1The Linux Kernel Device Model 2 3Patrick Mochel <mochel@digitalimplant.org> 4 5Drafted 26 August 2002 6Updated 31 January 2006 7 8 9Overview 10~~~~~~~~ 11 12The Linux Kernel Driver Model is a unification of all the disparate driver 13models that were previously used in the kernel. It is intended to augment the 14bus-specific drivers for bridges and devices by consolidating a set of data 15and operations into globally accessible data structures. 16 17Traditional driver models implemented some sort of tree-like structure 18(sometimes just a list) for the devices they control. There wasn't any 19uniformity across the different bus types. 20 21The current driver model provides a common, uniform data model for describing 22a bus and the devices that can appear under the bus. The unified bus 23model includes a set of common attributes which all busses carry, and a set 24of common callbacks, such as device discovery during bus probing, bus 25shutdown, bus power management, etc. 26 27The common device and bridge interface reflects the goals of the modern 28computer: namely the ability to do seamless device "plug and play", power 29management, and hot plug. In particular, the model dictated by Intel and 30Microsoft (namely ACPI) ensures that almost every device on almost any bus 31on an x86-compatible system can work within this paradigm. Of course, 32not every bus is able to support all such operations, although most 33buses support most of those operations. 34 35 36Downstream Access 37~~~~~~~~~~~~~~~~~ 38 39Common data fields have been moved out of individual bus layers into a common 40data structure. These fields must still be accessed by the bus layers, 41and sometimes by the device-specific drivers. 42 43Other bus layers are encouraged to do what has been done for the PCI layer. 44struct pci_dev now looks like this: 45 46struct pci_dev { 47 ... 48 49 struct device dev; /* Generic device interface */ 50 ... 51}; 52 53Note first that the struct device dev within the struct pci_dev is 54statically allocated. This means only one allocation on device discovery. 55 56Note also that that struct device dev is not necessarily defined at the 57front of the pci_dev structure. This is to make people think about what 58they're doing when switching between the bus driver and the global driver, 59and to discourage meaningless and incorrect casts between the two. 60 61The PCI bus layer freely accesses the fields of struct device. It knows about 62the structure of struct pci_dev, and it should know the structure of struct 63device. Individual PCI device drivers that have been converted to the current 64driver model generally do not and should not touch the fields of struct device, 65unless there is a compelling reason to do so. 66 67The above abstraction prevents unnecessary pain during transitional phases. 68If it were not done this way, then when a field was renamed or removed, every 69downstream driver would break. On the other hand, if only the bus layer 70(and not the device layer) accesses the struct device, it is only the bus 71layer that needs to change. 72 73 74User Interface 75~~~~~~~~~~~~~~ 76 77By virtue of having a complete hierarchical view of all the devices in the 78system, exporting a complete hierarchical view to userspace becomes relatively 79easy. This has been accomplished by implementing a special purpose virtual 80file system named sysfs. 81 82Almost all mainstream Linux distros mount this filesystem automatically; you 83can see some variation of the following in the output of the "mount" command: 84 85$ mount 86... 87none on /sys type sysfs (rw,noexec,nosuid,nodev) 88... 89$ 90 91The auto-mounting of sysfs is typically accomplished by an entry similar to 92the following in the /etc/fstab file: 93 94none /sys sysfs defaults 0 0 95 96or something similar in the /lib/init/fstab file on Debian-based systems: 97 98none /sys sysfs nodev,noexec,nosuid 0 0 99 100If sysfs is not automatically mounted, you can always do it manually with: 101 102# mount -t sysfs sysfs /sys 103 104Whenever a device is inserted into the tree, a directory is created for it. 105This directory may be populated at each layer of discovery - the global layer, 106the bus layer, or the device layer. 107 108The global layer currently creates two files - 'name' and 'power'. The 109former only reports the name of the device. The latter reports the 110current power state of the device. It will also be used to set the current 111power state. 112 113The bus layer may also create files for the devices it finds while probing the 114bus. For example, the PCI layer currently creates 'irq' and 'resource' files 115for each PCI device. 116 117A device-specific driver may also export files in its directory to expose 118device-specific data or tunable interfaces. 119 120More information about the sysfs directory layout can be found in 121the other documents in this directory and in the file 122Documentation/filesystems/sysfs.txt. 123 124