The Linux Kernel API

List Management Functions

void INIT_LIST_HEAD(struct list_head * list)

Initialize a list_head structure

Parameters

struct list_head * list
list_head structure to be initialized.

Description

Initializes the list_head to point to itself. If it is a list header, the result is an empty list.

void list_add(struct list_head * new, struct list_head * head)

add a new entry

Parameters

struct list_head * new
new entry to be added
struct list_head * head
list head to add it after

Description

Insert a new entry after the specified head. This is good for implementing stacks.

void list_add_tail(struct list_head * new, struct list_head * head)

add a new entry

Parameters

struct list_head * new
new entry to be added
struct list_head * head
list head to add it before

Description

Insert a new entry before the specified head. This is useful for implementing queues.

void list_del(struct list_head * entry)

deletes entry from list.

Parameters

struct list_head * entry
the element to delete from the list.

Note

list_empty() on entry does not return true after this, the entry is in an undefined state.

void list_replace(struct list_head * old, struct list_head * new)

replace old entry by new one

Parameters

struct list_head * old
the element to be replaced
struct list_head * new
the new element to insert

Description

If old was empty, it will be overwritten.

void list_replace_init(struct list_head * old, struct list_head * new)

replace old entry by new one and initialize the old one

Parameters

struct list_head * old
the element to be replaced
struct list_head * new
the new element to insert

Description

If old was empty, it will be overwritten.

void list_swap(struct list_head * entry1, struct list_head * entry2)

replace entry1 with entry2 and re-add entry1 at entry2’s position

Parameters

struct list_head * entry1
the location to place entry2
struct list_head * entry2
the location to place entry1
void list_del_init(struct list_head * entry)

deletes entry from list and reinitialize it.

Parameters

struct list_head * entry
the element to delete from the list.
void list_move(struct list_head * list, struct list_head * head)

delete from one list and add as another’s head

Parameters

struct list_head * list
the entry to move
struct list_head * head
the head that will precede our entry
void list_move_tail(struct list_head * list, struct list_head * head)

delete from one list and add as another’s tail

Parameters

struct list_head * list
the entry to move
struct list_head * head
the head that will follow our entry
void list_bulk_move_tail(struct list_head * head, struct list_head * first, struct list_head * last)

move a subsection of a list to its tail

Parameters

struct list_head * head
the head that will follow our entry
struct list_head * first
first entry to move
struct list_head * last
last entry to move, can be the same as first

Description

Move all entries between first and including last before head. All three entries must belong to the same linked list.

int list_is_first(const struct list_head * list, const struct list_head * head)
  • tests whether list is the first entry in list head

Parameters

const struct list_head * list
the entry to test
const struct list_head * head
the head of the list
int list_is_last(const struct list_head * list, const struct list_head * head)

tests whether list is the last entry in list head

Parameters

const struct list_head * list
the entry to test
const struct list_head * head
the head of the list
int list_empty(const struct list_head * head)

tests whether a list is empty

Parameters

const struct list_head * head
the list to test.
int list_empty_careful(const struct list_head * head)

tests whether a list is empty and not being modified

Parameters

const struct list_head * head
the list to test

Description

tests whether a list is empty _and_ checks that no other CPU might be in the process of modifying either member (next or prev)

NOTE

using list_empty_careful() without synchronization can only be safe if the only activity that can happen to the list entry is list_del_init(). Eg. it cannot be used if another CPU could re-list_add() it.

void list_rotate_left(struct list_head * head)

rotate the list to the left

Parameters

struct list_head * head
the head of the list
void list_rotate_to_front(struct list_head * list, struct list_head * head)

Rotate list to specific item.

Parameters

struct list_head * list
The desired new front of the list.
struct list_head * head
The head of the list.

Description

Rotates list so that list becomes the new front of the list.

int list_is_singular(const struct list_head * head)

tests whether a list has just one entry.

Parameters

const struct list_head * head
the list to test.
void list_cut_position(struct list_head * list, struct list_head * head, struct list_head * entry)

cut a list into two

Parameters

struct list_head * list
a new list to add all removed entries
struct list_head * head
a list with entries
struct list_head * entry
an entry within head, could be the head itself and if so we won’t cut the list

Description

This helper moves the initial part of head, up to and including entry, from head to list. You should pass on entry an element you know is on head. list should be an empty list or a list you do not care about losing its data.

void list_cut_before(struct list_head * list, struct list_head * head, struct list_head * entry)

cut a list into two, before given entry

Parameters

struct list_head * list
a new list to add all removed entries
struct list_head * head
a list with entries
struct list_head * entry
an entry within head, could be the head itself

Description

This helper moves the initial part of head, up to but excluding entry, from head to list. You should pass in entry an element you know is on head. list should be an empty list or a list you do not care about losing its data. If entry == head, all entries on head are moved to list.

void list_splice(const struct list_head * list, struct list_head * head)

join two lists, this is designed for stacks

Parameters

const struct list_head * list
the new list to add.
struct list_head * head
the place to add it in the first list.
void list_splice_tail(struct list_head * list, struct list_head * head)

join two lists, each list being a queue

Parameters

struct list_head * list
the new list to add.
struct list_head * head
the place to add it in the first list.
void list_splice_init(struct list_head * list, struct list_head * head)

join two lists and reinitialise the emptied list.

Parameters

struct list_head * list
the new list to add.
struct list_head * head
the place to add it in the first list.

Description

The list at list is reinitialised

void list_splice_tail_init(struct list_head * list, struct list_head * head)

join two lists and reinitialise the emptied list

Parameters

struct list_head * list
the new list to add.
struct list_head * head
the place to add it in the first list.

Description

Each of the lists is a queue. The list at list is reinitialised

list_entry(ptr, type, member)

get the struct for this entry

Parameters

ptr
the struct list_head pointer.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.
list_first_entry(ptr, type, member)

get the first element from a list

Parameters

ptr
the list head to take the element from.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

Note, that list is expected to be not empty.

list_last_entry(ptr, type, member)

get the last element from a list

Parameters

ptr
the list head to take the element from.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

Note, that list is expected to be not empty.

list_first_entry_or_null(ptr, type, member)

get the first element from a list

Parameters

ptr
the list head to take the element from.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

Note that if the list is empty, it returns NULL.

list_next_entry(pos, member)

get the next element in list

Parameters

pos
the type * to cursor
member
the name of the list_head within the struct.
list_prev_entry(pos, member)

get the prev element in list

Parameters

pos
the type * to cursor
member
the name of the list_head within the struct.
list_for_each(pos, head)

iterate over a list

Parameters

pos
the struct list_head to use as a loop cursor.
head
the head for your list.
list_for_each_continue(pos, head)

continue iteration over a list

Parameters

pos
the struct list_head to use as a loop cursor.
head
the head for your list.

Description

Continue to iterate over a list, continuing after the current position.

list_for_each_prev(pos, head)

iterate over a list backwards

Parameters

pos
the struct list_head to use as a loop cursor.
head
the head for your list.
list_for_each_safe(pos, n, head)

iterate over a list safe against removal of list entry

Parameters

pos
the struct list_head to use as a loop cursor.
n
another struct list_head to use as temporary storage
head
the head for your list.
list_for_each_prev_safe(pos, n, head)

iterate over a list backwards safe against removal of list entry

Parameters

pos
the struct list_head to use as a loop cursor.
n
another struct list_head to use as temporary storage
head
the head for your list.
list_for_each_entry(pos, head, member)

iterate over list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.
list_for_each_entry_reverse(pos, head, member)

iterate backwards over list of given type.

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.
list_prepare_entry(pos, head, member)

prepare a pos entry for use in list_for_each_entry_continue()

Parameters

pos
the type * to use as a start point
head
the head of the list
member
the name of the list_head within the struct.

Description

Prepares a pos entry for use as a start point in list_for_each_entry_continue().

list_for_each_entry_continue(pos, head, member)

continue iteration over list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.

Description

Continue to iterate over list of given type, continuing after the current position.

list_for_each_entry_continue_reverse(pos, head, member)

iterate backwards from the given point

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.

Description

Start to iterate over list of given type backwards, continuing after the current position.

list_for_each_entry_from(pos, head, member)

iterate over list of given type from the current point

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.

Description

Iterate over list of given type, continuing from current position.

list_for_each_entry_from_reverse(pos, head, member)

iterate backwards over list of given type from the current point

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.

Description

Iterate backwards over list of given type, continuing from current position.

list_for_each_entry_safe(pos, n, head, member)

iterate over list of given type safe against removal of list entry

Parameters

pos
the type * to use as a loop cursor.
n
another type * to use as temporary storage
head
the head for your list.
member
the name of the list_head within the struct.
list_for_each_entry_safe_continue(pos, n, head, member)

continue list iteration safe against removal

Parameters

pos
the type * to use as a loop cursor.
n
another type * to use as temporary storage
head
the head for your list.
member
the name of the list_head within the struct.

Description

Iterate over list of given type, continuing after current point, safe against removal of list entry.

list_for_each_entry_safe_from(pos, n, head, member)

iterate over list from current point safe against removal

Parameters

pos
the type * to use as a loop cursor.
n
another type * to use as temporary storage
head
the head for your list.
member
the name of the list_head within the struct.

Description

Iterate over list of given type from current point, safe against removal of list entry.

list_for_each_entry_safe_reverse(pos, n, head, member)

iterate backwards over list safe against removal

Parameters

pos
the type * to use as a loop cursor.
n
another type * to use as temporary storage
head
the head for your list.
member
the name of the list_head within the struct.

Description

Iterate backwards over list of given type, safe against removal of list entry.

list_safe_reset_next(pos, n, member)

reset a stale list_for_each_entry_safe loop

Parameters

pos
the loop cursor used in the list_for_each_entry_safe loop
n
temporary storage used in list_for_each_entry_safe
member
the name of the list_head within the struct.

Description

list_safe_reset_next is not safe to use in general if the list may be modified concurrently (eg. the lock is dropped in the loop body). An exception to this is if the cursor element (pos) is pinned in the list, and list_safe_reset_next is called after re-taking the lock and before completing the current iteration of the loop body.

int hlist_unhashed(const struct hlist_node * h)

Has node been removed from list and reinitialized?

Parameters

const struct hlist_node * h
Node to be checked

Description

Not that not all removal functions will leave a node in unhashed state. For example, hlist_nulls_del_init_rcu() does leave the node in unhashed state, but hlist_nulls_del() does not.

int hlist_unhashed_lockless(const struct hlist_node * h)

Version of hlist_unhashed for lockless use

Parameters

const struct hlist_node * h
Node to be checked

Description

This variant of hlist_unhashed() must be used in lockless contexts to avoid potential load-tearing. The READ_ONCE() is paired with the various WRITE_ONCE() in hlist helpers that are defined below.

int hlist_empty(const struct hlist_head * h)

Is the specified hlist_head structure an empty hlist?

Parameters

const struct hlist_head * h
Structure to check.
void hlist_del(struct hlist_node * n)

Delete the specified hlist_node from its list

Parameters

struct hlist_node * n
Node to delete.

Description

Note that this function leaves the node in hashed state. Use hlist_del_init() or similar instead to unhash n.

void hlist_del_init(struct hlist_node * n)

Delete the specified hlist_node from its list and initialize

Parameters

struct hlist_node * n
Node to delete.

Description

Note that this function leaves the node in unhashed state.

void hlist_add_head(struct hlist_node * n, struct hlist_head * h)

add a new entry at the beginning of the hlist

Parameters

struct hlist_node * n
new entry to be added
struct hlist_head * h
hlist head to add it after

Description

Insert a new entry after the specified head. This is good for implementing stacks.

void hlist_add_before(struct hlist_node * n, struct hlist_node * next)

add a new entry before the one specified

Parameters

struct hlist_node * n
new entry to be added
struct hlist_node * next
hlist node to add it before, which must be non-NULL
void hlist_add_behind(struct hlist_node * n, struct hlist_node * prev)

add a new entry after the one specified

Parameters

struct hlist_node * n
new entry to be added
struct hlist_node * prev
hlist node to add it after, which must be non-NULL
void hlist_add_fake(struct hlist_node * n)

create a fake hlist consisting of a single headless node

Parameters

struct hlist_node * n
Node to make a fake list out of

Description

This makes n appear to be its own predecessor on a headless hlist. The point of this is to allow things like hlist_del() to work correctly in cases where there is no list.

bool hlist_fake(struct hlist_node * h)

Parameters

struct hlist_node * h
Node to check for being a self-referential fake hlist.
bool hlist_is_singular_node(struct hlist_node * n, struct hlist_head * h)

is node the only element of the specified hlist?

Parameters

struct hlist_node * n
Node to check for singularity.
struct hlist_head * h
Header for potentially singular list.

Description

Check whether the node is the only node of the head without accessing head, thus avoiding unnecessary cache misses.

void hlist_move_list(struct hlist_head * old, struct hlist_head * new)

Move an hlist

Parameters

struct hlist_head * old
hlist_head for old list.
struct hlist_head * new
hlist_head for new list.

Description

Move a list from one list head to another. Fixup the pprev reference of the first entry if it exists.

hlist_for_each_entry(pos, head, member)

iterate over list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the hlist_node within the struct.
hlist_for_each_entry_continue(pos, member)

iterate over a hlist continuing after current point

Parameters

pos
the type * to use as a loop cursor.
member
the name of the hlist_node within the struct.
hlist_for_each_entry_from(pos, member)

iterate over a hlist continuing from current point

Parameters

pos
the type * to use as a loop cursor.
member
the name of the hlist_node within the struct.
hlist_for_each_entry_safe(pos, n, head, member)

iterate over list of given type safe against removal of list entry

Parameters

pos
the type * to use as a loop cursor.
n
another struct hlist_node to use as temporary storage
head
the head for your list.
member
the name of the hlist_node within the struct.

Basic C Library Functions

When writing drivers, you cannot in general use routines which are from the C Library. Some of the functions have been found generally useful and they are listed below. The behaviour of these functions may vary slightly from those defined by ANSI, and these deviations are noted in the text.

String Conversions

unsigned long long simple_strtoull(const char * cp, char ** endp, unsigned int base)

convert a string to an unsigned long long

Parameters

const char * cp
The start of the string
char ** endp
A pointer to the end of the parsed string will be placed here
unsigned int base
The number base to use

Description

This function is obsolete. Please use kstrtoull instead.

unsigned long simple_strtoul(const char * cp, char ** endp, unsigned int base)

convert a string to an unsigned long

Parameters

const char * cp
The start of the string
char ** endp
A pointer to the end of the parsed string will be placed here
unsigned int base
The number base to use

Description

This function is obsolete. Please use kstrtoul instead.

long simple_strtol(const char * cp, char ** endp, unsigned int base)

convert a string to a signed long

Parameters

const char * cp
The start of the string
char ** endp
A pointer to the end of the parsed string will be placed here
unsigned int base
The number base to use

Description

This function is obsolete. Please use kstrtol instead.

long long simple_strtoll(const char * cp, char ** endp, unsigned int base)

convert a string to a signed long long

Parameters

const char * cp
The start of the string
char ** endp
A pointer to the end of the parsed string will be placed here
unsigned int base
The number base to use

Description

This function is obsolete. Please use kstrtoll instead.

int vsnprintf(char * buf, size_t size, const char * fmt, va_list args)

Format a string and place it in a buffer

Parameters

char * buf
The buffer to place the result into
size_t size
The size of the buffer, including the trailing null space
const char * fmt
The format string to use
va_list args
Arguments for the format string

Description

This function generally follows C99 vsnprintf, but has some extensions and a few limitations:

  • ``n`` is unsupported
  • ``p``* is handled by pointer()

See pointer() or Documentation/core-api/printk-formats.rst for more extensive description.

Please update the documentation in both places when making changes

The return value is the number of characters which would be generated for the given input, excluding the trailing ‘0’, as per ISO C99. If you want to have the exact number of characters written into buf as return value (not including the trailing ‘0’), use vscnprintf(). If the return is greater than or equal to size, the resulting string is truncated.

If you’re not already dealing with a va_list consider using snprintf().

int vscnprintf(char * buf, size_t size, const char * fmt, va_list args)

Format a string and place it in a buffer

Parameters

char * buf
The buffer to place the result into
size_t size
The size of the buffer, including the trailing null space
const char * fmt
The format string to use
va_list args
Arguments for the format string

Description

The return value is the number of characters which have been written into the buf not including the trailing ‘0’. If size is == 0 the function returns 0.

If you’re not already dealing with a va_list consider using scnprintf().

See the vsnprintf() documentation for format string extensions over C99.

int snprintf(char * buf, size_t size, const char * fmt, ...)

Format a string and place it in a buffer

Parameters

char * buf
The buffer to place the result into
size_t size
The size of the buffer, including the trailing null space
const char * fmt
The format string to use
...
Arguments for the format string

Description

The return value is the number of characters which would be generated for the given input, excluding the trailing null, as per ISO C99. If the return is greater than or equal to size, the resulting string is truncated.

See the vsnprintf() documentation for format string extensions over C99.

int scnprintf(char * buf, size_t size, const char * fmt, ...)

Format a string and place it in a buffer

Parameters

char * buf
The buffer to place the result into
size_t size
The size of the buffer, including the trailing null space
const char * fmt
The format string to use
...
Arguments for the format string

Description

The return value is the number of characters written into buf not including the trailing ‘0’. If size is == 0 the function returns 0.

int vsprintf(char * buf, const char * fmt, va_list args)

Format a string and place it in a buffer

Parameters

char * buf
The buffer to place the result into
const char * fmt
The format string to use
va_list args
Arguments for the format string

Description

The function returns the number of characters written into buf. Use vsnprintf() or vscnprintf() in order to avoid buffer overflows.

If you’re not already dealing with a va_list consider using sprintf().

See the vsnprintf() documentation for format string extensions over C99.

int sprintf(char * buf, const char * fmt, ...)

Format a string and place it in a buffer

Parameters

char * buf
The buffer to place the result into
const char * fmt
The format string to use
...
Arguments for the format string

Description

The function returns the number of characters written into buf. Use snprintf() or scnprintf() in order to avoid buffer overflows.

See the vsnprintf() documentation for format string extensions over C99.

int vbin_printf(u32 * bin_buf, size_t size, const char * fmt, va_list args)

Parse a format string and place args’ binary value in a buffer

Parameters

u32 * bin_buf
The buffer to place args’ binary value
size_t size
The size of the buffer(by words(32bits), not characters)
const char * fmt
The format string to use
va_list args
Arguments for the format string

Description

The format follows C99 vsnprintf, except n is ignored, and its argument is skipped.

The return value is the number of words(32bits) which would be generated for the given input.

NOTE

If the return value is greater than size, the resulting bin_buf is NOT valid for bstr_printf().

int bstr_printf(char * buf, size_t size, const char * fmt, const u32 * bin_buf)

Format a string from binary arguments and place it in a buffer

Parameters

char * buf
The buffer to place the result into
size_t size
The size of the buffer, including the trailing null space
const char * fmt
The format string to use
const u32 * bin_buf
Binary arguments for the format string

Description

This function like C99 vsnprintf, but the difference is that vsnprintf gets arguments from stack, and bstr_printf gets arguments from bin_buf which is a binary buffer that generated by vbin_printf.

The format follows C99 vsnprintf, but has some extensions:
see vsnprintf comment for details.

The return value is the number of characters which would be generated for the given input, excluding the trailing ‘0’, as per ISO C99. If you want to have the exact number of characters written into buf as return value (not including the trailing ‘0’), use vscnprintf(). If the return is greater than or equal to size, the resulting string is truncated.

int bprintf(u32 * bin_buf, size_t size, const char * fmt, ...)

Parse a format string and place args’ binary value in a buffer

Parameters

u32 * bin_buf
The buffer to place args’ binary value
size_t size
The size of the buffer(by words(32bits), not characters)
const char * fmt
The format string to use
...
Arguments for the format string

Description

The function returns the number of words(u32) written into bin_buf.

int vsscanf(const char * buf, const char * fmt, va_list args)

Unformat a buffer into a list of arguments

Parameters

const char * buf
input buffer
const char * fmt
format of buffer
va_list args
arguments
int sscanf(const char * buf, const char * fmt, ...)

Unformat a buffer into a list of arguments

Parameters

const char * buf
input buffer
const char * fmt
formatting of buffer
...
resulting arguments
int kstrtol(const char * s, unsigned int base, long * res)

convert a string to a long

Parameters

const char * s
The start of the string. The string must be null-terminated, and may also include a single newline before its terminating null. The first character may also be a plus sign or a minus sign.
unsigned int base
The number base to use. The maximum supported base is 16. If base is given as 0, then the base of the string is automatically detected with the conventional semantics - If it begins with 0x the number will be parsed as a hexadecimal (case insensitive), if it otherwise begins with 0, it will be parsed as an octal number. Otherwise it will be parsed as a decimal.
long * res
Where to write the result of the conversion on success.

Description

Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error. Used as a replacement for the simple_strtoull. Return code must be checked.

int kstrtoul(const char * s, unsigned int base, unsigned long * res)

convert a string to an unsigned long

Parameters

const char * s
The start of the string. The string must be null-terminated, and may also include a single newline before its terminating null. The first character may also be a plus sign, but not a minus sign.
unsigned int base
The number base to use. The maximum supported base is 16. If base is given as 0, then the base of the string is automatically detected with the conventional semantics - If it begins with 0x the number will be parsed as a hexadecimal (case insensitive), if it otherwise begins with 0, it will be parsed as an octal number. Otherwise it will be parsed as a decimal.
unsigned long * res
Where to write the result of the conversion on success.

Description

Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error. Used as a replacement for the simple_strtoull. Return code must be checked.

int kstrtoull(const char * s, unsigned int base, unsigned long long * res)

convert a string to an unsigned long long

Parameters

const char * s
The start of the string. The string must be null-terminated, and may also include a single newline before its terminating null. The first character may also be a plus sign, but not a minus sign.
unsigned int base
The number base to use. The maximum supported base is 16. If base is given as 0, then the base of the string is automatically detected with the conventional semantics - If it begins with 0x the number will be parsed as a hexadecimal (case insensitive), if it otherwise begins with 0, it will be parsed as an octal number. Otherwise it will be parsed as a decimal.
unsigned long long * res
Where to write the result of the conversion on success.

Description

Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error. Used as a replacement for the obsolete simple_strtoull. Return code must be checked.

int kstrtoll(const char * s, unsigned int base, long long * res)

convert a string to a long long

Parameters

const char * s
The start of the string. The string must be null-terminated, and may also include a single newline before its terminating null. The first character may also be a plus sign or a minus sign.
unsigned int base
The number base to use. The maximum supported base is 16. If base is given as 0, then the base of the string is automatically detected with the conventional semantics - If it begins with 0x the number will be parsed as a hexadecimal (case insensitive), if it otherwise begins with 0, it will be parsed as an octal number. Otherwise it will be parsed as a decimal.
long long * res
Where to write the result of the conversion on success.

Description

Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error. Used as a replacement for the obsolete simple_strtoull. Return code must be checked.

int kstrtouint(const char * s, unsigned int base, unsigned int * res)

convert a string to an unsigned int

Parameters

const char * s
The start of the string. The string must be null-terminated, and may also include a single newline before its terminating null. The first character may also be a plus sign, but not a minus sign.
unsigned int base
The number base to use. The maximum supported base is 16. If base is given as 0, then the base of the string is automatically detected with the conventional semantics - If it begins with 0x the number will be parsed as a hexadecimal (case insensitive), if it otherwise begins with 0, it will be parsed as an octal number. Otherwise it will be parsed as a decimal.
unsigned int * res
Where to write the result of the conversion on success.

Description

Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error. Used as a replacement for the obsolete simple_strtoull. Return code must be checked.

int kstrtoint(const char * s, unsigned int base, int * res)

convert a string to an int

Parameters

const char * s
The start of the string. The string must be null-terminated, and may also include a single newline before its terminating null. The first character may also be a plus sign or a minus sign.
unsigned int base
The number base to use. The maximum supported base is 16. If base is given as 0, then the base of the string is automatically detected with the conventional semantics - If it begins with 0x the number will be parsed as a hexadecimal (case insensitive), if it otherwise begins with 0, it will be parsed as an octal number. Otherwise it will be parsed as a decimal.
int * res
Where to write the result of the conversion on success.

Description

Returns 0 on success, -ERANGE on overflow and -EINVAL on parsing error. Used as a replacement for the obsolete simple_strtoull. Return code must be checked.

int kstrtobool(const char * s, bool * res)

convert common user inputs into boolean values

Parameters

const char * s
input string
bool * res
result

Description

This routine returns 0 iff the first character is one of ‘Yy1Nn0’, or [oO][NnFf] for “on” and “off”. Otherwise it will return -EINVAL. Value pointed to by res is updated upon finding a match.

void string_get_size(u64 size, u64 blk_size, const enum string_size_units units, char * buf, int len)

get the size in the specified units

Parameters

u64 size
The size to be converted in blocks
u64 blk_size
Size of the block (use 1 for size in bytes)
const enum string_size_units units
units to use (powers of 1000 or 1024)
char * buf
buffer to format to
int len
length of buffer

Description

This function returns a string formatted to 3 significant figures giving the size in the required units. buf should have room for at least 9 bytes and will always be zero terminated.

int string_unescape(char * src, char * dst, size_t size, unsigned int flags)

unquote characters in the given string

Parameters

char * src
source buffer (escaped)
char * dst
destination buffer (unescaped)
size_t size
size of the destination buffer (0 to unlimit)
unsigned int flags
combination of the flags.

Description

The function unquotes characters in the given string.

Because the size of the output will be the same as or less than the size of the input, the transformation may be performed in place.

Caller must provide valid source and destination pointers. Be aware that destination buffer will always be NULL-terminated. Source string must be NULL-terminated as well. The supported flags are:

UNESCAPE_SPACE:
        '\f' - form feed
        '\n' - new line
        '\r' - carriage return
        '\t' - horizontal tab
        '\v' - vertical tab
UNESCAPE_OCTAL:
        '\NNN' - byte with octal value NNN (1 to 3 digits)
UNESCAPE_HEX:
        '\xHH' - byte with hexadecimal value HH (1 to 2 digits)
UNESCAPE_SPECIAL:
        '\"' - double quote
        '\\' - backslash
        '\a' - alert (BEL)
        '\e' - escape
UNESCAPE_ANY:
        all previous together

Return

The amount of the characters processed to the destination buffer excluding trailing ‘0’ is returned.

int string_escape_mem(const char * src, size_t isz, char * dst, size_t osz, unsigned int flags, const char * only)

quote characters in the given memory buffer

Parameters

const char * src
source buffer (unescaped)
size_t isz
source buffer size
char * dst
destination buffer (escaped)
size_t osz
destination buffer size
unsigned int flags
combination of the flags
const char * only
NULL-terminated string containing characters used to limit the selected escape class. If characters are included in only that would not normally be escaped by the classes selected in flags, they will be copied to dst unescaped.

Description

The process of escaping byte buffer includes several parts. They are applied in the following sequence.

  1. The character is matched to the printable class, if asked, and in case of match it passes through to the output.
  2. The character is not matched to the one from only string and thus must go as-is to the output.
  3. The character is checked if it falls into the class given by flags. ESCAPE_OCTAL and ESCAPE_HEX are going last since they cover any character. Note that they actually can’t go together, otherwise ESCAPE_HEX will be ignored.

Caller must provide valid source and destination pointers. Be aware that destination buffer will not be NULL-terminated, thus caller have to append it if needs. The supported flags are:

%ESCAPE_SPACE: (special white space, not space itself)
        '\f' - form feed
        '\n' - new line
        '\r' - carriage return
        '\t' - horizontal tab
        '\v' - vertical tab
%ESCAPE_SPECIAL:
        '\\' - backslash
        '\a' - alert (BEL)
        '\e' - escape
%ESCAPE_NULL:
        '\0' - null
%ESCAPE_OCTAL:
        '\NNN' - byte with octal value NNN (3 digits)
%ESCAPE_ANY:
        all previous together
%ESCAPE_NP:
        escape only non-printable characters (checked by isprint)
%ESCAPE_ANY_NP:
        all previous together
%ESCAPE_HEX:
        '\xHH' - byte with hexadecimal value HH (2 digits)

Return

The total size of the escaped output that would be generated for the given input and flags. To check whether the output was truncated, compare the return value to osz. There is room left in dst for a ‘0’ terminator if and only if ret < osz.

String Manipulation

int strncasecmp(const char * s1, const char * s2, size_t len)

Case insensitive, length-limited string comparison

Parameters

const char * s1
One string
const char * s2
The other string
size_t len
the maximum number of characters to compare
char * strcpy(char * dest, const char * src)

Copy a NUL terminated string

Parameters

char * dest
Where to copy the string to
const char * src
Where to copy the string from
char * strncpy(char * dest, const char * src, size_t count)

Copy a length-limited, C-string

Parameters

char * dest
Where to copy the string to
const char * src
Where to copy the string from
size_t count
The maximum number of bytes to copy

Description

The result is not NUL-terminated if the source exceeds count bytes.

In the case where the length of src is less than that of count, the remainder of dest will be padded with NUL.

size_t strlcpy(char * dest, const char * src, size_t size)

Copy a C-string into a sized buffer

Parameters

char * dest
Where to copy the string to
const char * src
Where to copy the string from
size_t size
size of destination buffer

Description

Compatible with *BSD: the result is always a valid NUL-terminated string that fits in the buffer (unless, of course, the buffer size is zero). It does not pad out the result like strncpy() does.

ssize_t strscpy(char * dest, const char * src, size_t count)

Copy a C-string into a sized buffer

Parameters

char * dest
Where to copy the string to
const char * src
Where to copy the string from
size_t count
Size of destination buffer

Description

Copy the string, or as much of it as fits, into the dest buffer. The behavior is undefined if the string buffers overlap. The destination buffer is always NUL terminated, unless it’s zero-sized.

Preferred to strlcpy() since the API doesn’t require reading memory from the src string beyond the specified “count” bytes, and since the return value is easier to error-check than strlcpy()’s. In addition, the implementation is robust to the string changing out from underneath it, unlike the current strlcpy() implementation.

Preferred to strncpy() since it always returns a valid string, and doesn’t unnecessarily force the tail of the destination buffer to be zeroed. If zeroing is desired please use strscpy_pad().

Return

  • The number of characters copied (not including the trailing NUL)
  • -E2BIG if count is 0 or src was truncated.
ssize_t strscpy_pad(char * dest, const char * src, size_t count)

Copy a C-string into a sized buffer

Parameters

char * dest
Where to copy the string to
const char * src
Where to copy the string from
size_t count
Size of destination buffer

Description

Copy the string, or as much of it as fits, into the dest buffer. The behavior is undefined if the string buffers overlap. The destination buffer is always NUL terminated, unless it’s zero-sized.

If the source string is shorter than the destination buffer, zeros the tail of the destination buffer.

For full explanation of why you may want to consider using the ‘strscpy’ functions please see the function docstring for strscpy().

