fe3ff217dd
enclosures in iovctl.conf(5), iovctl(8), pci(9), and pci_iov_schema(9). Differential Revision: https://reviews.freebsd.org/D3000 Reviewed by: wblock Approved by: jmallett (mentor)
786 lines
20 KiB
Groff
786 lines
20 KiB
Groff
.\"
|
|
.\" Copyright (c) 2005 Bruce M Simpson <bms@FreeBSD.org>
|
|
.\" All rights reserved.
|
|
.\"
|
|
.\" Redistribution and use in source and binary forms, with or without
|
|
.\" modification, are permitted provided that the following conditions
|
|
.\" are met:
|
|
.\" 1. Redistributions of source code must retain the above copyright
|
|
.\" notice, this list of conditions and the following disclaimer.
|
|
.\" 2. Redistributions in binary form must reproduce the above copyright
|
|
.\" notice, this list of conditions and the following disclaimer in the
|
|
.\" documentation and/or other materials provided with the distribution.
|
|
.\"
|
|
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
.\" SUCH DAMAGE.
|
|
.\"
|
|
.\" $FreeBSD$
|
|
.\"
|
|
.Dd July 8, 2015
|
|
.Dt PCI 9
|
|
.Os
|
|
.Sh NAME
|
|
.Nm pci ,
|
|
.Nm pci_alloc_msi ,
|
|
.Nm pci_alloc_msix ,
|
|
.Nm pci_disable_busmaster ,
|
|
.Nm pci_disable_io ,
|
|
.Nm pci_enable_busmaster ,
|
|
.Nm pci_enable_io ,
|
|
.Nm pci_find_bsf ,
|
|
.Nm pci_find_cap ,
|
|
.Nm pci_find_dbsf ,
|
|
.Nm pci_find_device ,
|
|
.Nm pci_find_extcap ,
|
|
.Nm pci_find_htcap ,
|
|
.Nm pci_get_max_read_req ,
|
|
.Nm pci_get_powerstate ,
|
|
.Nm pci_get_vpd_ident ,
|
|
.Nm pci_get_vpd_readonly ,
|
|
.Nm pci_iov_attach ,
|
|
.Nm pci_iov_detach ,
|
|
.Nm pci_msi_count ,
|
|
.Nm pci_msix_count ,
|
|
.Nm pci_pending_msix ,
|
|
.Nm pci_read_config ,
|
|
.Nm pci_release_msi ,
|
|
.Nm pci_remap_msix ,
|
|
.Nm pci_restore_state ,
|
|
.Nm pci_save_state ,
|
|
.Nm pci_set_max_read_req ,
|
|
.Nm pci_set_powerstate ,
|
|
.Nm pci_write_config
|
|
.Nd PCI bus interface
|
|
.Sh SYNOPSIS
|
|
.In sys/bus.h
|
|
.In dev/pci/pcireg.h
|
|
.In dev/pci/pcivar.h
|
|
.Ft int
|
|
.Fn pci_alloc_msi "device_t dev" "int *count"
|
|
.Ft int
|
|
.Fn pci_alloc_msix "device_t dev" "int *count"
|
|
.Ft int
|
|
.Fn pci_disable_busmaster "device_t dev"
|
|
.Ft int
|
|
.Fn pci_disable_io "device_t dev" "int space"
|
|
.Ft int
|
|
.Fn pci_enable_busmaster "device_t dev"
|
|
.Ft int
|
|
.Fn pci_enable_io "device_t dev" "int space"
|
|
.Ft device_t
|
|
.Fn pci_find_bsf "uint8_t bus" "uint8_t slot" "uint8_t func"
|
|
.Ft int
|
|
.Fn pci_find_cap "device_t dev" "int capability" "int *capreg"
|
|
.Ft device_t
|
|
.Fn pci_find_dbsf "uint32_t domain" "uint8_t bus" "uint8_t slot" "uint8_t func"
|
|
.Ft device_t
|
|
.Fn pci_find_device "uint16_t vendor" "uint16_t device"
|
|
.Ft int
|
|
.Fn pci_find_extcap "device_t dev" "int capability" "int *capreg"
|
|
.Ft int
|
|
.Fn pci_find_htcap "device_t dev" "int capability" "int *capreg"
|
|
.Ft int
|
|
.Fn pci_get_max_read_req "device_t dev"
|
|
.Ft int
|
|
.Fn pci_get_powerstate "device_t dev"
|
|
.Ft int
|
|
.Fn pci_get_vpd_ident "device_t dev" "const char **identptr"
|
|
.Ft int
|
|
