disabled in the
Root Port.
- Save and restore the ASPM L1 PM Substates configuration for
suspend/ resume. Previously this configuration was lost, so L1.x
states likely stopped working after resume.
- Check whether the L1 PM Substates Capability exists. If it didn't
exist, we previously read junk and tried to configure L1 Substates
based on that.
- Fix the LTR_L1.2_THRESHOLD computation, which previously set a
threshold for entering L1.2 that was too low in some cases.
- Reduce the delay after transitions to or from D3cold by using
usleep_range() rather than msleep(), which often slept for ~19ms
instead of the 10ms normally required. The spec says 10ms is
enough, but it's possible we could trip over devices that need a
little more.
Error handling:
- Work around a BIOS bug that caused Intel Root Ports to advertise a
Root Port Programmed I/O (RP PIO) log size of zero, which caused
annoying warnings and prevented the kernel from dumping log
registers for DPC errors.
Qualcomm PCIe controller driver:
- Add support for SC8280XP and SA8540P host controllers and SM8450
endpoint controller.
- Disable Master AXI clock on endpoint controllers to save power when
link is idle or in L1.x.
- Expose link state transition counts via debugfs to help debug
issues with low-power states.
- Add auto-loading module support.
Synopsys DesignWare PCIe controller driver:
- Remove a dependency on ZONE_DMA32 by allocating the MSI target page
differently. There's more work to do related to eDMA controllers,
so it's not completely settled"
* tag 'pci-v6.1-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (71 commits)
PCI: qcom-ep: Check platform_get_resource_byname() return value
PCI: qcom-ep: Add support for SM8450 SoC
dt-bindings: PCI: qcom-ep: Add support for SM8450 SoC
dt-bindings: PCI: qcom-ep: Define clocks per platform
PCI: qcom-ep: Make PERST separation optional
dt-bindings: PCI: qcom-ep: Make PERST separation optional
PCI: qcom-ep: Disable Master AXI Clock when there is no PCIe traffic
PCI: Expose PCIe Resizable BAR support via sysfs
PCI/ASPM: Correct LTR_L1.2_THRESHOLD computation
PCI/ASPM: Ignore L1 PM Substates if device lacks capability
PCI/ASPM: Factor out L1 PM Substates configuration
PCI: qcom-ep: Gate Master AXI clock to MHI bus during L1SS
PCI: qcom-ep: Expose link transition counts via debugfs
PCI: qcom-ep: Disable IRQs during driver remove
PCI/ASPM: Save L1 PM Substates Capability for suspend/resume
PCI/ASPM: Refactor L1 PM Substates Control Register programming
PCI: qcom-ep: Make use of the cached dev pointer
PCI: qcom-ep: Rely on the clocks supplied by devicetree
PCI: qcom-ep: Add kernel-doc for qcom_pcie_ep structure
phy: freescale: imx8m-pcie: Fix the wrong order of phy_init() and phy_power_on()
...
pcifront_try_connect() and pcifront_attach_devices() share a large
chunk of duplicated code for reading the config information from
Xenstore, which only differs regarding calling pcifront_rescan_root()
or pcifront_scan_root().
Put that code into a new sub-function. It is fine to always call
pcifront_rescan_root() from that common function, as it will fallback
to pcifront_scan_root() if the domain/bus combination isn't known
yet (and pcifront_scan_root() should never be called for an already
known domain/bus combination anyway). In order to avoid duplicate
messages for the fallback case move the check for domain/bus not known
to the beginning of pcifront_rescan_root().
While at it fix the error reporting in case the root-xx node had the
wrong format.
As the return value of pcifront_try_connect() and
pcifront_attach_devices() are not used anywhere make those functions
return void. As an additional bonus this removes the dubious return
of -EFAULT in case of an unexpected driver state.
