Device number

A device number identifies a device interface using a major device number and a minor device number. In the Linux interfaces discussed here, the combined value has type dev_t; makedev combines the two components and major and minor recover them. The documented interpretation associates the major value with a device class and the minor value with an instance within that class.^[ref]

A device node also has an inode number, and its containing file system has a device identity of its own. The fields st_ino, st_dev and st_rdev therefore answer different questions: which inode, which containing device, and which represented device for a device-special file.^[ref]

For an original metadata comparison, stipulate two character-special nodes in one file system with inode numbers 40 and 90, each representing major 240, minor 7. They are different named file-system objects representing the same interface in this example. A third node with major 240, minor 8 represents a different minor interface. The file type remains relevant: character and block device numbers are not equated solely by matching two integers.

Combining the components is not decimal concatenation. The pairs $(1,23)$ and $(12,3)$ both become the character string 123 if their decimal spellings are joined without a separator. The structured pairs are nevertheless different. The appropriate dev_t conversion preserves their distinct components rather than imposing that ambiguous textual encoding.

For valid supported component values, the intended round trips areUse the defined conversion interface rather than assuming a particular in-memory packing. The device number is an interface identifier; it is not automatically a serial number, a stable physical-device identity or the process-local file descriptor returned by opening the node.

Linux device-number allocation assigns or reserves the numerical identifiers through which device-special interfaces are registered. The relevant identity includes device type together with major device number and minor device number. It is distinct from naming a device node, assigning its inode number, or obtaining a process-local file descriptor.

The kernel's allocated-device list records established assignments. For example, character major 1 includes minor 3 for /dev/null and minor 5 for /dev/zero. The same integer major 1 has a separate block-device allocation for RAM disks. These entries concern different interface namespaces, not interchangeable interpretations of a single untyped integer.^[ref]

For character devices, the kernel API separates requesting a particular range with register_chrdev_region from dynamic character-device number allocation with alloc_chrdev_region. The latter receives a first minor and count, chooses the major dynamically, and returns the first assigned combined device number. The caller must check whether reservation succeeded.^[ref]

An original allocation example requests four minor values starting at 20. If the returned major is $M$, the reserved range in this example isIt does not contain $(M,24)$. Code should use the returned $M$ rather than guess a major from an earlier machine or boot. A numerical allocation can serve software interfaces as well as physical hardware, so four minors do not necessarily imply four physical peripherals.

Reserving numbers is only one step. The character-device object must also be associated with its operations and added to the kernel; cdev_add makes it live. Separately, userspace needs an appropriate device node or other naming arrangement. The registration name used for bookkeeping is not itself a pathname under /dev.^[ref]

This separation explains a common diagnostic trap. Creating a node containing a chosen major/minor pair does not reserve that range or supply a working device driver. Conversely, allocating a range does not make every invented pathname reach it. Inspect the actual registered interface, file type and stored device number together when identifying a device.

The allocation is also not a persistent hardware serial number. Dynamic assignment can differ between configurations or registration histories. Releasing a number range with unregister_chrdev_region returns the reservation; a complete driver teardown must also handle the device object's lifetime and existing users. The kernel documentation notes that deleting a character-device object prevents new opens while operations on already open descriptors can remain possible.^[ref]

Dynamic character-device number allocation asks the Linux kernel to reserve a range of character-device numbers while choosing its major device number at runtime. The API is alloc_chrdev_region(dev_t *dev, unsigned baseminor, unsigned count, const char *name). It returns zero on success or a negative error code; the first assigned combined device number is written through dev. The return status is not the allocated major number.^[ref]

For an original request, suppose alloc_chrdev_region(&first, 5, 3, "sample") succeeds and the returned first has major $M$. Its reserved identifiers are $(M,5)$, $(M,6)$ and $(M,7)$, not $(M,8)$. If one illustrative successful run returned major 243, the caller would use 243; another allocation can return a different major. This is a hypothetical range example, not a kernel operation executed here.

Keep first and the successful reservation count for cleanup. If a later initialization step fails, unregister_chrdev_region(first, 3) releases this reservation; it must not be treated as ownership of every minor under the returned major. Unwinding a failed character-device registration releases only resources actually acquired.

Number reservation is separate from associating file operations with a live character-device object and from creating a device node. The related cdev_add operation makes a registered object live; its name or bookkeeping label does not itself establish a pathname under /dev.^[ref]

Removal must also respect existing users. cdev_del prevents new opens, but already open file descriptors can still call device operations, so deleting the registration alone does not prove that backing state can be freed immediately.^[ref] A dynamic number is a reserved interface identifier, not a persistent hardware identity or an inode number.

Unwinding a failed character-device registration releases resources acquired before an initialization failure while preserving the lifetime of anything that concurrent users can still access. In Linux, reserving device numbers, publishing the character-device object and creating userspace-visible names are distinct steps. The cleanup path must know which steps actually succeeded.