.Fn pci_get_vpd_readonly "device_t dev" "const char *kw" "const char **vptr"
|
|
.Ft int
|
|
.Fn pci_msi_count "device_t dev"
|
|
.Ft int
|
|
.Fn pci_msix_count "device_t dev"
|
|
.Ft int
|
|
.Fn pci_pending_msix "device_t dev" "u_int index"
|
|
.Ft uint32_t
|
|
.Fn pci_read_config "device_t dev" "int reg" "int width"
|
|
.Ft int
|
|
.Fn pci_release_msi "device_t dev"
|
|
.Ft int
|
|
.Fn pci_remap_msix "device_t dev" "int count" "const u_int *vectors"
|
|
.Ft void
|
|
.Fn pci_restore_state "device_t dev"
|
|
.Ft void
|
|
.Fn pci_save_state "device_t dev"
|
|
.Ft int
|
|
.Fn pci_set_max_read_req "device_t dev" "int size"
|
|
.Ft int
|
|
.Fn pci_set_powerstate "device_t dev" "int state"
|
|
.Ft void
|
|
.Fn pci_write_config "device_t dev" "int reg" "uint32_t val" "int width"
|
|
.In dev/pci/pci_iov.h
|
|
.Ft int
|
|
.Fn pci_iov_attach "device_t dev" "nvlist_t *pf_schema" "nvlist_t *vf_schema"
|
|
.Ft int
|
|
.Fn pci_iov_detach "device_t dev"
|
|
.Sh DESCRIPTION
|
|
The
|
|
.Nm
|
|
set of functions are used for managing PCI devices.
|
|
The functions are split into several groups:
|
|
raw configuration access,
|
|
locating devices,
|
|
device information,
|
|
device configuration,
|
|
and
|
|
message signaled interrupts.
|
|
.Ss Raw Configuration Access
|
|
The
|
|
.Fn pci_read_config
|
|
function is used to read data from the PCI configuration
|
|
space of the device
|
|
.Fa dev ,
|
|
at offset
|
|
.Fa reg ,
|
|
with
|
|
.Fa width
|
|
specifying the size of the access.
|
|
.Pp
|
|
The
|
|
.Fn pci_write_config
|
|
function is used to write the value
|
|
.Fa val
|
|
to the PCI configuration
|
|
space of the device
|
|
.Fa dev ,
|
|
at offset
|
|
.Fa reg ,
|
|
with
|
|
.Fa width
|
|
specifying the size of the access.
|
|
.Pp
|
|
.Em NOTE :
|
|
Device drivers should only use these functions for functionality that
|
|
is not available via another
|
|
.Fn pci
|
|
function.
|
|
.Ss Locating Devices
|
|
The
|
|
.Fn pci_find_bsf
|
|
function looks up the
|
|
.Vt device_t
|
|
of a PCI device, given its
|
|
.Fa bus ,
|
|
.Fa slot ,
|
|
and
|
|
.Fa func .
|
|
The
|
|
.Fa slot
|
|
number actually refers to the number of the device on the bus,
|
|
which does not necessarily indicate its geographic location
|
|
in terms of a physical slot.
|
|
Note that in case the system has multiple PCI domains,
|
|
the
|
|
.Fn pci_find_bsf
|
|
function only searches the first one.
|
|
Actually, it is equivalent to:
|
|
.Bd -literal -offset indent
|
|
pci_find_dbsf(0, bus, slot, func);
|
|
.Ed
|
|
.Pp
|
|
The
|
|
.Fn pci_find_dbsf
|
|
function looks up the
|
|
.Vt device_t
|
|
of a PCI device, given its
|
|
.Fa domain ,
|
|
.Fa bus ,
|
|
.Fa slot ,
|
|
and
|
|
.Fa func .
|
|
The
|
|
.Fa slot
|
|
number actually refers to the number of the device on the bus,
|
|
which does not necessarily indicate its geographic location
|
|
in terms of a physical slot.
|
|
.Pp
|
|
The
|
|
.Fn pci_find_device
|
|
function looks up the
|
|
.Vt device_t
|
|
of a PCI device, given its
|
|
.Fa vendor
|
|
and
|
|
.Fa device
|
|
IDs.