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jason Andryuk <jandryuk@gmail.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
- Add macros for PCI Configuration Mechanism #1 and use them in the
ftpci100, mt7621, and tegra drivers (Pali Rohár)
* remotes/lorenzo/pci/misc:
PCI: tegra: Use PCI_CONF1_EXT_ADDRESS() macro
PCI: mt7621: Use PCI_CONF1_EXT_ADDRESS() macro
PCI: ftpci100: Use PCI_CONF1_ADDRESS() macro
PCI: Add standard PCI Config Address macros
- List platforms that use a single MSI host interrupt in qcom DT (Johan
Hovold)
- Add SC8280XP, SA8540P support to qcom DT binding and driver(Johan Hovold)
- Make all optional clocks truly optional in the driver (Johan Hovold)
- Rename per-IP structs to reflect the IP version (Johan Hovold)
- Sort device ID match table by compatible string (Johan Hovold)
- Add MODULE_DEVICE_TABLE to enable module autoloading (Dmitry Baryshkov)
- Drop the unused .post_deinit() callback (Johan Hovold)
- Rely on DT for clock information instead of hard-coding it in the driver
(Manivannan Sadhasivam)
- Disable IRQs when removing driver to avoid spurious IRQs later
(Manivannan Sadhasivam)
- Expose link transition counts via debugfs to help debug issues with
low-power states (Manivannan Sadhasivam)
- Gate Master AXI clock to the MHI bus while in L1 substates to save power
(Manivannan Sadhasivam)
- Disable Master AXI clock to save power when there is no traffic on PCIe
(Manivannan Sadhasivam)
- Make the "PERST separation" debug feature optional in the DT and the
driver (Manivannan Sadhasivam)
- Define clocks to be per-platform in DT to prepare for future SoCs
(Manivannan Sadhasivam)
- Add SM8450 SoC support (Manivannan Sadhasivam)
- Check for platform_get_resource_byname() to avoid a NULL pointer
dereference (Yang Yingliang)
* pci/qcom:
PCI: qcom-ep: Check platform_get_resource_byname() return value
PCI: qcom-ep: Add support for SM8450 SoC
dt-bindings: PCI: qcom-ep: Add support for SM8450 SoC
dt-bindings: PCI: qcom-ep: Define clocks per platform
PCI: qcom-ep: Make PERST separation optional
dt-bindings: PCI: qcom-ep: Make PERST separation optional
PCI: qcom-ep: Disable Master AXI Clock when there is no PCIe traffic
PCI: qcom-ep: Gate Master AXI clock to MHI bus during L1SS
PCI: qcom-ep: Expose link transition counts via debugfs
PCI: qcom-ep: Disable IRQs during driver remove
PCI: qcom-ep: Make use of the cached dev pointer
PCI: qcom-ep: Rely on the clocks supplied by devicetree
PCI: qcom-ep: Add kernel-doc for qcom_pcie_ep structure
PCI: qcom: Rename host-init error label
PCI: qcom: Drop unused post_deinit callback
PCI: qcom-ep: Add MODULE_DEVICE_TABLE
PCI: qcom: Sort device-id table
PCI: qcom: Clean up IP configurations
PCI: qcom: Make all optional clocks optional
PCI: qcom: Add support for SA8540P
PCI: qcom: Add support for SC8280XP
dt-bindings: PCI: qcom: Add SA8540P to binding
dt-bindings: PCI: qcom: Add SC8280XP to binding
dt-bindings: PCI: qcom: Enumerate platforms with single msi interrupt
- Rename the pcie-mediatek-gen3 driver from 'mtk-pcie' to 'mtk-pcie-gen3'
so it can coexist with the pcie-mediatek driver, which also uses
'mtk-pcie' (Felix Fietkau)
* remotes/lorenzo/pci/mediatek:
PCI: mediatek-gen3: Change driver name to mtk-pcie-gen3
- Use dmam_alloc_coherent() instead of dma_map_page() to allocate the MSI
target page, which means dwc drivers will work even when ZONE_DMA32 is
disabled (Will McVicker)
- If we can't allocate an MSI target page with a 32-bit address, try
allocating one with a 64-bit address (Will McVicker)
- Switch from of_gpio_named_count() to generic gpiod_count() (Andy