Return

  • The number of characters copied (not including the trailing NUL)
  • -E2BIG if count is 0 or src was truncated.
char * strcat(char * dest, const char * src)

Append one NUL-terminated string to another

Parameters

char * dest
The string to be appended to
const char * src
The string to append to it
char * strncat(char * dest, const char * src, size_t count)

Append a length-limited, C-string to another

Parameters

char * dest
The string to be appended to
const char * src
The string to append to it
size_t count
The maximum numbers of bytes to copy

Description

Note that in contrast to strncpy(), strncat() ensures the result is terminated.

size_t strlcat(char * dest, const char * src, size_t count)

Append a length-limited, C-string to another

Parameters

char * dest
The string to be appended to
const char * src
The string to append to it
size_t count
The size of the destination buffer.
int strcmp(const char * cs, const char * ct)

Compare two strings

Parameters

const char * cs
One string
const char * ct
Another string
int strncmp(const char * cs, const char * ct, size_t count)

Compare two length-limited strings

Parameters

const char * cs
One string
const char * ct
Another string
size_t count
The maximum number of bytes to compare
char * strchr(const char * s, int c)

Find the first occurrence of a character in a string

Parameters

const char * s
The string to be searched
int c
The character to search for

Description

Note that the NUL-terminator is considered part of the string, and can be searched for.

char * strchrnul(const char * s, int c)

Find and return a character in a string, or end of string

Parameters

const char * s
The string to be searched
int c
The character to search for

Description

Returns pointer to first occurrence of ‘c’ in s. If c is not found, then return a pointer to the null byte at the end of s.

char * strrchr(const char * s, int c)

Find the last occurrence of a character in a string

Parameters

const char * s
The string to be searched
int c
The character to search for
char * strnchr(const char * s, size_t count, int c)

Find a character in a length limited string

Parameters

const char * s
The string to be searched
size_t count
The number of characters to be searched
int c
The character to search for

Description

Note that the NUL-terminator is considered part of the string, and can be searched for.

char * skip_spaces(const char * str)

Removes leading whitespace from str.

Parameters

const char * str
The string to be stripped.

Description

Returns a pointer to the first non-whitespace character in str.

char * strim(char * s)

Removes leading and trailing whitespace from s.

Parameters

char * s
The string to be stripped.

Description

Note that the first trailing whitespace is replaced with a NUL-terminator in the given string s. Returns a pointer to the first non-whitespace character in s.

size_t strlen(const char * s)

Find the length of a string

Parameters

const char * s
The string to be sized
size_t strnlen(const char * s, size_t count)

Find the length of a length-limited string

Parameters

const char * s
The string to be sized
size_t count
The maximum number of bytes to search
size_t strspn(const char * s, const char * accept)

Calculate the length of the initial substring of s which only contain letters in accept

Parameters

const char * s
The string to be searched
const char * accept
The string to search for
size_t strcspn(const char * s, const char * reject)

Calculate the length of the initial substring of s which does not contain letters in reject

Parameters

const char * s
The string to be searched
const char * reject
The string to avoid
char * strpbrk(const char * cs, const char * ct)

Find the first occurrence of a set of characters

Parameters

const char * cs
The string to be searched
const char * ct
The characters to search for
char * strsep(char ** s, const char * ct)

Split a string into tokens

Parameters

char ** s
The string to be searched
const char * ct
The characters to search for

Description

strsep() updates s to point after the token, ready for the next call.

It returns empty tokens, too, behaving exactly like the libc function of that name. In fact, it was stolen from glibc2 and de-fancy-fied. Same semantics, slimmer shape. ;)

bool sysfs_streq(const char * s1, const char * s2)

return true if strings are equal, modulo trailing newline

Parameters

const char * s1
one string
const char * s2
another string

Description

This routine returns true iff two strings are equal, treating both NUL and newline-then-NUL as equivalent string terminations. It’s geared for use with sysfs input strings, which generally terminate with newlines but are compared against values without newlines.

int match_string(const char *const * array, size_t n, const char * string)

matches given string in an array

Parameters

const char *const * array
array of strings
size_t n
number of strings in the array or -1 for NULL terminated arrays
const char * string
string to match with

Description

This routine will look for a string in an array of strings up to the n-th element in the array or until the first NULL element.

Historically the value of -1 for n, was used to search in arrays that are NULL terminated. However, the function does not make a distinction when finishing the search: either n elements have been compared OR the first NULL element was found.

Return

index of a string in the array if matches, or -EINVAL otherwise.

int __sysfs_match_string(const char *const * array, size_t n, const char * str)

matches given string in an array

Parameters

const char *const * array
array of strings
size_t n
number of strings in the array or -1 for NULL terminated arrays
const char * str
string to match with

Description

Returns index of str in the array or -EINVAL, just like match_string(). Uses sysfs_streq instead of strcmp for matching.

This routine will look for a string in an array of strings up to the n-th element in the array or until the first NULL element.

Historically the value of -1 for n, was used to search in arrays that are NULL terminated. However, the function does not make a distinction when finishing the search: either n elements have been compared OR the first NULL element was found.

void * memset(void * s, int c, size_t count)

Fill a region of memory with the given value

Parameters

void * s
Pointer to the start of the area.
int c
The byte to fill the area with
size_t count
The size of the area.

Description

Do not use memset() to access IO space, use memset_io() instead.

void * memset16(uint16_t * s, uint16_t v, size_t count)

Fill a memory area with a uint16_t

Parameters

uint16_t * s
Pointer to the start of the area.
uint16_t v
The value to fill the area with
size_t count
The number of values to store

Description

Differs from memset() in that it fills with a uint16_t instead of a byte. Remember that count is the number of uint16_ts to store, not the number of bytes.

void * memset32(uint32_t * s, uint32_t v, size_t count)

Fill a memory area with a uint32_t

Parameters

uint32_t * s
Pointer to the start of the area.
uint32_t v
The value to fill the area with
size_t count
The number of values to store

Description

Differs from memset() in that it fills with a uint32_t instead of a byte. Remember that count is the number of uint32_ts to store, not the number of bytes.

void * memset64(uint64_t * s, uint64_t v, size_t count)

Fill a memory area with a uint64_t

Parameters

uint64_t * s
Pointer to the start of the area.
uint64_t v
The value to fill the area with
size_t count
The number of values to store

Description

Differs from memset() in that it fills with a uint64_t instead of a byte. Remember that count is the number of uint64_ts to store, not the number of bytes.

void * memcpy(void * dest, const void * src, size_t count)

Copy one area of memory to another

Parameters

void * dest
Where to copy to
const void * src
Where to copy from
size_t count
The size of the area.

Description

You should not use this function to access IO space, use memcpy_toio() or memcpy_fromio() instead.

void * memmove(void * dest, const void * src, size_t count)

Copy one area of memory to another

Parameters

void * dest
Where to copy to
const void * src
Where to copy from
size_t count
The size of the area.

Description

Unlike memcpy(), memmove() copes with overlapping areas.

__visible int memcmp(const void * cs, const void * ct, size_t count)

Compare two areas of memory

Parameters

const void * cs
One area of memory
const void * ct
Another area of memory
size_t count
The size of the area.
int bcmp(const void * a, const void * b, size_t len)

returns 0 if and only if the buffers have identical contents.

Parameters

const void * a
pointer to first buffer.
const void * b
pointer to second buffer.
size_t len
size of buffers.

Description

The sign or magnitude of a non-zero return value has no particular meaning, and architectures may implement their own more efficient bcmp(). So while this particular implementation is a simple (tail) call to memcmp, do not rely on anything but whether the return value is zero or non-zero.

void * memscan(void * addr, int c, size_t size)

Find a character in an area of memory.

Parameters

void * addr
The memory area
int c
The byte to search for
size_t size
The size of the area.

Description

returns the address of the first occurrence of c, or 1 byte past the area if c is not found

char * strstr(const char * s1, const char * s2)

Find the first substring in a NUL terminated string

Parameters

const char * s1
The string to be searched
const char * s2
The string to search for
char * strnstr(const char * s1, const char * s2, size_t len)

Find the first substring in a length-limited string

Parameters

const char * s1
The string to be searched
const char * s2
The string to search for
size_t len
the maximum number of characters to search
void * memchr(const void * s, int c, size_t n)

Find a character in an area of memory.

Parameters

const void * s
The memory area
int c
The byte to search for
size_t n
The size of the area.

Description

returns the address of the first occurrence of c, or NULL if c is not found

void * memchr_inv(const void * start, int c, size_t bytes)

Find an unmatching character in an area of memory.

Parameters

const void * start
The memory area
int c
Find a character other than c
size_t bytes
The size of the area.

Description

returns the address of the first character other than c, or NULL if the whole buffer contains just c.

char * strreplace(char * s, char old, char new)

Replace all occurrences of character in string.

Parameters

char * s
The string to operate on.
char old
The character being replaced.
char new
The character old is replaced with.

Description

Returns pointer to the nul byte at the end of s.

sysfs_match_string(_a, _s)

matches given string in an array

Parameters

_a
array of strings
_s
string to match with

Description

Helper for __sysfs_match_string(). Calculates the size of a automatically.

bool strstarts(const char * str, const char * prefix)

does str start with prefix?

Parameters

const char * str
string to examine
const char * prefix
prefix to look for.
void memzero_explicit(void * s, size_t count)

Fill a region of memory (e.g. sensitive keying data) with 0s.

Parameters

void * s
Pointer to the start of the area.
size_t count
The size of the area.

Note

usually using memset() is just fine (!), but in cases where clearing out _local_ data at the end of a scope is necessary, memzero_explicit() should be used instead in order to prevent the compiler from optimising away zeroing.

memzero_explicit() doesn’t need an arch-specific version as it just invokes the one of memset() implicitly.

const char * kbasename(const char * path)

return the last part of a pathname.

Parameters

const char * path
path to extract the filename from.
void memcpy_and_pad(void * dest, size_t dest_len, const void * src, size_t count, int pad)

Copy one buffer to another with padding

Parameters

void * dest
Where to copy to
size_t dest_len
The destination buffer size
const void * src
Where to copy from
size_t count
The number of bytes to copy
int pad
Character to use for padding if space is left in destination.
size_t str_has_prefix(const char * str, const char * prefix)

Test if a string has a given prefix

Parameters

const char * str
The string to test
const char * prefix
The string to see if str starts with

Description

A common way to test a prefix of a string is to do:
strncmp(str, prefix, sizeof(prefix) - 1)

But this can lead to bugs due to typos, or if prefix is a pointer and not a constant. Instead use str_has_prefix().

Return

  • strlen(prefix) if str starts with prefix
  • 0 if str does not start with prefix
char * kstrdup(const char * s, gfp_t gfp)

allocate space for and copy an existing string

Parameters

const char * s
the string to duplicate
gfp_t gfp
the GFP mask used in the kmalloc() call when allocating memory

Return

newly allocated copy of s or NULL in case of error

const char * kstrdup_const(const char * s, gfp_t gfp)

conditionally duplicate an existing const string

Parameters

const char * s
the string to duplicate
gfp_t gfp
the GFP mask used in the kmalloc() call when allocating memory

Note

Strings allocated by kstrdup_const should be freed by kfree_const.

Return

source string if it is in .rodata section otherwise fallback to kstrdup.

char * kstrndup(const char * s, size_t max, gfp_t gfp)

allocate space for and copy an existing string

Parameters

const char * s
the string to duplicate
size_t max
read at most max chars from s
gfp_t gfp
the GFP mask used in the kmalloc() call when allocating memory

Note

Use kmemdup_nul() instead if the size is known exactly.

Return

newly allocated copy of s or NULL in case of error

void * kmemdup(const void * src, size_t len, gfp_t gfp)

duplicate region of memory

Parameters

const void * src
memory region to duplicate
size_t len
memory region length
gfp_t gfp
GFP mask to use

Return

newly allocated copy of src or NULL in case of error

char * kmemdup_nul(const char * s, size_t len, gfp_t gfp)

Create a NUL-terminated string from unterminated data

Parameters

const char * s
The data to stringify
size_t len
The size of the data
gfp_t gfp
the GFP mask used in the kmalloc() call when allocating memory

Return

newly allocated copy of s with NUL-termination or NULL in case of error

void * memdup_user(const void __user * src, size_t len)

duplicate memory region from user space

Parameters

const void __user * src
source address in user space
size_t len
number of bytes to copy

Return

an ERR_PTR() on failure. Result is physically contiguous, to be freed by kfree().

void * vmemdup_user(const void __user * src, size_t len)

duplicate memory region from user space

Parameters

const void __user * src
source address in user space
size_t len
number of bytes to copy

Return

an ERR_PTR() on failure. Result may be not physically contiguous. Use kvfree() to free.

char * strndup_user(const char __user * s, long n)

duplicate an existing string from user space

Parameters

const char __user * s
The string to duplicate
long n
Maximum number of bytes to copy, including the trailing NUL.

Return

newly allocated copy of s or an ERR_PTR() in case of error

void * memdup_user_nul(const void __user * src, size_t len)

duplicate memory region from user space and NUL-terminate

Parameters

const void __user * src
source address in user space
size_t len
number of bytes to copy

Return

an ERR_PTR() on failure.

Basic Kernel Library Functions

The Linux kernel provides more basic utility functions.

Bit Operations

void set_bit(long nr, volatile unsigned long * addr)

Atomically set a bit in memory

Parameters

long nr
the bit to set
volatile unsigned long * addr
the address to start counting from

Description

This is a relaxed atomic operation (no implied memory barriers).

Note that nr may be almost arbitrarily large; this function is not restricted to acting on a single-word quantity.

void clear_bit(long nr, volatile unsigned long * addr)

Clears a bit in memory

Parameters

long nr
Bit to clear
volatile unsigned long * addr
Address to start counting from

Description

This is a relaxed atomic operation (no implied memory barriers).

void change_bit(long nr, volatile unsigned long * addr)

Toggle a bit in memory

Parameters

long nr
Bit to change
volatile unsigned long * addr
Address to start counting from

Description

This is a relaxed atomic operation (no implied memory barriers).

Note that nr may be almost arbitrarily large; this function is not restricted to acting on a single-word quantity.

bool test_and_set_bit(long nr, volatile unsigned long * addr)

Set a bit and return its old value

Parameters

long nr
Bit to set
volatile unsigned long * addr
Address to count from

Description

This is an atomic fully-ordered operation (implied full memory barrier).

bool test_and_clear_bit(long nr, volatile unsigned long * addr)

Clear a bit and return its old value

Parameters

long nr
Bit to clear
volatile unsigned long * addr
Address to count from

Description

This is an atomic fully-ordered operation (implied full memory barrier).

bool test_and_change_bit(long nr, volatile unsigned long * addr)

Change a bit and return its old value

Parameters

long nr
Bit to change
volatile unsigned long * addr
Address to count from

Description

This is an atomic fully-ordered operation (implied full memory barrier).

void __set_bit(long nr, volatile unsigned long * addr)

Set a bit in memory

Parameters

long nr
the bit to set
volatile unsigned long * addr
the address to start counting from

Description

Unlike set_bit(), this function is non-atomic. If it is called on the same region of memory concurrently, the effect may be that only one operation succeeds.

void __clear_bit(long nr, volatile unsigned long * addr)

Clears a bit in memory

Parameters

long nr
the bit to clear
volatile unsigned long * addr
the address to start counting from

Description

Unlike clear_bit(), this function is non-atomic. If it is called on the same region of memory concurrently, the effect may be that only one operation succeeds.

void __change_bit(long nr, volatile unsigned long * addr)

Toggle a bit in memory

Parameters

long nr
the bit to change
volatile unsigned long * addr
the address to start counting from

Description

Unlike change_bit(), this function is non-atomic. If it is called on the same region of memory concurrently, the effect may be that only one operation succeeds.

bool __test_and_set_bit(long nr, volatile unsigned long * addr)

Set a bit and return its old value

Parameters

long nr
Bit to set
volatile unsigned long * addr
Address to count from

Description

This operation is non-atomic. If two instances of this operation race, one can appear to succeed but actually fail.

bool __test_and_clear_bit(long nr, volatile unsigned long * addr)

Clear a bit and return its old value

Parameters

long nr
Bit to clear
volatile unsigned long * addr
Address to count from

Description

This operation is non-atomic. If two instances of this operation race, one can appear to succeed but actually fail.

bool __test_and_change_bit(long nr, volatile unsigned long * addr)

Change a bit and return its old value

Parameters

long nr
Bit to change
volatile unsigned long * addr
Address to count from

Description

This operation is non-atomic. If two instances of this operation race, one can appear to succeed but actually fail.

bool test_bit(long nr, const volatile unsigned long * addr)

Determine whether a bit is set

Parameters

long nr
bit number to test
const volatile unsigned long * addr
Address to start counting from
void clear_bit_unlock(long nr, volatile unsigned long * addr)

Clear a bit in memory, for unlock

Parameters

long nr
the bit to set
volatile unsigned long * addr
the address to start counting from

Description

This operation is atomic and provides release barrier semantics.

void __clear_bit_unlock(long nr, volatile unsigned long * addr)

Clears a bit in memory

Parameters

long nr
Bit to clear
volatile unsigned long * addr
Address to start counting from

Description

This is a non-atomic operation but implies a release barrier before the memory operation. It can be used for an unlock if no other CPUs can concurrently modify other bits in the word.

bool test_and_set_bit_lock(long nr, volatile unsigned long * addr)

Set a bit and return its old value, for lock

Parameters

long nr
Bit to set
volatile unsigned long * addr
Address to count from

Description

This operation is atomic and provides acquire barrier semantics if the returned value is 0. It can be used to implement bit locks.

bool clear_bit_unlock_is_negative_byte(long nr, volatile unsigned long * addr)

Clear a bit in memory and test if bottom byte is negative, for unlock.

Parameters

long nr
the bit to clear
volatile unsigned long * addr
the address to start counting from

Description

This operation is atomic and provides release barrier semantics.

This is a bit of a one-trick-pony for the filemap code, which clears PG_locked and tests PG_waiters,

Bitmap Operations

bitmaps provide an array of bits, implemented using an an array of unsigned longs. The number of valid bits in a given bitmap does _not_ need to be an exact multiple of BITS_PER_LONG.

The possible unused bits in the last, partially used word of a bitmap are ‘don’t care’. The implementation makes no particular effort to keep them zero. It ensures that their value will not affect the results of any operation. The bitmap operations that return Boolean (bitmap_empty, for example) or scalar (bitmap_weight, for example) results carefully filter out these unused bits from impacting their results.

The byte ordering of bitmaps is more natural on little endian architectures. See the big-endian headers include/asm-ppc64/bitops.h and include/asm-s390/bitops.h for the best explanations of this ordering.

The DECLARE_BITMAP(name,bits) macro, in linux/types.h, can be used to declare an array named ‘name’ of just enough unsigned longs to contain all bit positions from 0 to ‘bits’ - 1.

The available bitmap operations and their rough meaning in the case that the bitmap is a single unsigned long are thus:

The generated code is more efficient when nbits is known at compile-time and at most BITS_PER_LONG.

bitmap_zero(dst, nbits)                     *dst = 0UL
bitmap_fill(dst, nbits)                     *dst = ~0UL
bitmap_copy(dst, src, nbits)                *dst = *src
bitmap_and(dst, src1, src2, nbits)          *dst = *src1 & *src2
bitmap_or(dst, src1, src2, nbits)           *dst = *src1 | *src2
bitmap_xor(dst, src1, src2, nbits)          *dst = *src1 ^ *src2
bitmap_andnot(dst, src1, src2, nbits)       *dst = *src1 & ~(*src2)
bitmap_complement(dst, src, nbits)          *dst = ~(*src)
bitmap_equal(src1, src2, nbits)             Are *src1 and *src2 equal?
bitmap_intersects(src1, src2, nbits)        Do *src1 and *src2 overlap?
bitmap_subset(src1, src2, nbits)            Is *src1 a subset of *src2?
bitmap_empty(src, nbits)                    Are all bits zero in *src?
bitmap_full(src, nbits)                     Are all bits set in *src?
bitmap_weight(src, nbits)                   Hamming Weight: number set bits
bitmap_set(dst, pos, nbits)                 Set specified bit area
bitmap_clear(dst, pos, nbits)               Clear specified bit area
bitmap_find_next_zero_area(buf, len, pos, n, mask)  Find bit free area
bitmap_find_next_zero_area_off(buf, len, pos, n, mask)  as above
bitmap_shift_right(dst, src, n, nbits)      *dst = *src >> n
bitmap_shift_left(dst, src, n, nbits)       *dst = *src << n
bitmap_cut(dst, src, first, n, nbits)       Cut n bits from first, copy rest
bitmap_replace(dst, old, new, mask, nbits)  *dst = (*old & ~(*mask)) | (*new & *mask)
bitmap_remap(dst, src, old, new, nbits)     *dst = map(old, new)(src)
bitmap_bitremap(oldbit, old, new, nbits)    newbit = map(old, new)(oldbit)
bitmap_onto(dst, orig, relmap, nbits)       *dst = orig relative to relmap
bitmap_fold(dst, orig, sz, nbits)           dst bits = orig bits mod sz
bitmap_parse(buf, buflen, dst, nbits)       Parse bitmap dst from kernel buf
bitmap_parse_user(ubuf, ulen, dst, nbits)   Parse bitmap dst from user buf
bitmap_parselist(buf, dst, nbits)           Parse bitmap dst from kernel buf
bitmap_parselist_user(buf, dst, nbits)      Parse bitmap dst from user buf
bitmap_find_free_region(bitmap, bits, order)  Find and allocate bit region
bitmap_release_region(bitmap, pos, order)   Free specified bit region
bitmap_allocate_region(bitmap, pos, order)  Allocate specified bit region
bitmap_from_arr32(dst, buf, nbits)          Copy nbits from u32[] buf to dst
bitmap_to_arr32(buf, src, nbits)            Copy nbits from buf to u32[] dst
bitmap_get_value8(map, start)               Get 8bit value from map at start
bitmap_set_value8(map, value, start)        Set 8bit value to map at start

Note, bitmap_zero() and bitmap_fill() operate over the region of unsigned longs, that is, bits behind bitmap till the unsigned long boundary will be zeroed or filled as well. Consider to use bitmap_clear() or bitmap_set() to make explicit zeroing or filling respectively.

Also the following operations in asm/bitops.h apply to bitmaps.:

set_bit(bit, addr)                  *addr |= bit
clear_bit(bit, addr)                *addr &= ~bit
change_bit(bit, addr)               *addr ^= bit
test_bit(bit, addr)                 Is bit set in *addr?
test_and_set_bit(bit, addr)         Set bit and return old value
test_and_clear_bit(bit, addr)       Clear bit and return old value
test_and_change_bit(bit, addr)      Change bit and return old value
find_first_zero_bit(addr, nbits)    Position first zero bit in *addr
find_first_bit(addr, nbits)         Position first set bit in *addr
find_next_zero_bit(addr, nbits, bit)
                                    Position next zero bit in *addr >= bit
find_next_bit(addr, nbits, bit)     Position next set bit in *addr >= bit
find_next_and_bit(addr1, addr2, nbits, bit)
                                    Same as find_next_bit, but in
                                    (*addr1 & *addr2)
void __bitmap_shift_right(unsigned long * dst, const unsigned long * src, unsigned shift, unsigned nbits)

logical right shift of the bits in a bitmap

Parameters

unsigned long * dst
destination bitmap
const unsigned long * src
source bitmap
unsigned shift
shift by this many bits
unsigned nbits
bitmap size, in bits

Description

Shifting right (dividing) means moving bits in the MS -> LS bit direction. Zeros are fed into the vacated MS positions and the LS bits shifted off the bottom are lost.

void __bitmap_shift_left(unsigned long * dst, const unsigned long * src, unsigned int shift, unsigned int nbits)

logical left shift of the bits in a bitmap

Parameters

unsigned long * dst
destination bitmap
const unsigned long * src
source bitmap
unsigned int shift
shift by this many bits
unsigned int nbits
bitmap size, in bits

Description

Shifting left (multiplying) means moving bits in the LS -> MS direction. Zeros are fed into the vacated LS bit positions and those MS bits shifted off the top are lost.

void bitmap_cut(unsigned long * dst, const unsigned long * src, unsigned int first, unsigned int cut, unsigned int nbits)

remove bit region from bitmap and right shift remaining bits

Parameters

unsigned long * dst
destination bitmap, might overlap with src
const unsigned long * src

31 63 | | 10000000 11000001 11110010 00010101 10000000 11000001 01110010 00010101

| | |

16 14 0 32

unsigned int first
start bit of region to be removed
unsigned int cut
number of bits to remove
unsigned int nbits
bitmap size, in bits

Description

Set the n-th bit of dst iff the n-th bit of src is set and n is less than first, or the m-th bit of src is set for any m such that first <= n < nbits, and m = n + cut.

In pictures, example for a big-endian 32-bit architecture:

if cut is 3, and first is 14, bits 14-16 in src are cut and dst is:

31 63 | | 10110000 00011000 00110010 00010101 00010000 00011000 00101110 01000010

| |
14 (bit 17 0 32
from src)

Note that dst and src might overlap partially or entirely.

This is implemented in the obvious way, with a shift and carry step for each moved bit. Optimisation is left as an exercise for the compiler.

unsigned long bitmap_find_next_zero_area_off(unsigned long * map, unsigned long size, unsigned long start, unsigned int nr, unsigned long align_mask, unsigned long align_offset)

find a contiguous aligned zero area

Parameters

unsigned long * map
The address to base the search on
unsigned long size
The bitmap size in bits
unsigned long start
The bitnumber to start searching at
unsigned int nr
The number of zeroed bits we’re looking for
unsigned long align_mask
Alignment mask for zero area
unsigned long align_offset
Alignment offset for zero area.

Description

The align_mask should be one less than a power of 2; the effect is that the bit offset of all zero areas this function finds plus align_offset is multiple of that power of 2.

int bitmap_parse_user(const char __user * ubuf, unsigned int ulen, unsigned long * maskp, int nmaskbits)

convert an ASCII hex string in a user buffer into a bitmap

Parameters

const char __user * ubuf
pointer to user buffer containing string.
unsigned int ulen
buffer size in bytes. If string is smaller than this then it must be terminated with a 0.
unsigned long * maskp
pointer to bitmap array that will contain result.
int nmaskbits
size of bitmap, in bits.
int bitmap_print_to_pagebuf(bool list, char * buf, const unsigned long * maskp, int nmaskbits)

convert bitmap to list or hex format ASCII string

Parameters

bool list
indicates whether the bitmap must be list
char * buf
page aligned buffer into which string is placed
const unsigned long * maskp
pointer to bitmap to convert
int nmaskbits
size of bitmap, in bits

Description

Output format is a comma-separated list of decimal numbers and ranges if list is specified or hex digits grouped into comma-separated sets of 8 digits/set. Returns the number of characters written to buf.

It is assumed that buf is a pointer into a PAGE_SIZE, page-aligned area and that sufficient storage remains at buf to accommodate the bitmap_print_to_pagebuf() output. Returns the number of characters actually printed to buf, excluding terminating ‘0’.

int bitmap_parselist(const char * buf, unsigned long * maskp, int nmaskbits)

convert list format ASCII string to bitmap

Parameters

const char * buf
read user string from this buffer; must be terminated with a 0 or n.
unsigned long * maskp
write resulting mask here
int nmaskbits
number of bits in mask to be written

Description

Input format is a comma-separated list of decimal numbers and ranges. Consecutively set bits are shown as two hyphen-separated decimal numbers, the smallest and largest bit numbers set in the range. Optionally each range can be postfixed to denote that only parts of it should be set. The range will divided to groups of specific size. From each group will be used only defined amount of bits. Syntax: range:used_size/group_size

Example

0-1023:2/256 ==> 0,1,256,257,512,513,768,769

Return

0 on success, -errno on invalid input strings. Error values:

  • -EINVAL: wrong region format
  • -EINVAL: invalid character in string
  • -ERANGE: bit number specified too large for mask
  • -EOVERFLOW: integer overflow in the input parameters
int bitmap_parselist_user(const char __user * ubuf, unsigned int ulen, unsigned long * maskp, int nmaskbits)

Parameters

const char __user * ubuf
pointer to user buffer containing string.
unsigned int ulen
buffer size in bytes. If string is smaller than this then it must be terminated with a 0.
unsigned long * maskp
pointer to bitmap array that will contain result.
int nmaskbits
size of bitmap, in bits.

Description

Wrapper for bitmap_parselist(), providing it with user buffer.

int bitmap_parse(const char * start, unsigned int buflen, unsigned long * maskp, int nmaskbits)

convert an ASCII hex string into a bitmap.

Parameters

const char * start
pointer to buffer containing string.
unsigned int buflen
buffer size in bytes. If string is smaller than this then it must be terminated with a 0 or n. In that case, UINT_MAX may be provided instead of string length.
unsigned long * maskp
pointer to bitmap array that will contain result.
int nmaskbits
size of bitmap, in bits.

Description

Commas group hex digits into chunks. Each chunk defines exactly 32 bits of the resultant bitmask. No chunk may specify a value larger than 32 bits (-EOVERFLOW), and if a chunk specifies a smaller value then leading 0-bits are prepended. -EINVAL is returned for illegal characters. Grouping such as “1,,5”, “,44”, “,” or “” is allowed. Leading, embedded and trailing whitespace accepted.

int bitmap_find_free_region(unsigned long * bitmap, unsigned int bits, int order)

find a contiguous aligned mem region

Parameters

unsigned long * bitmap
array of unsigned longs corresponding to the bitmap
unsigned int bits
number of bits in the bitmap
int order
region size (log base 2 of number of bits) to find

Description

Find a region of free (zero) bits in a bitmap of bits bits and allocate them (set them to one). Only consider regions of length a power (order) of two, aligned to that power of two, which makes the search algorithm much faster.

Return the bit offset in bitmap of the allocated region, or -errno on failure.

void bitmap_release_region(unsigned long * bitmap, unsigned int pos, int order)

release allocated bitmap region

Parameters

unsigned long * bitmap
array of unsigned longs corresponding to the bitmap
unsigned int pos
beginning of bit region to release
int order
region size (log base 2 of number of bits) to release

Description

This is the complement to __bitmap_find_free_region() and releases the found region (by clearing it in the bitmap).

No return value.

int bitmap_allocate_region(unsigned long * bitmap, unsigned int pos, int order)

allocate bitmap region

Parameters

unsigned long * bitmap
array of unsigned longs corresponding to the bitmap
unsigned int pos
beginning of bit region to allocate
int order
region size (log base 2 of number of bits) to allocate

Description

Allocate (set bits in) a specified region of a bitmap.

Return 0 on success, or -EBUSY if specified region wasn’t free (not all bits were zero).

void bitmap_copy_le(unsigned long * dst, const unsigned long * src, unsigned int nbits)

copy a bitmap, putting the bits into little-endian order.