Consider an original abstract initialization sequence: reserve a number range, initialize a character-device object, publish it, then finish a later setup step. The following table tracks successful acquisitions; it is not a complete driver implementation.

alloc_chrdev_region reserves a range and reports failure through its return value. cdev_init initializes the object with its operations; cdev_add publishes it and makes it live immediately. unregister_chrdev_region releases a successful number reservation. Pair each release with the acquisition that actually succeeded, using the same first number and count.^[ref]

For a concrete bookkeeping example, suppose a successful reservation contains four interfaces starting at returned pair $(M,20)$. Failure before publication requires releasing that four-number range. Releasing only three leaves part of the reservation behind; releasing a guessed starting major or an unacquired range is a different error. The returned major from dynamic character-device number allocation must be retained rather than reconstructed from a pathname.

Publication changes the lifetime problem. cdev_del prevents new opens, but existing open file descriptors can still invoke their file operations. Therefore removing the device from lookup is not a proof that its backing state can immediately be freed. Character-device lifetime after deletion must account for the references and callbacks already in progress.^[ref]

The combined cdev_device_add helper has an especially important failure contract: callers must assume userspace could open the device and call operations even if the helper returns failure. A blanket error handler that immediately frees all backing state would violate that contract.^[ref] Design ownership and reference handling around the documented publication boundary, then release each owned resource once. The general lesson is that reverse acquisition order must be combined with concurrency-aware lifetime management.

Character-device lifetime after deletion is the interval during which a removed character device still needs valid backing state because existing users can invoke its operations. Linux cdev_del removes the device from new-open lookup; it does not invalidate already-open file descriptors. The kernel explicitly permits their file-operation callbacks after deletion returns.^[ref]

Consider an original event sequence. Process A opens an interface and retains a pointer to driver state. Removal calls cdev_del; process B can no longer obtain a new open through that registration. Process A then calls read. Freeing the state immediately after deletion would make that callback access dead storage: a use-after-free. The lookup lifetime and the backing allocation lifetime are different.

A useful ownership model gives the allocation one registration reference and one reference for each successful open. In a simplified reference counting design with $n$ opens, the reference count is $1+n$. Removal drops the registration reference; each final close drops its open reference. With two opens, the sequence is $3\to 2\to 1\to 0$, and storage is reclaimed only at zero. This is a model of ownership, not a complete driver implementation: taking references must itself be synchronized with removal, and asynchronous work can require additional references.

The character device may be disconnected from its hardware before its storage is reclaimed. Existing operations can observe a synchronized disconnected state and return an appropriate error while the allocation remains valid. Reference protection alone does not make removed hardware usable or serialize simultaneous operations.

For combined device objects, cdev_device_add and cdev_device_del have their own documented pairing and parent-reference rules. The add helper can expose callbacks even when it reports failure; the delete helper distinguishes stopped sysfs callbacks from still-callable operations on existing opens.^[ref] Therefore unwinding a failed character-device registration must follow the actual publication and ownership state. Releasing the device number reservation is separate from proving that every user of private storage has finished.

A minor device number is the second component of a device number. In the Linux interface, minor(d) extracts it from a value of type dev_t; the matching major device number supplies its device-class context.^[ref]

A concrete Linux allocation distinguishes two character-special interfaces sharing major1: /dev/null has minor3, while /dev/zero has minor5.^[ref] The former supplies end-of-file when read, while the latter supplies zero-valued bytes. Their different minor values select different interface behavior within the same major allocation.^[ref]

For an original pair comparison, let two registered character-device interfaces have numbers $(240,7)$ and $(241,7)$. Both have minor7, but their complete device numbers differ. Conversely, $(240,7)$ and $(240,8)$ share a major while selecting different minor values within it. A minor value alone is therefore insufficient to identify the interface.

Minor numbers are interpreted in the relevant device allocation; there is no universal rule that every minor counts disks, partitions or physical pieces of hardware. The device may also be a purely virtual interface. The conversion functions preserve the structured components:for supported component values. Extracting the component with the documented interface avoids assuming an implementation-specific bit packing.

A major device number is the first component of a device number. In Linux interfaces it identifies a device class or registered device-number allocation; the minor device number distinguishes interfaces within that context. major(d) extracts the major component from the combined dev_t value.^[ref]

The device type is also significant. The Linux registry assigns character major1 to memory-related interfaces, including /dev/null, while block major1 describes RAM disks. Reusing the integer1 across these two types does not equate their behavior.^[ref]

An original metadata example makes the components explicit. Suppose a character-special node represents $(\mathrm{major},\mathrm{minor})=(240,12)$. Renaming that node within the file system does not by itself change its stored pair. A second node can represent the same pair, while a node with pair $(241,12)$ has a different major context despite matching minor12.

Linux device-number allocation can be dynamic, so a major number should not automatically be treated as a permanent hardware serial number or a universally fixed driver label.^[ref] Its role is part of the device interface and registration scheme.

The structured conversion satisfiesfor supported values. An inode number identifies a file-system object, while the represented device's major/minor pair answers a different question. The containing file system's own device identity is another distinct field.^[ref]