Index: head/share/man/man9/pci.9 =================================================================== --- head/share/man/man9/pci.9 (revision 285272) +++ head/share/man/man9/pci.9 (revision 285273) @@ -1,785 +1,785 @@ .\" .\" Copyright (c) 2005 Bruce M Simpson .\" 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 May 28, 2015 +.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 -.Po SR-IOV Pc . +.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. Index: head/share/man/man9/pci_iov_schema.9 =================================================================== --- head/share/man/man9/pci_iov_schema.9 (revision 285272) +++ head/share/man/man9/pci_iov_schema.9 (revision 285273) @@ -1,265 +1,265 @@ .\" .\" Copyright (c) 2014 Sandvine Inc. .\" 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 June 18, 2015 +.Dd July 8, 2015 .Dt PCI_IOV_SCHEMA 9 .Os .Sh NAME .Nm pci_iov_schema , .Nm pci_iov_schema_alloc_node , .Nm pci_iov_schema_add_bool , .Nm pci_iov_schema_add_string , .Nm pci_iov_schema_add_uint8 , .Nm pci_iov_schema_add_uint16 , .Nm pci_iov_schema_add_uint32 , .Nm pci_iov_schema_add_uint64 , .Nm pci_iov_schema_add_unicast_mac .Nd PCI SR-IOV config schema interface .Sh SYNOPSIS .In machine/stdarg.h .In sys/nv.h .In sys/iov_schema.h .Ft nvlist_t * .Fn pci_iov_schema_alloc_node "void" .Ft void .Fn pci_iov_schema_add_bool "nvlist_t *schema" "const char *name" \ "uint32_t flags" "int defaultVal" .Ft void .Fn pci_iov_schema_add_string "nvlist_t *schema" "const char *name" \ "uint32_t flags" "const char *defaultVal" .Ft void .Fn pci_iov_schema_add_uint8 "nvlist_t *schema" "const char *name" \ "uint32_t flags" "uint8_t defaultVal" .Ft void .Fn pci_iov_schema_add_uint16 "nvlist_t *schema" "const char *name" \ "uint32_t flags" "uint16_t defaultVal" .Ft void .Fn pci_iov_schema_add_uint32 "nvlist_t *schema" "const char *name" \ "uint32_t flags" "uint32_t defaultVal" .Ft void .Fn pci_iov_schema_add_uint64 "nvlist_t *schema" "const char *name" \ "uint32_t flags" "uint64_t defaultVal" .Ft void .Fn pci_iov_schema_add_unicast_mac "nvlist_t *schema" "const char *name" \ "uint32_t flags" "const uint8_t *defaultVal" .Sh DESCRIPTION The PCI Single-Root I/O Virtualization .Pq SR-IOV configuration schema is a data structure that describes the device-specific configuration parameters that a PF driver will accept when SR-IOV is enabled on the PF device. Each PF driver defines two schema instances: the PF schema and the VF schema. The PF schema describes configuration that applies to the PF device as a whole. The VF schema describes configuration that applies to an individual VF device. Different VF devices may have different configuration applied to them, as long as the configuration for each VF conforms to the VF schema. .Pp A PF driver builds a configuration schema by first allocating a schema node and then adding configuration parameter specifications to the schema. The configuration parameter specification consists of a name and a value type. .Pp Configuration parameter names are case-insensitive. It is an error to specify two or more configuration parameters with the same name. It is also an error to specific a configuration parameter that uses the same name as a configuration parameter used by the SR-IOV infrastructure. See .Xr iovctl.conf 5 for documentation of all configuration parameters used by the SR-IOV infrastructure. .Pp The parameter type constrains the possible values that the configuration parameter may take. .Pp A configuration parameter may be specified as a required parameter by setting the .Dv IOV_SCHEMA_REQUIRED flag in the .Pa flags argument. Required parameters must be specified by the user when SR-IOV is enabled. If the user does not specify a required parameter, the SR-IOV infrastructure will abort the request to enable SR-IOV and return an error to the user. .Pp Alternatively, a configuration parameter may be given a default value by setting the .Dv IOV_SCHEMA_HASDEFAULT flag in the .Pa flags argument. If a configuration parameter has a default value but the user has not specified a value for that parameter, then the SR-IOV infrastructure