|
|
Note that there can be multiple matches for this search; this function
|
|
only returns the first matching device.
|
|
.Ss Device Information
|
|
The
|
|
.Fn pci_find_cap
|
|
function is used to locate the first instance of a PCI capability
|
|
register set for the device
|
|
.Fa dev .
|
|
The capability to locate is specified by ID via
|
|
.Fa capability .
|
|
Constant macros of the form
|
|
.Dv PCIY_xxx
|
|
for standard capability IDs are defined in
|
|
.In dev/pci/pcireg.h .
|
|
If the capability is found, then
|
|
.Fa *capreg
|
|
is set to the offset in configuration space of the capability register set,
|
|
and
|
|
.Fn pci_find_cap
|
|
returns zero.
|
|
If the capability is not found or the device does not support capabilities,
|
|
.Fn pci_find_cap
|
|
returns an error.
|
|
.Pp
|
|
The
|
|
.Fn pci_find_extcap
|
|
function is used to locate the first instance of a PCI-express
|
|
extended capability register set for the device
|
|
.Fa dev .
|
|
The extended capability to locate is specified by ID via
|
|
.Fa capability .
|
|
Constant macros of the form
|
|
.Dv PCIZ_xxx
|
|
for standard extended capability IDs are defined in
|
|
.In dev/pci/pcireg.h .
|
|
If the extended capability is found, then
|
|
.Fa *capreg
|
|
is set to the offset in configuration space of the extended capability
|
|
register set, and
|
|
.Fn pci_find_extcap
|
|
returns zero.
|
|
If the extended capability is not found or the device is not a
|
|
PCI-express device,
|
|
.Fn pci_find_extcap
|
|
returns an error.
|
|
.Pp
|
|
The
|
|
.Fn pci_find_htcap
|
|
function is used to locate the first instance of a HyperTransport capability
|
|
register set for the device
|
|
.Fa dev .
|
|
The capability to locate is specified by type via
|
|
.Fa capability .
|
|
Constant macros of the form
|
|
.Dv PCIM_HTCAP_xxx
|
|
for standard HyperTransport capability types are defined in
|
|
.In dev/pci/pcireg.h .
|
|
If the capability is found, then
|
|
.Fa *capreg
|
|
is set to the offset in configuration space of the capability register set,
|
|
and
|
|
.Fn pci_find_htcap
|
|
returns zero.
|
|
If the capability is not found or the device is not a HyperTransport device,
|
|
.Fn pci_find_htcap
|
|
returns an error.
|
|
.Pp
|
|
The
|
|
.Fn pci_get_vpd_ident
|
|
function is used to fetch a device's Vital Product Data
|
|
.Pq VPD
|
|
identifier string.
|
|
If the device
|
|
.Fa dev
|
|
supports VPD and provides an identifier string,
|
|
then
|
|
.Fa *identptr
|
|
is set to point at a read-only, null-terminated copy of the identifier
|
|
string,
|
|
and
|
|
.Fn pci_get_vpd_ident
|
|
returns zero.
|
|
If the device does not support VPD or does not provide an identifier
|
|
string,
|
|
then
|
|
.Fn pci_get_vpd_ident
|
|
returns an error.
|
|
.Pp
|
|
The
|
|
.Fn pci_get_vpd_readonly
|
|
function is used to fetch the value of a single VPD read-only keyword
|
|
for the device
|
|
.Fa dev .
|
|
The keyword to fetch is identified by the two character string
|
|
.Fa kw .
|
|
If the device supports VPD and provides a read-only value for the
|
|
requested keyword,
|
|
then
|
|
.Fa *vptr
|
|
is set to point at a read-only, null-terminated copy of the value,
|
|
and
|
|
.Fn pci_get_vpd_readonly
|
|
returns zero.
|
|
If the device does not support VPD or does not provide the requested
|
|
keyword,
|
|
then
|
|
.Fn pci_get_vpd_readonly
|
|
returns an error.
|
|
.Ss Device Configuration
|
|
The
|
|
.Fn pci_enable_busmaster
|
|
function enables PCI bus mastering for the device
|
|
.Fa dev ,
|
|
by setting the
|
|
.Dv PCIM_CMD_BUSMASTEREN
|
|
bit in the
|
|
.Dv PCIR_COMMAND
|
|
register.