Parameters

unsigned long * dst
destination buffer
const unsigned long * src
bitmap to copy
unsigned int nbits
number of bits in the bitmap

Description

Require nbits % BITS_PER_LONG == 0.

void bitmap_from_arr32(unsigned long * bitmap, const u32 * buf, unsigned int nbits)

copy the contents of u32 array of bits to bitmap

Parameters

unsigned long * bitmap
array of unsigned longs, the destination bitmap
const u32 * buf
array of u32 (in host byte order), the source bitmap
unsigned int nbits
number of bits in bitmap
void bitmap_to_arr32(u32 * buf, const unsigned long * bitmap, unsigned int nbits)

copy the contents of bitmap to a u32 array of bits

Parameters

u32 * buf
array of u32 (in host byte order), the dest bitmap
const unsigned long * bitmap
array of unsigned longs, the source bitmap
unsigned int nbits
number of bits in bitmap
int bitmap_pos_to_ord(const unsigned long * buf, unsigned int pos, unsigned int nbits)

find ordinal of set bit at given position in bitmap

Parameters

const unsigned long * buf
pointer to a bitmap
unsigned int pos
a bit position in buf (0 <= pos < nbits)
unsigned int nbits
number of valid bit positions in buf

Description

Map the bit at position pos in buf (of length nbits) to the ordinal of which set bit it is. If it is not set or if pos is not a valid bit position, map to -1.

If for example, just bits 4 through 7 are set in buf, then pos values 4 through 7 will get mapped to 0 through 3, respectively, and other pos values will get mapped to -1. When pos value 7 gets mapped to (returns) ord value 3 in this example, that means that bit 7 is the 3rd (starting with 0th) set bit in buf.

The bit positions 0 through bits are valid positions in buf.

unsigned int bitmap_ord_to_pos(const unsigned long * buf, unsigned int ord, unsigned int nbits)

find position of n-th set bit in bitmap

Parameters

const unsigned long * buf
pointer to bitmap
unsigned int ord
ordinal bit position (n-th set bit, n >= 0)
unsigned int nbits
number of valid bit positions in buf

Description

Map the ordinal offset of bit ord in buf to its position in buf. Value of ord should be in range 0 <= ord < weight(buf). If ord >= weight(buf), returns nbits.

If for example, just bits 4 through 7 are set in buf, then ord values 0 through 3 will get mapped to 4 through 7, respectively, and all other ord values returns nbits. When ord value 3 gets mapped to (returns) pos value 7 in this example, that means that the 3rd set bit (starting with 0th) is at position 7 in buf.

The bit positions 0 through nbits-1 are valid positions in buf.

void bitmap_remap(unsigned long * dst, const unsigned long * src, const unsigned long * old, const unsigned long * new, unsigned int nbits)

Apply map defined by a pair of bitmaps to another bitmap

Parameters

unsigned long * dst
remapped result
const unsigned long * src
subset to be remapped
const unsigned long * old
defines domain of map
const unsigned long * new
defines range of map
unsigned int nbits
number of bits in each of these bitmaps

Description

Let old and new define a mapping of bit positions, such that whatever position is held by the n-th set bit in old is mapped to the n-th set bit in new. In the more general case, allowing for the possibility that the weight ‘w’ of new is less than the weight of old, map the position of the n-th set bit in old to the position of the m-th set bit in new, where m == n % w.

If either of the old and new bitmaps are empty, or if src and dst point to the same location, then this routine copies src to dst.

The positions of unset bits in old are mapped to themselves (the identify map).

Apply the above specified mapping to src, placing the result in dst, clearing any bits previously set in dst.

For example, lets say that old has bits 4 through 7 set, and new has bits 12 through 15 set. This defines the mapping of bit position 4 to 12, 5 to 13, 6 to 14 and 7 to 15, and of all other bit positions unchanged. So if say src comes into this routine with bits 1, 5 and 7 set, then dst should leave with bits 1, 13 and 15 set.

int bitmap_bitremap(int oldbit, const unsigned long * old, const unsigned long * new, int bits)

Apply map defined by a pair of bitmaps to a single bit

Parameters

int oldbit
bit position to be mapped
const unsigned long * old
defines domain of map
const unsigned long * new
defines range of map
int bits
number of bits in each of these bitmaps

Description

Let old and new define a mapping of bit positions, such that whatever position is held by the n-th set bit in old is mapped to the n-th set bit in new. In the more general case, allowing for the possibility that the weight ‘w’ of new is less than the weight of old, map the position of the n-th set bit in old to the position of the m-th set bit in new, where m == n % w.

The positions of unset bits in old are mapped to themselves (the identify map).

Apply the above specified mapping to bit position oldbit, returning the new bit position.

For example, lets say that old has bits 4 through 7 set, and new has bits 12 through 15 set. This defines the mapping of bit position 4 to 12, 5 to 13, 6 to 14 and 7 to 15, and of all other bit positions unchanged. So if say oldbit is 5, then this routine returns 13.

void bitmap_onto(unsigned long * dst, const unsigned long * orig, const unsigned long * relmap, unsigned int bits)

translate one bitmap relative to another

Parameters

unsigned long * dst
resulting translated bitmap
const unsigned long * orig
original untranslated bitmap
const unsigned long * relmap
bitmap relative to which translated
unsigned int bits
number of bits in each of these bitmaps

Description

Set the n-th bit of dst iff there exists some m such that the n-th bit of relmap is set, the m-th bit of orig is set, and the n-th bit of relmap is also the m-th _set_ bit of relmap. (If you understood the previous sentence the first time your read it, you’re overqualified for your current job.)

In other words, orig is mapped onto (surjectively) dst, using the map { <n, m> | the n-th bit of relmap is the m-th set bit of relmap }.

Any set bits in orig above bit number W, where W is the weight of (number of set bits in) relmap are mapped nowhere. In particular, if for all bits m set in orig, m >= W, then dst will end up empty. In situations where the possibility of such an empty result is not desired, one way to avoid it is to use the bitmap_fold() operator, below, to first fold the orig bitmap over itself so that all its set bits x are in the range 0 <= x < W. The bitmap_fold() operator does this by setting the bit (m % W) in dst, for each bit (m) set in orig.

Example [1] for bitmap_onto():

Let’s say relmap has bits 30-39 set, and orig has bits 1, 3, 5, 7, 9 and 11 set. Then on return from this routine, dst will have bits 31, 33, 35, 37 and 39 set.

When bit 0 is set in orig, it means turn on the bit in dst corresponding to whatever is the first bit (if any) that is turned on in relmap. Since bit 0 was off in the above example, we leave off that bit (bit 30) in dst.

When bit 1 is set in orig (as in the above example), it means turn on the bit in dst corresponding to whatever is the second bit that is turned on in relmap. The second bit in relmap that was turned on in the above example was bit 31, so we turned on bit 31 in dst.

Similarly, we turned on bits 33, 35, 37 and 39 in dst, because they were the 4th, 6th, 8th and 10th set bits set in relmap, and the 4th, 6th, 8th and 10th bits of orig (i.e. bits 3, 5, 7 and 9) were also set.

When bit 11 is set in orig, it means turn on the bit in dst corresponding to whatever is the twelfth bit that is turned on in relmap. In the above example, there were only ten bits turned on in relmap (30..39), so that bit 11 was set in orig had no affect on dst.

Example [2] for bitmap_fold() + bitmap_onto():

Let’s say relmap has these ten bits set:

40 41 42 43 45 48 53 61 74 95

(for the curious, that’s 40 plus the first ten terms of the Fibonacci sequence.)

Further lets say we use the following code, invoking bitmap_fold() then bitmap_onto, as suggested above to avoid the possibility of an empty dst result:

unsigned long *tmp;     // a temporary bitmap's bits

bitmap_fold(tmp, orig, bitmap_weight(relmap, bits), bits);
bitmap_onto(dst, tmp, relmap, bits);

Then this table shows what various values of dst would be, for various orig’s. I list the zero-based positions of each set bit. The tmp column shows the intermediate result, as computed by using bitmap_fold() to fold the orig bitmap modulo ten (the weight of relmap):

orig tmp dst
0 0 40
1 1 41
9 9 95
10 0 40 [1]
1 3 5 7 1 3 5 7 41 43 48 61
0 1 2 3 4 0 1 2 3 4 40 41 42 43 45
0 9 18 27 0 9 8 7 40 61 74 95
0 10 20 30 0 40
0 11 22 33 0 1 2 3 40 41 42 43
0 12 24 36 0 2 4 6 40 42 45 53
78 102 211 1 2 8 41 42 74 [1]
[1](1, 2) For these marked lines, if we hadn’t first done bitmap_fold() into tmp, then the dst result would have been empty.

If either of orig or relmap is empty (no set bits), then dst will be returned empty.

If (as explained above) the only set bits in orig are in positions m where m >= W, (where W is the weight of relmap) then dst will once again be returned empty.

All bits in dst not set by the above rule are cleared.

void bitmap_fold(unsigned long * dst, const unsigned long * orig, unsigned int sz, unsigned int nbits)

fold larger bitmap into smaller, modulo specified size

Parameters

unsigned long * dst
resulting smaller bitmap
const unsigned long * orig
original larger bitmap
unsigned int sz
specified size
unsigned int nbits
number of bits in each of these bitmaps

Description

For each bit oldbit in orig, set bit oldbit mod sz in dst. Clear all other bits in dst. See further the comment and Example [2] for bitmap_onto() for why and how to use this.

unsigned long bitmap_find_next_zero_area(unsigned long * map, unsigned long size, unsigned long start, unsigned int nr, unsigned long align_mask)

find a contiguous aligned zero area

Parameters

unsigned long * map
The address to base the search on
unsigned long size
The bitmap size in bits
unsigned long start
The bitnumber to start searching at
unsigned int nr
The number of zeroed bits we’re looking for
unsigned long align_mask
Alignment mask for zero area

Description

The align_mask should be one less than a power of 2; the effect is that the bit offset of all zero areas this function finds is multiples of that power of 2. A align_mask of 0 means no alignment is required.

bool bitmap_or_equal(const unsigned long * src1, const unsigned long * src2, const unsigned long * src3, unsigned int nbits)

Check whether the or of two bitmaps is equal to a third

Parameters

const unsigned long * src1
Pointer to bitmap 1
const unsigned long * src2
Pointer to bitmap 2 will be or’ed with bitmap 1
const unsigned long * src3
Pointer to bitmap 3. Compare to the result of *src1 | *src2
unsigned int nbits
number of bits in each of these bitmaps

Return

True if (*src1 | *src2) == *src3, false otherwise

BITMAP_FROM_U64(n)

Represent u64 value in the format suitable for bitmap.

Parameters

n
u64 value

Description

Linux bitmaps are internally arrays of unsigned longs, i.e. 32-bit integers in 32-bit environment, and 64-bit integers in 64-bit one.

There are four combinations of endianness and length of the word in linux ABIs: LE64, BE64, LE32 and BE32.

On 64-bit kernels 64-bit LE and BE numbers are naturally ordered in bitmaps and therefore don’t require any special handling.

On 32-bit kernels 32-bit LE ABI orders lo word of 64-bit number in memory prior to hi, and 32-bit BE orders hi word prior to lo. The bitmap on the other hand is represented as an array of 32-bit words and the position of bit N may therefore be calculated as: word #(N/32) and bit #(N``32``) in that word. For example, bit #42 is located at 10th position of 2nd word. It matches 32-bit LE ABI, and we can simply let the compiler store 64-bit values in memory as it usually does. But for BE we need to swap hi and lo words manually.

With all that, the macro BITMAP_FROM_U64() does explicit reordering of hi and lo parts of u64. For LE32 it does nothing, and for BE environment it swaps hi and lo words, as is expected by bitmap.

void bitmap_from_u64(unsigned long * dst, u64 mask)

Check and swap words within u64.

Parameters

unsigned long * dst
destination bitmap
u64 mask
source bitmap

Description

In 32-bit Big Endian kernel, when using (u32 *)(:c:type:`val`)[*] to read u64 mask, we will get the wrong word. That is (u32 *)(:c:type:`val`)[0] gets the upper 32 bits, but we expect the lower 32-bits of u64.

unsigned long bitmap_get_value8(const unsigned long * map, unsigned long start)

get an 8-bit value within a memory region

Parameters

const unsigned long * map
address to the bitmap memory region
unsigned long start
bit offset of the 8-bit value; must be a multiple of 8

Description

Returns the 8-bit value located at the start bit offset within the src memory region.

void bitmap_set_value8(unsigned long * map, unsigned long value, unsigned long start)

set an 8-bit value within a memory region

Parameters

unsigned long * map
address to the bitmap memory region
unsigned long value
the 8-bit value; values wider than 8 bits may clobber bitmap
unsigned long start
bit offset of the 8-bit value; must be a multiple of 8

Command-line Parsing

int get_option(char ** str, int * pint)

Parse integer from an option string

Parameters

char ** str
option string
int * pint
(output) integer value parsed from str

Description

Read an int from an option string; if available accept a subsequent comma as well.

Return values: 0 - no int in string 1 - int found, no subsequent comma 2 - int found including a subsequent comma 3 - hyphen found to denote a range

char * get_options(const char * str, int nints, int * ints)

Parse a string into a list of integers

Parameters

const char * str
String to be parsed
int nints
size of integer array
int * ints
integer array

Description

This function parses a string containing a comma-separated list of integers, a hyphen-separated range of _positive_ integers, or a combination of both. The parse halts when the array is full, or when no more numbers can be retrieved from the string.

Return value is the character in the string which caused the parse to end (typically a null terminator, if str is completely parseable).

unsigned long long memparse(const char * ptr, char ** retptr)

parse a string with mem suffixes into a number

Parameters

const char * ptr
Where parse begins
char ** retptr
(output) Optional pointer to next char after parse completes

Description

Parses a string into a number. The number stored at ptr is potentially suffixed with K, M, G, T, P, E.

Sorting

void sort_r(void * base, size_t num, size_t size, cmp_r_func_t cmp_func, swap_func_t swap_func, const void * priv)

sort an array of elements

Parameters

void * base
pointer to data to sort
size_t num
number of elements
size_t size
size of each element
cmp_r_func_t cmp_func
pointer to comparison function
swap_func_t swap_func
pointer to swap function or NULL
const void * priv
third argument passed to comparison function

Description

This function does a heapsort on the given array. You may provide a swap_func function if you need to do something more than a memory copy (e.g. fix up pointers or auxiliary data), but the built-in swap avoids a slow retpoline and so is significantly faster.

Sorting time is O(n log n) both on average and worst-case. While quicksort is slightly faster on average, it suffers from exploitable O(n*n) worst-case behavior and extra memory requirements that make it less suitable for kernel use.

void list_sort(void * priv, struct list_head * head, int (*cmp) (void *priv, struct list_head *a, struct list_head *b)

sort a list

Parameters

void * priv
private data, opaque to list_sort(), passed to cmp
struct list_head * head
the list to sort
int (*)(void *priv, struct list_head *a, struct list_head *b) cmp
the elements comparison function

Description

The comparison funtion cmp must return > 0 if a should sort after b (“a > b” if you want an ascending sort), and <= 0 if a should sort before b or their original order should be preserved. It is always called with the element that came first in the input in a, and list_sort is a stable sort, so it is not necessary to distinguish the a < b and a == b cases.

This is compatible with two styles of cmp function: - The traditional style which returns <0 / =0 / >0, or - Returning a boolean 0/1. The latter offers a chance to save a few cycles in the comparison (which is used by e.g. plug_ctx_cmp() in block/blk-mq.c).

A good way to write a multi-word comparison is:

if (a->high != b->high)
        return a->high > b->high;
if (a->middle != b->middle)
        return a->middle > b->middle;
return a->low > b->low;

This mergesort is as eager as possible while always performing at least 2:1 balanced merges. Given two pending sublists of size 2^k, they are merged to a size-2^(k+1) list as soon as we have 2^k following elements.

Thus, it will avoid cache thrashing as long as 3*2^k elements can fit into the cache. Not quite as good as a fully-eager bottom-up mergesort, but it does use 0.2*n fewer comparisons, so is faster in the common case that everything fits into L1.

The merging is controlled by “count”, the number of elements in the pending lists. This is beautiully simple code, but rather subtle.

Each time we increment “count”, we set one bit (bit k) and clear bits k-1 .. 0. Each time this happens (except the very first time for each bit, when count increments to 2^k), we merge two lists of size 2^k into one list of size 2^(k+1).

This merge happens exactly when the count reaches an odd multiple of 2^k, which is when we have 2^k elements pending in smaller lists, so it’s safe to merge away two lists of size 2^k.

After this happens twice, we have created two lists of size 2^(k+1), which will be merged into a list of size 2^(k+2) before we create a third list of size 2^(k+1), so there are never more than two pending.

The number of pending lists of size 2^k is determined by the state of bit k of “count” plus two extra pieces of information:

  • The state of bit k-1 (when k == 0, consider bit -1 always set), and
  • Whether the higher-order bits are zero or non-zero (i.e. is count >= 2^(k+1)).

There are six states we distinguish. “x” represents some arbitrary bits, and “y” represents some arbitrary non-zero bits: 0: 00x: 0 pending of size 2^k; x pending of sizes < 2^k 1: 01x: 0 pending of size 2^k; 2^(k-1) + x pending of sizes < 2^k 2: x10x: 0 pending of size 2^k; 2^k + x pending of sizes < 2^k 3: x11x: 1 pending of size 2^k; 2^(k-1) + x pending of sizes < 2^k 4: y00x: 1 pending of size 2^k; 2^k + x pending of sizes < 2^k 5: y01x: 2 pending of size 2^k; 2^(k-1) + x pending of sizes < 2^k (merge and loop back to state 2)

We gain lists of size 2^k in the 2->3 and 4->5 transitions (because bit k-1 is set while the more significant bits are non-zero) and merge them away in the 5->2 transition. Note in particular that just before the 5->2 transition, all lower-order bits are 11 (state 3), so there is one list of each smaller size.

When we reach the end of the input, we merge all the pending lists, from smallest to largest. If you work through cases 2 to 5 above, you can see that the number of elements we merge with a list of size 2^k varies from 2^(k-1) (cases 3 and 5 when x == 0) to 2^(k+1) - 1 (second merge of case 5 when x == 2^(k-1) - 1).

Text Searching

INTRODUCTION

The textsearch infrastructure provides text searching facilities for both linear and non-linear data. Individual search algorithms are implemented in modules and chosen by the user.

ARCHITECTURE

  User
  +----------------+
  |        finish()|<--------------(6)-----------------+
  |get_next_block()|<--------------(5)---------------+ |
  |                |                     Algorithm   | |
  |                |                    +------------------------------+
  |                |                    |  init()   find()   destroy() |
  |                |                    +------------------------------+
  |                |       Core API           ^       ^          ^
  |                |      +---------------+  (2)     (4)        (8)
  |             (1)|----->| prepare()     |---+       |          |
  |             (3)|----->| find()/next() |-----------+          |
  |             (7)|----->| destroy()     |----------------------+
  +----------------+      +---------------+

(1) User configures a search by calling textsearch_prepare() specifying
    the search parameters such as the pattern and algorithm name.
(2) Core requests the algorithm to allocate and initialize a search
    configuration according to the specified parameters.
(3) User starts the search(es) by calling textsearch_find() or
    textsearch_next() to fetch subsequent occurrences. A state variable
    is provided to the algorithm to store persistent variables.
(4) Core eventually resets the search offset and forwards the find()
    request to the algorithm.
(5) Algorithm calls get_next_block() provided by the user continuously
    to fetch the data to be searched in block by block.
(6) Algorithm invokes finish() after the last call to get_next_block
    to clean up any leftovers from get_next_block. (Optional)
(7) User destroys the configuration by calling textsearch_destroy().
(8) Core notifies the algorithm to destroy algorithm specific
    allocations. (Optional)

USAGE

Before a search can be performed, a configuration must be created by calling textsearch_prepare() specifying the searching algorithm, the pattern to look for and flags. As a flag, you can set TS_IGNORECASE to perform case insensitive matching. But it might slow down performance of algorithm, so you should use it at own your risk. The returned configuration may then be used for an arbitrary amount of times and even in parallel as long as a separate struct ts_state variable is provided to every instance.

The actual search is performed by either calling textsearch_find_continuous() for linear data or by providing an own get_next_block() implementation and calling textsearch_find(). Both functions return the position of the first occurrence of the pattern or UINT_MAX if no match was found. Subsequent occurrences can be found by calling textsearch_next() regardless of the linearity of the data.

Once you’re done using a configuration it must be given back via textsearch_destroy.

EXAMPLE:

int pos;
struct ts_config *conf;
struct ts_state state;
const char *pattern = "chicken";
const char *example = "We dance the funky chicken";

conf = textsearch_prepare("kmp", pattern, strlen(pattern),
                          GFP_KERNEL, TS_AUTOLOAD);
if (IS_ERR(conf)) {
    err = PTR_ERR(conf);
    goto errout;
}

pos = textsearch_find_continuous(conf, &state, example, strlen(example));
if (pos != UINT_MAX)
    panic("Oh my god, dancing chickens at %d\n", pos);

textsearch_destroy(conf);
int textsearch_register(struct ts_ops * ops)

register a textsearch module

Parameters

struct ts_ops * ops
operations lookup table

Description

This function must be called by textsearch modules to announce their presence. The specified &**ops** must have name set to a unique identifier and the callbacks find(), init(), get_pattern(), and get_pattern_len() must be implemented.

Returns 0 or -EEXISTS if another module has already registered with same name.

int textsearch_unregister(struct ts_ops * ops)

unregister a textsearch module

Parameters

struct ts_ops * ops
operations lookup table

Description

This function must be called by textsearch modules to announce their disappearance for examples when the module gets unloaded. The ops parameter must be the same as the one during the registration.

Returns 0 on success or -ENOENT if no matching textsearch registration was found.

unsigned int textsearch_find_continuous(struct ts_config * conf, struct ts_state * state, const void * data, unsigned int len)

search a pattern in continuous/linear data

Parameters

struct ts_config * conf
search configuration
struct ts_state * state
search state
const void * data
data to search in
unsigned int len
length of data

Description

A simplified version of textsearch_find() for continuous/linear data. Call textsearch_next() to retrieve subsequent matches.

Returns the position of first occurrence of the pattern or UINT_MAX if no occurrence was found.

struct ts_config * textsearch_prepare(const char * algo, const void * pattern, unsigned int len, gfp_t gfp_mask, int flags)

Prepare a search

Parameters

const char * algo
name of search algorithm
const void * pattern
pattern data
unsigned int len
length of pattern
gfp_t gfp_mask
allocation mask
int flags
search flags

Description

Looks up the search algorithm module and creates a new textsearch configuration for the specified pattern.

Note

The format of the pattern may not be compatible between
the various search algorithms.

Returns a new textsearch configuration according to the specified parameters or a ERR_PTR(). If a zero length pattern is passed, this function returns EINVAL.

void textsearch_destroy(struct ts_config * conf)

destroy a search configuration

Parameters

struct ts_config * conf
search configuration

Description

Releases all references of the configuration and frees up the memory.

unsigned int textsearch_next(struct ts_config * conf, struct ts_state * state)

continue searching for a pattern

Parameters

struct ts_config * conf
search configuration
struct ts_state * state
search state

Description

Continues a search looking for more occurrences of the pattern. textsearch_find() must be called to find the first occurrence in order to reset the state.

Returns the position of the next occurrence of the pattern or UINT_MAX if not match was found.

unsigned int textsearch_find(struct ts_config * conf, struct ts_state * state)

start searching for a pattern

Parameters

struct ts_config * conf
search configuration
struct ts_state * state
search state

Description

Returns the position of first occurrence of the pattern or UINT_MAX if no match was found.

void * textsearch_get_pattern(struct ts_config * conf)

return head of the pattern

Parameters

struct ts_config * conf
search configuration
unsigned int textsearch_get_pattern_len(struct ts_config * conf)

return length of the pattern

Parameters

struct ts_config * conf
search configuration

CRC and Math Functions in Linux

CRC Functions

uint8_t crc4(uint8_t c, uint64_t x, int bits)

calculate the 4-bit crc of a value.

Parameters

uint8_t c
starting crc4
uint64_t x
value to checksum
int bits
number of bits in x to checksum

Description

Returns the crc4 value of x, using polynomial 0b10111.

The x value is treated as left-aligned, and bits above bits are ignored in the crc calculations.

u8 crc7_be(u8 crc, const u8 * buffer, size_t len)

update the CRC7 for the data buffer

Parameters

u8 crc
previous CRC7 value
const u8 * buffer
data pointer
size_t len
number of bytes in the buffer

Context

any

Description

Returns the updated CRC7 value. The CRC7 is left-aligned in the byte (the lsbit is always 0), as that makes the computation easier, and all callers want it in that form.

void crc8_populate_msb(u8 table, u8 polynomial)

fill crc table for given polynomial in reverse bit order.

Parameters

u8 table
table to be filled.
u8 polynomial
polynomial for which table is to be filled.
void crc8_populate_lsb(u8 table, u8 polynomial)

fill crc table for given polynomial in regular bit order.

Parameters

u8 table
table to be filled.
u8 polynomial
polynomial for which table is to be filled.
u8 crc8(const u8 table, u8 * pdata, size_t nbytes, u8 crc)

calculate a crc8 over the given input data.

Parameters

const u8 table
crc table used for calculation.
u8 * pdata
pointer to data buffer.
size_t nbytes
number of bytes in data buffer.
u8 crc
previous returned crc8 value.
u16 crc16(u16 crc, u8 const * buffer, size_t len)

compute the CRC-16 for the data buffer

Parameters

u16 crc
previous CRC value
u8 const * buffer
data pointer
size_t len
number of bytes in the buffer

Description

Returns the updated CRC value.

u32 __pure crc32_le_generic(u32 crc, unsigned char const * p, size_t len, const u32 ( * tab, u32 polynomial)

Calculate bitwise little-endian Ethernet AUTODIN II CRC32/CRC32C

Parameters

u32 crc
seed value for computation. ~0 for Ethernet, sometimes 0 for other uses, or the previous crc32/crc32c value if computing incrementally.
unsigned char const * p
pointer to buffer over which CRC32/CRC32C is run
size_t len
length of buffer p
const u32 ( * tab
little-endian Ethernet table
u32 polynomial
CRC32/CRC32c LE polynomial
u32 __attribute_const__ crc32_generic_shift(u32 crc, size_t len, u32 polynomial)

Append len 0 bytes to crc, in logarithmic time

Parameters

u32 crc
The original little-endian CRC (i.e. lsbit is x^31 coefficient)
size_t len
The number of bytes. crc is multiplied by x^(8***len**)
u32 polynomial
The modulus used to reduce the result to 32 bits.

Description

It’s possible to parallelize CRC computations by computing a CRC over separate ranges of a buffer, then summing them. This shifts the given CRC by 8*len bits (i.e. produces the same effect as appending len bytes of zero to the data), in time proportional to log(len).

u32 __pure crc32_be_generic(u32 crc, unsigned char const * p, size_t len, const u32 ( * tab, u32 polynomial)

Calculate bitwise big-endian Ethernet AUTODIN II CRC32

Parameters

u32 crc
seed value for computation. ~0 for Ethernet, sometimes 0 for other uses, or the previous crc32 value if computing incrementally.
unsigned char const * p
pointer to buffer over which CRC32 is run
size_t len
length of buffer p
const u32 ( * tab
big-endian Ethernet table
u32 polynomial
CRC32 BE polynomial
u16 crc_ccitt(u16 crc, u8 const * buffer, size_t len)

recompute the CRC (CRC-CCITT variant) for the data buffer

Parameters

u16 crc
previous CRC value
u8 const * buffer
data pointer
size_t len
number of bytes in the buffer
u16 crc_ccitt_false(u16 crc, u8 const * buffer, size_t len)

recompute the CRC (CRC-CCITT-FALSE variant) for the data buffer

Parameters

u16 crc
previous CRC value
u8 const * buffer
data pointer
size_t len
number of bytes in the buffer
u16 crc_itu_t(u16 crc, const u8 * buffer, size_t len)

Compute the CRC-ITU-T for the data buffer

Parameters

u16 crc
previous CRC value
const u8 * buffer
data pointer
size_t len
number of bytes in the buffer

Description

Returns the updated CRC value

Base 2 log and power Functions

bool is_power_of_2(unsigned long n)

check if a value is a power of two

Parameters

unsigned long n
the value to check

Description

Determine whether some value is a power of two, where zero is not considered a power of two.

Return

true if n is a power of 2, otherwise false.

unsigned long __roundup_pow_of_two(unsigned long n)

round up to nearest power of two

Parameters

unsigned long n
value to round up
unsigned long __rounddown_pow_of_two(unsigned long n)

round down to nearest power of two

Parameters

unsigned long n
value to round down
const_ilog2(n)

log base 2 of 32-bit or a 64-bit constant unsigned value

Parameters

n
parameter

Description

Use this where sparse expects a true constant expression, e.g. for array indices.

ilog2(n)

log base 2 of 32-bit or a 64-bit unsigned value

Parameters

n
parameter

Description

constant-capable log of base 2 calculation - this can be used to initialise global variables from constant data, hence the massive ternary operator construction

selects the appropriately-sized optimised version depending on sizeof(n)

roundup_pow_of_two(n)

round the given value up to nearest power of two

Parameters

n
parameter

Description

round the given value up to the nearest power of two - the result is undefined when n == 0 - this can be used to initialise global variables from constant data

rounddown_pow_of_two(n)

round the given value down to nearest power of two

Parameters

n
parameter

Description

round the given value down to the nearest power of two - the result is undefined when n == 0 - this can be used to initialise global variables from constant data

order_base_2(n)

calculate the (rounded up) base 2 order of the argument

Parameters

n
parameter

Description

The first few values calculated by this routine:
ob2(0) = 0 ob2(1) = 0 ob2(2) = 1 ob2(3) = 2 ob2(4) = 2 ob2(5) = 3 … and so on.
bits_per(n)

calculate the number of bits required for the argument

Parameters

n
parameter

Description

This is constant-capable and can be used for compile time initializations, e.g bitfields.

The first few values calculated by this routine: bf(0) = 1 bf(1) = 1 bf(2) = 2 bf(3) = 2 bf(4) = 3 … and so on.

Integer power Functions

u64 int_pow(u64 base, unsigned int exp)

computes the exponentiation of the given base and exponent

Parameters

u64 base
base which will be raised to the given power
unsigned int exp
power to be raised to

Description

Computes: pow(base, exp), i.e. base raised to the exp power

unsigned long int_sqrt(unsigned long x)

computes the integer square root

Parameters

unsigned long x
integer of which to calculate the sqrt

Description

Computes: floor(sqrt(x))

u32 int_sqrt64(u64 x)

strongly typed int_sqrt function when minimum 64 bit input is expected.

