rte_vfio.h(3)

📦

version 25.11.3 · 9 versions

25.11.3 · ubuntu resolute, ubuntu stonking (latest)
25.11 · centos-stream 9
26.03 · nixos 26.05, nixpkgs unstable
24.11.8 · debian trixie
24.11.3 · ubuntu questing
24.11.2 · fedora 43
24.11.4 · fedora 43, fedora 44
25.11 · centos-stream 10, ubuntu resolute
24.11.4 · debian trixie

🌸 `rte_vfio.h` — the VFIO bestie handbook 🌸

okay so bestie, welcome to rte_vfio.h, straight outta section 3 of the Library Functions Manual 📚. This is DPDK's VFIO utility header and it's giving big kernel-adjacent energy. Buckle up, we're going through every single function, no skipping, no skimming, periodt. 💅✨

📛 NAME

rte_vfio.h

✨ SYNOPSIS ✨

First, the little pre-party — the stuff you gotta pull in:

#include <stdbool.h>
#include <stdint.h>
#include <rte_compat.h>

🦋 Functions

Here's the whole squad of function signatures, exactly as they come:

int rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr, int *vfio_dev_fd, struct vfio_device_info *device_info)
int rte_vfio_release_device(const char *sysfs_base, const char *dev_addr, int fd)
int rte_vfio_enable(const char *modname)
int rte_vfio_is_enabled(const char *modname)
int rte_vfio_noiommu_is_enabled(void)
int rte_vfio_clear_group(int vfio_group_fd)
int rte_vfio_get_group_num(const char *sysfs_base, const char *dev_addr, int *iommu_group_num)
__rte_experimental int rte_vfio_get_device_info(const char *sysfs_base, const char *dev_addr, int *vfio_dev_fd, struct vfio_device_info *device_info)
int rte_vfio_get_container_fd(void)
int rte_vfio_get_group_fd(int iommu_group_num)
int rte_vfio_container_create(void)
int rte_vfio_container_destroy(int container_fd)
int rte_vfio_container_group_bind(int container_fd, int iommu_group_num)
int rte_vfio_container_group_unbind(int container_fd, int iommu_group_num)
int rte_vfio_container_dma_map(int container_fd, uint64_t vaddr, uint64_t iova, uint64_t len)
int rte_vfio_container_dma_unmap(int container_fd, uint64_t vaddr, uint64_t iova, uint64_t len)

💖 Detailed Description 💖

RTE VFIO. This library hands you a whole lil bundle of VFIO-related utility functions so you can do the device-passthrough thing. 🫶

Definition in file rte_vfio.h.

🧚 Function Documentation 🧚

`rte_vfio_setup_device`

int rte_vfio_setup_device(const char *sysfs_base, const char *dev_addr,
    int *vfio_dev_fd, struct vfio_device_info *device_info)

Sets up vfio_cfg for the device that you identify by its address. It goes and discovers the I/O MMU groups that are already configured, or it sets up a brand new one for this device. And if a new group gets assigned? Then the DMA mapping happens too. Iconic multitasking. ✨

heads up bestie: this function is only relevant to linux and will return an error on BSD. no BSD crumbs for this one. 😭

Parameters

sysfs_base — sysfs path prefix.
dev_addr — device location.
vfio_dev_fd — VFIO fd.
device_info — Device information.

Returns

0 on success.
<0 on failure.
>1 if the device cannot be managed this way.

`rte_vfio_release_device`

int rte_vfio_release_device(const char *sysfs_base, const char *dev_addr,
    int fd)

Releases a device that's mapped to a VFIO-managed I/O MMU group. Say bye 👋

same deal: only relevant to linux, returns an error on BSD.

Parameters

sysfs_base — sysfs path prefix.
dev_addr — device location.
fd — VFIO fd.

Returns

0 on success.
<0 on failure.

`rte_vfio_enable`

int rte_vfio_enable(const char *modname)

Turns on a VFIO-related kmod. basically the "go go go" button for your kernel module. 🔥

you guessed it: only relevant to linux, returns an error on BSD.

Parameters

modname — kernel module name.

Returns

0 on success.
<0 on failure.

`rte_vfio_is_enabled`

int rte_vfio_is_enabled(const char *modname)

Checks whether a VFIO-related kmod is enabled. vibe check on your kernel module, fr fr. 🕵️‍♀️

note the capital L: only relevant to Linux.

Parameters

modname — kernel module name.

Returns

1 if true.
0 otherwise.

`rte_vfio_noiommu_is_enabled`

int rte_vfio_noiommu_is_enabled(void)

Tells you whether VFIO <cyanish>NOIOMMU mode is enabled. (are we doing the no-iommu thing or not, bestie?)

Linux only for this one.

Returns

1 if true.
0 if false.
<0 for errors.

`rte_vfio_clear_group`

int rte_vfio_clear_group(int vfio_group_fd)

Removes a group fd from the internal VFIO group fd array/ — yep, that trailing slash is how it's written and I am not touching it. 💅

linux only, and it will return an error on BSD.

Parameters

vfio_group_fd — VFIO Group FD.

Returns

0 on success.
<0 on failure.

`rte_vfio_get_group_num`

int rte_vfio_get_group_num(const char *sysfs_base, const char *dev_addr,
    int *iommu_group_num)

Parses the IOMMU group number for a device. 🔢

linux only, error on BSD.

Parameters

sysfs_base — sysfs path prefix.
dev_addr — device location.
iommu_group_num — iommu group number.

Returns

>0 on success.
0 for non-existent group or VFIO.
<0 for errors.