Shevchenko)
- Add support for i.MX8MP PCIe (Richard Zhu)
- Fix the Freescale i.MX8 PHY driver, which had interchanged the phy_init()
and phy_power_on() interfaces (Richard Zhu)
* remotes/lorenzo/pci/dwc:
phy: freescale: imx8m-pcie: Fix the wrong order of phy_init() and phy_power_on()
PCI: imx6: Add i.MX8MP PCIe support
PCI: dwc: Replace of_gpio_named_count() by gpiod_count()
PCI: dwc: Drop dependency on ZONE_DMA32
- In an emulated PCI bridge, set Capability offsets so they match the
hardware offsets shown by U-Boot (Pali Rohár)
* remotes/lorenzo/pci/bridge-emul:
PCI: pci-bridge-emul: Set position of PCI capabilities to real HW value
- Switch from gpiod_get_from_of_node() to generic devm GPIO API (Dmitry
Torokhov)
* remotes/lorenzo/pci/apple:
PCI: apple: Do not leak reset GPIO on unbind/unload/error
- Emulate the PCI Bridge Subsystem Vendor ID (Pali Rohár)
* remotes/lorenzo/pci/aardvark:
PCI: aardvark: Add support for PCI Bridge Subsystem Vendor ID on emulated bridge
- Distribute resources to unconfigured hotplug bridges at boot-time (not
just when hot-adding such a bridge), which makes hot-adding devices to
docks work (Mika Westerberg)
- Fix the "revert to firmware assignment" code so we do the revert only if
the address is actually reachable. Previously we sometimes assigned
addresses that could not be reached via upstream bridges (Maciej W.
Rozycki)
* pci/resource:
PCI: Sanitise firmware BAR assignments behind a PCI-PCI bridge
PCI: Fix typo in pci_scan_child_bus_extend()
PCI: Fix whitespace and indentation
PCI: Distribute available resources for root buses, too
PCI: Move pci_assign_unassigned_root_bus_resources()
PCI: Pass available buses even if the bridge is already configured
PCI: Fix used_buses calculation in pci_scan_child_bus_extend()
- Expose a sysfs interface for configuring Resizable BARs so we can resize
BARs before assigning devices to a VM through VFIO (Alex Williamson)
* pci/rebar:
PCI: Expose PCIe Resizable BAR support via sysfs
- Cache the PTM capability offset instead of searching for it every time
(Bjorn Helgaas)
- Separate PTM configuration from PTM enable (Bjorn Helgaas)
- Add pci_suspend_ptm() and pci_resume_ptm() to disable and re-enable PTM
on suspend/resume so some Root Ports can safely enter a lower-power PM
state (Bjorn Helgaas)
- Disable PTM for all devices during suspend; previously we only did this
for Root Ports and even then only in certain cases (Bjorn Helgaas)
- Simplify pci_pm_suspend_noirq() (Rajvi Jingar)
- Reduce the delay after transitions to/from D3hot by using usleep_range()
instead of msleep(), which reduces the typical delay from 19ms to 10ms
(Sajid Dalvi, Will McVicker)
* pci/pm:
PCI/PM: Reduce D3hot delay with usleep_range()
PCI/PM: Simplify pci_pm_suspend_noirq()
PCI/PM: Always disable PTM for all devices during suspend
PCI/PTM: Consolidate PTM interface declarations
PCI/PTM: Reorder functions in logical order
PCI/PTM: Preserve RsvdP bits in PTM Control register
PCI/PTM: Move pci_ptm_info() body into its only caller
PCI/PTM: Add pci_suspend_ptm() and pci_resume_ptm()
PCI/PTM: Separate configuration and enable
PCI/PTM: Add pci_upstream_ptm() helper
PCI/PTM: Cache PTM Capability offset
- Correct a typo in 71020a3c0d ('PCI/MSI: Use msi_add_msi_desc()') that
reversed the sense of 'can_mask' in msi_add_msi_desc() (Josef Johansson)
* pci/msi:
PCI/MSI: Correct 'can_mask' test in msi_add_msi_desc()
- Work around a BIOS defect that makes some Intel Root Ports report an RP
PIO log size of zero (Mika Westerberg)
* pci/dpc:
PCI/DPC: Quirk PIO log size for certain Intel Root Ports