Parameters

u64 x
64bit integer of which to calculate the sqrt

Division Functions

do_div(n, base)

returns 2 values: calculate remainder and update new dividend

Parameters

n
uint64_t dividend (will be updated)
base
uint32_t divisor

Description

Summary: uint32_t remainder = n % base; n = n / base;

Return

(uint32_t)remainder

NOTE

macro parameter n is evaluated multiple times, beware of side effects!

u64 div_u64_rem(u64 dividend, u32 divisor, u32 * remainder)

unsigned 64bit divide with 32bit divisor with remainder

Parameters

u64 dividend
unsigned 64bit dividend
u32 divisor
unsigned 32bit divisor
u32 * remainder
pointer to unsigned 32bit remainder

Return

sets *remainder, then returns dividend / divisor

This is commonly provided by 32bit archs to provide an optimized 64bit divide.

s64 div_s64_rem(s64 dividend, s32 divisor, s32 * remainder)

signed 64bit divide with 32bit divisor with remainder

Parameters

s64 dividend
signed 64bit dividend
s32 divisor
signed 32bit divisor
s32 * remainder
pointer to signed 32bit remainder

Return

sets *remainder, then returns dividend / divisor

u64 div64_u64_rem(u64 dividend, u64 divisor, u64 * remainder)

unsigned 64bit divide with 64bit divisor and remainder

Parameters

u64 dividend
unsigned 64bit dividend
u64 divisor
unsigned 64bit divisor
u64 * remainder
pointer to unsigned 64bit remainder

Return

sets *remainder, then returns dividend / divisor

u64 div64_u64(u64 dividend, u64 divisor)

unsigned 64bit divide with 64bit divisor

Parameters

u64 dividend
unsigned 64bit dividend
u64 divisor
unsigned 64bit divisor

Return

dividend / divisor

s64 div64_s64(s64 dividend, s64 divisor)

signed 64bit divide with 64bit divisor

Parameters

s64 dividend
signed 64bit dividend
s64 divisor
signed 64bit divisor

Return

dividend / divisor

u64 div_u64(u64 dividend, u32 divisor)

unsigned 64bit divide with 32bit divisor

Parameters

u64 dividend
unsigned 64bit dividend
u32 divisor
unsigned 32bit divisor

Description

This is the most common 64bit divide and should be used if possible, as many 32bit archs can optimize this variant better than a full 64bit divide.

s64 div_s64(s64 dividend, s32 divisor)

signed 64bit divide with 32bit divisor

Parameters

s64 dividend
signed 64bit dividend
s32 divisor
signed 32bit divisor
DIV64_U64_ROUND_CLOSEST(dividend, divisor)

unsigned 64bit divide with 64bit divisor rounded to nearest integer

Parameters

dividend
unsigned 64bit dividend
divisor
unsigned 64bit divisor

Description

Divide unsigned 64bit dividend by unsigned 64bit divisor and round to closest integer.

Return

dividend / divisor rounded to nearest integer

s64 div_s64_rem(s64 dividend, s32 divisor, s32 * remainder)

signed 64bit divide with 64bit divisor and remainder

Parameters

s64 dividend
64bit dividend
s32 divisor
64bit divisor
s32 * remainder
64bit remainder
u64 div64_u64_rem(u64 dividend, u64 divisor, u64 * remainder)

unsigned 64bit divide with 64bit divisor and remainder

Parameters

u64 dividend
64bit dividend
u64 divisor
64bit divisor
u64 * remainder
64bit remainder

Description

This implementation is a comparable to algorithm used by div64_u64. But this operation, which includes math for calculating the remainder, is kept distinct to avoid slowing down the div64_u64 operation on 32bit systems.

u64 div64_u64(u64 dividend, u64 divisor)

unsigned 64bit divide with 64bit divisor

Parameters

u64 dividend
64bit dividend
u64 divisor
64bit divisor

Description

This implementation is a modified version of the algorithm proposed by the book ‘Hacker’s Delight’. The original source and full proof can be found here and is available for use without restriction.

http://www.hackersdelight.org/hdcodetxt/divDouble.c.txt

s64 div64_s64(s64 dividend, s64 divisor)

signed 64bit divide with 64bit divisor

Parameters

s64 dividend
64bit dividend
s64 divisor
64bit divisor
unsigned long gcd(unsigned long a, unsigned long b)

calculate and return the greatest common divisor of 2 unsigned longs

Parameters

unsigned long a
first value
unsigned long b
second value

UUID/GUID

void generate_random_uuid(unsigned char uuid)

generate a random UUID

Parameters

unsigned char uuid
where to put the generated UUID

Description

Random UUID interface

Used to create a Boot ID or a filesystem UUID/GUID, but can be useful for other kernel drivers.

bool uuid_is_valid(const char * uuid)

checks if a UUID string is valid

Parameters

const char * uuid
UUID string to check

Description

It checks if the UUID string is following the format:
xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

where x is a hex digit.

Return

true if input is valid UUID string.

Kernel IPC facilities

IPC utilities

int ipc_init(void)

initialise ipc subsystem

Parameters

void
no arguments

Description

The various sysv ipc resources (semaphores, messages and shared memory) are initialised.

A callback routine is registered into the memory hotplug notifier chain: since msgmni scales to lowmem this callback routine will be called upon successful memory add / remove to recompute msmgni.

void ipc_init_ids(struct ipc_ids * ids)

initialise ipc identifiers

Parameters

struct ipc_ids * ids
ipc identifier set

Description

Set up the sequence range to use for the ipc identifier range (limited below ipc_mni) then initialise the keys hashtable and ids idr.

void ipc_init_proc_interface(const char * path, const char * header, int ids, int (*show) (struct seq_file *, void *)

create a proc interface for sysipc types using a seq_file interface.

Parameters

const char * path
Path in procfs
const char * header
Banner to be printed at the beginning of the file.
int ids
ipc id table to iterate.
int (*)(struct seq_file *, void *) show
show routine.
struct kern_ipc_perm * ipc_findkey(struct ipc_ids * ids, key_t key)

find a key in an ipc identifier set

Parameters

struct ipc_ids * ids
ipc identifier set
key_t key
key to find

Description

Returns the locked pointer to the ipc structure if found or NULL otherwise. If key is found ipc points to the owning ipc structure

Called with writer ipc_ids.rwsem held.

int ipc_addid(struct ipc_ids * ids, struct kern_ipc_perm * new, int limit)

add an ipc identifier

Parameters

struct ipc_ids * ids
ipc identifier set
struct kern_ipc_perm * new
new ipc permission set
int limit
limit for the number of used ids

Description

Add an entry ‘new’ to the ipc ids idr. The permissions object is initialised and the first free entry is set up and the index assigned is returned. The ‘new’ entry is returned in a locked state on success.

On failure the entry is not locked and a negative err-code is returned. The caller must use ipc_rcu_putref() to free the identifier.

Called with writer ipc_ids.rwsem held.

int ipcget_new(struct ipc_namespace * ns, struct ipc_ids * ids, const struct ipc_ops * ops, struct ipc_params * params)

create a new ipc object

Parameters

struct ipc_namespace * ns
ipc namespace
struct ipc_ids * ids
ipc identifier set
const struct ipc_ops * ops
the actual creation routine to call
struct ipc_params * params
its parameters

Description

This routine is called by sys_msgget, sys_semget() and sys_shmget() when the key is IPC_PRIVATE.

int ipc_check_perms(struct ipc_namespace * ns, struct kern_ipc_perm * ipcp, const struct ipc_ops * ops, struct ipc_params * params)

check security and permissions for an ipc object

Parameters

struct ipc_namespace * ns
ipc namespace
struct kern_ipc_perm * ipcp
ipc permission set
const struct ipc_ops * ops
the actual security routine to call
struct ipc_params * params
its parameters

Description

This routine is called by sys_msgget(), sys_semget() and sys_shmget() when the key is not IPC_PRIVATE and that key already exists in the ds IDR.

On success, the ipc id is returned.

It is called with ipc_ids.rwsem and ipcp->lock held.

int ipcget_public(struct ipc_namespace * ns, struct ipc_ids * ids, const struct ipc_ops * ops, struct ipc_params * params)

get an ipc object or create a new one

Parameters

struct ipc_namespace * ns
ipc namespace
struct ipc_ids * ids
ipc identifier set
const struct ipc_ops * ops
the actual creation routine to call
struct ipc_params * params
its parameters

Description

This routine is called by sys_msgget, sys_semget() and sys_shmget() when the key is not IPC_PRIVATE. It adds a new entry if the key is not found and does some permission / security checkings if the key is found.

On success, the ipc id is returned.

void ipc_kht_remove(struct ipc_ids * ids, struct kern_ipc_perm * ipcp)

remove an ipc from the key hashtable

Parameters

struct ipc_ids * ids
ipc identifier set
struct kern_ipc_perm * ipcp
ipc perm structure containing the key to remove

Description

ipc_ids.rwsem (as a writer) and the spinlock for this ID are held before this function is called, and remain locked on the exit.

void ipc_rmid(struct ipc_ids * ids, struct kern_ipc_perm * ipcp)

remove an ipc identifier

Parameters

struct ipc_ids * ids
ipc identifier set
struct kern_ipc_perm * ipcp
ipc perm structure containing the identifier to remove

Description

ipc_ids.rwsem (as a writer) and the spinlock for this ID are held before this function is called, and remain locked on the exit.

void ipc_set_key_private(struct ipc_ids * ids, struct kern_ipc_perm * ipcp)

switch the key of an existing ipc to IPC_PRIVATE

Parameters

struct ipc_ids * ids
ipc identifier set
struct kern_ipc_perm * ipcp
ipc perm structure containing the key to modify

Description

ipc_ids.rwsem (as a writer) and the spinlock for this ID are held before this function is called, and remain locked on the exit.

int ipcperms(struct ipc_namespace * ns, struct kern_ipc_perm * ipcp, short flag)

check ipc permissions

Parameters

struct ipc_namespace * ns
ipc namespace
struct kern_ipc_perm * ipcp
ipc permission set
short flag
desired permission set

Description

Check user, group, other permissions for access to ipc resources. return 0 if allowed

flag will most probably be 0 or S_...UGO from <linux/stat.h>

void kernel_to_ipc64_perm(struct kern_ipc_perm * in, struct ipc64_perm * out)

convert kernel ipc permissions to user

Parameters

struct kern_ipc_perm * in
kernel permissions
struct ipc64_perm * out
new style ipc permissions

Description

Turn the kernel object in into a set of permissions descriptions for returning to userspace (out).

void ipc64_perm_to_ipc_perm(struct ipc64_perm * in, struct ipc_perm * out)

convert new ipc permissions to old

Parameters

struct ipc64_perm * in
new style ipc permissions
struct ipc_perm * out
old style ipc permissions

Description

Turn the new style permissions object in into a compatibility object and store it into the out pointer.

struct kern_ipc_perm * ipc_obtain_object_idr(struct ipc_ids * ids, int id)

Parameters

struct ipc_ids * ids
ipc identifier set
int id
ipc id to look for

Description

Look for an id in the ipc ids idr and return associated ipc object.

Call inside the RCU critical section. The ipc object is not locked on exit.

struct kern_ipc_perm * ipc_obtain_object_check(struct ipc_ids * ids, int id)

Parameters

struct ipc_ids * ids
ipc identifier set
int id
ipc id to look for

Description

Similar to ipc_obtain_object_idr() but also checks the ipc object sequence number.

Call inside the RCU critical section. The ipc object is not locked on exit.

int ipcget(struct ipc_namespace * ns, struct ipc_ids * ids, const struct ipc_ops * ops, struct ipc_params * params)

Common sys_*get() code

Parameters

struct ipc_namespace * ns
namespace
struct ipc_ids * ids
ipc identifier set
const struct ipc_ops * ops
operations to be called on ipc object creation, permission checks and further checks
struct ipc_params * params
the parameters needed by the previous operations.

Description

Common routine called by sys_msgget(), sys_semget() and sys_shmget().

int ipc_update_perm(struct ipc64_perm * in, struct kern_ipc_perm * out)

update the permissions of an ipc object

Parameters

struct ipc64_perm * in
the permission given as input.
struct kern_ipc_perm * out
the permission of the ipc to set.
struct kern_ipc_perm * ipcctl_obtain_check(struct ipc_namespace * ns, struct ipc_ids * ids, int id, int cmd, struct ipc64_perm * perm, int extra_perm)

retrieve an ipc object and check permissions

Parameters

struct ipc_namespace * ns
ipc namespace
struct ipc_ids * ids
the table of ids where to look for the ipc
int id
the id of the ipc to retrieve
int cmd
the cmd to check
struct ipc64_perm * perm
the permission to set
int extra_perm
one extra permission parameter used by msq

Description

This function does some common audit and permissions check for some IPC_XXX cmd and is called from semctl_down, shmctl_down and msgctl_down.

It:
  • retrieves the ipc object with the given id in the given table.
  • performs some audit and permission check, depending on the given cmd
  • returns a pointer to the ipc object or otherwise, the corresponding error.

Call holding the both the rwsem and the rcu read lock.

int ipc_parse_version(int * cmd)

ipc call version

Parameters

int * cmd
pointer to command

Description

Return IPC_64 for new style IPC and IPC_OLD for old style IPC. The cmd value is turned from an encoding command and version into just the command code.

FIFO Buffer

kfifo interface

DECLARE_KFIFO_PTR(fifo, type)

macro to declare a fifo pointer object

Parameters

fifo
name of the declared fifo
type
type of the fifo elements
DECLARE_KFIFO(fifo, type, size)

macro to declare a fifo object

Parameters

fifo
name of the declared fifo
type
type of the fifo elements
size
the number of elements in the fifo, this must be a power of 2
INIT_KFIFO(fifo)

Initialize a fifo declared by DECLARE_KFIFO

Parameters

fifo
name of the declared fifo datatype
DEFINE_KFIFO(fifo, type, size)

macro to define and initialize a fifo

Parameters

fifo
name of the declared fifo datatype
type
type of the fifo elements
size
the number of elements in the fifo, this must be a power of 2

Note

the macro can be used for global and local fifo data type variables.

kfifo_initialized(fifo)

Check if the fifo is initialized

Parameters

fifo
address of the fifo to check

Description

Return true if fifo is initialized, otherwise false. Assumes the fifo was 0 before.

kfifo_esize(fifo)

returns the size of the element managed by the fifo

Parameters

fifo
address of the fifo to be used
kfifo_recsize(fifo)

returns the size of the record length field

Parameters

fifo
address of the fifo to be used
kfifo_size(fifo)

returns the size of the fifo in elements

Parameters

fifo
address of the fifo to be used
kfifo_reset(fifo)

removes the entire fifo content

Parameters

fifo
address of the fifo to be used

Note

usage of kfifo_reset() is dangerous. It should be only called when the fifo is exclusived locked or when it is secured that no other thread is accessing the fifo.

kfifo_reset_out(fifo)

skip fifo content

Parameters

fifo
address of the fifo to be used

Note

The usage of kfifo_reset_out() is safe until it will be only called from the reader thread and there is only one concurrent reader. Otherwise it is dangerous and must be handled in the same way as kfifo_reset().

kfifo_len(fifo)

returns the number of used elements in the fifo

Parameters

fifo
address of the fifo to be used
kfifo_is_empty(fifo)

returns true if the fifo is empty

Parameters

fifo
address of the fifo to be used
kfifo_is_full(fifo)

returns true if the fifo is full

Parameters

fifo
address of the fifo to be used
kfifo_avail(fifo)

returns the number of unused elements in the fifo

Parameters

fifo
address of the fifo to be used
kfifo_skip(fifo)

skip output data

Parameters

fifo
address of the fifo to be used
kfifo_peek_len(fifo)

gets the size of the next fifo record

Parameters

fifo
address of the fifo to be used

Description

This function returns the size of the next fifo record in number of bytes.

kfifo_alloc(fifo, size, gfp_mask)

dynamically allocates a new fifo buffer

Parameters

fifo
pointer to the fifo
size
the number of elements in the fifo, this must be a power of 2
gfp_mask
get_free_pages mask, passed to kmalloc()

Description

This macro dynamically allocates a new fifo buffer.

The number of elements will be rounded-up to a power of 2. The fifo will be release with kfifo_free(). Return 0 if no error, otherwise an error code.

kfifo_free(fifo)

frees the fifo

Parameters

fifo
the fifo to be freed
kfifo_init(fifo, buffer, size)

initialize a fifo using a preallocated buffer

Parameters

fifo
the fifo to assign the buffer
buffer
the preallocated buffer to be used
size
the size of the internal buffer, this have to be a power of 2

Description

This macro initializes a fifo using a preallocated buffer.

The number of elements will be rounded-up to a power of 2. Return 0 if no error, otherwise an error code.

kfifo_put(fifo, val)

put data into the fifo

Parameters

fifo
address of the fifo to be used
val
the data to be added

Description

This macro copies the given value into the fifo. It returns 0 if the fifo was full. Otherwise it returns the number processed elements.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_get(fifo, val)

get data from the fifo

Parameters

fifo
address of the fifo to be used
val
address where to store the data

Description

This macro reads the data from the fifo. It returns 0 if the fifo was empty. Otherwise it returns the number processed elements.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_peek(fifo, val)

get data from the fifo without removing

Parameters

fifo
address of the fifo to be used
val
address where to store the data

Description

This reads the data from the fifo without removing it from the fifo. It returns 0 if the fifo was empty. Otherwise it returns the number processed elements.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_in(fifo, buf, n)

put data into the fifo

Parameters

fifo
address of the fifo to be used
buf
the data to be added
n
number of elements to be added

Description

This macro copies the given buffer into the fifo and returns the number of copied elements.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_in_spinlocked(fifo, buf, n, lock)

put data into the fifo using a spinlock for locking

Parameters

fifo
address of the fifo to be used
buf
the data to be added
n
number of elements to be added
lock
pointer to the spinlock to use for locking

Description

This macro copies the given values buffer into the fifo and returns the number of copied elements.

kfifo_out(fifo, buf, n)

get data from the fifo

Parameters

fifo
address of the fifo to be used
buf
pointer to the storage buffer
n
max. number of elements to get

Description

This macro get some data from the fifo and return the numbers of elements copied.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_out_spinlocked(fifo, buf, n, lock)

get data from the fifo using a spinlock for locking

Parameters

fifo
address of the fifo to be used
buf
pointer to the storage buffer
n
max. number of elements to get
lock
pointer to the spinlock to use for locking

Description

This macro get the data from the fifo and return the numbers of elements copied.

kfifo_from_user(fifo, from, len, copied)

puts some data from user space into the fifo

Parameters

fifo
address of the fifo to be used
from
pointer to the data to be added
len
the length of the data to be added
copied
pointer to output variable to store the number of copied bytes

Description

This macro copies at most len bytes from the from into the fifo, depending of the available space and returns -EFAULT/0.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_to_user(fifo, to, len, copied)

copies data from the fifo into user space

Parameters

fifo
address of the fifo to be used
to
where the data must be copied
len
the size of the destination buffer
copied
pointer to output variable to store the number of copied bytes

Description

This macro copies at most len bytes from the fifo into the to buffer and returns -EFAULT/0.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

kfifo_dma_in_prepare(fifo, sgl, nents, len)

setup a scatterlist for DMA input

Parameters

fifo
address of the fifo to be used
sgl
pointer to the scatterlist array
nents
number of entries in the scatterlist array
len
number of elements to transfer

Description

This macro fills a scatterlist for DMA input. It returns the number entries in the scatterlist array.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macros.

kfifo_dma_in_finish(fifo, len)

finish a DMA IN operation

Parameters

fifo
address of the fifo to be used
len
number of bytes to received

Description

This macro finish a DMA IN operation. The in counter will be updated by the len parameter. No error checking will be done.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macros.

kfifo_dma_out_prepare(fifo, sgl, nents, len)

setup a scatterlist for DMA output

Parameters

fifo
address of the fifo to be used
sgl
pointer to the scatterlist array
nents
number of entries in the scatterlist array
len
number of elements to transfer

Description

This macro fills a scatterlist for DMA output which at most len bytes to transfer. It returns the number entries in the scatterlist array. A zero means there is no space available and the scatterlist is not filled.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macros.

kfifo_dma_out_finish(fifo, len)

finish a DMA OUT operation

Parameters

fifo
address of the fifo to be used
len
number of bytes transferred

Description

This macro finish a DMA OUT operation. The out counter will be updated by the len parameter. No error checking will be done.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macros.

kfifo_out_peek(fifo, buf, n)

gets some data from the fifo

Parameters

fifo
address of the fifo to be used
buf
pointer to the storage buffer
n
max. number of elements to get

Description

This macro get the data from the fifo and return the numbers of elements copied. The data is not removed from the fifo.

Note that with only one concurrent reader and one concurrent writer, you don’t need extra locking to use these macro.

relay interface support

Relay interface support is designed to provide an efficient mechanism for tools and facilities to relay large amounts of data from kernel space to user space.

relay interface

int relay_buf_full(struct rchan_buf * buf)

boolean, is the channel buffer full?

Parameters

struct rchan_buf * buf
channel buffer

Description

Returns 1 if the buffer is full, 0 otherwise.
void relay_reset(struct rchan * chan)

reset the channel

Parameters

struct rchan * chan
the channel

Description

This has the effect of erasing all data from all channel buffers and restarting the channel in its initial state. The buffers are not freed, so any mappings are still in effect.

NOTE. Care should be taken that the channel isn’t actually being used by anything when this call is made.

struct rchan * relay_open(const char * base_filename, struct dentry * parent, size_t subbuf_size, size_t n_subbufs, struct rchan_callbacks * cb, void * private_data)

create a new relay channel

Parameters

const char * base_filename
base name of files to create, NULL for buffering only
struct dentry * parent
dentry of parent directory, NULL for root directory or buffer
size_t subbuf_size
size of sub-buffers
size_t n_subbufs
number of sub-buffers
struct rchan_callbacks * cb
client callback functions
void * private_data
user-defined data

Description

Returns channel pointer if successful, NULL otherwise.

Creates a channel buffer for each cpu using the sizes and attributes specified. The created channel buffer files will be named base_filename0…base_filenameN-1. File permissions will be S_IRUSR.

If opening a buffer (parent = NULL) that you later wish to register in a filesystem, call relay_late_setup_files() once the parent dentry is available.

int relay_late_setup_files(struct rchan * chan, const char * base_filename, struct dentry * parent)

triggers file creation

Parameters

struct rchan * chan
channel to operate on
const char * base_filename
base name of files to create
struct dentry * parent
dentry of parent directory, NULL for root directory

Description

Returns 0 if successful, non-zero otherwise.

Use to setup files for a previously buffer-only channel created by relay_open() with a NULL parent dentry.

For example, this is useful for perfomring early tracing in kernel, before VFS is up and then exposing the early results once the dentry is available.

size_t relay_switch_subbuf(struct rchan_buf * buf, size_t length)

switch to a new sub-buffer

Parameters

struct rchan_buf * buf
channel buffer
size_t length
size of current event

Description

Returns either the length passed in or 0 if full.

Performs sub-buffer-switch tasks such as invoking callbacks, updating padding counts, waking up readers, etc.

void relay_subbufs_consumed(struct rchan * chan, unsigned int cpu, size_t subbufs_consumed)

update the buffer’s sub-buffers-consumed count

Parameters

struct rchan * chan
the channel
unsigned int cpu
the cpu associated with the channel buffer to update
size_t subbufs_consumed
number of sub-buffers to add to current buf’s count

Description

Adds to the channel buffer’s consumed sub-buffer count. subbufs_consumed should be the number of sub-buffers newly consumed, not the total consumed.

NOTE. Kernel clients don’t need to call this function if the channel mode is ‘overwrite’.

void relay_close(struct rchan * chan)

close the channel

Parameters

struct rchan * chan
the channel

Description

Closes all channel buffers and frees the channel.
void relay_flush(struct rchan * chan)

close the channel

Parameters

struct rchan * chan
the channel

Description

Flushes all channel buffers, i.e. forces buffer switch.
int relay_mmap_buf(struct rchan_buf * buf, struct vm_area_struct * vma)

mmap channel buffer to process address space

Parameters

struct rchan_buf * buf
relay channel buffer
struct vm_area_struct * vma
vm_area_struct describing memory to be mapped

Description

Returns 0 if ok, negative on error

Caller should already have grabbed mmap_sem.

void * relay_alloc_buf(struct rchan_buf * buf, size_t * size)

allocate a channel buffer

Parameters

struct rchan_buf * buf
the buffer struct
size_t * size
total size of the buffer

Description

Returns a pointer to the resulting buffer, NULL if unsuccessful. The passed in size will get page aligned, if it isn’t already.
struct rchan_buf * relay_create_buf(struct rchan * chan)

allocate and initialize a channel buffer

Parameters

struct rchan * chan
the relay channel

Description

Returns channel buffer if successful, NULL otherwise.
void relay_destroy_channel(struct kref * kref)

free the channel struct

Parameters

struct kref * kref
target kernel reference that contains the relay channel

Description

Should only be called from kref_put().
void relay_destroy_buf(struct rchan_buf * buf)

destroy an rchan_buf struct and associated buffer

Parameters

struct rchan_buf * buf
the buffer struct
void relay_remove_buf(struct kref * kref)

remove a channel buffer

Parameters

struct kref * kref
target kernel reference that contains the relay buffer

Description

Removes the file from the filesystem, which also frees the rchan_buf_struct and the channel buffer. Should only be called from kref_put().
int relay_buf_empty(struct rchan_buf * buf)

boolean, is the channel buffer empty?

Parameters

struct rchan_buf * buf
channel buffer

Description

Returns 1 if the buffer is empty, 0 otherwise.
void wakeup_readers(struct irq_work * work)

wake up readers waiting on a channel

Parameters

struct irq_work * work
contains the channel buffer

Description

This is the function used to defer reader waking
void __relay_reset(struct rchan_buf * buf, unsigned int init)

reset a channel buffer

Parameters

struct rchan_buf * buf
the channel buffer
unsigned int init
1 if this is a first-time initialization

Description

See relay_reset() for description of effect.
void relay_close_buf(struct rchan_buf * buf)

close a channel buffer

Parameters

struct rchan_buf * buf
channel buffer

Description

Marks the buffer finalized and restores the default callbacks. The channel buffer and channel buffer data structure are then freed automatically when the last reference is given up.
int relay_file_open(struct inode * inode, struct file * filp)

open file op for relay files

Parameters

struct inode * inode
the inode
struct file * filp
the file

Description

Increments the channel buffer refcount.
int relay_file_mmap(struct file * filp, struct vm_area_struct * vma)

mmap file op for relay files

Parameters

struct file * filp
the file
struct vm_area_struct * vma
the vma describing what to map

Description

Calls upon relay_mmap_buf() to map the file into user space.
__poll_t relay_file_poll(struct file * filp, poll_table * wait)

poll file op for relay files

Parameters

struct file * filp
the file
poll_table * wait
poll table

Description

Poll implemention.
int relay_file_release(struct inode * inode, struct file * filp)

release file op for relay files

Parameters

struct inode * inode
the inode
struct file * filp
the file

Description

Decrements the channel refcount, as the filesystem is no longer using it.
size_t relay_file_read_subbuf_avail(size_t read_pos, struct rchan_buf * buf)

return bytes available in sub-buffer

Parameters

size_t read_pos
file read position
struct rchan_buf * buf
relay channel buffer
size_t relay_file_read_start_pos(size_t read_pos, struct rchan_buf * buf)

find the first available byte to read

Parameters

size_t read_pos
file read position
struct rchan_buf * buf
relay channel buffer

Description

If the read_pos is in the middle of padding, return the position of the first actually available byte, otherwise return the original value.
size_t relay_file_read_end_pos(struct rchan_buf * buf, size_t read_pos, size_t count)

return the new read position

Parameters

struct rchan_buf * buf
relay channel buffer
size_t read_pos
file read position
size_t count
number of bytes to be read

Module Support

Module Loading

int __request_module(bool wait, const char * fmt, ...)

try to load a kernel module

Parameters

bool wait
wait (or not) for the operation to complete
const char * fmt
printf style format string for the name of the module
...
arguments as specified in the format string

Description

Load a module using the user mode module loader. The function returns zero on success or a negative errno code or positive exit code from “modprobe” on failure. Note that a successful module load does not mean the module did not then unload and exit on an error of its own. Callers must check that the service they requested is now available not blindly invoke it.

If module auto-loading support is disabled then this function becomes a no-operation.

Inter Module support

Refer to the file kernel/module.c for more information.

Hardware Interfaces

Interrupt Handling

bool synchronize_hardirq(unsigned int irq)

wait for pending hard IRQ handlers (on other CPUs)

Parameters

unsigned int irq
interrupt number to wait for

Description

This function waits for any pending hard IRQ handlers for this interrupt to complete before returning. If you use this function while holding a resource the IRQ handler may need you will deadlock. It does not take associated threaded handlers into account.

Do not use this for shutdown scenarios where you must be sure that all parts (hardirq and threaded handler) have completed.

Return

false if a threaded handler is active.

This function may be called - with care - from IRQ context.

It does not check whether there is an interrupt in flight at the hardware level, but not serviced yet, as this might deadlock when called with interrupts disabled and the target CPU of the interrupt is the current CPU.

void synchronize_irq(unsigned int irq)

wait for pending IRQ handlers (on other CPUs)

Parameters

unsigned int irq
interrupt number to wait for

Description

This function waits for any pending IRQ handlers for this interrupt to complete before returning. If you use this function while holding a resource the IRQ handler may need you will deadlock.

Can only be called from preemptible code as it might sleep when an interrupt thread is associated to irq.

It optionally makes sure (when the irq chip supports that method) that the interrupt is not pending in any CPU and waiting for service.

int irq_set_affinity_notifier(unsigned int irq, struct irq_affinity_notify * notify)

control notification of IRQ affinity changes

Parameters

unsigned int irq
Interrupt for which to enable/disable notification
struct irq_affinity_notify * notify
Context for notification, or NULL to disable notification. Function pointers must be initialised; the other fields will be initialised by this function.

Description

Must be called in process context. Notification may only be enabled after the IRQ is allocated and must be disabled before the IRQ is freed using free_irq().
int irq_set_vcpu_affinity(unsigned int irq, void * vcpu_info)

Set vcpu affinity for the interrupt

Parameters

unsigned int irq
interrupt number to set affinity
void * vcpu_info
vCPU specific data or pointer to a percpu array of vCPU specific data for percpu_devid interrupts

Description

This function uses the vCPU specific data to set the vCPU affinity for an irq. The vCPU specific data is passed from outside, such as KVM. One example code path is as below: KVM -> IOMMU -> irq_set_vcpu_affinity().
void disable_irq_nosync(unsigned int irq)

disable an irq without waiting

Parameters

unsigned int irq
Interrupt to disable

Description

Disable the selected interrupt line. Disables and Enables are nested. Unlike disable_irq(), this function does not ensure existing instances of the IRQ handler have completed before returning.

This function may be called from IRQ context.

void disable_irq(unsigned int irq)

disable an irq and wait for completion

Parameters

unsigned int irq
Interrupt to disable

Description

Disable the selected interrupt line. Enables and Disables are nested. This function waits for any pending IRQ handlers for this interrupt to complete before returning. If you use this function while holding a resource the IRQ handler may need you will deadlock.

This function may be called - with care - from IRQ context.

bool disable_hardirq(unsigned int irq)

disables an irq and waits for hardirq completion

Parameters

unsigned int irq
Interrupt to disable

Description

Disable the selected interrupt line. Enables and Disables are nested. This function waits for any pending hard IRQ handlers for this interrupt to complete before returning. If you use this function while holding a resource the hard IRQ handler may need you will deadlock.