|
|
The
|
|
.Fn pci_disable_busmaster
|
|
function clears this bit.
|
|
.Pp
|
|
The
|
|
.Fn pci_enable_io
|
|
function enables memory or I/O port address decoding for the device
|
|
.Fa dev ,
|
|
by setting the
|
|
.Dv PCIM_CMD_MEMEN
|
|
or
|
|
.Dv PCIM_CMD_PORTEN
|
|
bit in the
|
|
.Dv PCIR_COMMAND
|
|
register appropriately.
|
|
The
|
|
.Fn pci_disable_io
|
|
function clears the appropriate bit.
|
|
The
|
|
.Fa space
|
|
argument specifies which resource is affected; this can be either
|
|
.Dv SYS_RES_MEMORY
|
|
or
|
|
.Dv SYS_RES_IOPORT
|
|
as appropriate.
|
|
Device drivers should generally not use these routines directly.
|
|
The PCI bus will enable decoding automatically when a
|
|
.Dv SYS_RES_MEMORY
|
|
or
|
|
.Dv SYS_RES_IOPORT
|
|
resource is activated via
|
|
.Xr bus_alloc_resource 9
|
|
or
|
|
.Xr bus_activate_resource 9 .
|
|
.Pp
|
|
The
|
|
.Fn pci_get_max_read_req
|
|
function returns the current maximum read request size in bytes for a
|
|
PCI-express device.
|
|
If the
|
|
.Fa dev
|
|
device is not a PCI-express device,
|
|
.Fn pci_get_max_read_req
|
|
returns zero.
|
|
.Pp
|
|
The
|
|
.Fn pci_set_max_read_req
|
|
sets the PCI-express maximum read request size for
|
|
.Fa dev .
|
|
The requested
|
|
.Fa size
|
|
may be adjusted,
|
|
and
|
|
.Fn pci_set_max_read_req
|
|
returns the actual size set in bytes.
|
|
If the
|
|
.Fa dev
|
|
device is not a PCI-express device,
|
|
.Fn pci_set_max_read_req
|
|
returns zero.
|
|
.Pp
|
|
The
|
|
.Fn pci_get_powerstate
|
|
function returns the current power state of the device
|
|
.Fa dev .
|
|
If the device does not support power management capabilities, then the default
|
|
state of
|
|
.Dv PCI_POWERSTATE_D0
|
|
is returned.
|
|
The following power states are defined by PCI:
|
|
.Bl -hang -width ".Dv PCI_POWERSTATE_UNKNOWN"
|
|
.It Dv PCI_POWERSTATE_D0
|
|
State in which device is on and running.
|
|
It is receiving full power from the system and delivering
|
|
full functionality to the user.
|
|
.It Dv PCI_POWERSTATE_D1
|
|
Class-specific low-power state in which device context may or
|
|
may not be lost.
|
|
Busses in this state cannot do anything to the bus, to
|
|
force devices to lose context.
|
|
.It Dv PCI_POWERSTATE_D2
|
|
Class-specific low-power state in which device context may or
|
|
may not be lost.
|
|
Attains greater power savings than
|
|
.Dv PCI_POWERSTATE_D1 .
|
|
Busses in this state can cause devices to lose some context.
|
|
Devices
|
|
.Em must
|
|
be prepared for the bus to be in this state or higher.
|
|
.It Dv PCI_POWERSTATE_D3
|
|
State in which the device is off and not running.
|
|
Device context is lost, and power from the device can
|
|
be removed.
|
|
.It Dv PCI_POWERSTATE_UNKNOWN
|
|
State of the device is unknown.
|
|
.El
|
|
.Pp
|
|
The
|
|
.Fn pci_set_powerstate
|
|
function is used to transition the device
|
|
.Fa dev
|
|
to the PCI power state
|
|
.Fa state .
|
|
If the device does not support power management capabilities or
|
|
it does not support the specific power state
|
|
.Fa state ,
|
|
then the function will fail with
|
|
.Er EOPNOTSUPP .
|
|
.Pp
|
|
The
|
|
.Fn pci_iov_attach
|
|
function is used to advertise that the given device
|
|
.Pq and associated device driver
|
|
supports PCI Single-Root I/O Virtualization
|
|
.Pq SR-IOV .