`rte_vfio_get_device_info`

__rte_experimental int rte_vfio_get_device_info(const char *sysfs_base,
    const char *dev_addr, int *vfio_dev_fd,
    struct vfio_device_info *device_info)

Gets device information. 👀 (yes, it's marked __rte_experimental — experimental things can wiggle around, so don't get too attached.)

Linux only, and returns an error on BSD.

Parameters

sysfs_base — sysfs path prefix.
dev_addr — device location.
vfio_dev_fd — VFIO fd.
device_info — Device information.

Returns

0 on success.
<0 on failure.

`rte_vfio_get_container_fd`

int rte_vfio_get_container_fd(void)

Gets the default VFIO container fd. the OG container, if you will. 🧺

linux only, error on BSD.

Returns

0 — default container fd.
< 0 if VFIO is not enabled or not supported.

`rte_vfio_get_group_fd`

int rte_vfio_get_group_fd(int iommu_group_num)

Opens a VFIO group fd, or grabs an already-existing one. 🚪

linux only, error on BSD.

Parameters

iommu_group_num — iommu group number.

Returns

0 — group fd.
< 0 for errors.

`rte_vfio_container_create`

int rte_vfio_container_create(void)

Creates a shiny new container for device binding. ✨🧺

Note (read this twice, bestie): - Any newly allocated DPDK memory will not be mapped into these containers by default, user needs to manage DMA mappings for any container created by this API. - When creating containers using this API, the container will only be available in the process that has created it. Sharing containers and devices between multiple processes is not supported. 😭 no group-chat sharing here.

Returns

the container fd if successful.
<0 if failed.

`rte_vfio_container_destroy`

int rte_vfio_container_destroy(int container_fd)

Destroys the container, and unbinds all vfio groups within it. bye forever 🫡

Parameters

container_fd — the container fd to destroy.

Returns

0 if successful.
<0 if failed.

`rte_vfio_container_group_bind`

int rte_vfio_container_group_bind(int container_fd, int iommu_group_num)

Binds an IOMMU group to a container. it's giving arranged marriage but for fds. 💍

Parameters

container_fd — the container's fd.
iommu_group_num — the iommu group number to bind to container.

Returns

group fd if successful.
<0 if failed.

`rte_vfio_container_group_unbind`

int rte_vfio_container_group_unbind(int container_fd, int iommu_group_num)

Unbinds an IOMMU group from a container. 💔

Parameters

container_fd — the container fd of container.
iommu_group_num — the iommu group number to delete from container.

Returns

0 if successful.
<0 if failed.

`rte_vfio_container_dma_map`

int rte_vfio_container_dma_map(int container_fd, uint64_t vaddr,
    uint64_t iova, uint64_t len)

Performs DMA mapping for devices in a container. 🗺️

Parameters

container_fd — the specified container fd. Use RTE_VFIO_DEFAULT_CONTAINER_FD to use the default container.
vaddr — Starting virtual address of memory to be mapped.
iova — Starting IOVA address of memory to be mapped.
len — Length of memory segment being mapped.

Returns

0 if successful.
<0 if failed.

`rte_vfio_container_dma_unmap`

int rte_vfio_container_dma_unmap(int container_fd, uint64_t vaddr,
    uint64_t iova, uint64_t len)

Performs DMA unmapping for devices in a container. 🧹 (reverse of the map, duh)

Parameters

container_fd — the specified container fd. Use RTE_VFIO_DEFAULT_CONTAINER_FD to use the default container.
vaddr — Starting virtual address of memory to be unmapped.
iova — Starting IOVA address of memory to be unmapped.
len — Length of memory segment being unmapped.

Returns

0 if successful.
<0 if failed.

👑 Author

Generated automatically by Doxygen for DPDK from the source code.

and that's the whole page bestie, you ate it 💅✨

this was DPDK, Version 25.11.3, from the rte_vfio.h(3) page 🎀 — go forth and map your DMA, no cap 🫶🌸

📦 where this page comes from

ubuntu resolute: dpdk-doc 25.11.3-0ubuntu0.26.04.1 · BSD-3-Clause
ubuntu stonking: dpdk-doc 25.11.3-1 · BSD-3-Clause
💖 girlified by deepseek-ai/DeepSeek-V4.1-Flash on 2026-11-03: a language model's rewrite of the original, under its terms

rte_vfio.h(3)

🌸 rte_vfio.h — the VFIO bestie handbook 🌸

📛 NAME

✨ SYNOPSIS ✨

🦋 Functions

💖 Detailed Description 💖

🧚 Function Documentation 🧚

rte_vfio_setup_device

rte_vfio_release_device

rte_vfio_enable

rte_vfio_is_enabled

rte_vfio_noiommu_is_enabled

rte_vfio_clear_group

rte_vfio_get_group_num

rte_vfio_get_device_info

rte_vfio_get_container_fd

rte_vfio_get_group_fd

rte_vfio_container_create

rte_vfio_container_destroy

rte_vfio_container_group_bind

rte_vfio_container_group_unbind

rte_vfio_container_dma_map

rte_vfio_container_dma_unmap

👑 Author

🌸 `rte_vfio.h` — the VFIO bestie handbook 🌸

`rte_vfio_setup_device`

`rte_vfio_release_device`

`rte_vfio_enable`

`rte_vfio_is_enabled`

`rte_vfio_noiommu_is_enabled`

`rte_vfio_clear_group`

`rte_vfio_get_group_num`

`rte_vfio_get_device_info`

`rte_vfio_get_container_fd`

`rte_vfio_get_group_fd`

`rte_vfio_container_create`

`rte_vfio_container_destroy`

`rte_vfio_container_group_bind`

`rte_vfio_container_group_unbind`

`rte_vfio_container_dma_map`

`rte_vfio_container_dma_unmap`