will apply .Pa defaultVal for that parameter in the configuration before passing it to the PF driver. It is an error for the value of the .Pa defaultVal parameter to not conform to the restrictions of the specified type. If this flag is not specified then the .Pa defaultVal argument is ignored. This flag is not compatible with the .Dv IOV_SCHEMA_REQUIRED flag; it is an error to specify both on the same parameter. .Pp The SR-IOV infrastructure guarantees that all configuration parameters that are either specified as required or given a default value will be present in the configuration passed to the PF driver. Configuration parameters that are neither specified as required nor given a default value are optional and may or may not be present in the configuration passed to the PF driver. .Pp It is highly recommended that a PF driver reserve the use of optional parameters for configuration that is truly optional. For example, a Network Interface PF device might have the option to encapsulate all traffic to and from a VF device in a vlan tag. The PF driver could expose that option as a "vlan" parameter accepting an integer argument specifying the vlan tag. In this case, it would be appropriate to set the "vlan" parameter as an optional parameter as it would be legitimate for a VF to be configured to have no vlan tagging enabled at all. .Pp Alternatively, if the PF device had an boolean option that controlled whether the VF was allowed to change its MAC address, it would not be appropriate to set this parameter as optional. The PF driver must either allow the MAC to change or not, so it would be more appropriate for the PF driver to document the default behaviour by specifying a default value in the schema .Po or potentially force the user to make the choice by setting the parameter to be required .Pc . .Pp Configuration parameters that have security implications must default to the most secure configuration possible. .Pp All device-specific configuration parameters must be documented in the manual page for the PF driver, or in a separate manual page that is cross-referenced from the main driver manual page. .Pp It is not necessary for a PF driver to check for failure from any of these functions. If an error occurs, it is flagged in the schema. The .Xr pci_iov_attach 9 function checks for this error and will fail to initialize SR-IOV on the PF device if an error is set in the schema. If this occurs, it is recommended that the PF driver still succeed in attaching and run with SR-IOV disabled on the device. .Pp The .Fn pci_iov_schema_alloc_node function is used to allocate an empty configuration schema. It is not necessary to check for failure from this function. The SR-IOV infrastructure will gracefully handle failure to allocate a schema and will simply not enable SR-IOV on the PF device. .Pp The .Fn pci_iov_schema_add_bool function is used to specify a configuration parameter in the given schema with the name .Pa name and having a boolean type. Boolean values can only take the value true or false (1 or 0, respectively). .Pp The .Fn pci_iov_schema_add_string function is used to specify a configuration parameter in the given schema with the name .Pa name and having a string type. String values are standard C strings. .Pp The .Fn pci_iov_schema_add_uint8 function is used to specify a configuration parameter in the given schema with the name .Pa name and having a .Vt uint8_t type. Values of type .Vt uint8_t are unsigned integers in the range 0 to 255, inclusive. .Pp The .Fn pci_iov_schema_add_uint16 function is used to specify a configuration parameter in the given schema with the name .Pa name and having a .Vt uint16_t type. Values of type .Vt uint16_t are unsigned integers in the range 0 to 65535, inclusive. .Pp The .Fn pci_iov_schema_add_uint32 function is used to specify a configuration parameter in the given schema with the name .Pa name and having a .Vt uint32_t type. Values of type .Vt uint32_t are unsigned integers in the range 0 to -.Po 2**32 - 1 Pc , +.Pq 2**32 - 1 , inclusive. .Pp The .Fn pci_iov_schema_add_uint64 function is used to specify a configuration