|
|
A driver that supports SR-IOV must implement the
|
|
.Xr PCI_IOV_INIT 9 ,
|
|
.Xr PCI_IOV_ADD_VF 9
|
|
and
|
|
.Xr PCI_IOV_UNINIT 9
|
|
methods.
|
|
This function should be called during the
|
|
.Xr DEVICE_ATTACH 9
|
|
method.
|
|
If this function returns an error, it is recommended that the device driver
|
|
still successfully attaches, but runs with SR-IOV disabled.
|
|
The
|
|
.Fa pf_schema
|
|
and
|
|
.Fa vf_schema
|
|
parameters are used to define what device-specific configuration parameters the
|
|
device driver accepts when SR-IOV is enabled for the Physical Function
|
|
.Pq PF
|
|
and for individual Virtual Functions
|
|
.Pq VFs
|
|
respectively.
|
|
See
|
|
.Xr pci_iov_schema 9
|
|
for details on how to construct the schema.
|
|
If either the
|
|
.Pa pf_schema
|
|
or
|
|
.Pa vf_schema
|
|
is invalid or specifies parameter names that conflict with parameter names that
|
|
are already in use,
|
|
.Fn pci_iov_attach
|
|
will return an error and SR-IOV will not be available on the PF device.
|
|
If a driver does not accept configuration parameters for either the PF device
|
|
or the VF devices, the driver must pass an empty schema for that device.
|
|
The SR-IOV infrastructure takes ownership of the
|
|
.Fa pf_schema
|
|
and
|
|
.Fa vf_schema
|
|
and is responsible for freeing them.
|
|
The driver must never free the schemas itself.
|
|
.Pp
|
|
The
|
|
.Fn pci_iov_detach
|
|
function is used to advise the SR-IOV infrastructure that the driver for the
|
|
given device is attempting to detach and that all SR-IOV resources for the
|
|
device must be released.
|
|
This function must be called during the
|
|
.Xr DEVICE_DETACH 9
|
|
method if
|
|
.Fn pci_iov_attach
|
|
was successfully called on the device and
|
|
.Fn pci_iov_detach
|
|
has not subsequently been called on the device and returned no error.
|
|
If this function returns an error, the
|
|
.Xr DEVICE_DETACH 9
|
|
method must fail and return an error, as detaching the PF driver while VF
|
|
devices are active would cause system instability.
|
|
This function is safe to call and will always succeed if
|
|
.Fn pci_iov_attach
|
|
previously failed with an error on the given device, or if
|
|
.Fn pci_iov_attach
|
|
was never called on the device.
|
|
.Pp
|
|
The
|
|
.Fn pci_save_state
|
|
and
|
|
.Fn pci_restore_state
|
|
functions can be used by a device driver to save and restore standard PCI
|
|
config registers.
|
|
The
|
|
.Fn pci_save_state
|
|
function must be invoked while the device has valid state before
|
|
.Fn pci_restore_state
|
|
can be used.
|
|
If the device is not in the fully-powered state
|
|
.Pq Dv PCI_POWERSTATE_D0
|
|
when
|
|
.Fn pci_restore_state
|
|
is invoked,
|
|
then the device will be transitioned to
|
|
.Dv PCI_POWERSTATE_D0
|
|
before any config registers are restored.
|
|
.Ss Message Signaled Interrupts
|
|
Message Signaled Interrupts
|
|
.Pq MSI
|
|
and
|
|
Enhanced Message Signaled Interrupts
|
|
.Pq MSI-X
|
|
are PCI capabilities that provide an alternate method for PCI
|
|
devices to signal interrupts.
|
|
The legacy INTx interrupt is available to PCI devices as a
|
|
.Dv SYS_RES_IRQ
|
|
resource with a resource ID of zero.
|
|
MSI and MSI-X interrupts are available to PCI devices as one or more
|
|
.Dv SYS_RES_IRQ
|
|
resources with resource IDs greater than zero.
|
|
A driver must ask the PCI bus to allocate MSI or MSI-X interrupts
|
|
using
|
|
.Fn pci_alloc_msi
|
|
or
|
|
.Fn pci_alloc_msix
|
|
before it can use MSI or MSI-X
|
|
.Dv SYS_RES_IRQ
|
|
resources.