When used to optimistically disable an interrupt from atomic context the return value must be checked.

Return

false if a threaded handler is active.

This function may be called - with care - from IRQ context.
void enable_irq(unsigned int irq)

enable handling of an irq

Parameters

unsigned int irq
Interrupt to enable

Description

Undoes the effect of one call to disable_irq(). If this matches the last disable, processing of interrupts on this IRQ line is re-enabled.

This function may be called from IRQ context only when desc->irq_data.chip->bus_lock and desc->chip->bus_sync_unlock are NULL !

int irq_set_irq_wake(unsigned int irq, unsigned int on)

control irq power management wakeup

Parameters

unsigned int irq
interrupt to control
unsigned int on
enable/disable power management wakeup

Description

Enable/disable power management wakeup mode, which is disabled by default. Enables and disables must match, just as they match for non-wakeup mode support.

Wakeup mode lets this IRQ wake the system from sleep states like “suspend to RAM”.

Note

irq enable/disable state is completely orthogonal
to the enable/disable state of irq wake. An irq can be disabled with disable_irq() and still wake the system as long as the irq has wake enabled. If this does not hold, then the underlying irq chip and the related driver need to be investigated.
void irq_wake_thread(unsigned int irq, void * dev_id)

wake the irq thread for the action identified by dev_id

Parameters

unsigned int irq
Interrupt line
void * dev_id
Device identity for which the thread should be woken
int setup_irq(unsigned int irq, struct irqaction * act)

setup an interrupt

Parameters

unsigned int irq
Interrupt line to setup
struct irqaction * act
irqaction for the interrupt

Description

Used to statically setup interrupts in the early boot process.

void remove_irq(unsigned int irq, struct irqaction * act)

free an interrupt

Parameters

unsigned int irq
Interrupt line to free
struct irqaction * act
irqaction for the interrupt

Description

Used to remove interrupts statically setup by the early boot process.

const void * free_irq(unsigned int irq, void * dev_id)

free an interrupt allocated with request_irq

Parameters

unsigned int irq
Interrupt line to free
void * dev_id
Device identity to free

Description

Remove an interrupt handler. The handler is removed and if the interrupt line is no longer in use by any driver it is disabled. On a shared IRQ the caller must ensure the interrupt is disabled on the card it drives before calling this function. The function does not return until any executing interrupts for this IRQ have completed.

This function must not be called from interrupt context.

Returns the devname argument passed to request_irq.

int request_threaded_irq(unsigned int irq, irq_handler_t handler, irq_handler_t thread_fn, unsigned long irqflags, const char * devname, void * dev_id)

allocate an interrupt line

Parameters

unsigned int irq
Interrupt line to allocate
irq_handler_t handler
Function to be called when the IRQ occurs. Primary handler for threaded interrupts If NULL and thread_fn != NULL the default primary handler is installed
irq_handler_t thread_fn
Function called from the irq handler thread If NULL, no irq thread is created
unsigned long irqflags
Interrupt type flags
const char * devname
An ascii name for the claiming device
void * dev_id
A cookie passed back to the handler function

Description

This call allocates interrupt resources and enables the interrupt line and IRQ handling. From the point this call is made your handler function may be invoked. Since your handler function must clear any interrupt the board raises, you must take care both to initialise your hardware and to set up the interrupt handler in the right order.

If you want to set up a threaded irq handler for your device then you need to supply handler and thread_fn. handler is still called in hard interrupt context and has to check whether the interrupt originates from the device. If yes it needs to disable the interrupt on the device and return IRQ_WAKE_THREAD which will wake up the handler thread and run thread_fn. This split handler design is necessary to support shared interrupts.

Dev_id must be globally unique. Normally the address of the device data structure is used as the cookie. Since the handler receives this value it makes sense to use it.

If your interrupt is shared you must pass a non NULL dev_id as this is required when freeing the interrupt.

Flags:

IRQF_SHARED Interrupt is shared IRQF_TRIGGER_* Specify active edge(s) or level

int request_any_context_irq(unsigned int irq, irq_handler_t handler, unsigned long flags, const char * name, void * dev_id)

allocate an interrupt line

Parameters

unsigned int irq
Interrupt line to allocate
irq_handler_t handler
Function to be called when the IRQ occurs. Threaded handler for threaded interrupts.
unsigned long flags
Interrupt type flags
const char * name
An ascii name for the claiming device
void * dev_id
A cookie passed back to the handler function

Description

This call allocates interrupt resources and enables the interrupt line and IRQ handling. It selects either a hardirq or threaded handling method depending on the context.

On failure, it returns a negative value. On success, it returns either IRQC_IS_HARDIRQ or IRQC_IS_NESTED.

bool irq_percpu_is_enabled(unsigned int irq)

Check whether the per cpu irq is enabled

Parameters

unsigned int irq
Linux irq number to check for

Description

Must be called from a non migratable context. Returns the enable state of a per cpu interrupt on the current cpu.

void free_percpu_irq(unsigned int irq, void __percpu * dev_id)

free an interrupt allocated with request_percpu_irq

Parameters

unsigned int irq
Interrupt line to free
void __percpu * dev_id
Device identity to free

Description

Remove a percpu interrupt handler. The handler is removed, but the interrupt line is not disabled. This must be done on each CPU before calling this function. The function does not return until any executing interrupts for this IRQ have completed.

This function must not be called from interrupt context.

int __request_percpu_irq(unsigned int irq, irq_handler_t handler, unsigned long flags, const char * devname, void __percpu * dev_id)

allocate a percpu interrupt line

Parameters

unsigned int irq
Interrupt line to allocate
irq_handler_t handler
Function to be called when the IRQ occurs.
unsigned long flags
Interrupt type flags (IRQF_TIMER only)
const char * devname
An ascii name for the claiming device
void __percpu * dev_id
A percpu cookie passed back to the handler function

Description

This call allocates interrupt resources and enables the interrupt on the local CPU. If the interrupt is supposed to be enabled on other CPUs, it has to be done on each CPU using enable_percpu_irq().

Dev_id must be globally unique. It is a per-cpu variable, and the handler gets called with the interrupted CPU’s instance of that variable.

int irq_get_irqchip_state(unsigned int irq, enum irqchip_irq_state which, bool * state)

returns the irqchip state of a interrupt.

Parameters

unsigned int irq
Interrupt line that is forwarded to a VM
enum irqchip_irq_state which
One of IRQCHIP_STATE_* the caller wants to know about
bool * state
a pointer to a boolean where the state is to be storeed

Description

This call snapshots the internal irqchip state of an interrupt, returning into state the bit corresponding to stage which

This function should be called with preemption disabled if the interrupt controller has per-cpu registers.

int irq_set_irqchip_state(unsigned int irq, enum irqchip_irq_state which, bool val)

set the state of a forwarded interrupt.

Parameters

unsigned int irq
Interrupt line that is forwarded to a VM
enum irqchip_irq_state which
State to be restored (one of IRQCHIP_STATE_*)
bool val
Value corresponding to which

Description

This call sets the internal irqchip state of an interrupt, depending on the value of which.

This function should be called with preemption disabled if the interrupt controller has per-cpu registers.

DMA Channels

int request_dma(unsigned int dmanr, const char * device_id)

request and reserve a system DMA channel

Parameters

unsigned int dmanr
DMA channel number
const char * device_id
reserving device ID string, used in /proc/dma
void free_dma(unsigned int dmanr)

free a reserved system DMA channel

Parameters

unsigned int dmanr
DMA channel number

Resources Management

struct resource * request_resource_conflict(struct resource * root, struct resource * new)

request and reserve an I/O or memory resource

Parameters

struct resource * root
root resource descriptor
struct resource * new
resource descriptor desired by caller

Description

Returns 0 for success, conflict resource on error.

int find_next_iomem_res(resource_size_t start, resource_size_t end, unsigned long flags, unsigned long desc, bool first_lvl, struct resource * res)

Parameters

resource_size_t start
start address of the resource searched for
resource_size_t end
end address of same resource
unsigned long flags
flags which the resource must have
unsigned long desc
descriptor the resource must have
bool first_lvl
walk only the first level children, if set
struct resource * res
return ptr, if resource found

Description

caller must specify start, end, flags, and desc (which may be IORES_DESC_NONE).

If a resource is found, returns 0 and ***res is overwritten with the part of the resource that’s within [**start..**end**]; if none is found, returns -ENODEV. Returns -EINVAL for invalid parameters.

This function walks the whole tree and not just first level children unless first_lvl is true.

int reallocate_resource(struct resource * root, struct resource * old, resource_size_t newsize, struct resource_constraint * constraint)

allocate a slot in the resource tree given range & alignment. The resource will be relocated if the new size cannot be reallocated in the current location.

Parameters

struct resource * root
root resource descriptor
struct resource * old
resource descriptor desired by caller
resource_size_t newsize
new size of the resource descriptor
struct resource_constraint * constraint
the size and alignment constraints to be met.
struct resource * lookup_resource(struct resource * root, resource_size_t start)

find an existing resource by a resource start address

Parameters

struct resource * root
root resource descriptor
resource_size_t start
resource start address

Description

Returns a pointer to the resource if found, NULL otherwise

struct resource * insert_resource_conflict(struct resource * parent, struct resource * new)

Inserts resource in the resource tree

Parameters

struct resource * parent
parent of the new resource
struct resource * new
new resource to insert

Description

Returns 0 on success, conflict resource if the resource can’t be inserted.

This function is equivalent to request_resource_conflict when no conflict happens. If a conflict happens, and the conflicting resources entirely fit within the range of the new resource, then the new resource is inserted and the conflicting resources become children of the new resource.

This function is intended for producers of resources, such as FW modules and bus drivers.

void insert_resource_expand_to_fit(struct resource * root, struct resource * new)

Insert a resource into the resource tree

Parameters

struct resource * root
root resource descriptor
struct resource * new
new resource to insert

Description

Insert a resource into the resource tree, possibly expanding it in order to make it encompass any conflicting resources.

resource_size_t resource_alignment(struct resource * res)

calculate resource’s alignment

Parameters

struct resource * res
resource pointer

Description

Returns alignment on success, 0 (invalid alignment) on failure.

int release_mem_region_adjustable(struct resource * parent, resource_size_t start, resource_size_t size)

release a previously reserved memory region

Parameters

struct resource * parent
parent resource descriptor
resource_size_t start
resource start address
resource_size_t size
resource region size

Description

This interface is intended for memory hot-delete. The requested region is released from a currently busy memory resource. The requested region must either match exactly or fit into a single busy resource entry. In the latter case, the remaining resource is adjusted accordingly. Existing children of the busy memory resource must be immutable in the request.

Note

  • Additional release conditions, such as overlapping region, can be supported after they are confirmed as valid cases.
  • When a busy memory resource gets split into two entries, the code assumes that all children remain in the lower address entry for simplicity. Enhance this logic when necessary.
int request_resource(struct resource * root, struct resource * new)

request and reserve an I/O or memory resource

Parameters

struct resource * root
root resource descriptor
struct resource * new
resource descriptor desired by caller

Description

Returns 0 for success, negative error code on error.

int release_resource(struct resource * old)

release a previously reserved resource

Parameters

struct resource * old
resource pointer
int walk_iomem_res_desc(unsigned long desc, unsigned long flags, u64 start, u64 end, void * arg, int (*func) (struct resource *, void *)

Parameters

unsigned long desc
I/O resource descriptor. Use IORES_DESC_NONE to skip desc check.
unsigned long flags
I/O resource flags
u64 start
start addr
u64 end
end addr
void * arg
function argument for the callback func
int (*)(struct resource *, void *) func
callback function that is called for each qualifying resource area

Description

ranges. This walks through whole tree and not just first level children. All the memory ranges which overlap start,end and also match flags and desc are valid candidates.

NOTE

For a new descriptor search, define a new IORES_DESC in <linux/ioport.h> and set it in ‘desc’ of a target resource entry.

int region_intersects(resource_size_t start, size_t size, unsigned long flags, unsigned long desc)

determine intersection of region with known resources

Parameters

resource_size_t start
region start address
size_t size
size of region
unsigned long flags
flags of resource (in iomem_resource)
unsigned long desc
descriptor of resource (in iomem_resource) or IORES_DESC_NONE

Description

Check if the specified region partially overlaps or fully eclipses a resource identified by flags and desc (optional with IORES_DESC_NONE). Return REGION_DISJOINT if the region does not overlap flags/desc, return REGION_MIXED if the region overlaps flags/desc and another resource, and return REGION_INTERSECTS if the region overlaps flags/desc and no other defined resource. Note that REGION_INTERSECTS is also returned in the case when the specified region overlaps RAM and undefined memory holes.

region_intersect() is used by memory remapping functions to ensure the user is not remapping RAM and is a vast speed up over walking through the resource table page by page.

int allocate_resource(struct resource * root, struct resource * new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, resource_size_t (*alignf) (void *, const struct resource *, resource_size_t, resource_size_t, void * alignf_data)

allocate empty slot in the resource tree given range & alignment. The resource will be reallocated with a new size if it was already allocated

Parameters

struct resource * root
root resource descriptor
struct resource * new
resource descriptor desired by caller
resource_size_t size
requested resource region size
resource_size_t min
minimum boundary to allocate
resource_size_t max
maximum boundary to allocate
resource_size_t align
alignment requested, in bytes
resource_size_t (*)(void *, const struct resource *, resource_size_t, resource_size_t) alignf
alignment function, optional, called if not NULL
void * alignf_data
arbitrary data to pass to the alignf function
int insert_resource(struct resource * parent, struct resource * new)

Inserts a resource in the resource tree

Parameters

struct resource * parent
parent of the new resource
struct resource * new
new resource to insert

Description

Returns 0 on success, -EBUSY if the resource can’t be inserted.

This function is intended for producers of resources, such as FW modules and bus drivers.

int remove_resource(struct resource * old)

Remove a resource in the resource tree

Parameters

struct resource * old
resource to remove

Description

Returns 0 on success, -EINVAL if the resource is not valid.

This function removes a resource previously inserted by insert_resource() or insert_resource_conflict(), and moves the children (if any) up to where they were before. insert_resource() and insert_resource_conflict() insert a new resource, and move any conflicting resources down to the children of the new resource.

insert_resource(), insert_resource_conflict() and remove_resource() are intended for producers of resources, such as FW modules and bus drivers.

int adjust_resource(struct resource * res, resource_size_t start, resource_size_t size)

modify a resource’s start and size

Parameters

struct resource * res
resource to modify
resource_size_t start
new start value
resource_size_t size
new size

Description

Given an existing resource, change its start and size to match the arguments. Returns 0 on success, -EBUSY if it can’t fit. Existing children of the resource are assumed to be immutable.

struct resource * __request_region(struct resource * parent, resource_size_t start, resource_size_t n, const char * name, int flags)

create a new busy resource region

Parameters

struct resource * parent
parent resource descriptor
resource_size_t start
resource start address
resource_size_t n
resource region size
const char * name
reserving caller’s ID string
int flags
IO resource flags
void __release_region(struct resource * parent, resource_size_t start, resource_size_t n)

release a previously reserved resource region

Parameters

struct resource * parent
parent resource descriptor
resource_size_t start
resource start address
resource_size_t n
resource region size

Description

The described resource region must match a currently busy region.

int devm_request_resource(struct device * dev, struct resource * root, struct resource * new)

request and reserve an I/O or memory resource

Parameters

struct device * dev
device for which to request the resource
struct resource * root
root of the resource tree from which to request the resource
struct resource * new
descriptor of the resource to request

Description

This is a device-managed version of request_resource(). There is usually no need to release resources requested by this function explicitly since that will be taken care of when the device is unbound from its driver. If for some reason the resource needs to be released explicitly, because of ordering issues for example, drivers must call devm_release_resource() rather than the regular release_resource().

When a conflict is detected between any existing resources and the newly requested resource, an error message will be printed.

Returns 0 on success or a negative error code on failure.

void devm_release_resource(struct device * dev, struct resource * new)

release a previously requested resource

Parameters

struct device * dev
device for which to release the resource
struct resource * new
descriptor of the resource to release

Description

Releases a resource previously requested using devm_request_resource().

struct resource * devm_request_free_mem_region(struct device * dev, struct resource * base, unsigned long size)

find free region for device private memory

Parameters

struct device * dev
device struct to bind the resource to
struct resource * base
resource tree to look in
unsigned long size
size in bytes of the device memory to add

Description

This function tries to find an empty range of physical address big enough to contain the new resource, so that it can later be hotplugged as ZONE_DEVICE memory, which in turn allocates struct pages.

MTRR Handling

int arch_phys_wc_add(unsigned long base, unsigned long size)

add a WC MTRR and handle errors if PAT is unavailable

Parameters

unsigned long base
Physical base address
unsigned long size
Size of region

Description

If PAT is available, this does nothing. If PAT is unavailable, it attempts to add a WC MTRR covering size bytes starting at base and logs an error if this fails.

The called should provide a power of two size on an equivalent power of two boundary.

Drivers must store the return value to pass to mtrr_del_wc_if_needed, but drivers should not try to interpret that return value.

Security Framework

int security_init(void)

initializes the security framework

Parameters

void
no arguments

Description

This should be called early in the kernel initialization sequence.

void security_add_hooks(struct security_hook_list * hooks, int count, char * lsm)

Add a modules hooks to the hook lists.

Parameters

struct security_hook_list * hooks
the hooks to add
int count
the number of hooks to add
char * lsm
the name of the security module

Description

Each LSM has to register its hooks with the infrastructure.

int lsm_cred_alloc(struct cred * cred, gfp_t gfp)

allocate a composite cred blob

Parameters

struct cred * cred
the cred that needs a blob
gfp_t gfp
allocation type

Description

Allocate the cred blob for all the modules

Returns 0, or -ENOMEM if memory can’t be allocated.

void lsm_early_cred(struct cred * cred)

during initialization allocate a composite cred blob

Parameters

struct cred * cred
the cred that needs a blob

Description

Allocate the cred blob for all the modules

int lsm_file_alloc(struct file * file)

allocate a composite file blob

Parameters

struct file * file
the file that needs a blob

Description

Allocate the file blob for all the modules

Returns 0, or -ENOMEM if memory can’t be allocated.

int lsm_inode_alloc(struct inode * inode)

allocate a composite inode blob

Parameters

struct inode * inode
the inode that needs a blob

Description

Allocate the inode blob for all the modules

Returns 0, or -ENOMEM if memory can’t be allocated.

int lsm_task_alloc(struct task_struct * task)

allocate a composite task blob

Parameters

struct task_struct * task
the task that needs a blob

Description

Allocate the task blob for all the modules

Returns 0, or -ENOMEM if memory can’t be allocated.

int lsm_ipc_alloc(struct kern_ipc_perm * kip)

allocate a composite ipc blob

Parameters

struct kern_ipc_perm * kip
the ipc that needs a blob

Description

Allocate the ipc blob for all the modules

Returns 0, or -ENOMEM if memory can’t be allocated.

int lsm_msg_msg_alloc(struct msg_msg * mp)

allocate a composite msg_msg blob

Parameters

struct msg_msg * mp
the msg_msg that needs a blob

Description

Allocate the ipc blob for all the modules

Returns 0, or -ENOMEM if memory can’t be allocated.

void lsm_early_task(struct task_struct * task)

during initialization allocate a composite task blob

Parameters

struct task_struct * task
the task that needs a blob

Description

Allocate the task blob for all the modules

struct dentry * securityfs_create_file(const char * name, umode_t mode, struct dentry * parent, void * data, const struct file_operations * fops)

create a file in the securityfs filesystem

Parameters

const char * name
a pointer to a string containing the name of the file to create.
umode_t mode
the permission that the file should have
struct dentry * parent
a pointer to the parent dentry for this file. This should be a directory dentry if set. If this parameter is NULL, then the file will be created in the root of the securityfs filesystem.
void * data
a pointer to something that the caller will want to get to later on. The inode.i_private pointer will point to this value on the open() call.
const struct file_operations * fops
a pointer to a struct file_operations that should be used for this file.

Description

This function creates a file in securityfs with the given name.

This function returns a pointer to a dentry if it succeeds. This pointer must be passed to the securityfs_remove() function when the file is to be removed (no automatic cleanup happens if your module is unloaded, you are responsible here). If an error occurs, the function will return the error value (via ERR_PTR).

If securityfs is not enabled in the kernel, the value -ENODEV is returned.

struct dentry * securityfs_create_dir(const char * name, struct dentry * parent)

create a directory in the securityfs filesystem

Parameters

const char * name
a pointer to a string containing the name of the directory to create.
struct dentry * parent
a pointer to the parent dentry for this file. This should be a directory dentry if set. If this parameter is NULL, then the directory will be created in the root of the securityfs filesystem.

Description

This function creates a directory in securityfs with the given name.

This function returns a pointer to a dentry if it succeeds. This pointer must be passed to the securityfs_remove() function when the file is to be removed (no automatic cleanup happens if your module is unloaded, you are responsible here). If an error occurs, the function will return the error value (via ERR_PTR).

If securityfs is not enabled in the kernel, the value -ENODEV is returned.

create a symlink in the securityfs filesystem

Parameters

const char * name
a pointer to a string containing the name of the symlink to create.
struct dentry * parent
a pointer to the parent dentry for the symlink. This should be a directory dentry if set. If this parameter is NULL, then the directory will be created in the root of the securityfs filesystem.
const char * target
a pointer to a string containing the name of the symlink’s target. If this parameter is NULL, then the iops parameter needs to be setup to handle .readlink and .get_link inode_operations.
const struct inode_operations * iops
a pointer to the struct inode_operations to use for the symlink. If this parameter is NULL, then the default simple_symlink_inode operations will be used.

Description

This function creates a symlink in securityfs with the given name.

This function returns a pointer to a dentry if it succeeds. This pointer must be passed to the securityfs_remove() function when the file is to be removed (no automatic cleanup happens if your module is unloaded, you are responsible here). If an error occurs, the function will return the error value (via ERR_PTR).

If securityfs is not enabled in the kernel, the value -ENODEV is returned.

void securityfs_remove(struct dentry * dentry)

removes a file or directory from the securityfs filesystem

Parameters

struct dentry * dentry
a pointer to a the dentry of the file or directory to be removed.

Description

This function removes a file or directory in securityfs that was previously created with a call to another securityfs function (like securityfs_create_file() or variants thereof.)

This function is required to be called in order for the file to be removed. No automatic cleanup of files will happen when a module is removed; you are responsible here.

Audit Interfaces

struct audit_buffer * audit_log_start(struct audit_context * ctx, gfp_t gfp_mask, int type)

obtain an audit buffer

Parameters

struct audit_context * ctx
audit_context (may be NULL)
gfp_t gfp_mask
type of allocation
int type
audit message type

Description

Returns audit_buffer pointer on success or NULL on error.

Obtain an audit buffer. This routine does locking to obtain the audit buffer, but then no locking is required for calls to audit_log_*format. If the task (ctx) is a task that is currently in a syscall, then the syscall is marked as auditable and an audit record will be written at syscall exit. If there is no associated task, then task context (ctx) should be NULL.

void audit_log_format(struct audit_buffer * ab, const char * fmt, ...)

format a message into the audit buffer.

Parameters

struct audit_buffer * ab
audit_buffer
const char * fmt
format string
...
optional parameters matching fmt string

Description

All the work is done in audit_log_vformat.

void audit_log_end(struct audit_buffer * ab)

end one audit record

Parameters

struct audit_buffer * ab
the audit_buffer

Description

We can not do a netlink send inside an irq context because it blocks (last arg, flags, is not set to MSG_DONTWAIT), so the audit buffer is placed on a queue and a tasklet is scheduled to remove them from the queue outside the irq context. May be called in any context.

void audit_log(struct audit_context * ctx, gfp_t gfp_mask, int type, const char * fmt, ...)

Log an audit record

Parameters

struct audit_context * ctx
audit context
gfp_t gfp_mask
type of allocation
int type
audit message type
const char * fmt
format string to use
...
variable parameters matching the format string

Description

This is a convenience function that calls audit_log_start, audit_log_vformat, and audit_log_end. It may be called in any context.

int audit_alloc(struct task_struct * tsk)

allocate an audit context block for a task

Parameters

struct task_struct * tsk
task

Description

Filter on the task information and allocate a per-task audit context if necessary. Doing so turns on system call auditing for the specified task. This is called from copy_process, so no lock is needed.

void __audit_free(struct task_struct * tsk)

free a per-task audit context

Parameters

struct task_struct * tsk
task whose audit context block to free

Description

Called from copy_process and do_exit

void __audit_syscall_entry(int major, unsigned long a1, unsigned long a2, unsigned long a3, unsigned long a4)

fill in an audit record at syscall entry

Parameters

int major
major syscall type (function)
unsigned long a1
additional syscall register 1
unsigned long a2
additional syscall register 2
unsigned long a3
additional syscall register 3
unsigned long a4
additional syscall register 4

Description

Fill in audit context at syscall entry. This only happens if the audit context was created when the task was created and the state or filters demand the audit context be built. If the state from the per-task filter or from the per-syscall filter is AUDIT_RECORD_CONTEXT, then the record will be written at syscall exit time (otherwise, it will only be written if another part of the kernel requests that it be written).

void __audit_syscall_exit(int success, long return_code)

deallocate audit context after a system call

Parameters

int success
success value of the syscall
long return_code
return value of the syscall

Description

Tear down after system call. If the audit context has been marked as auditable (either because of the AUDIT_RECORD_CONTEXT state from filtering, or because some other part of the kernel wrote an audit message), then write out the syscall information. In call cases, free the names stored from getname().

struct filename * __audit_reusename(const __user char * uptr)

fill out filename with info from existing entry

Parameters

const __user char * uptr
userland ptr to pathname

Description

Search the audit_names list for the current audit context. If there is an existing entry with a matching “uptr” then return the filename associated with that audit_name. If not, return NULL.

void __audit_getname(struct filename * name)

add a name to the list

Parameters

struct filename * name
name to add

Description

Add a name to the list of audit names for this context. Called from fs/namei.c:getname().

void __audit_inode(struct filename * name, const struct dentry * dentry, unsigned int flags)

store the inode and device from a lookup

Parameters

struct filename * name
name being audited
const struct dentry * dentry
dentry being audited
unsigned int flags
attributes for this particular entry
int auditsc_get_stamp(struct audit_context * ctx, struct timespec64 * t, unsigned int * serial)

get local copies of audit_context values

Parameters

struct audit_context * ctx
audit_context for the task
struct timespec64 * t
timespec64 to store time recorded in the audit_context
unsigned int * serial
serial value that is recorded in the audit_context

Description

Also sets the context as auditable.

void __audit_mq_open(int oflag, umode_t mode, struct mq_attr * attr)

record audit data for a POSIX MQ open

Parameters

int oflag
open flag
umode_t mode
mode bits
struct mq_attr * attr
queue attributes
void __audit_mq_sendrecv(mqd_t mqdes, size_t msg_len, unsigned int msg_prio, const struct timespec64 * abs_timeout)

record audit data for a POSIX MQ timed send/receive

Parameters

mqd_t mqdes
MQ descriptor
size_t msg_len
Message length
unsigned int msg_prio
Message priority
const struct timespec64 * abs_timeout
Message timeout in absolute time
void __audit_mq_notify(mqd_t mqdes, const struct sigevent * notification)

record audit data for a POSIX MQ notify

Parameters

mqd_t mqdes
MQ descriptor
const struct sigevent * notification
Notification event
void __audit_mq_getsetattr(mqd_t mqdes, struct mq_attr * mqstat)

record audit data for a POSIX MQ get/set attribute

Parameters

mqd_t mqdes
MQ descriptor
struct mq_attr * mqstat
MQ flags
void __audit_ipc_obj(struct kern_ipc_perm * ipcp)

record audit data for ipc object

Parameters

struct kern_ipc_perm * ipcp
ipc permissions
void __audit_ipc_set_perm(unsigned long qbytes, uid_t uid, gid_t gid, umode_t mode)

record audit data for new ipc permissions

Parameters

unsigned long qbytes
msgq bytes
uid_t uid
msgq user id
gid_t gid
msgq group id
umode_t mode
msgq mode (permissions)

Description

Called only after audit_ipc_obj().

int __audit_socketcall(int nargs, unsigned long * args)

record audit data for sys_socketcall

Parameters

int nargs
number of args, which should not be more than AUDITSC_ARGS.
unsigned long * args
args array
void __audit_fd_pair(int fd1, int fd2)

record audit data for pipe and socketpair

Parameters

int fd1
the first file descriptor
int fd2
the second file descriptor
int __audit_sockaddr(int len, void * a)

record audit data for sys_bind, sys_connect, sys_sendto

Parameters

int len
data length in user space
void * a
data address in kernel space

Description

Returns 0 for success or NULL context or < 0 on error.

int audit_signal_info_syscall(struct task_struct * t)

record signal info for syscalls

Parameters

struct task_struct * t
task being signaled

Description

If the audit subsystem is being terminated, record the task (pid) and uid that is doing that.

int __audit_log_bprm_fcaps(struct linux_binprm * bprm, const struct cred * new, const struct cred * old)

store information about a loading bprm and relevant fcaps

Parameters

struct linux_binprm * bprm
pointer to the bprm being processed
const struct cred * new
the proposed new credentials
const struct cred * old
the old credentials

Description

Simply check if the proc already has the caps given by the file and if not store the priv escalation info for later auditing at the end of the syscall

-Eric

void __audit_log_capset(const struct cred * new, const struct cred * old)

store information about the arguments to the capset syscall

Parameters

const struct cred * new
the new credentials
const struct cred * old
the old (current) credentials

Description

Record the arguments userspace sent to sys_capset for later printing by the audit system if applicable

void audit_core_dumps(long signr)

record information about processes that end abnormally

Parameters

long signr
signal value

Description

If a process ends with a core dump, something fishy is going on and we should record the event for investigation.

void audit_seccomp(unsigned long syscall, long signr, int code)

record information about a seccomp action

Parameters

unsigned long syscall
syscall number
long signr
signal value
int code
the seccomp action

Description

Record the information associated with a seccomp action. Event filtering for seccomp actions that are not to be logged is done in seccomp_log(). Therefore, this function forces auditing independent of the audit_enabled and dummy context state because seccomp actions should be logged even when audit is not in use.

int audit_rule_change(int type, int seq, void * data, size_t datasz)

apply all rules to the specified message type

Parameters

int type
audit message type
int seq
netlink audit message sequence (serial) number
void * data
payload data
size_t datasz
size of payload data
int audit_list_rules_send(struct sk_buff * request_skb, int seq)

list the audit rules

Parameters

struct sk_buff * request_skb
skb of request we are replying to (used to target the reply)
int seq
netlink audit message sequence (serial) number
int parent_len(const char * path)

find the length of the parent portion of a pathname

Parameters

const char * path
pathname of which to determine length
int audit_compare_dname_path(const struct qstr * dname, const char * path, int parentlen)

compare given dentry name with last component in given path. Return of 0 indicates a match.

Parameters

const struct qstr * dname
dentry name that we’re comparing
const char * path
full pathname that we’re comparing
int parentlen
length of the parent if known. Passing in AUDIT_NAME_FULL here indicates that we must compute this value.

