pci_iov_attach (9)

Leading comments

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
   docum...

(The comments found at the beginning of the groff file "man9/pci_iov_attach.9freebsd".)

NAME

pci pci_alloc_msi pci_alloc_msix pci_disable_busmaster pci_disable_io pci_enable_busmaster pci_enable_io pci_find_bsf pci_find_cap pci_find_dbsf pci_find_device pci_find_extcap pci_find_htcap pci_find_pcie_root_port pci_get_id pci_get_max_payload pci_get_max_read_req pci_get_powerstate pci_get_vpd_ident pci_get_vpd_readonly pci_iov_attach pci_iov_detach pci_msi_count pci_msix_count pci_msix_pba_bar pci_msix_table_bar pci_pending_msix pci_read_config pci_release_msi pci_remap_msix pci_restore_state pci_save_state pci_set_max_read_req pci_set_powerstate pci_write_config pcie_adjust_config pcie_read_config pcie_write_config - PCI bus interface

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 device_t Fn pci_find_pcie_root_port device_t dev Ft int Fn pci_get_id device_t dev enum pci_id_type type uintptr_t *id Ft int Fn pci_get_max_payload device_t dev 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_msix_pba_bar device_t dev Ft int Fn pci_msix_table_bar 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 Ft uint32_t Fo pcie_adjust_config Fa device_t dev Fa int reg Fa uint32_t mask Fa uint32_t val Fa int width Fc Ft uint32_t Fn pcie_read_config device_t dev int reg int width Ft void Fn pcie_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

DESCRIPTION

The 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.

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.

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.

The Fn pcie_adjust_config function is used to modify the value of a register in the PCI-express capability register set of device Fa dev . The offset Fa reg specifies a relative offset in the register set with Fa width specifying the size of the access. The new value of the register is computed by modifying bits set in Fa mask to the value in Fa val . Any bits not specified in Fa mask are preserved. The previous value of the register is returned.

The Fn pcie_read_config function is used to read the value of a register in the PCI-express capability register set of device Fa dev . The offset Fa reg specifies a relative offset in the register set with Fa width specifying the size of the access.

The Fn pcie_write_config function is used to write the value Fa val to a register in the PCI-express capability register set of device Fa dev . The offset Fa reg specifies a relative offset in the register set with Fa width specifying the size of the access.

NOTE Device drivers should only use these functions for functionality that is not available via another Fn pci function.

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:

pci_find_dbsf(0, bus, slot, func);

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.

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.

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 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.

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 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.

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 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.

The Fn pci_find_pcie_root_port function walks up the PCI device hierarchy to locate the PCI-express root port upstream of Fa dev . If a root port is not found, Fn pci_find_pcie_root_port returns NULL

The Fn pci_get_id function is used to read an identifier from a device. The Fa type flag is used to specify which identifier to read. The following flags are supported:

PCI_ID_RID

Read the routing identifier for the device.

PCI_ID_MSI

Read the MSI routing ID. This is needed by some interrupt controllers to route MSI and MSI-X interrupts.

The Fn pci_get_vpd_ident function is used to fetch a device's Vital Product Data (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.

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.

Device Configuration

The Fn pci_enable_busmaster function enables PCI bus mastering for the device Fa dev , by setting the PCIM_CMD_BUSMASTEREN bit in the PCIR_COMMAND register. The Fn pci_disable_busmaster function clears this bit.

The Fn pci_enable_io function enables memory or I/O port address decoding for the device Fa dev , by setting the PCIM_CMD_MEMEN or PCIM_CMD_PORTEN bit in the 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 SYS_RES_MEMORY or SYS_RES_IOPORT as appropriate. Device drivers should generally not use these routines directly. The PCI bus will enable decoding automatically when a SYS_RES_MEMORY or SYS_RES_IOPORT resource is activated via bus_alloc_resource9 or bus_activate_resource9.

The Fn pci_get_max_payload function returns the current maximum TLP payload size in bytes for a PCI-express device. If the Fa dev device is not a PCI-express device, Fn pci_get_max_payload returns zero.

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.

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.

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 PCI_POWERSTATE_D0 is returned. The following power states are defined by PCI:

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.

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.

PCI_POWERSTATE_D2

Class-specific low-power state in which device context may or may not be lost. Attains greater power savings than PCI_POWERSTATE_D1 Busses in this state can cause devices to lose some context. Devices must be prepared for the bus to be in this state or higher.

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.

PCI_POWERSTATE_UNKNOWN

State of the device is unknown.

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 .

The Fn pci_iov_attach function is used to advertise that the given device (and associated device driver) supports PCI Single-Root I/O Virtualization (SR-IOV) A driver that supports SR-IOV must implement the PCI_IOV_INIT9, PCI_IOV_ADD_VF9 and PCI_IOV_UNINIT9 methods. This function should be called during the DEVICE_ATTACH9 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 (PF) and for individual Virtual Functions (VFs respectively. ) See pci_iov_schema9 for details on how to construct the schema. If either the pf_schema or 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.

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 DEVICE_DETACH9 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 DEVICE_DETACH9 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.

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 (PCI_POWERSTATE_D0 ) when Fn pci_restore_state is invoked, then the device will be transitioned to PCI_POWERSTATE_D0 before any config registers are restored.

Message Signaled Interrupts

Message Signaled Interrupts (MSI) and Enhanced Message Signaled Interrupts (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 SYS_RES_IRQ resource with a resource ID of zero. MSI and MSI-X interrupts are available to PCI devices as one or more 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 SYS_RES_IRQ resources. A driver is not allowed to use the legacy INTx 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 SYS_RES_IRQ resource. A driver is only allowed to use either MSI or MSI-X, but not both.

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.

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 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.

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 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.

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.

The Fn pci_msix_pba_bar function returns the offset in configuration space of the Base Address Register (BAR) containing the MSI-X Pending Bit Array (PBA) for device Fa dev . The returned value can be used as the resource ID with bus_alloc_resource9 and bus_release_resource9 to allocate the BAR. If the device does not support MSI-X, then Fn pci_msix_pba_bar returns -1.

The Fn pci_msix_table_bar function returns the offset in configuration space of the BAR containing the MSI-X vector table for device Fa dev . The returned value can be used as the resource ID with bus_alloc_resource9 and bus_release_resource9 to allocate the BAR. If the device does not support MSI-X, then Fn pci_msix_table_bar returns -1.

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 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 indices. 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.

The BARs containing the MSI-X vector table and PBA must be allocated via bus_alloc_resource9 before calling Fn pci_alloc_msix and must not be released until after calling Fn pci_release_msi . Note that the vector table and PBA may be stored in the same BAR or in different BARs.

The Fn pci_pending_msix function examines the Fa dev device's 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.

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 SYS_RES_IRQ resources are allocated. The Fn pci_remap_msix function returns zero on success, or an error on failure.

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.

If Fn pci_remap_msix succeeds, each MSI-X table entry with a non-zero vector will have an associated 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 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.

IMPLEMENTATION NOTES

The Vt pci_addr_t type varies according to the size of the PCI bus address space on the target architecture.

SEE ALSO

pci(4), pciconf(8), bus_alloc_resource9, bus_dma9, bus_release_resource9, bus_setup_intr9, bus_teardown_intr9, devclass(9), device(9), driver(9), rman(9)

FreeBSD Developers' Handbook NewBus

Shanley Anderson PCI System Architecture 2nd Edition ISBN 0-201-30974-2

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 .

BUGS

The kernel PCI code has a number of references to ``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.

The PCI bus driver should allocate the MSI-X vector table and PBA internally as necessary rather than requiring the caller to do so.