|
|
A driver is not allowed to use the legacy INTx
|
|
.Dv SYS_RES_IRQ
|
|
resource if MSI or MSI-X interrupts have been allocated,
|
|
and attempts to allocate MSI or MSI-X interrupts will fail if the
|
|
driver is currently using the legacy INTx
|
|
.Dv SYS_RES_IRQ
|
|
resource.
|
|
A driver is only allowed to use either MSI or MSI-X,
|
|
but not both.
|
|
.Pp
|
|
The
|
|
.Fn pci_msi_count
|
|
function returns the maximum number of MSI messages supported by the
|
|
device
|
|
.Fa dev .
|
|
If the device does not support MSI,
|
|
then
|
|
.Fn pci_msi_count
|
|
returns zero.
|
|
.Pp
|
|
The
|
|
.Fn pci_alloc_msi
|
|
function attempts to allocate
|
|
.Fa *count
|
|
MSI messages for the device
|
|
.Fa dev .
|
|
The
|
|
.Fn pci_alloc_msi
|
|
function may allocate fewer messages than requested for various
|
|
reasons including requests for more messages than the device
|
|
.Fa dev
|
|
supports,
|
|
or if the system has a shortage of available MSI messages.
|
|
On success,
|
|
.Fa *count
|
|
is set to the number of messages allocated and
|
|
.Fn pci_alloc_msi
|
|
returns zero.
|
|
The
|
|
.Dv SYS_RES_IRQ
|
|
resources for the allocated messages will be available at consecutive
|
|
resource IDs beginning with one.
|
|
If
|
|
.Fn pci_alloc_msi
|
|
is not able to allocate any messages,
|
|
it returns an error.
|
|
Note that MSI only supports message counts that are powers of two;
|
|
requests to allocate a non-power of two count of messages will fail.
|
|
.Pp
|
|
The
|
|
.Fn pci_release_msi
|
|
function is used to release any allocated MSI or MSI-X messages back
|
|
to the system.
|
|
If any MSI or MSI-X
|
|
.Dv SYS_RES_IRQ
|
|
resources are allocated by the driver or have a configured interrupt
|
|
handler,
|
|
this function will fail with
|
|
.Er EBUSY .
|
|
The
|
|
.Fn pci_release_msi
|
|
function returns zero on success and an error on failure.
|
|
.Pp
|
|
The
|
|
.Fn pci_msix_count
|
|
function returns the maximum number of MSI-X messages supported by the
|
|
device
|
|
.Fa dev .
|
|
If the device does not support MSI-X,
|
|
then
|
|
.Fn pci_msix_count
|
|
returns zero.
|
|
.Pp
|
|
The
|
|
.Fn pci_alloc_msix
|
|
function attempts to allocate
|
|
.Fa *count
|
|
MSI-X messages for the device
|
|
.Fa dev .
|
|
The
|
|
.Fn pci_alloc_msix
|
|
function may allocate fewer messages than requested for various
|
|
reasons including requests for more messages than the device
|
|
.Fa dev
|
|
supports,
|
|
or if the system has a shortage of available MSI-X messages.
|
|
On success,
|
|
.Fa *count
|
|
is set to the number of messages allocated and
|
|
.Fn pci_alloc_msix
|
|
returns zero.
|
|
For MSI-X messages,
|
|
the resource ID for each
|
|
.Dv SYS_RES_IRQ
|
|
resource identifies the index in the MSI-X table of the
|
|
corresponding message.
|
|
A resource ID of one maps to the first index of the MSI-X table;
|
|
a resource ID two identifies the second index in the table, etc.
|
|
The
|
|
.Fn pci_alloc_msix
|
|
function assigns the
|
|
.Fa *count
|
|
messages allocated to the first
|
|
.Fa *count
|
|
table indicies.
|
|
If
|
|
.Fn pci_alloc_msix
|
|
is not able to allocate any messages,
|
|
it returns an error.
|
|
Unlike MSI,
|
|
MSI-X does not require message counts that are powers of two.
|
|
.Pp
|
|
The
|
|
.Fn pci_pending_msix
|
|
function examines the
|
|
.Fa dev
|
|
device's Pending Bit Array
|
|
.Pq PBA
|
|
to determine the pending status of the MSI-X message at table index
|
|
.Fa index .
|
|
If the indicated message is pending,
|
|
this function returns a non-zero value;
|
|
otherwise,
|
|
it returns zero.
|
|
Passing an invalid
|
|
.Fa index
|
|
to this function will result in undefined behavior.