Accounting Framework

long sys_acct(const char __user * name)

enable/disable process accounting

Parameters

const char __user * name
file name for accounting records or NULL to shutdown accounting

Description

Returns 0 for success or negative errno values for failure.

sys_acct() is the only system call needed to implement process accounting. It takes the name of the file where accounting records should be written. If the filename is NULL, accounting will be shutdown.

void acct_collect(long exitcode, int group_dead)

collect accounting information into pacct_struct

Parameters

long exitcode
task exit code
int group_dead
not 0, if this thread is the last one in the process.
void acct_process(void)

Parameters

void
no arguments

Description

handles process accounting for an exiting task

Block Devices

void blk_queue_flag_set(unsigned int flag, struct request_queue * q)

atomically set a queue flag

Parameters

unsigned int flag
flag to be set
struct request_queue * q
request queue
void blk_queue_flag_clear(unsigned int flag, struct request_queue * q)

atomically clear a queue flag

Parameters

unsigned int flag
flag to be cleared
struct request_queue * q
request queue
bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue * q)

atomically test and set a queue flag

Parameters

unsigned int flag
flag to be set
struct request_queue * q
request queue

Description

Returns the previous value of flag - 0 if the flag was not set and 1 if the flag was already set.

const char * blk_op_str(unsigned int op)

Return string XXX in the REQ_OP_XXX.

Parameters

unsigned int op
REQ_OP_XXX.

Description

Centralize block layer function to convert REQ_OP_XXX into string format. Useful in the debugging and tracing bio or request. For invalid REQ_OP_XXX it returns string “UNKNOWN”.

void blk_sync_queue(struct request_queue * q)

cancel any pending callbacks on a queue

Parameters

struct request_queue * q
the queue

Description

The block layer may perform asynchronous callback activity on a queue, such as calling the unplug function after a timeout. A block device may call blk_sync_queue to ensure that any such activity is cancelled, thus allowing it to release resources that the callbacks might use. The caller must already have made sure that its ->make_request_fn will not re-add plugging prior to calling this function.

This function does not cancel any asynchronous activity arising out of elevator or throttling code. That would require elevator_exit() and blkcg_exit_queue() to be called with queue lock initialized.

void blk_set_pm_only(struct request_queue * q)

increment pm_only counter

Parameters

struct request_queue * q
request queue pointer
void blk_cleanup_queue(struct request_queue * q)

shutdown a request queue

Parameters

struct request_queue * q
request queue to shutdown

Description

Mark q DYING, drain all pending requests, mark q DEAD, destroy and put it. All future requests will be failed immediately with -ENODEV.

struct request_queue * blk_alloc_queue_node(gfp_t gfp_mask, int node_id)

allocate a request queue

Parameters

gfp_t gfp_mask
memory allocation flags
int node_id
NUMA node to allocate memory from
struct request * blk_get_request(struct request_queue * q, unsigned int op, blk_mq_req_flags_t flags)

allocate a request

Parameters

struct request_queue * q
request queue to allocate a request for
unsigned int op
operation (REQ_OP_*) and REQ_* flags, e.g. REQ_SYNC.
blk_mq_req_flags_t flags
BLK_MQ_REQ_* flags, e.g. BLK_MQ_REQ_NOWAIT.
blk_qc_t generic_make_request(struct bio * bio)

hand a buffer to its device driver for I/O

Parameters

struct bio * bio
The bio describing the location in memory and on the device.

Description

generic_make_request() is used to make I/O requests of block devices. It is passed a struct bio, which describes the I/O that needs to be done.

generic_make_request() does not return any status. The success/failure status of the request, along with notification of completion, is delivered asynchronously through the bio->bi_end_io function described (one day) else where.

The caller of generic_make_request must make sure that bi_io_vec are set to describe the memory buffer, and that bi_dev and bi_sector are set to describe the device address, and the bi_end_io and optionally bi_private are set to describe how completion notification should be signaled.

generic_make_request and the drivers it calls may use bi_next if this bio happens to be merged with someone else, and may resubmit the bio to a lower device by calling into generic_make_request recursively, which means the bio should NOT be touched after the call to ->make_request_fn.

blk_qc_t direct_make_request(struct bio * bio)

hand a buffer directly to its device driver for I/O

Parameters

struct bio * bio
The bio describing the location in memory and on the device.

Description

This function behaves like generic_make_request(), but does not protect against recursion. Must only be used if the called driver is known to not call generic_make_request (or direct_make_request) again from its make_request function. (Calling direct_make_request again from a workqueue is perfectly fine as that doesn’t recurse).

blk_qc_t submit_bio(struct bio * bio)

submit a bio to the block device layer for I/O

Parameters

struct bio * bio
The struct bio which describes the I/O

Description

submit_bio() is very similar in purpose to generic_make_request(), and uses that function to do most of the work. Both are fairly rough interfaces; bio must be presetup and ready for I/O.

blk_status_t blk_insert_cloned_request(struct request_queue * q, struct request * rq)

Helper for stacking drivers to submit a request

Parameters

struct request_queue * q
the queue to submit the request
struct request * rq
the request being queued
unsigned int blk_rq_err_bytes(const struct request * rq)

determine number of bytes till the next failure boundary

Parameters

const struct request * rq
request to examine

Description

A request could be merge of IOs which require different failure handling. This function determines the number of bytes which can be failed from the beginning of the request without crossing into area which need to be retried further.

Return

The number of bytes to fail.
bool blk_update_request(struct request * req, blk_status_t error, unsigned int nr_bytes)

Special helper function for request stacking drivers

Parameters

struct request * req
the request being processed
blk_status_t error
block status code
unsigned int nr_bytes
number of bytes to complete req

Description

Ends I/O on a number of bytes attached to req, but doesn’t complete the request structure even if req doesn’t have leftover. If req has leftover, sets it up for the next range of segments.

This special helper function is only for request stacking drivers (e.g. request-based dm) so that they can handle partial completion. Actual device drivers should use blk_mq_end_request instead.

Passing the result of blk_rq_bytes() as nr_bytes guarantees false return from this function.

Note

The RQF_SPECIAL_PAYLOAD flag is ignored on purpose in both blk_rq_bytes() and in blk_update_request().

Return

false - this request doesn’t have any more data true - this request has more data
void rq_flush_dcache_pages(struct request * rq)

Helper function to flush all pages in a request

Parameters

struct request * rq
the request to be flushed

Description

Flush all pages in rq.
int blk_lld_busy(struct request_queue * q)

Check if underlying low-level drivers of a device are busy

Parameters

struct request_queue * q
the queue of the device being checked

Description

Check if underlying low-level drivers of a device are busy. If the drivers want to export their busy state, they must set own exporting function using blk_queue_lld_busy() first.

Basically, this function is used only by request stacking drivers to stop dispatching requests to underlying devices when underlying devices are busy. This behavior helps more I/O merging on the queue of the request stacking driver and prevents I/O throughput regression on burst I/O load.

Return

0 - Not busy (The request stacking driver should dispatch request) 1 - Busy (The request stacking driver should stop dispatching request)
void blk_rq_unprep_clone(struct request * rq)

Helper function to free all bios in a cloned request

Parameters

struct request * rq
the clone request to be cleaned up

Description

Free all bios in rq for a cloned request.
int blk_rq_prep_clone(struct request * rq, struct request * rq_src, struct bio_set * bs, gfp_t gfp_mask, int (*bio_ctr) (struct bio *, struct bio *, void *, void * data)

Helper function to setup clone request

Parameters

struct request * rq
the request to be setup
struct request * rq_src
original request to be cloned
struct bio_set * bs
bio_set that bios for clone are allocated from
gfp_t gfp_mask
memory allocation mask for bio
int (*)(struct bio *, struct bio *, void *) bio_ctr
setup function to be called for each clone bio. Returns 0 for success, non 0 for failure.
void * data
private data to be passed to bio_ctr

Description

Clones bios in rq_src to rq, and copies attributes of rq_src to rq. The actual data parts of rq_src (e.g. ->cmd, ->sense) are not copied, and copying such parts is the caller’s responsibility. Also, pages which the original bios are pointing to are not copied and the cloned bios just point same pages. So cloned bios must be completed before original bios, which means the caller must complete rq before rq_src.
void blk_start_plug(struct blk_plug * plug)

initialize blk_plug and track it inside the task_struct

Parameters

struct blk_plug * plug
The struct blk_plug that needs to be initialized

Description

blk_start_plug() indicates to the block layer an intent by the caller to submit multiple I/O requests in a batch. The block layer may use this hint to defer submitting I/Os from the caller until blk_finish_plug() is called. However, the block layer may choose to submit requests before a call to blk_finish_plug() if the number of queued I/Os exceeds BLK_MAX_REQUEST_COUNT, or if the size of the I/O is larger than BLK_PLUG_FLUSH_SIZE. The queued I/Os may also be submitted early if the task schedules (see below).

Tracking blk_plug inside the task_struct will help with auto-flushing the pending I/O should the task end up blocking between blk_start_plug() and blk_finish_plug(). This is important from a performance perspective, but also ensures that we don’t deadlock. For instance, if the task is blocking for a memory allocation, memory reclaim could end up wanting to free a page belonging to that request that is currently residing in our private plug. By flushing the pending I/O when the process goes to sleep, we avoid this kind of deadlock.

void blk_finish_plug(struct blk_plug * plug)

mark the end of a batch of submitted I/O

Parameters

struct blk_plug * plug
The struct blk_plug passed to blk_start_plug()

Description

Indicate that a batch of I/O submissions is complete. This function must be paired with an initial call to blk_start_plug(). The intent is to allow the block layer to optimize I/O submission. See the documentation for blk_start_plug() for more information.

int blk_queue_enter(struct request_queue * q, blk_mq_req_flags_t flags)

try to increase q->q_usage_counter

Parameters

struct request_queue * q
request queue pointer
blk_mq_req_flags_t flags
BLK_MQ_REQ_NOWAIT and/or BLK_MQ_REQ_PREEMPT
bool blk_attempt_plug_merge(struct request_queue * q, struct bio * bio, unsigned int nr_segs, struct request ** same_queue_rq)

try to merge with current’s plugged list

Parameters

struct request_queue * q
request_queue new bio is being queued at
struct bio * bio
new bio being queued
unsigned int nr_segs
number of segments in bio
struct request ** same_queue_rq
pointer to struct request that gets filled in when another request associated with q is found on the plug list (optional, may be NULL)

Description

Determine whether bio being queued on q can be merged with a request on current’s plugged list. Returns true if merge was successful, otherwise false.

Plugging coalesces IOs from the same issuer for the same purpose without going through q->queue_lock. As such it’s more of an issuing mechanism than scheduling, and the request, while may have elvpriv data, is not added on the elevator at this point. In addition, we don’t have reliable access to the elevator outside queue lock. Only check basic merging parameters without querying the elevator.

Caller must ensure !blk_queue_nomerges(q) beforehand.

int blk_cloned_rq_check_limits(struct request_queue * q, struct request * rq)

Helper function to check a cloned request for new the queue limits

Parameters

struct request_queue * q
the queue
struct request * rq
the request being checked

Description

rq may have been made based on weaker limitations of upper-level queues in request stacking drivers, and it may violate the limitation of q. Since the block layer and the underlying device driver trust rq after it is inserted to q, it should be checked against q before the insertion using this generic function.

Request stacking drivers like request-based dm may change the queue limits when retrying requests on other queues. Those requests need to be checked against the new queue limits again during dispatch.

int blk_rq_map_user_iov(struct request_queue * q, struct request * rq, struct rq_map_data * map_data, const struct iov_iter * iter, gfp_t gfp_mask)

map user data to a request, for passthrough requests

Parameters

struct request_queue * q
request queue where request should be inserted
struct request * rq
request to map data to
struct rq_map_data * map_data
pointer to the rq_map_data holding pages (if necessary)
const struct iov_iter * iter
iovec iterator
gfp_t gfp_mask
memory allocation flags

Description

Data will be mapped directly for zero copy I/O, if possible. Otherwise a kernel bounce buffer is used.

A matching blk_rq_unmap_user() must be issued at the end of I/O, while still in process context.

Note

The mapped bio may need to be bounced through blk_queue_bounce()
before being submitted to the device, as pages mapped may be out of reach. It’s the callers responsibility to make sure this happens. The original bio must be passed back in to blk_rq_unmap_user() for proper unmapping.
int blk_rq_unmap_user(struct bio * bio)

unmap a request with user data

Parameters

struct bio * bio
start of bio list

Description

Unmap a rq previously mapped by blk_rq_map_user(). The caller must supply the original rq->bio from the blk_rq_map_user() return, since the I/O completion may have changed rq->bio.
int blk_rq_map_kern(struct request_queue * q, struct request * rq, void * kbuf, unsigned int len, gfp_t gfp_mask)

map kernel data to a request, for passthrough requests

Parameters

struct request_queue * q
request queue where request should be inserted
struct request * rq
request to fill
void * kbuf
the kernel buffer
unsigned int len
length of user data
gfp_t gfp_mask
memory allocation flags

Description

Data will be mapped directly if possible. Otherwise a bounce buffer is used. Can be called multiple times to append multiple buffers.
void __blk_release_queue(struct work_struct * work)

release a request queue

Parameters

struct work_struct * work
pointer to the release_work member of the request queue to be released

Description

This function is called when a block device is being unregistered. The process of releasing a request queue starts with blk_cleanup_queue, which set the appropriate flags and then calls blk_put_queue, that decrements the reference counter of the request queue. Once the reference counter of the request queue reaches zero, blk_release_queue is called to release all allocated resources of the request queue.
void blk_unregister_queue(struct gendisk * disk)

counterpart of blk_register_queue()

Parameters

struct gendisk * disk
Disk of which the request queue should be unregistered from sysfs.

Note

the caller is responsible for guaranteeing that this function is called after blk_register_queue() has finished.

void blk_set_default_limits(struct queue_limits * lim)

reset limits to default values

Parameters

struct queue_limits * lim
the queue_limits structure to reset

Description

Returns a queue_limit struct to its default state.
void blk_set_stacking_limits(struct queue_limits * lim)

set default limits for stacking devices

Parameters

struct queue_limits * lim
the queue_limits structure to reset

Description

Returns a queue_limit struct to its default state. Should be used by stacking drivers like DM that have no internal limits.
void blk_queue_make_request(struct request_queue * q, make_request_fn * mfn)

define an alternate make_request function for a device

Parameters

struct request_queue * q
the request queue for the device to be affected
make_request_fn * mfn
the alternate make_request function

Description

The normal way for struct bios to be passed to a device driver is for them to be collected into requests on a request queue, and then to allow the device driver to select requests off that queue when it is ready. This works well for many block devices. However some block devices (typically virtual devices such as md or lvm) do not benefit from the processing on the request queue, and are served best by having the requests passed directly to them. This can be achieved by providing a function to blk_queue_make_request().
Caveat:
The driver that does this must be able to deal appropriately with buffers in “highmemory”. This can be accomplished by either calling kmap_atomic() to get a temporary kernel mapping, or by calling blk_queue_bounce() to create a buffer in normal memory.
void blk_queue_bounce_limit(struct request_queue * q, u64 max_addr)

set bounce buffer limit for queue

Parameters

struct request_queue * q
the request queue for the device
u64 max_addr
the maximum address the device can handle

Description

Different hardware can have different requirements as to what pages it can do I/O directly to. A low level driver can call blk_queue_bounce_limit to have lower memory pages allocated as bounce buffers for doing I/O to pages residing above max_addr.
void blk_queue_max_hw_sectors(struct request_queue * q, unsigned int max_hw_sectors)

set max sectors for a request for this queue

Parameters

struct request_queue * q
the request queue for the device
unsigned int max_hw_sectors
max hardware sectors in the usual 512b unit

Description

Enables a low level driver to set a hard upper limit, max_hw_sectors, on the size of requests. max_hw_sectors is set by the device driver based upon the capabilities of the I/O controller.

max_dev_sectors is a hard limit imposed by the storage device for READ/WRITE requests. It is set by the disk driver.

max_sectors is a soft limit imposed by the block layer for filesystem type requests. This value can be overridden on a per-device basis in /sys/block/<device>/queue/max_sectors_kb. The soft limit can not exceed max_hw_sectors.

void blk_queue_chunk_sectors(struct request_queue * q, unsigned int chunk_sectors)

set size of the chunk for this queue

Parameters

struct request_queue * q
the request queue for the device
unsigned int chunk_sectors
chunk sectors in the usual 512b unit

Description

If a driver doesn’t want IOs to cross a given chunk size, it can set this limit and prevent merging across chunks. Note that the chunk size must currently be a power-of-2 in sectors. Also note that the block layer must accept a page worth of data at any offset. So if the crossing of chunks is a hard limitation in the driver, it must still be prepared to split single page bios.
void blk_queue_max_discard_sectors(struct request_queue * q, unsigned int max_discard_sectors)

set max sectors for a single discard

Parameters

struct request_queue * q
the request queue for the device
unsigned int max_discard_sectors
maximum number of sectors to discard
void blk_queue_max_write_same_sectors(struct request_queue * q, unsigned int max_write_same_sectors)

set max sectors for a single write same

Parameters

struct request_queue * q
the request queue for the device
unsigned int max_write_same_sectors
maximum number of sectors to write per command
void blk_queue_max_write_zeroes_sectors(struct request_queue * q, unsigned int max_write_zeroes_sectors)

set max sectors for a single write zeroes

Parameters

struct request_queue * q
the request queue for the device
unsigned int max_write_zeroes_sectors
maximum number of sectors to write per command
void blk_queue_max_segments(struct request_queue * q, unsigned short max_segments)

set max hw segments for a request for this queue

Parameters

struct request_queue * q
the request queue for the device
unsigned short max_segments
max number of segments

Description

Enables a low level driver to set an upper limit on the number of hw data segments in a request.
void blk_queue_max_discard_segments(struct request_queue * q, unsigned short max_segments)

set max segments for discard requests

Parameters

struct request_queue * q
the request queue for the device
unsigned short max_segments
max number of segments

Description

Enables a low level driver to set an upper limit on the number of segments in a discard request.
void blk_queue_max_segment_size(struct request_queue * q, unsigned int max_size)

set max segment size for blk_rq_map_sg

Parameters

struct request_queue * q
the request queue for the device
unsigned int max_size
max size of segment in bytes

Description

Enables a low level driver to set an upper limit on the size of a coalesced segment
void blk_queue_logical_block_size(struct request_queue * q, unsigned int size)

set logical block size for the queue

Parameters

struct request_queue * q
the request queue for the device
unsigned int size
the logical block size, in bytes

Description

This should be set to the lowest possible block size that the storage device can address. The default of 512 covers most hardware.
void blk_queue_physical_block_size(struct request_queue * q, unsigned int size)

set physical block size for the queue

Parameters

struct request_queue * q
the request queue for the device
unsigned int size
the physical block size, in bytes

Description

This should be set to the lowest possible sector size that the hardware can operate on without reverting to read-modify-write operations.
void blk_queue_alignment_offset(struct request_queue * q, unsigned int offset)

set physical block alignment offset

Parameters

struct request_queue * q
the request queue for the device
unsigned int offset
alignment offset in bytes

Description

Some devices are naturally misaligned to compensate for things like the legacy DOS partition table 63-sector offset. Low-level drivers should call this function for devices whose first sector is not naturally aligned.
void blk_limits_io_min(struct queue_limits * limits, unsigned int min)

set minimum request size for a device

Parameters

struct queue_limits * limits
the queue limits
unsigned int min
smallest I/O size in bytes

Description

Some devices have an internal block size bigger than the reported hardware sector size. This function can be used to signal the smallest I/O the device can perform without incurring a performance penalty.
void blk_queue_io_min(struct request_queue * q, unsigned int min)

set minimum request size for the queue

Parameters

struct request_queue * q
the request queue for the device
unsigned int min
smallest I/O size in bytes

Description

Storage devices may report a granularity or preferred minimum I/O size which is the smallest request the device can perform without incurring a performance penalty. For disk drives this is often the physical block size. For RAID arrays it is often the stripe chunk size. A properly aligned multiple of minimum_io_size is the preferred request size for workloads where a high number of I/O operations is desired.
void blk_limits_io_opt(struct queue_limits * limits, unsigned int opt)

set optimal request size for a device

Parameters

struct queue_limits * limits
the queue limits
unsigned int opt
smallest I/O size in bytes

Description

Storage devices may report an optimal I/O size, which is the device’s preferred unit for sustained I/O. This is rarely reported for disk drives. For RAID arrays it is usually the stripe width or the internal track size. A properly aligned multiple of optimal_io_size is the preferred request size for workloads where sustained throughput is desired.
void blk_queue_io_opt(struct request_queue * q, unsigned int opt)

set optimal request size for the queue

Parameters

struct request_queue * q
the request queue for the device
unsigned int opt
optimal request size in bytes

Description

Storage devices may report an optimal I/O size, which is the device’s preferred unit for sustained I/O. This is rarely reported for disk drives. For RAID arrays it is usually the stripe width or the internal track size. A properly aligned multiple of optimal_io_size is the preferred request size for workloads where sustained throughput is desired.
void blk_queue_stack_limits(struct request_queue * t, struct request_queue * b)

inherit underlying queue limits for stacked drivers

Parameters

struct request_queue * t
the stacking driver (top)
struct request_queue * b
the underlying device (bottom)
int blk_stack_limits(struct queue_limits * t, struct queue_limits * b, sector_t start)

adjust queue_limits for stacked devices

Parameters

struct queue_limits * t
the stacking driver limits (top device)
struct queue_limits * b
the underlying queue limits (bottom, component device)
sector_t start
first data sector within component device

Description

This function is used by stacking drivers like MD and DM to ensure that all component devices have compatible block sizes and alignments. The stacking driver must provide a queue_limits struct (top) and then iteratively call the stacking function for all component (bottom) devices. The stacking function will attempt to combine the values and ensure proper alignment.

Returns 0 if the top and bottom queue_limits are compatible. The top device’s block sizes and alignment offsets may be adjusted to ensure alignment with the bottom device. If no compatible sizes and alignments exist, -1 is returned and the resulting top queue_limits will have the misaligned flag set to indicate that the alignment_offset is undefined.

int bdev_stack_limits(struct queue_limits * t, struct block_device * bdev, sector_t start)

adjust queue limits for stacked drivers

Parameters

struct queue_limits * t
the stacking driver limits (top device)
struct block_device * bdev
the component block_device (bottom)
sector_t start
first data sector within component device

Description

Merges queue limits for a top device and a block_device. Returns 0 if alignment didn’t change. Returns -1 if adding the bottom device caused misalignment.
void disk_stack_limits(struct gendisk * disk, struct block_device * bdev, sector_t offset)

adjust queue limits for stacked drivers

Parameters

struct gendisk * disk
MD/DM gendisk (top)
struct block_device * bdev
the underlying block device (bottom)
sector_t offset
offset to beginning of data within component device

Description

Merges the limits for a top level gendisk and a bottom level block_device.
void blk_queue_update_dma_pad(struct request_queue * q, unsigned int mask)

update pad mask

Parameters

struct request_queue * q
the request queue for the device
unsigned int mask
pad mask

Description

Update dma pad mask.

Appending pad buffer to a request modifies the last entry of a scatter list such that it includes the pad buffer.

int blk_queue_dma_drain(struct request_queue * q, dma_drain_needed_fn * dma_drain_needed, void * buf, unsigned int size)

Set up a drain buffer for excess dma.

Parameters

struct request_queue * q
the request queue for the device
dma_drain_needed_fn * dma_drain_needed
fn which returns non-zero if drain is necessary
void * buf
physically contiguous buffer
unsigned int size
size of the buffer in bytes

Description

Some devices have excess DMA problems and can’t simply discard (or zero fill) the unwanted piece of the transfer. They have to have a real area of memory to transfer it into. The use case for this is ATAPI devices in DMA mode. If the packet command causes a transfer bigger than the transfer size some HBAs will lock up if there aren’t DMA elements to contain the excess transfer. What this API does is adjust the queue so that the buf is always appended silently to the scatterlist.

Note

This routine adjusts max_hw_segments to make room for appending the drain buffer. If you call blk_queue_max_segments() after calling this routine, you must set the limit to one fewer than your device can support otherwise there won’t be room for the drain buffer.

void blk_queue_segment_boundary(struct request_queue * q, unsigned long mask)

set boundary rules for segment merging

Parameters

struct request_queue * q
the request queue for the device
unsigned long mask
the memory boundary mask
void blk_queue_virt_boundary(struct request_queue * q, unsigned long mask)

set boundary rules for bio merging

Parameters

struct request_queue * q
the request queue for the device
unsigned long mask
the memory boundary mask
void blk_queue_dma_alignment(struct request_queue * q, int mask)

set dma length and memory alignment

Parameters

struct request_queue * q
the request queue for the device
int mask
alignment mask

Description

set required memory and length alignment for direct dma transactions. this is used when building direct io requests for the queue.
void blk_queue_update_dma_alignment(struct request_queue * q, int mask)

update dma length and memory alignment

Parameters

struct request_queue * q
the request queue for the device
int mask
alignment mask

Description

update required memory and length alignment for direct dma transactions. If the requested alignment is larger than the current alignment, then the current queue alignment is updated to the new value, otherwise it is left alone. The design of this is to allow multiple objects (driver, device, transport etc) to set their respective alignments without having them interfere.
void blk_set_queue_depth(struct request_queue * q, unsigned int depth)

tell the block layer about the device queue depth

Parameters

struct request_queue * q
the request queue for the device
unsigned int depth
queue depth
void blk_queue_write_cache(struct request_queue * q, bool wc, bool fua)

configure queue’s write cache

Parameters

struct request_queue * q
the request queue for the device
bool wc
write back cache on or off
bool fua
device supports FUA writes, if true

Description

Tell the block layer about the write cache of q.

void blk_queue_required_elevator_features(struct request_queue * q, unsigned int features)

Set a queue required elevator features

Parameters

struct request_queue * q
the request queue for the target device
unsigned int features
Required elevator features OR’ed together

Description

Tell the block layer that for the device controlled through q, only the only elevators that can be used are those that implement at least the set of features specified by features.

bool blk_queue_can_use_dma_map_merging(struct request_queue * q, struct device * dev)

configure queue for merging segments.

Parameters

struct request_queue * q
the request queue for the device
struct device * dev
the device pointer for dma

Description

Tell the block layer about merging the segments by dma map of q.

void blk_execute_rq_nowait(struct request_queue * q, struct gendisk * bd_disk, struct request * rq, int at_head, rq_end_io_fn * done)

insert a request into queue for execution

Parameters

struct request_queue * q
queue to insert the request in
struct gendisk * bd_disk
matching gendisk
struct request * rq
request to insert
int at_head
insert request at head or tail of queue
rq_end_io_fn * done
I/O completion handler

Description

Insert a fully prepared request at the back of the I/O scheduler queue for execution. Don’t wait for completion.

Note

This function will invoke done directly if the queue is dead.
void blk_execute_rq(struct request_queue * q, struct gendisk * bd_disk, struct request * rq, int at_head)

insert a request into queue for execution

Parameters

struct request_queue * q
queue to insert the request in
struct gendisk * bd_disk
matching gendisk
struct request * rq
request to insert
int at_head
insert request at head or tail of queue

Description

Insert a fully prepared request at the back of the I/O scheduler queue for execution and wait for completion.
int blkdev_issue_flush(struct block_device * bdev, gfp_t gfp_mask, sector_t * error_sector)

queue a flush

Parameters

struct block_device * bdev
blockdev to issue flush for
gfp_t gfp_mask
memory allocation flags (for bio_alloc)
sector_t * error_sector
error sector

Description

Issue a flush for the block device in question. Caller can supply room for storing the error offset in case of a flush error, if they wish to.
int blkdev_issue_discard(struct block_device * bdev, sector_t sector, sector_t nr_sects, gfp_t gfp_mask, unsigned long flags)

queue a discard

Parameters

struct block_device * bdev
blockdev to issue discard for
sector_t sector
start sector
sector_t nr_sects
number of sectors to discard
gfp_t gfp_mask
memory allocation flags (for bio_alloc)
unsigned long flags
BLKDEV_DISCARD_* flags to control behaviour

Description

Issue a discard request for the sectors in question.
int blkdev_issue_write_same(struct block_device * bdev, sector_t sector, sector_t nr_sects, gfp_t gfp_mask, struct page * page)

queue a write same operation

Parameters

struct block_device * bdev
target blockdev
sector_t sector
start sector
sector_t nr_sects
number of sectors to write
gfp_t gfp_mask
memory allocation flags (for bio_alloc)
struct page * page
page containing data

Description

Issue a write same request for the sectors in question.
int __blkdev_issue_zeroout(struct block_device * bdev, sector_t sector, sector_t nr_sects, gfp_t gfp_mask, struct bio ** biop, unsigned flags)

generate number of zero filed write bios

Parameters

struct block_device * bdev
blockdev to issue
sector_t sector
start sector
sector_t nr_sects
number of sectors to write
gfp_t gfp_mask
memory allocation flags (for bio_alloc)
struct bio ** biop
pointer to anchor bio
unsigned flags
controls detailed behavior

Description

Zero-fill a block range, either using hardware offload or by explicitly writing zeroes to the device.

If a device is using logical block provisioning, the underlying space will not be released if flags contains BLKDEV_ZERO_NOUNMAP.