parameter in the given schema with the name .Pa name and having a .Vt uint64_t type. Values of type .Vt uint64_t are unsigned integers in the range 0 to -.Po 2**64 - 1 Pc , +.Pq 2**64 - 1 , inclusive. .Pp The .Fn pci_iov_schema_add_unicast_mac function is used to specify a configuration parameter in the given schema with the name .Pa name and having a unicast-mac type. Values of type unicast-mac are binary values exactly 6 bytes long. The MAC address is guaranteed to not be a multicast or broadcast address. .Sh RETURN VALUES The .Fn pci_iov_schema_alloc_node function returns a pointer to the allocated schema, or NULL if a failure occurs. .Sh SEE ALSO .Xr pci 9 , .Xr PCI_IOV_ADD_VF 9 , .Xr PCI_IOV_INIT 9 .Sh AUTHORS This manual page was written by .An Ryan Stone Aq rstone@FreeBSD.org . Index: head/usr.sbin/iovctl/iovctl.8 =================================================================== --- head/usr.sbin/iovctl/iovctl.8 (revision 285272) +++ head/usr.sbin/iovctl/iovctl.8 (revision 285273) @@ -1,123 +1,123 @@ .\" .\" Copyright (c) 2014 Sandvine Inc. .\" 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 May 21, 2014 +.Dd July 8, 2015 .Dt IOVCTL 8 .Os .Sh NAME .Nm iovctl .Nd "PCI SR-IOV configuration utility" .Sh SYNOPSIS .Nm .Fl C .Op Fl f Ar config-file .Op Fl n .Nm .Fl D .Op Fl f Ar config-file | Fl d Ar device .Op Fl n .Nm .Fl S .Op Fl f Ar config-file | Fl d Ar device .Sh DESCRIPTION The .Nm utility creates or destroys PCI Single-Root I/O Virtualization .Pq SR-IOV Virtual Functions -.Po VFs Pc . +.Pq VFs . When invoked with the .Fl C flag, .Nm creates VFs as children of the Physical Function .Pq PF configured in the specified configuration file. When invoked with the .Fl D flag, .Nm destroys all VFs that are children of the specified device. Available PF devices can be seen in .Pa /dev/iov/ . .Pp The following options are available: .Bl -tag -width indent .It Fl C Enable SR-IOV on the specified PF device and create VF children. This operation will fail if the PF already has VF children. This option must be used in conjunction with the .Fl f option. .It Fl d Ar device Specify the PF device to use for the given operation. .Ar device may either be the name of a PF device, or a full path name to a node in .Pa /dev/iov/ . This option may not be used with the .Fl C option. .It Fl D Delete all VF children of the specified PF device. This operation will fail if SR-IOV is not currently enabled on the specified device. .It Fl f Ar config-file Specify the pathname of the configuration file. For the .Fl C option, this file will be used to specify all configuration values. For the .Fl D and .Fl S options, this file will only be used to specify the name of the PF device. .Pp See .Xr iovctl.conf for a description of the config file format and documentation of the configuration parameters that apply to all PF drivers. See the PF driver manual page for configuration parameters specific to particular hardware. .It Fl n Perform a dry-run. Perform all validation of the specified action and print what would be done, but do not perform the actual creation or destruction of VFs. This option may not be used with the .Fl S flag. .It Fl S Read the configuration schema from the specified device and print its contents to stdout. This action may be used to discover the configuration parameters supported on a given PF device. .El .Sh SEE ALSO .Xr iovctl.conf 5 , .Xr rc.conf 5 .Sh AUTHORS This manual page was written by .An Ryan Stone Aq Mt rstone@FreeBSD.org . Index: head/usr.sbin/iovctl/iovctl.conf.5 =================================================================== --- head/usr.sbin/iovctl/iovctl.conf.5 (revision 285272) +++ head/usr.sbin/iovctl/iovctl.conf.5 (revision 285273) @@ -1,171 +1,171 @@ .\" .\" Copyright (c) 2014 Sandvine Inc. .