|
|
.Pp
|
|
As mentioned in the description of
|
|
.Fn pci_alloc_msix ,
|
|
MSI-X messages are initially assigned to the first N table entries.
|
|
A driver may use a different distribution of available messages to
|
|
table entries via the
|
|
.Fn pci_remap_msix
|
|
function.
|
|
Note that this function must be called after a successful call to
|
|
.Fn pci_alloc_msix
|
|
but before any of the
|
|
.Dv SYS_RES_IRQ
|
|
resources are allocated.
|
|
The
|
|
.Fn pci_remap_msix
|
|
function returns zero on success,
|
|
or an error on failure.
|
|
.Pp
|
|
The
|
|
.Fa vectors
|
|
array should contain
|
|
.Fa count
|
|
message vectors.
|
|
The array maps directly to the MSI-X table in that the first entry in
|
|
the array specifies the message used for the first entry in the MSI-X
|
|
table,
|
|
the second entry in the array corresponds to the second entry in the
|
|
MSI-X table,
|
|
etc.
|
|
The vector value in each array index can either be zero to indicate
|
|
that no message should be assigned to the corresponding MSI-X table entry,
|
|
or it can be a number from one to N
|
|
.Po
|
|
where N is the count returned from the previous call to
|
|
.Fn pci_alloc_msix
|
|
.Pc
|
|
to indicate which of the allocated messages should be assigned to the
|
|
corresponding MSI-X table entry.
|
|
.Pp
|
|
If
|
|
.Fn pci_remap_msix
|
|
succeeds,
|
|
each MSI-X table entry with a non-zero vector will have an associated
|
|
.Dv SYS_RES_IRQ
|
|
resource whose resource ID corresponds to the table index as described
|
|
above for
|
|
.Fn pci_alloc_msix .
|
|
MSI-X table entries that with a vector of zero will not have an
|
|
associated
|
|
.Dv SYS_RES_IRQ
|
|
resource.
|
|
Additionally,
|
|
if any of the original messages allocated by
|
|
.Fn pci_alloc_msix
|
|
are not used in the new distribution of messages in the MSI-X table,
|
|
they will be released automatically.
|
|
Note that if a driver wishes to use fewer messages than were allocated by
|
|
.Fn pci_alloc_msix ,
|
|
the driver must use a single, contiguous range of messages beginning
|
|
with one in the new distribution.
|
|
The
|
|
.Fn pci_remap_msix
|
|
function will fail if this condition is not met.
|
|
.Sh IMPLEMENTATION NOTES
|
|
The
|
|
.Vt pci_addr_t
|
|
type varies according to the size of the PCI bus address
|
|
space on the target architecture.
|
|
.Sh SEE ALSO
|
|
.Xr pci 4 ,
|
|
.Xr pciconf 8 ,
|
|
.Xr bus_alloc_resource 9 ,
|
|
.Xr bus_dma 9 ,
|
|
.Xr bus_release_resource 9 ,
|
|
.Xr bus_setup_intr 9 ,
|
|
.Xr bus_teardown_intr 9 ,
|
|
.Xr devclass 9 ,
|
|
.Xr device 9 ,
|
|
.Xr driver 9 ,
|
|
.Xr rman 9
|
|
.Rs
|
|
.%B FreeBSD Developers' Handbook
|
|
.%T NewBus
|
|
.%U http://www.FreeBSD.org/doc/en_US.ISO8859-1/books/developers-handbook/
|
|
.Re
|
|
.Rs
|
|
.%A Shanley
|
|
.%A Anderson
|
|
.%B PCI System Architecture
|
|
.%N 2nd Edition
|
|
.%I Addison-Wesley
|
|
.%O ISBN 0-201-30974-2
|
|
.Re
|
|
.Sh AUTHORS
|
|
.An -nosplit
|
|
This manual page was written by
|
|
.An Bruce M Simpson Aq Mt bms@FreeBSD.org
|
|
and
|
|
.An John Baldwin Aq Mt jhb@FreeBSD.org .
|
|
.Sh BUGS
|
|
The kernel PCI code has a number of references to
|
|
.Dq "slot numbers" .
|
|
These do not refer to the geographic location of PCI devices,
|
|
but to the device number assigned by the combination of the PCI IDSEL
|
|
mechanism and the platform firmware.
|
|
This should be taken note of when working with the kernel PCI code.
|