If flags contains BLKDEV_ZERO_NOFALLBACK, the function will return -EOPNOTSUPP if no explicit hardware offload for zeroing is provided.

int blkdev_issue_zeroout(struct block_device * bdev, sector_t sector, sector_t nr_sects, gfp_t gfp_mask, unsigned flags)

zero-fill a block range

Parameters

struct block_device * bdev
blockdev to write
sector_t sector
start sector
sector_t nr_sects
number of sectors to write
gfp_t gfp_mask
memory allocation flags (for bio_alloc)
unsigned flags
controls detailed behavior

Description

Zero-fill a block range, either using hardware offload or by explicitly writing zeroes to the device. See __blkdev_issue_zeroout() for the valid values for flags.
int blk_rq_count_integrity_sg(struct request_queue * q, struct bio * bio)

Count number of integrity scatterlist elements

Parameters

struct request_queue * q
request queue
struct bio * bio
bio with integrity metadata attached

Description

Returns the number of elements required in a scatterlist corresponding to the integrity metadata in a bio.

int blk_rq_map_integrity_sg(struct request_queue * q, struct bio * bio, struct scatterlist * sglist)

Map integrity metadata into a scatterlist

Parameters

struct request_queue * q
request queue
struct bio * bio
bio with integrity metadata attached
struct scatterlist * sglist
target scatterlist

Description

Map the integrity vectors in request into a scatterlist. The scatterlist must be big enough to hold all elements. I.e. sized using blk_rq_count_integrity_sg().

int blk_integrity_compare(struct gendisk * gd1, struct gendisk * gd2)

Compare integrity profile of two disks

Parameters

struct gendisk * gd1
Disk to compare
struct gendisk * gd2
Disk to compare

Description

Meta-devices like DM and MD need to verify that all sub-devices use the same integrity format before advertising to upper layers that they can send/receive integrity metadata. This function can be used to check whether two gendisk devices have compatible integrity formats.

void blk_integrity_register(struct gendisk * disk, struct blk_integrity * template)

Register a gendisk as being integrity-capable

Parameters

struct gendisk * disk
struct gendisk pointer to make integrity-aware
struct blk_integrity * template
block integrity profile to register

Description

When a device needs to advertise itself as being able to send/receive integrity metadata it must use this function to register the capability with the block layer. The template is a blk_integrity struct with values appropriate for the underlying hardware. See Documentation/block/data-integrity.rst.

void blk_integrity_unregister(struct gendisk * disk)

Unregister block integrity profile

Parameters

struct gendisk * disk
disk whose integrity profile to unregister

Description

This function unregisters the integrity capability from a block device.

int blk_trace_ioctl(struct block_device * bdev, unsigned cmd, char __user * arg)

handle the ioctls associated with tracing

Parameters

struct block_device * bdev
the block device
unsigned cmd
the ioctl cmd
char __user * arg
the argument data, if any
void blk_trace_shutdown(struct request_queue * q)

stop and cleanup trace structures

Parameters

struct request_queue * q
the request queue associated with the device
void blk_add_trace_rq(struct request * rq, int error, unsigned int nr_bytes, u32 what, u64 cgid)

Add a trace for a request oriented action

Parameters

struct request * rq
the source request
int error
return status to log
unsigned int nr_bytes
number of completed bytes
u32 what
the action
u64 cgid
the cgroup info

Description

Records an action against a request. Will log the bio offset + size.
void blk_add_trace_bio(struct request_queue * q, struct bio * bio, u32 what, int error)

Add a trace for a bio oriented action

Parameters

struct request_queue * q
queue the io is for
struct bio * bio
the source bio
u32 what
the action
int error
error, if any

Description

Records an action against a bio. Will log the bio offset + size.
void blk_add_trace_bio_remap(void * ignore, struct request_queue * q, struct bio * bio, dev_t dev, sector_t from)

Add a trace for a bio-remap operation

Parameters

void * ignore
trace callback data parameter (not used)
struct request_queue * q
queue the io is for
struct bio * bio
the source bio
dev_t dev
target device
sector_t from
source sector

Description

Device mapper or raid target sometimes need to split a bio because it spans a stripe (or similar). Add a trace for that action.
void blk_add_trace_rq_remap(void * ignore, struct request_queue * q, struct request * rq, dev_t dev, sector_t from)

Add a trace for a request-remap operation

Parameters

void * ignore
trace callback data parameter (not used)
struct request_queue * q
queue the io is for
struct request * rq
the source request
dev_t dev
target device
sector_t from
source sector

Description

Device mapper remaps request to other devices. Add a trace for that action.
int blk_mangle_minor(int minor)

scatter minor numbers apart

Parameters

int minor
minor number to mangle

Description

Scatter consecutively allocated minor number apart if MANGLE_DEVT is enabled. Mangling twice gives the original value.

Return

Mangled value.

Context

Don’t care.

int blk_alloc_devt(struct hd_struct * part, dev_t * devt)

allocate a dev_t for a partition

Parameters

struct hd_struct * part
partition to allocate dev_t for
dev_t * devt
out parameter for resulting dev_t

Description

Allocate a dev_t for block device.

Return

0 on success, allocated dev_t is returned in *devt. -errno on failure.

Context

Might sleep.

void blk_free_devt(dev_t devt)

free a dev_t

Parameters

dev_t devt
dev_t to free

Description

Free devt which was allocated using blk_alloc_devt().

Context

Might sleep.

void __device_add_disk(struct device * parent, struct gendisk * disk, const struct attribute_group ** groups, bool register_queue)

add disk information to kernel list

Parameters

struct device * parent
parent device for the disk
struct gendisk * disk
per-device partitioning information
const struct attribute_group ** groups
Additional per-device sysfs groups
bool register_queue
register the queue if set to true

Description

This function registers the partitioning information in disk with the kernel.

FIXME: error handling

void disk_replace_part_tbl(struct gendisk * disk, struct disk_part_tbl * new_ptbl)

replace disk->part_tbl in RCU-safe way

Parameters

struct gendisk * disk
disk to replace part_tbl for
struct disk_part_tbl * new_ptbl
new part_tbl to install

Description

Replace disk->part_tbl with new_ptbl in RCU-safe way. The original ptbl is freed using RCU callback.

LOCKING: Matching bd_mutex locked or the caller is the only user of disk.

int disk_expand_part_tbl(struct gendisk * disk, int partno)

expand disk->part_tbl

Parameters

struct gendisk * disk
disk to expand part_tbl for
int partno
expand such that this partno can fit in

Description

Expand disk->part_tbl such that partno can fit in. disk->part_tbl uses RCU to allow unlocked dereferencing for stats and other stuff.

LOCKING: Matching bd_mutex locked or the caller is the only user of disk. Might sleep.

Return

0 on success, -errno on failure.

void disk_block_events(struct gendisk * disk)

block and flush disk event checking

Parameters

struct gendisk * disk
disk to block events for

Description

On return from this function, it is guaranteed that event checking isn’t in progress and won’t happen until unblocked by disk_unblock_events(). Events blocking is counted and the actual unblocking happens after the matching number of unblocks are done.

Note that this intentionally does not block event checking from disk_clear_events().

Context

Might sleep.

void disk_unblock_events(struct gendisk * disk)

unblock disk event checking

Parameters

struct gendisk * disk
disk to unblock events for

Description

Undo disk_block_events(). When the block count reaches zero, it starts events polling if configured.

Context

Don’t care. Safe to call from irq context.

void disk_flush_events(struct gendisk * disk, unsigned int mask)

schedule immediate event checking and flushing

Parameters

struct gendisk * disk
disk to check and flush events for
unsigned int mask
events to flush

Description

Schedule immediate event checking on disk if not blocked. Events in mask are scheduled to be cleared from the driver. Note that this doesn’t clear the events from disk->ev.

Context

If mask is non-zero must be called with bdev->bd_mutex held.

unsigned int disk_clear_events(struct gendisk * disk, unsigned int mask)

synchronously check, clear and return pending events

Parameters

struct gendisk * disk
disk to fetch and clear events from
unsigned int mask
mask of events to be fetched and cleared

Description

Disk events are synchronously checked and pending events in mask are cleared and returned. This ignores the block count.

Context

Might sleep.

struct hd_struct * disk_get_part(struct gendisk * disk, int partno)

get partition

Parameters

struct gendisk * disk
disk to look partition from
int partno
partition number

Description

Look for partition partno from disk. If found, increment reference count and return it.

Context

Don’t care.

Return

Pointer to the found partition on success, NULL if not found.

void disk_part_iter_init(struct disk_part_iter * piter, struct gendisk * disk, unsigned int flags)

initialize partition iterator

Parameters

struct disk_part_iter * piter
iterator to initialize
struct gendisk * disk
disk to iterate over
unsigned int flags
DISK_PITER_* flags

Description

Initialize piter so that it iterates over partitions of disk.

Context

Don’t care.

struct hd_struct * disk_part_iter_next(struct disk_part_iter * piter)

proceed iterator to the next partition and return it

Parameters

struct disk_part_iter * piter
iterator of interest

Description

Proceed piter to the next partition and return it.

Context

Don’t care.

void disk_part_iter_exit(struct disk_part_iter * piter)

finish up partition iteration

Parameters

struct disk_part_iter * piter
iter of interest

Description

Called when iteration is over. Cleans up piter.

Context

Don’t care.

struct hd_struct * disk_map_sector_rcu(struct gendisk * disk, sector_t sector)

map sector to partition

Parameters

struct gendisk * disk
gendisk of interest
sector_t sector
sector to map

Description

Find out which partition sector maps to on disk. This is primarily used for stats accounting.

Context

RCU read locked. The returned partition pointer is valid only while preemption is disabled.

Return

Found partition on success, part0 is returned if no partition matches

bool disk_has_partitions(struct gendisk * disk)

Parameters

struct gendisk * disk
gendisk of interest

Description

Walk through the partition table and check if valid partition exists.

Context

Don’t care.

Return

True if the gendisk has at least one valid non-zero size partition. Otherwise false.

int register_blkdev(unsigned int major, const char * name)

register a new block device

Parameters

unsigned int major
the requested major device number [1..BLKDEV_MAJOR_MAX-1]. If major = 0, try to allocate any unused major number.
const char * name
the name of the new block device as a zero terminated string

Description

The name must be unique within the system.

The return value depends on the major input parameter:

  • if a major device number was requested in range [1..BLKDEV_MAJOR_MAX-1] then the function returns zero on success, or a negative error code
  • if any unused major number was requested with major = 0 parameter then the return value is the allocated major number in range [1..BLKDEV_MAJOR_MAX-1] or a negative error code otherwise

See Documentation/admin-guide/devices.txt for the list of allocated major numbers.

struct gendisk * get_gendisk(dev_t devt, int * partno)

get partitioning information for a given device

Parameters

dev_t devt
device to get partitioning information for
int * partno
returned partition index

Description

This function gets the structure containing partitioning information for the given device devt.

struct block_device * bdget_disk(struct gendisk * disk, int partno)

do bdget() by gendisk and partition number

Parameters

struct gendisk * disk
gendisk of interest
int partno
partition number

Description

Find partition partno from disk, do bdget() on it.

Context

Don’t care.

Return

Resulting block_device on success, NULL on failure.

Char devices

int register_chrdev_region(dev_t from, unsigned count, const char * name)

register a range of device numbers

Parameters

dev_t from
the first in the desired range of device numbers; must include the major number.
unsigned count
the number of consecutive device numbers required
const char * name
the name of the device or driver.

Description

Return value is zero on success, a negative error code on failure.

int alloc_chrdev_region(dev_t * dev, unsigned baseminor, unsigned count, const char * name)

register a range of char device numbers

Parameters

dev_t * dev
output parameter for first assigned number
unsigned baseminor
first of the requested range of minor numbers
unsigned count
the number of minor numbers required
const char * name
the name of the associated device or driver

Description

Allocates a range of char device numbers. The major number will be chosen dynamically, and returned (along with the first minor number) in dev. Returns zero or a negative error code.

int __register_chrdev(unsigned int major, unsigned int baseminor, unsigned int count, const char * name, const struct file_operations * fops)

create and register a cdev occupying a range of minors

Parameters

unsigned int major
major device number or 0 for dynamic allocation
unsigned int baseminor
first of the requested range of minor numbers
unsigned int count
the number of minor numbers required
const char * name
name of this range of devices
const struct file_operations * fops
file operations associated with this devices

Description

If major == 0 this functions will dynamically allocate a major and return its number.

If major > 0 this function will attempt to reserve a device with the given major number and will return zero on success.

Returns a -ve errno on failure.

The name of this device has nothing to do with the name of the device in /dev. It only helps to keep track of the different owners of devices. If your module name has only one type of devices it’s ok to use e.g. the name of the module here.

void unregister_chrdev_region(dev_t from, unsigned count)

unregister a range of device numbers

Parameters

dev_t from
the first in the range of numbers to unregister
unsigned count
the number of device numbers to unregister

Description

This function will unregister a range of count device numbers, starting with from. The caller should normally be the one who allocated those numbers in the first place…

void __unregister_chrdev(unsigned int major, unsigned int baseminor, unsigned int count, const char * name)

unregister and destroy a cdev

Parameters

unsigned int major
major device number
unsigned int baseminor
first of the range of minor numbers
unsigned int count
the number of minor numbers this cdev is occupying
const char * name
name of this range of devices

Description

Unregister and destroy the cdev occupying the region described by major, baseminor and count. This function undoes what __register_chrdev() did.

int cdev_add(struct cdev * p, dev_t dev, unsigned count)

add a char device to the system

Parameters

struct cdev * p
the cdev structure for the device
dev_t dev
the first device number for which this device is responsible
unsigned count
the number of consecutive minor numbers corresponding to this device

Description

cdev_add() adds the device represented by p to the system, making it live immediately. A negative error code is returned on failure.

void cdev_set_parent(struct cdev * p, struct kobject * kobj)

set the parent kobject for a char device

Parameters

struct cdev * p
the cdev structure
struct kobject * kobj
the kobject to take a reference to

Description

cdev_set_parent() sets a parent kobject which will be referenced appropriately so the parent is not freed before the cdev. This should be called before cdev_add.

int cdev_device_add(struct cdev * cdev, struct device * dev)

add a char device and it’s corresponding struct device, linkink

Parameters

struct cdev * cdev
the cdev structure
struct device * dev
the device structure

Description

cdev_device_add() adds the char device represented by cdev to the system, just as cdev_add does. It then adds dev to the system using device_add The dev_t for the char device will be taken from the struct device which needs to be initialized first. This helper function correctly takes a reference to the parent device so the parent will not get released until all references to the cdev are released.

This helper uses dev->devt for the device number. If it is not set it will not add the cdev and it will be equivalent to device_add.

This function should be used whenever the struct cdev and the struct device are members of the same structure whose lifetime is managed by the struct device.

NOTE

Callers must assume that userspace was able to open the cdev and can call cdev fops callbacks at any time, even if this function fails.

void cdev_device_del(struct cdev * cdev, struct device * dev)

inverse of cdev_device_add

Parameters

struct cdev * cdev
the cdev structure
struct device * dev
the device structure

Description

cdev_device_del() is a helper function to call cdev_del and device_del. It should be used whenever cdev_device_add is used.

If dev->devt is not set it will not remove the cdev and will be equivalent to device_del.

NOTE

This guarantees that associated sysfs callbacks are not running or runnable, however any cdevs already open will remain and their fops will still be callable even after this function returns.

void cdev_del(struct cdev * p)

remove a cdev from the system

Parameters

struct cdev * p
the cdev structure to be removed

Description

cdev_del() removes p from the system, possibly freeing the structure itself.

NOTE

This guarantees that cdev device will no longer be able to be opened, however any cdevs already open will remain and their fops will still be callable even after cdev_del returns.

struct cdev * cdev_alloc(void)

allocate a cdev structure

Parameters

void
no arguments

Description

Allocates and returns a cdev structure, or NULL on failure.

void cdev_init(struct cdev * cdev, const struct file_operations * fops)

initialize a cdev structure

Parameters

struct cdev * cdev
the structure to initialize
const struct file_operations * fops
the file_operations for this device

Description

Initializes cdev, remembering fops, making it ready to add to the system with cdev_add().

Clock Framework

The clock framework defines programming interfaces to support software management of the system clock tree. This framework is widely used with System-On-Chip (SOC) platforms to support power management and various devices which may need custom clock rates. Note that these “clocks” don’t relate to timekeeping or real time clocks (RTCs), each of which have separate frameworks. These struct clk instances may be used to manage for example a 96 MHz signal that is used to shift bits into and out of peripherals or busses, or otherwise trigger synchronous state machine transitions in system hardware.

Power management is supported by explicit software clock gating: unused clocks are disabled, so the system doesn’t waste power changing the state of transistors that aren’t in active use. On some systems this may be backed by hardware clock gating, where clocks are gated without being disabled in software. Sections of chips that are powered but not clocked may be able to retain their last state. This low power state is often called a retention mode. This mode still incurs leakage currents, especially with finer circuit geometries, but for CMOS circuits power is mostly used by clocked state changes.

Power-aware drivers only enable their clocks when the device they manage is in active use. Also, system sleep states often differ according to which clock domains are active: while a “standby” state may allow wakeup from several active domains, a “mem” (suspend-to-RAM) state may require a more wholesale shutdown of clocks derived from higher speed PLLs and oscillators, limiting the number of possible wakeup event sources. A driver’s suspend method may need to be aware of system-specific clock constraints on the target sleep state.

Some platforms support programmable clock generators. These can be used by external chips of various kinds, such as other CPUs, multimedia codecs, and devices with strict requirements for interface clocking.

struct clk_notifier

associate a clk with a notifier

Definition

struct clk_notifier {
  struct clk                      *clk;
  struct srcu_notifier_head       notifier_head;
  struct list_head                node;
};

Members

clk
struct clk * to associate the notifier with
notifier_head
a blocking_notifier_head for this clk
node
linked list pointers

Description

A list of struct clk_notifier is maintained by the notifier code. An entry is created whenever code registers the first notifier on a particular clk. Future notifiers on that clk are added to the notifier_head.

struct clk_notifier_data

rate data to pass to the notifier callback

Definition

struct clk_notifier_data {
  struct clk              *clk;
  unsigned long           old_rate;
  unsigned long           new_rate;
};

Members

clk
struct clk * being changed
old_rate
previous rate of this clk
new_rate
new rate of this clk

Description

For a pre-notifier, old_rate is the clk’s rate before this rate change, and new_rate is what the rate will be in the future. For a post-notifier, old_rate and new_rate are both set to the clk’s current rate (this was done to optimize the implementation).

struct clk_bulk_data

Data used for bulk clk operations.

Definition

struct clk_bulk_data {
  const char              *id;
  struct clk              *clk;
};

Members

id
clock consumer ID
clk
struct clk * to store the associated clock

Description

The CLK APIs provide a series of clk_bulk_() API calls as a convenience to consumers which require multiple clks. This structure is used to manage data for these calls.

int clk_notifier_register(struct clk * clk, struct notifier_block * nb)

change notifier callback

Parameters

struct clk * clk
clock whose rate we are interested in
struct notifier_block * nb
notifier block with callback function pointer

Description

ProTip: debugging across notifier chains can be frustrating. Make sure that your notifier callback function prints a nice big warning in case of failure.

int clk_notifier_unregister(struct clk * clk, struct notifier_block * nb)

change notifier callback

Parameters

struct clk * clk
clock whose rate we are no longer interested in
struct notifier_block * nb
notifier block which will be unregistered
long clk_get_accuracy(struct clk * clk)

obtain the clock accuracy in ppb (parts per billion) for a clock source.

Parameters

struct clk * clk
clock source

Description

This gets the clock source accuracy expressed in ppb. A perfect clock returns 0.

int clk_set_phase(struct clk * clk, int degrees)

adjust the phase shift of a clock signal

Parameters

struct clk * clk
clock signal source
int degrees
number of degrees the signal is shifted

Description

Shifts the phase of a clock signal by the specified degrees. Returns 0 on success, -EERROR otherwise.

int clk_get_phase(struct clk * clk)

return the phase shift of a clock signal

Parameters

struct clk * clk
clock signal source

Description

Returns the phase shift of a clock node in degrees, otherwise returns -EERROR.

int clk_set_duty_cycle(struct clk * clk, unsigned int num, unsigned int den)

adjust the duty cycle ratio of a clock signal

Parameters

struct clk * clk
clock signal source
unsigned int num
numerator of the duty cycle ratio to be applied
unsigned int den
denominator of the duty cycle ratio to be applied

Description

Adjust the duty cycle of a clock signal by the specified ratio. Returns 0 on success, -EERROR otherwise.

int clk_get_scaled_duty_cycle(struct clk * clk, unsigned int scale)

return the duty cycle ratio of a clock signal

Parameters

struct clk * clk
clock signal source
unsigned int scale
scaling factor to be applied to represent the ratio as an integer

Description

Returns the duty cycle ratio multiplied by the scale provided, otherwise returns -EERROR.

bool clk_is_match(const struct clk * p, const struct clk * q)

check if two clk’s point to the same hardware clock

Parameters

const struct clk * p
clk compared against q
const struct clk * q
clk compared against p

Description

Returns true if the two struct clk pointers both point to the same hardware clock node. Put differently, returns true if p and q share the same struct clk_core object.

Returns false otherwise. Note that two NULL clks are treated as matching.

int clk_prepare(struct clk * clk)

prepare a clock source

Parameters

struct clk * clk
clock source

Description

This prepares the clock source for use.

Must not be called from within atomic context.

void clk_unprepare(struct clk * clk)

undo preparation of a clock source

Parameters

struct clk * clk
clock source

Description

This undoes a previously prepared clock. The caller must balance the number of prepare and unprepare calls.

Must not be called from within atomic context.

struct clk * clk_get(struct device * dev, const char * id)

lookup and obtain a reference to a clock producer.

Parameters

struct device * dev
device for clock “consumer”
const char * id
clock consumer ID

Description

Returns a struct clk corresponding to the clock producer, or valid IS_ERR() condition containing errno. The implementation uses dev and id to determine the clock consumer, and thereby the clock producer. (IOW, id may be identical strings, but clk_get may return different clock producers depending on dev.)

Drivers must assume that the clock source is not enabled.

clk_get should not be called from within interrupt context.

int clk_bulk_get(struct device * dev, int num_clks, struct clk_bulk_data * clks)

lookup and obtain a number of references to clock producer.

Parameters

struct device * dev
device for clock “consumer”
int num_clks
the number of clk_bulk_data
struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Description

This helper function allows drivers to get several clk consumers in one operation. If any of the clk cannot be acquired then any clks that were obtained will be freed before returning to the caller.

Returns 0 if all clocks specified in clk_bulk_data table are obtained successfully, or valid IS_ERR() condition containing errno. The implementation uses dev and clk_bulk_data.id to determine the clock consumer, and thereby the clock producer. The clock returned is stored in each clk_bulk_data.clk field.

Drivers must assume that the clock source is not enabled.

clk_bulk_get should not be called from within interrupt context.

int clk_bulk_get_all(struct device * dev, struct clk_bulk_data ** clks)

lookup and obtain all available references to clock producer.

Parameters

struct device * dev
device for clock “consumer”
struct clk_bulk_data ** clks
pointer to the clk_bulk_data table of consumer

Description

This helper function allows drivers to get all clk consumers in one operation. If any of the clk cannot be acquired then any clks that were obtained will be freed before returning to the caller.

Returns a positive value for the number of clocks obtained while the clock references are stored in the clk_bulk_data table in clks field. Returns 0 if there’re none and a negative value if something failed.

Drivers must assume that the clock source is not enabled.

clk_bulk_get should not be called from within interrupt context.

int clk_bulk_get_optional(struct device * dev, int num_clks, struct clk_bulk_data * clks)

lookup and obtain a number of references to clock producer

Parameters

struct device * dev
device for clock “consumer”
int num_clks
the number of clk_bulk_data
struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Description

Behaves the same as clk_bulk_get() except where there is no clock producer. In this case, instead of returning -ENOENT, the function returns 0 and NULL for a clk for which a clock producer could not be determined.

int devm_clk_bulk_get(struct device * dev, int num_clks, struct clk_bulk_data * clks)

managed get multiple clk consumers

Parameters

struct device * dev
device for clock “consumer”
int num_clks
the number of clk_bulk_data
struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Description

Return 0 on success, an errno on failure.

This helper function allows drivers to get several clk consumers in one operation with management, the clks will automatically be freed when the device is unbound.

int devm_clk_bulk_get_optional(struct device * dev, int num_clks, struct clk_bulk_data * clks)

managed get multiple optional consumer clocks

Parameters

struct device * dev
device for clock “consumer”
int num_clks
the number of clk_bulk_data
struct clk_bulk_data * clks
pointer to the clk_bulk_data table of consumer

Description

Behaves the same as devm_clk_bulk_get() except where there is no clock producer. In this case, instead of returning -ENOENT, the function returns NULL for given clk. It is assumed all clocks in clk_bulk_data are optional.

Returns 0 if all clocks specified in clk_bulk_data table are obtained successfully or for any clk there was no clk provider available, otherwise returns valid IS_ERR() condition containing errno. The implementation uses dev and clk_bulk_data.id to determine the clock consumer, and thereby the clock producer. The clock returned is stored in each clk_bulk_data.clk field.

Drivers must assume that the clock source is not enabled.

clk_bulk_get should not be called from within interrupt context.

int devm_clk_bulk_get_all(struct device * dev, struct clk_bulk_data ** clks)

managed get multiple clk consumers

Parameters

struct device * dev
device for clock “consumer”
struct clk_bulk_data ** clks
pointer to the clk_bulk_data table of consumer

Description

Returns a positive value for the number of clocks obtained while the clock references are stored in the clk_bulk_data table in clks field. Returns 0 if there’re none and a negative value if something failed.

This helper function allows drivers to get several clk consumers in one operation with management, the clks will automatically be freed when the device is unbound.

struct clk * devm_clk_get(struct device * dev, const char * id)

lookup and obtain a managed reference to a clock producer.

Parameters

struct device * dev
device for clock “consumer”
const char * id
clock consumer ID

Description

Returns a struct clk corresponding to the clock producer, or valid IS_ERR() condition containing errno. The implementation uses dev and id to determine the clock consumer, and thereby the clock producer. (IOW, id may be identical strings, but clk_get may return different clock producers depending on dev.)

Drivers must assume that the clock source is not enabled.

devm_clk_get should not be called from within interrupt context.

The clock will automatically be freed when the device is unbound from the bus.

struct clk * devm_clk_get_optional(struct device * dev, const char * id)

lookup and obtain a managed reference to an optional clock producer.

Parameters

struct device * dev
device for clock “consumer”
const char * id
clock consumer ID

Description

Behaves the same as devm_clk_get() except where there is no clock producer. In this case, instead of returning -ENOENT, the function returns NULL.

struct clk * devm_get_clk_from_child(struct device * dev, struct device_node * np, const char * con_id)

lookup and obtain a managed reference to a clock producer from child node.

Parameters

struct device * dev
device for clock “consumer”
struct device_node * np
pointer to clock consumer node
const char * con_id
clock consumer ID

Description

This function parses the clocks, and uses them to look up the struct clk from the registered list of clock providers by using np and con_id

The clock will automatically be freed when the device is unbound from the bus.

int clk_rate_exclusive_get(struct clk * clk)

get exclusivity over the rate control of a producer

Parameters

struct clk * clk
clock source

Description

This function allows drivers to get exclusive control over the rate of a provider. It prevents any other consumer to execute, even indirectly, opereation which could alter the rate of the provider or cause glitches

If exlusivity is claimed more than once on clock, even by the same driver, the rate effectively gets locked as exclusivity can’t be preempted.

Must not be called from within atomic context.

Returns success (0) or negative errno.

void clk_rate_exclusive_put(struct clk * clk)

release exclusivity over the rate control of a producer

Parameters

struct clk * clk
clock source

Description

This function allows drivers to release the exclusivity it previously got from clk_rate_exclusive_get()

The caller must balance the number of clk_rate_exclusive_get() and clk_rate_exclusive_put() calls.

Must not be called from within atomic context.

int clk_enable(struct clk * clk)

inform the system when the clock source should be running.

Parameters

struct clk * clk
clock source

Description

If the clock can not be enabled/disabled, this should return success.

May be called from atomic contexts.

Returns success (0) or negative errno.

int clk_bulk_enable(int num_clks, const struct clk_bulk_data * clks)

inform the system when the set of clks should be running.

Parameters

int num_clks
the number of clk_bulk_data
const struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Description

May be called from atomic contexts.

Returns success (0) or negative errno.

void clk_disable(struct clk * clk)

inform the system when the clock source is no longer required.

Parameters

struct clk * clk
clock source

Description

Inform the system that a clock source is no longer required by a driver and may be shut down.

May be called from atomic contexts.

Implementation detail: if the clock source is shared between multiple drivers, clk_enable() calls must be balanced by the same number of clk_disable() calls for the clock source to be disabled.

void clk_bulk_disable(int num_clks, const struct clk_bulk_data * clks)

inform the system when the set of clks is no longer required.

Parameters

int num_clks
the number of clk_bulk_data
const struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Description

Inform the system that a set of clks is no longer required by a driver and may be shut down.

May be called from atomic contexts.

Implementation detail: if the set of clks is shared between multiple drivers, clk_bulk_enable() calls must be balanced by the same number of clk_bulk_disable() calls for the clock source to be disabled.

unsigned long clk_get_rate(struct clk * clk)

obtain the current clock rate (in Hz) for a clock source. This is only valid once the clock source has been enabled.

Parameters

struct clk * clk
clock source
void clk_put(struct clk * clk)

“free” the clock source

Parameters

struct clk * clk
clock source

Note

drivers must ensure that all clk_enable calls made on this clock source are balanced by clk_disable calls prior to calling this function.

clk_put should not be called from within interrupt context.

void clk_bulk_put(int num_clks, struct clk_bulk_data * clks)

“free” the clock source

Parameters

int num_clks
the number of clk_bulk_data
struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Note

drivers must ensure that all clk_bulk_enable calls made on this clock source are balanced by clk_bulk_disable calls prior to calling this function.

clk_bulk_put should not be called from within interrupt context.

void clk_bulk_put_all(int num_clks, struct clk_bulk_data * clks)

“free” all the clock source

Parameters

int num_clks
the number of clk_bulk_data
struct clk_bulk_data * clks
the clk_bulk_data table of consumer

Note

drivers must ensure that all clk_bulk_enable calls made on this clock source are balanced by clk_bulk_disable calls prior to calling this function.

clk_bulk_put_all should not be called from within interrupt context.

void devm_clk_put(struct device * dev, struct clk * clk)

“free” a managed clock source

Parameters

struct device * dev
device used to acquire the clock
struct clk * clk
clock source acquired with devm_clk_get()

Note

drivers must ensure that all clk_enable calls made on this clock source are balanced by clk_disable calls prior to calling this function.

clk_put should not be called from within interrupt context.

long clk_round_rate(struct clk * clk, unsigned long rate)

adjust a rate to the exact rate a clock can provide

Parameters

struct clk * clk
clock source
unsigned long rate
desired clock rate in Hz

Description

This answers the question “if I were to pass rate to clk_set_rate(), what clock rate would I end up with?” without changing the hardware in any way. In other words:

rate = clk_round_rate(clk, r);

and:

clk_set_rate(clk, r); rate = clk_get_rate(clk);

are equivalent except the former does not modify the clock hardware in any way.

Returns rounded clock rate in Hz, or negative errno.

int clk_set_rate(struct clk * clk, unsigned long rate)

set the clock rate for a clock source

Parameters

struct clk * clk
clock source
unsigned long rate
desired clock rate in Hz

Description

Updating the rate starts at the top-most affected clock and then walks the tree down to the bottom-most clock that needs updating.

Returns success (0) or negative errno.

int clk_set_rate_exclusive(struct clk * clk, unsigned long rate)

set the clock rate and claim exclusivity over clock source

Parameters

struct clk * clk
clock source
unsigned long rate
desired clock rate in Hz

Description

This helper function allows drivers to atomically set the rate of a producer and claim exclusivity over the rate control of the producer.

It is essentially a combination of clk_set_rate() and clk_rate_exclusite_get(). Caller must balance this call with a call to clk_rate_exclusive_put()

Returns success (0) or negative errno.

bool clk_has_parent(struct clk * clk, struct clk * parent)

check if a clock is a possible parent for another

Parameters

struct clk * clk
clock source
struct clk * parent
parent clock source

Description

This function can be used in drivers that need to check that a clock can be the parent of another without actually changing the parent.

