Virtio PCI设备规格草案0.9.5：标准、高效虚拟环境接口

需积分: 18 42 浏览量更新于2024-07-14 收藏 365KB PDF 举报

Virtio PCI Card Specification v0.9.5 DRAFT 是关于 Virtio 技术在虚拟环境中的PCI设备规范。Virtio 是一种设计目标为简化虚拟环境中设备交互的解决方案，它旨在提供一个通用、高效、标准且可扩展的机制，使得虚拟设备能够无缝地与物理世界中的标准PCI设备协同工作。该文档的第1章阐述了Virtio设备的目的和描述。其核心理念是，尽管这些设备运行在虚拟环境中，但它们的设计尽可能接近物理设备，以便于支持标准的PCI驱动程序和发现机制。这样做的好处在于，无论是设备驱动开发者还是操作系统，都可以利用熟悉的中断处理和DMA（Direct Memory Access）技术来操作这些设备，降低了学习曲线和定制需求。 "Efficient"这一特性体现在 Virtio PCI 设备的内部结构上，通过输入和输出描述符环的使用，减少了主机和设备之间数据交换时的缓存冲突。这些描述符环的组织方式确保了数据传输的高效性，避免了由于数据共享造成的性能瓶颈。 "Standard"是指Virtio PCI 设计遵循标准的PCI协议，不引入复杂的页面映射或Copy-On-Write（CoW）机制，保持了与传统PCI设备的兼容性，使得硬件和软件的集成更加便捷。 Virtio PCI Card Specification v0.9.5 DRAFT 是为了推动虚拟化环境下的设备管理标准化，通过简化技术实现，提高性能效率，并确保跨平台的兼容性。这对于构建灵活、高性能的虚拟化基础设施至关重要。这份规格文档不仅定义了设备的接口和行为，还为开发人员提供了如何在虚拟环境中实现类似物理设备功能的指导原则，促进了整个行业的技术进步和发展。

u16 f l a g s ;

u16 idx ;

u16 ring [ qsz ] ; /

∗

qsz i s the Queue Size f i e l d read from device

∗

u16 used_event ;

};

2.3.5 Used Ring

The used ring is where the device returns buers once it is done with them. The

ags eld can be used by the device to hint that no notication is necessary when

the guest adds to the

available

ring. Alternatively, the avail_event eld can be

used by the device to hint that no notication is necessary until an entry with an

index specied by the avail_event is written in the available ring (equivalently,

until the

idx

eld in the available ring will reach the value

avail_event + 1

The method employed by the device is controlled by the guest through the

VIRTIO_RING_F_EVENT_IDX feature bit (see 3).

Each entry in the ring is a pair: the head entry of the descriptor chain describing

the buer (this matches an entry placed in the available ring by the guest

earlier), and the total of bytes written into the buer. The latter is extremely

useful for guests using untrusted buers: if you do not know exactly how much

has been written by the device, you usually have to zero the buer to ensure no

data leakage occurs.

∗

u32 i s used here f o r ids for padding reasons .

∗

stru c t vring_used_elem {

∗

Index of s ta r t of used descri p t o r chain .

∗

u32 id ;

∗

Total length of the desc r i p t o r chain which was used ( written to )

∗

u32 len ;

};

stru c t vring_used {

#defin e VRING_USED_F_NO_NOTIFY 1

u16 f l a g s ;

u16 idx ;

stru c t vring_used_elem ring [ qsz ] ;

u16 avail_event ;

};

These elds are kept here because this is the only part of the virtqueue written by the

device

2.3.6 Helpers for Managing Virtqueues

The Linux Kernel Source code contains the denitions above and helper rou-

tines in a more usable form, in include/linux/virtio_ring.h. This was explicitly

licensed by IBM and Red Hat under the (3-clause) BSD license so that it can

be freely used by all other projects, and is reproduced (with slight variation to

remove Linux assumptions) in Appendix A.

2.4 Device Operation

There are two parts to device operation: supplying new buers to the device,

and processing used buers from the device. As an example, the virtio network

device has two virtqueues: the transmit virtqueue and the receive virtqueue.

The driver adds outgoing (read-only) packets to the transmit virtqueue, and

then frees them after they are used. Similarly, incoming (write-only) buers are

added to the receive virtqueue, and processed after they are used.

2.4.1 Supplying Buers to The Device

Actual transfer of buers from the guest OS to the device operates as follows:

1. Place the buer(s) into free descriptor(s).

(a) If there are no free descriptors, the guest may choose to notify the

device even if notications are suppressed (to reduce latency).

2. Place the id of the buer in the next ring entry of the available ring.

3. The steps (1) and (2) may be performed repeatedly if batching is possible.

4. A memory barrier should be executed to ensure the device sees the updated

descriptor table and available ring before the next step.

5. The available idx eld should be increased by the number of entries

added to the available ring.

6. A memory barrier should be executed to ensure that we update the idx

eld before checking for notication suppression.

7. If notications are not suppressed, the device should be notied of the

new buers.

The Linux drivers do this only for read-only buers: for write-only buers, it is assumed

that the driver is merely trying to keep the receive buer ring full, and no notication of this

expected condition is necessary.

剩余61页未读，继续阅读

rtoax

粉丝: 2739
资源: 216

Virtio PCI设备规格草案0.9.5：标准、高效虚拟环境接口

virtio协议

Virtio PCI Card Specification

virtio-v1.1.pdf

virtio-v1.0-cs04.pdf

virtio-v1.1-csprd01.pdf

Virtio An IO virtualization framework for Linux.pdf

caif_virtio.rar_If...

virtio驱动virtio-win-0.1.240版本

virtio-win-0.1.171.iso

virtio-win-0.1.204.iso

最新资源