As a programmer, you will be able to make your own choices about your driver, choosing an acceptable tradeoff between the programming time required and the flexibility of the result. Though it may appear strange to say that a driver is ``flexible,'' I like this word because it emphasizes that the role of a device driver is providing mechanisms, not policies.

The distinction between mechanism and policy is one of the best ideas behind the Unix design. Most programming problems can indeed be split into two parts: ``what needs to be done'' (the mechanism) and ``how can the program be used'' (the policy). If the two issues are addressed by different parts of the program, or even by different programs altogether, the software package is much easier to develop and to adapt to particular needs.

For example, Unix management of the graphic display is split between the X server, which knows the hardware and offers a unified interface to user programs, and the window manager, which implements a particular policy without knowing anything about the hardware. People can use the same window manager on different hardware, and different users can run different configurations on the same workstation. Another example is the layered structure of TCP/IP networking: the operating system offers the socket abstraction, which is policy-free, while different servers are in charge of the services. Moreover, a server like ftpd provides the file transfer mechanism, while users can use whatever client they prefer; both command-line and graphic clients exist, and anyone can write a new user interface to transfer files.

Where drivers are concerned, the same role-splitting applies. The floppy driver is policy-free—its role is only to show the diskette as a continuous byte array. How to use the device is the role of the application: tar writes it sequentially, while mkfs prepares the device to be mounted, and mcopy relies on the existence of a specific data structure on the device.

When writing drivers, a programmer should pay particular attention to this fundamental problem: we need to write kernel code to access the hardware, but we shouldn’t force particular policies on the user, since different users have different needs. The driver should only deal with hardware handling, leaving all the issues about how to use the hardware to the applications. A driver, then, is ``flexible'' if it offers access to the hardware capabilities without adding constraints. Sometimes, however, some policy decisions must be made.

You can also look at your driver from a different perspective: it is a software layer that lies between the applications and the actual device. This privileged role of the driver allows the driver programmer to choose exactly how the device should appear: different drivers can offer different capabilities, even for the same device. The actual driver design should be a balance between many different considerations. For instance, a single device may be used concurrently by different programs, and the driver programmer has complete freedom to determine how to handle concurrency. You could implement memory mapping on the device independently of its hardware capabilities, or you could provide a user library to help application programmers implement new policies on top of the available primitives, and so forth. One major consideration is the tradeoff between the desire to present the user with as many options as possible, balanced against the time you have to do the writing and the need to keep things simple so that errors don’t creep in.

If a driver is designed for both synchronous and asynchronous operations, if it allows itself to be opened multiple times, and if it is able to exploit all the hardware capabilities without adding a software layer ``to simplify things''—like converting binary data to text or other policy-related operations—then it will turn out to be easier to write and to maintain. Being ``policy-free'' is actually a common target for software designers.

Most device drivers, indeed, are released together with user programs to help with configuration and access to the target device. Those programs can range from simple configuring utilities to complete graphical applications. Usually a client library is provided as well.

The scope of this book is the kernel, so we’ll try not to deal with policy issues, nor with application programs or support libraries. Sometimes I’ll talk about different policies and how to support them, but I won’t go into much detail about programs using the device or the policies they enforce. You should understand, however, that user programs are an integral part of a software package and that even policy-free packages are distributed with configuration files that apply a default behavior to the underlying mechanisms.