\" 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 5, 2015 +.Dd July 8, 2015 .Dt IOVCTL.CONF 5 .Os .Sh NAME .Nm iovctl.conf .Nd IOVCTL configuration file .Sh DESCRIPTION The .Nm file is the configuration file for the .Xr iovctl 8 program. This file specifies configuration parameters for a single Physical Function .Pq PF device. To configure SR-IOV on multiple PF devices, use one configuration file for each PF. The locations of all .Xr iovctl 9 configuration files are specified in .Xr rc.conf 5 . .Pp The .Nm file uses UCL format. UCL syntax is documented at the official UCL website: http://github.com/vstakhov/libucl. .Pp There are three types of sections in the .Nm file. A section is a key at the top level of the file with a list as its value. The list may contain the keys specified in the .Sx OPTIONS section of this manual page. Individual PF driver implementations may specify additional device-specific configuration keys that they will accept. The order in which sections appear in .Nm is ignored. No two sections may have the same key. For example, two sections for VF-1 must not be defined. .Pp The first section type is the PF section. This section always has the key "PF"; therefore, only one such section may be defined. This section defines configuration parameters that apply to the PF as a whole. .Pp The second section type is the VF section. This section has the key "VF-" followed by a VF index. VF indices start at 0 and always increment by 1. Valid VF indices are in the range of 0 to -.Po num_vfs - 1 Pc . +.Pq num_vfs - 1 . The VF index must be given as a decimal integer with no leading zeros. This section defines configuration parameters that apply to a single VF. .Pp The third section type is the default section. This section always has the key "DEFAULT"; therefore, only one such section may be specified. This section defines default configuration parameters that apply to all VFs. All configuration keys that are valid to be applied to a VF are valid in this section. An individual VF section may override a default specified in this section by providing a different value for the configuration parameter. Note that the default section applies to ALL VFs. The default section must appear before any VF sections. The default section may appear before or after the PF section. .Pp The following option types are supported: .Bl -tag -width indent .It boolean Accepts a boolean value of true or false. .It mac-addr Accepts a unicast MAC address specified as a string of the form xx:xx:xx:xx:xx:xx, where xx is one or two hexadecimal digits. .It string Accepts any string value. .It uint8_t Accepts any integer in the range 0 to 255, inclusive. .It uint16_t Accepts any integer in the range 0 to 65535, inclusive. .It uint32_t Accepts any integer in the range 0 to -.Po 2**32 - 1 Pc , +.Pq 2**32 - 1 , inclusive. .It uint64_t Accepts any integer in the range 0 to -.Po 2**64 - 1 Pc , +.Pq 2**64 - 1 , inclusive. .El .Sh OPTIONS The following parameters are accepted by all PF drivers: .Bl -tag -width indent .It device Pq string This parameter specifies the name of the PF device. This parameter is required to be specified. .It num_vfs Pq uint16_t This parameter specifies the number of VF children to create. This parameter may not be zero. The maximum value of this parameter is device-specific. .El .Pp The following parameters are accepted by all VFs: .Bl -tag -width indent .It passthrough Pq boolean This parameter controls whether the VF is reserved for the use of the .Xr bhyve 8 hypervisor as a PCI passthrough device. If this parameter is set to true, then the VF will be reserved as a PCI passthrough device and it will not be accessible from the host OS. The default value of this parameter is false. .El .Pp See the PF driver manual page for configuration parameters specific to particular hardware. .Sh EXAMPLES This sample file will create 3 VFs as children of the ix0 device. VF-1 and VF-2 are set as .Xr bhyve 8 passthrough devices through the use of the default section. VF-0 is not configured as a passthrough device as it explicitly overrides the default. VF-0 also sets a device-specific parameter named mac-addr. .Bd -literal -offset ident PF { device : "ix0"; num_vfs : 3; } DEFAULT { passthrough : true; } VF-0 { mac-addr : "02:56:48:7e:d9:f7"; passthrough : false; } .Ed .Sh SEE ALSO .Xr rc.conf 5 , .Xr iovctl 8 .Sh AUTHORS This manual page was written by .An Ryan Stone Aq Mt rstone@FreeBSD.org .