Returns true if parent is a possible parent for clk, false otherwise.

int clk_set_rate_range(struct clk * clk, unsigned long min, unsigned long max)

set a rate range for a clock source

Parameters

struct clk * clk
clock source
unsigned long min
desired minimum clock rate in Hz, inclusive
unsigned long max
desired maximum clock rate in Hz, inclusive

Description

Returns success (0) or negative errno.

int clk_set_min_rate(struct clk * clk, unsigned long rate)

set a minimum clock rate for a clock source

Parameters

struct clk * clk
clock source
unsigned long rate
desired minimum clock rate in Hz, inclusive

Description

Returns success (0) or negative errno.

int clk_set_max_rate(struct clk * clk, unsigned long rate)

set a maximum clock rate for a clock source

Parameters

struct clk * clk
clock source
unsigned long rate
desired maximum clock rate in Hz, inclusive

Description

Returns success (0) or negative errno.

int clk_set_parent(struct clk * clk, struct clk * parent)

set the parent clock source for this clock

Parameters

struct clk * clk
clock source
struct clk * parent
parent clock source

Description

Returns success (0) or negative errno.

struct clk * clk_get_parent(struct clk * clk)

get the parent clock source for this clock

Parameters

struct clk * clk
clock source

Description

Returns struct clk corresponding to parent clock source, or valid IS_ERR() condition containing errno.

struct clk * clk_get_sys(const char * dev_id, const char * con_id)

get a clock based upon the device name

Parameters

const char * dev_id
device name
const char * con_id
connection ID

Description

Returns a struct clk corresponding to the clock producer, or valid IS_ERR() condition containing errno. The implementation uses dev_id and con_id to determine the clock consumer, and thereby the clock producer. In contrast to clk_get() this function takes the device name instead of the device itself for identification.

Drivers must assume that the clock source is not enabled.

clk_get_sys should not be called from within interrupt context.

int clk_save_context(void)

save clock context for poweroff

Parameters

void
no arguments

Description

Saves the context of the clock register for powerstates in which the contents of the registers will be lost. Occurs deep within the suspend code so locking is not necessary.

void clk_restore_context(void)

restore clock context after poweroff

Parameters

void
no arguments

Description

This occurs with all clocks enabled. Occurs deep within the resume code so locking is not necessary.

struct clk * clk_get_optional(struct device * dev, const char * id)

lookup and obtain a reference to an optional clock producer.

Parameters

struct device * dev
device for clock “consumer”
const char * id
clock consumer ID

Description

Behaves the same as clk_get() except where there is no clock producer. In this case, instead of returning -ENOENT, the function returns NULL.

Synchronization Primitives

Read-Copy Update (RCU)

RCU_NONIDLE(a)

Indicate idle-loop code that needs RCU readers

Parameters

a
Code that RCU needs to pay attention to.

Description

RCU read-side critical sections are forbidden in the inner idle loop, that is, between the rcu_idle_enter() and the rcu_idle_exit() – RCU will happily ignore any such read-side critical sections. However, things like powertop need tracepoints in the inner idle loop.

This macro provides the way out: RCU_NONIDLE(do_something_with_RCU()) will tell RCU that it needs to pay attention, invoke its argument (in this example, calling the do_something_with_RCU() function), and then tell RCU to go back to ignoring this CPU. It is permissible to nest RCU_NONIDLE() wrappers, but not indefinitely (but the limit is on the order of a million or so, even on 32-bit systems). It is not legal to block within RCU_NONIDLE(), nor is it permissible to transfer control either into or out of RCU_NONIDLE()’s statement.

cond_resched_tasks_rcu_qs()

Report potential quiescent states to RCU

Parameters

Description

This macro resembles cond_resched(), except that it is defined to report potential quiescent states to RCU-tasks even if the cond_resched() machinery were to be shut off, as some advocate for PREEMPTION kernels.

RCU_LOCKDEP_WARN(c, s)

emit lockdep splat if specified condition is met

Parameters

c
condition to check
s
informative message
RCU_INITIALIZER(v)

statically initialize an RCU-protected global variable

Parameters

v
The value to statically initialize with.
rcu_assign_pointer(p, v)

assign to RCU-protected pointer

Parameters

p
pointer to assign to
v
value to assign (publish)

Description

Assigns the specified value to the specified RCU-protected pointer, ensuring that any concurrent RCU readers will see any prior initialization.

Inserts memory barriers on architectures that require them (which is most of them), and also prevents the compiler from reordering the code that initializes the structure after the pointer assignment. More importantly, this call documents which pointers will be dereferenced by RCU read-side code.

In some special cases, you may use RCU_INIT_POINTER() instead of rcu_assign_pointer(). RCU_INIT_POINTER() is a bit faster due to the fact that it does not constrain either the CPU or the compiler. That said, using RCU_INIT_POINTER() when you should have used rcu_assign_pointer() is a very bad thing that results in impossible-to-diagnose memory corruption. So please be careful. See the RCU_INIT_POINTER() comment header for details.

Note that rcu_assign_pointer() evaluates each of its arguments only once, appearances notwithstanding. One of the “extra” evaluations is in typeof() and the other visible only to sparse (__CHECKER__), neither of which actually execute the argument. As with most cpp macros, this execute-arguments-only-once property is important, so please be careful when making changes to rcu_assign_pointer() and the other macros that it invokes.

rcu_replace_pointer(rcu_ptr, ptr, c)

replace an RCU pointer, returning its old value

Parameters

rcu_ptr
RCU pointer, whose old value is returned
ptr
regular pointer
c
the lockdep conditions under which the dereference will take place

Description

Perform a replacement, where rcu_ptr is an RCU-annotated pointer and c is the lockdep argument that is passed to the rcu_dereference_protected() call used to read that pointer. The old value of rcu_ptr is returned, and rcu_ptr is set to ptr.

rcu_access_pointer(p)

fetch RCU pointer with no dereferencing

Parameters

p
The pointer to read

Description

Return the value of the specified RCU-protected pointer, but omit the lockdep checks for being in an RCU read-side critical section. This is useful when the value of this pointer is accessed, but the pointer is not dereferenced, for example, when testing an RCU-protected pointer against NULL. Although rcu_access_pointer() may also be used in cases where update-side locks prevent the value of the pointer from changing, you should instead use rcu_dereference_protected() for this use case.

It is also permissible to use rcu_access_pointer() when read-side access to the pointer was removed at least one grace period ago, as is the case in the context of the RCU callback that is freeing up the data, or after a synchronize_rcu() returns. This can be useful when tearing down multi-linked structures after a grace period has elapsed.

rcu_dereference_check(p, c)

rcu_dereference with debug checking

Parameters

p
The pointer to read, prior to dereferencing
c
The conditions under which the dereference will take place

Description

Do an rcu_dereference(), but check that the conditions under which the dereference will take place are correct. Typically the conditions indicate the various locking conditions that should be held at that point. The check should return true if the conditions are satisfied. An implicit check for being in an RCU read-side critical section (rcu_read_lock()) is included.

For example:

bar = rcu_dereference_check(foo->bar, lockdep_is_held(foo->lock));

could be used to indicate to lockdep that foo->bar may only be dereferenced if either rcu_read_lock() is held, or that the lock required to replace the bar struct at foo->bar is held.

Note that the list of conditions may also include indications of when a lock need not be held, for example during initialisation or destruction of the target struct:

bar = rcu_dereference_check(foo->bar, lockdep_is_held(foo->lock) ||
atomic_read(foo->usage) == 0);

Inserts memory barriers on architectures that require them (currently only the Alpha), prevents the compiler from refetching (and from merging fetches), and, more importantly, documents exactly which pointers are protected by RCU and checks that the pointer is annotated as __rcu.

rcu_dereference_bh_check(p, c)

rcu_dereference_bh with debug checking

Parameters

p
The pointer to read, prior to dereferencing
c
The conditions under which the dereference will take place

Description

This is the RCU-bh counterpart to rcu_dereference_check().

rcu_dereference_sched_check(p, c)

rcu_dereference_sched with debug checking

Parameters

p
The pointer to read, prior to dereferencing
c
The conditions under which the dereference will take place

Description

This is the RCU-sched counterpart to rcu_dereference_check().

rcu_dereference_protected(p, c)

fetch RCU pointer when updates prevented

Parameters

p
The pointer to read, prior to dereferencing
c
The conditions under which the dereference will take place

Description

Return the value of the specified RCU-protected pointer, but omit the READ_ONCE(). This is useful in cases where update-side locks prevent the value of the pointer from changing. Please note that this primitive does not prevent the compiler from repeating this reference or combining it with other references, so it should not be used without protection of appropriate locks.

This function is only for update-side use. Using this function when protected only by rcu_read_lock() will result in infrequent but very ugly failures.

rcu_dereference(p)

fetch RCU-protected pointer for dereferencing

Parameters

p
The pointer to read, prior to dereferencing

Description

This is a simple wrapper around rcu_dereference_check().

rcu_dereference_bh(p)

fetch an RCU-bh-protected pointer for dereferencing

Parameters

p
The pointer to read, prior to dereferencing

Description

Makes rcu_dereference_check() do the dirty work.

rcu_dereference_sched(p)

fetch RCU-sched-protected pointer for dereferencing

Parameters

p
The pointer to read, prior to dereferencing

Description

Makes rcu_dereference_check() do the dirty work.

rcu_pointer_handoff(p)

Hand off a pointer from RCU to other mechanism

Parameters

p
The pointer to hand off

Description

This is simply an identity function, but it documents where a pointer is handed off from RCU to some other synchronization mechanism, for example, reference counting or locking. In C11, it would map to kill_dependency(). It could be used as follows:

rcu_read_lock();
p = rcu_dereference(gp);
long_lived = is_long_lived(p);
if (long_lived) {
        if (!atomic_inc_not_zero(p->refcnt))
                long_lived = false;
        else
                p = rcu_pointer_handoff(p);
}
rcu_read_unlock();
void rcu_read_lock(void)

mark the beginning of an RCU read-side critical section

Parameters

void
no arguments

Description

When synchronize_rcu() is invoked on one CPU while other CPUs are within RCU read-side critical sections, then the synchronize_rcu() is guaranteed to block until after all the other CPUs exit their critical sections. Similarly, if call_rcu() is invoked on one CPU while other CPUs are within RCU read-side critical sections, invocation of the corresponding RCU callback is deferred until after the all the other CPUs exit their critical sections.

Note, however, that RCU callbacks are permitted to run concurrently with new RCU read-side critical sections. One way that this can happen is via the following sequence of events: (1) CPU 0 enters an RCU read-side critical section, (2) CPU 1 invokes call_rcu() to register an RCU callback, (3) CPU 0 exits the RCU read-side critical section, (4) CPU 2 enters a RCU read-side critical section, (5) the RCU callback is invoked. This is legal, because the RCU read-side critical section that was running concurrently with the call_rcu() (and which therefore might be referencing something that the corresponding RCU callback would free up) has completed before the corresponding RCU callback is invoked.

RCU read-side critical sections may be nested. Any deferred actions will be deferred until the outermost RCU read-side critical section completes.

You can avoid reading and understanding the next paragraph by following this rule: don’t put anything in an rcu_read_lock() RCU read-side critical section that would block in a !PREEMPTION kernel. But if you want the full story, read on!

In non-preemptible RCU implementations (pure TREE_RCU and TINY_RCU), it is illegal to block while in an RCU read-side critical section. In preemptible RCU implementations (PREEMPT_RCU) in CONFIG_PREEMPTION kernel builds, RCU read-side critical sections may be preempted, but explicit blocking is illegal. Finally, in preemptible RCU implementations in real-time (with -rt patchset) kernel builds, RCU read-side critical sections may be preempted and they may also block, but only when acquiring spinlocks that are subject to priority inheritance.

void rcu_read_unlock(void)

marks the end of an RCU read-side critical section.

Parameters

void
no arguments

Description

In most situations, rcu_read_unlock() is immune from deadlock. However, in kernels built with CONFIG_RCU_BOOST, rcu_read_unlock() is responsible for deboosting, which it does via rt_mutex_unlock(). Unfortunately, this function acquires the scheduler’s runqueue and priority-inheritance spinlocks. This means that deadlock could result if the caller of rcu_read_unlock() already holds one of these locks or any lock that is ever acquired while holding them.

That said, RCU readers are never priority boosted unless they were preempted. Therefore, one way to avoid deadlock is to make sure that preemption never happens within any RCU read-side critical section whose outermost rcu_read_unlock() is called with one of rt_mutex_unlock()’s locks held. Such preemption can be avoided in a number of ways, for example, by invoking preempt_disable() before critical section’s outermost rcu_read_lock().

Given that the set of locks acquired by rt_mutex_unlock() might change at any time, a somewhat more future-proofed approach is to make sure that that preemption never happens within any RCU read-side critical section whose outermost rcu_read_unlock() is called with irqs disabled. This approach relies on the fact that rt_mutex_unlock() currently only acquires irq-disabled locks.

The second of these two approaches is best in most situations, however, the first approach can also be useful, at least to those developers willing to keep abreast of the set of locks acquired by rt_mutex_unlock().

See rcu_read_lock() for more information.

void rcu_read_lock_bh(void)

mark the beginning of an RCU-bh critical section

Parameters

void
no arguments

Description

This is equivalent of rcu_read_lock(), but also disables softirqs. Note that anything else that disables softirqs can also serve as an RCU read-side critical section.

Note that rcu_read_lock_bh() and the matching rcu_read_unlock_bh() must occur in the same context, for example, it is illegal to invoke rcu_read_unlock_bh() from one task if the matching rcu_read_lock_bh() was invoked from some other task.

void rcu_read_lock_sched(void)

mark the beginning of a RCU-sched critical section

Parameters

void
no arguments

Description

This is equivalent of rcu_read_lock(), but disables preemption. Read-side critical sections can also be introduced by anything else that disables preemption, including local_irq_disable() and friends.

Note that rcu_read_lock_sched() and the matching rcu_read_unlock_sched() must occur in the same context, for example, it is illegal to invoke rcu_read_unlock_sched() from process context if the matching rcu_read_lock_sched() was invoked from an NMI handler.

RCU_INIT_POINTER(p, v)

initialize an RCU protected pointer

Parameters

p
The pointer to be initialized.
v
The value to initialized the pointer to.

Description

Initialize an RCU-protected pointer in special cases where readers do not need ordering constraints on the CPU or the compiler. These special cases are:

  1. This use of RCU_INIT_POINTER() is NULLing out the pointer or
  2. The caller has taken whatever steps are required to prevent RCU readers from concurrently accessing this pointer or
  3. The referenced data structure has already been exposed to readers either at compile time or via rcu_assign_pointer() and
    1. You have not made any reader-visible changes to this structure since then or
    2. It is OK for readers accessing this structure from its new location to see the old state of the structure. (For example, the changes were to statistical counters or to other state where exact synchronization is not required.)

Failure to follow these rules governing use of RCU_INIT_POINTER() will result in impossible-to-diagnose memory corruption. As in the structures will look OK in crash dumps, but any concurrent RCU readers might see pre-initialized values of the referenced data structure. So please be very careful how you use RCU_INIT_POINTER()!!!

If you are creating an RCU-protected linked structure that is accessed by a single external-to-structure RCU-protected pointer, then you may use RCU_INIT_POINTER() to initialize the internal RCU-protected pointers, but you must use rcu_assign_pointer() to initialize the external-to-structure pointer after you have completely initialized the reader-accessible portions of the linked structure.

Note that unlike rcu_assign_pointer(), RCU_INIT_POINTER() provides no ordering guarantees for either the CPU or the compiler.

RCU_POINTER_INITIALIZER(p, v)

statically initialize an RCU protected pointer

Parameters

p
The pointer to be initialized.
v
The value to initialized the pointer to.

Description

GCC-style initialization for an RCU-protected pointer in a structure field.

kfree_rcu(ptr, rhf)

kfree an object after a grace period.

Parameters

ptr
pointer to kfree
rhf
the name of the struct rcu_head within the type of ptr.

Description

Many rcu callbacks functions just call kfree() on the base structure. These functions are trivial, but their size adds up, and furthermore when they are used in a kernel module, that module must invoke the high-latency rcu_barrier() function at module-unload time.

The kfree_rcu() function handles this issue. Rather than encoding a function address in the embedded rcu_head structure, kfree_rcu() instead encodes the offset of the rcu_head structure within the base structure. Because the functions are not allowed in the low-order 4096 bytes of kernel virtual memory, offsets up to 4095 bytes can be accommodated. If the offset is larger than 4095 bytes, a compile-time error will be generated in __kfree_rcu(). If this error is triggered, you can either fall back to use of call_rcu() or rearrange the structure to position the rcu_head structure into the first 4096 bytes.

Note that the allowable offset might decrease in the future, for example, to allow something like kmem_cache_free_rcu().

The BUILD_BUG_ON check must not involve any function calls, hence the checks are done in macros here.

void rcu_head_init(struct rcu_head * rhp)

Initialize rcu_head for rcu_head_after_call_rcu()

Parameters

struct rcu_head * rhp
The rcu_head structure to initialize.

Description

If you intend to invoke rcu_head_after_call_rcu() to test whether a given rcu_head structure has already been passed to call_rcu(), then you must also invoke this rcu_head_init() function on it just after allocating that structure. Calls to this function must not race with calls to call_rcu(), rcu_head_after_call_rcu(), or callback invocation.

bool rcu_head_after_call_rcu(struct rcu_head * rhp, rcu_callback_t f)

Has this rcu_head been passed to call_rcu()?

Parameters

struct rcu_head * rhp
The rcu_head structure to test.
rcu_callback_t f
The function passed to call_rcu() along with rhp.

Description

Returns true if the rhp has been passed to call_rcu() with func, and false otherwise. Emits a warning in any other case, including the case where rhp has already been invoked after a grace period. Calls to this function must not race with callback invocation. One way to avoid such races is to enclose the call to rcu_head_after_call_rcu() in an RCU read-side critical section that includes a read-side fetch of the pointer to the structure containing rhp.

int rcu_is_cpu_rrupt_from_idle(void)

see if interrupted from idle

Parameters

void
no arguments

Description

If the current CPU is idle and running at a first-level (not nested) interrupt from idle, return true. The caller must have at least disabled preemption.

void rcu_idle_enter(void)

inform RCU that current CPU is entering idle

Parameters

void
no arguments

Description

Enter idle mode, in other words, -leave- the mode in which RCU read-side critical sections can occur. (Though RCU read-side critical sections can occur in irq handlers in idle, a possibility handled by irq_enter() and irq_exit().)

If you add or remove a call to rcu_idle_enter(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_user_enter(void)

inform RCU that we are resuming userspace.

Parameters

void
no arguments

Description

Enter RCU idle mode right before resuming userspace. No use of RCU is permitted between this call and rcu_user_exit(). This way the CPU doesn’t need to maintain the tick for RCU maintenance purposes when the CPU runs in userspace.

If you add or remove a call to rcu_user_enter(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_nmi_exit(void)

inform RCU of exit from NMI context

Parameters

void
no arguments

Description

If you add or remove a call to rcu_nmi_exit(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_irq_exit(void)

inform RCU that current CPU is exiting irq towards idle

Parameters

void
no arguments

Description

Exit from an interrupt handler, which might possibly result in entering idle mode, in other words, leaving the mode in which read-side critical sections can occur. The caller must have disabled interrupts.

This code assumes that the idle loop never does anything that might result in unbalanced calls to irq_enter() and irq_exit(). If your architecture’s idle loop violates this assumption, RCU will give you what you deserve, good and hard. But very infrequently and irreproducibly.

Use things like work queues to work around this limitation.

You have been warned.

If you add or remove a call to rcu_irq_exit(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_idle_exit(void)

inform RCU that current CPU is leaving idle

Parameters

void
no arguments

Description

Exit idle mode, in other words, -enter- the mode in which RCU read-side critical sections can occur.

If you add or remove a call to rcu_idle_exit(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_user_exit(void)

inform RCU that we are exiting userspace.

Parameters

void
no arguments

Description

Exit RCU idle mode while entering the kernel because it can run a RCU read side critical section anytime.

If you add or remove a call to rcu_user_exit(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_nmi_enter_common(bool irq)

inform RCU of entry to NMI context

Parameters

bool irq
Is this call from rcu_irq_enter?

Description

If the CPU was idle from RCU’s viewpoint, update rdp->dynticks and rdp->dynticks_nmi_nesting to let the RCU grace-period handling know that the CPU is active. This implementation permits nested NMIs, as long as the nesting level does not overflow an int. (You will probably run out of stack space first.)

If you add or remove a call to rcu_nmi_enter_common(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

void rcu_nmi_enter(void)

inform RCU of entry to NMI context

Parameters

void
no arguments
void rcu_irq_enter(void)

inform RCU that current CPU is entering irq away from idle

Parameters

void
no arguments

Description

Enter an interrupt handler, which might possibly result in exiting idle mode, in other words, entering the mode in which read-side critical sections can occur. The caller must have disabled interrupts.

Note that the Linux kernel is fully capable of entering an interrupt handler that it never exits, for example when doing upcalls to user mode! This code assumes that the idle loop never does upcalls to user mode. If your architecture’s idle loop does do upcalls to user mode (or does anything else that results in unbalanced calls to the irq_enter() and irq_exit() functions), RCU will give you what you deserve, good and hard. But very infrequently and irreproducibly.

Use things like work queues to work around this limitation.

You have been warned.

If you add or remove a call to rcu_irq_enter(), be sure to test with CONFIG_RCU_EQS_DEBUG=y.

bool notrace rcu_is_watching(void)

see if RCU thinks that the current CPU is not idle

Parameters

void
no arguments

Description

Return true if RCU is watching the running CPU, which means that this CPU can safely enter RCU read-side critical sections. In other words, if the current CPU is not in its idle loop or is in an interrupt or NMI handler, return true.

void call_rcu(struct rcu_head * head, rcu_callback_t func)

Queue an RCU callback for invocation after a grace period.

Parameters

struct rcu_head * head
structure to be used for queueing the RCU updates.
rcu_callback_t func
actual callback function to be invoked after the grace period

Description

The callback function will be invoked some time after a full grace period elapses, in other words after all pre-existing RCU read-side critical sections have completed. However, the callback function might well execute concurrently with RCU read-side critical sections that started after call_rcu() was invoked. RCU read-side critical sections are delimited by rcu_read_lock() and rcu_read_unlock(), and may be nested. In addition, regions of code across which interrupts, preemption, or softirqs have been disabled also serve as RCU read-side critical sections. This includes hardware interrupt handlers, softirq handlers, and NMI handlers.

Note that all CPUs must agree that the grace period extended beyond all pre-existing RCU read-side critical section. On systems with more than one CPU, this means that when “func()” is invoked, each CPU is guaranteed to have executed a full memory barrier since the end of its last RCU read-side critical section whose beginning preceded the call to call_rcu(). It also means that each CPU executing an RCU read-side critical section that continues beyond the start of “func()” must have executed a memory barrier after the call_rcu() but before the beginning of that RCU read-side critical section. Note that these guarantees include CPUs that are offline, idle, or executing in user mode, as well as CPUs that are executing in the kernel.

Furthermore, if CPU A invoked call_rcu() and CPU B invoked the resulting RCU callback function “func()”, then both CPU A and CPU B are guaranteed to execute a full memory barrier during the time interval between the call to call_rcu() and the invocation of “func()” – even if CPU A and CPU B are the same CPU (but again only if the system has more than one CPU).

struct kfree_rcu_cpu_work

single batch of kfree_rcu() requests

Definition

struct kfree_rcu_cpu_work {
  struct rcu_work rcu_work;
  struct rcu_head *head_free;
  struct kfree_rcu_cpu *krcp;
};

Members

rcu_work
Let queue_rcu_work() invoke workqueue handler after grace period
head_free
List of kfree_rcu() objects waiting for a grace period
krcp
Pointer to kfree_rcu_cpu structure
struct kfree_rcu_cpu

batch up kfree_rcu() requests for RCU grace period

Definition

struct kfree_rcu_cpu {
  struct rcu_head *head;
  struct kfree_rcu_cpu_work krw_arr[KFREE_N_BATCHES];
  spinlock_t lock;
  struct delayed_work monitor_work;
  bool monitor_todo;
  bool initialized;
};

Members

head
List of kfree_rcu() objects not yet waiting for a grace period
krw_arr
Array of batches of kfree_rcu() objects waiting for a grace period
lock
Synchronize access to this structure
monitor_work
Promote head to head_free after KFREE_DRAIN_JIFFIES
monitor_todo
Tracks whether a monitor_work delayed work is pending
initialized
The lock and rcu_work fields have been initialized

Description

This is a per-CPU structure. The reason that it is not included in the rcu_data structure is to permit this code to be extracted from the RCU files. Such extraction could allow further optimization of the interactions with the slab allocators.

void synchronize_rcu(void)

wait until a grace period has elapsed.

Parameters

void
no arguments

Description

Control will return to the caller some time after a full grace period has elapsed, in other words after all currently executing RCU read-side critical sections have completed. Note, however, that upon return from synchronize_rcu(), the caller might well be executing concurrently with new RCU read-side critical sections that began while synchronize_rcu() was waiting. RCU read-side critical sections are delimited by rcu_read_lock() and rcu_read_unlock(), and may be nested. In addition, regions of code across which interrupts, preemption, or softirqs have been disabled also serve as RCU read-side critical sections. This includes hardware interrupt handlers, softirq handlers, and NMI handlers.

Note that this guarantee implies further memory-ordering guarantees. On systems with more than one CPU, when synchronize_rcu() returns, each CPU is guaranteed to have executed a full memory barrier since the end of its last RCU read-side critical section whose beginning preceded the call to synchronize_rcu(). In addition, each CPU having an RCU read-side critical section that extends beyond the return from synchronize_rcu() is guaranteed to have executed a full memory barrier after the beginning of synchronize_rcu() and before the beginning of that RCU read-side critical section. Note that these guarantees include CPUs that are offline, idle, or executing in user mode, as well as CPUs that are executing in the kernel.

Furthermore, if CPU A invoked synchronize_rcu(), which returned to its caller on CPU B, then both CPU A and CPU B are guaranteed to have executed a full memory barrier during the execution of synchronize_rcu() – even if CPU A and CPU B are the same CPU (but again only if the system has more than one CPU).

unsigned long get_state_synchronize_rcu(void)

Snapshot current RCU state

Parameters

void
no arguments

Description

Returns a cookie that is used by a later call to cond_synchronize_rcu() to determine whether or not a full grace period has elapsed in the meantime.

void cond_synchronize_rcu(unsigned long oldstate)

Conditionally wait for an RCU grace period

Parameters

unsigned long oldstate
return value from earlier call to get_state_synchronize_rcu()

Description

If a full RCU grace period has elapsed since the earlier call to get_state_synchronize_rcu(), just return. Otherwise, invoke synchronize_rcu() to wait for a full grace period.

Yes, this function does not take counter wrap into account. But counter wrap is harmless. If the counter wraps, we have waited for more than 2 billion grace periods (and way more on a 64-bit system!), so waiting for one additional grace period should be just fine.

void rcu_barrier(void)

Wait until all in-flight call_rcu() callbacks complete.

Parameters

void
no arguments

Description

Note that this primitive does not necessarily wait for an RCU grace period to complete. For example, if there are no RCU callbacks queued anywhere in the system, then rcu_barrier() is within its rights to return immediately, without waiting for anything, much less an RCU grace period.

void synchronize_rcu_expedited(void)

Brute-force RCU grace period

Parameters

void
no arguments

Description

Wait for an RCU grace period, but expedite it. The basic idea is to IPI all non-idle non-nohz online CPUs. The IPI handler checks whether the CPU is in an RCU critical section, and if so, it sets a flag that causes the outermost rcu_read_unlock() to report the quiescent state for RCU-preempt or asks the scheduler for help for RCU-sched. On the other hand, if the CPU is not in an RCU read-side critical section, the IPI handler reports the quiescent state immediately.

Although this is a great improvement over previous expedited implementations, it is still unfriendly to real-time workloads, so is thus not recommended for any sort of common-case code. In fact, if you are using synchronize_rcu_expedited() in a loop, please restructure your code to batch your updates, and then use a single synchronize_rcu() instead.

This has the same semantics as (but is more brutal than) synchronize_rcu().

bool rcu_read_lock_held_common(bool * ret)

might we be in RCU-sched read-side critical section?

Parameters

bool * ret
Best guess answer if lockdep cannot be relied on

Description

Returns true if lockdep must be ignored, in which case *ret contains the best guess described below. Otherwise returns false, in which case *ret tells the caller nothing and the caller should instead consult lockdep.

If CONFIG_DEBUG_LOCK_ALLOC is selected, set *ret to nonzero iff in an RCU-sched read-side critical section. In absence of CONFIG_DEBUG_LOCK_ALLOC, this assumes we are in an RCU-sched read-side critical section unless it can prove otherwise. Note that disabling of preemption (including disabling irqs) counts as an RCU-sched read-side critical section. This is useful for debug checks in functions that required that they be called within an RCU-sched read-side critical section.

Check debug_lockdep_rcu_enabled() to prevent false positives during boot and while lockdep is disabled.

Note that if the CPU is in the idle loop from an RCU point of view (ie: that we are in the section between rcu_idle_enter() and rcu_idle_exit()) then rcu_read_lock_held() sets *ret to false even if the CPU did an rcu_read_lock(). The reason for this is that RCU ignores CPUs that are in such a section, considering these as in extended quiescent state, so such a CPU is effectively never in an RCU read-side critical section regardless of what RCU primitives it invokes. This state of affairs is required — we need to keep an RCU-free window in idle where the CPU may possibly enter into low power mode. This way we can notice an extended quiescent state to other CPUs that started a grace period. Otherwise we would delay any grace period as long as we run in the idle task.

Similarly, we avoid claiming an RCU read lock held if the current CPU is offline.

void rcu_expedite_gp(void)

Expedite future RCU grace periods

Parameters

void
no arguments

Description

After a call to this function, future calls to synchronize_rcu() and friends act as the corresponding synchronize_rcu_expedited() function had instead been called.

void rcu_unexpedite_gp(void)

Cancel prior rcu_expedite_gp() invocation

Parameters

void
no arguments

Description

Undo a prior call to rcu_expedite_gp(). If all prior calls to rcu_expedite_gp() are undone by a subsequent call to rcu_unexpedite_gp(), and if the rcu_expedited sysfs/boot parameter is not set, then all subsequent calls to synchronize_rcu() and friends will return to their normal non-expedited behavior.

int rcu_read_lock_held(void)

might we be in RCU read-side critical section?

Parameters

void
no arguments

Description

If CONFIG_DEBUG_LOCK_ALLOC is selected, returns nonzero iff in an RCU read-side critical section. In absence of CONFIG_DEBUG_LOCK_ALLOC, this assumes we are in an RCU read-side critical section unless it can prove otherwise. This is useful for debug checks in functions that require that they be called within an RCU read-side critical section.

Checks debug_lockdep_rcu_enabled() to prevent false positives during boot and while lockdep is disabled.

Note that rcu_read_lock() and the matching rcu_read_unlock() must occur in the same context, for example, it is illegal to invoke rcu_read_unlock() in process context if the matching rcu_read_lock() was invoked from within an irq handler.

Note that rcu_read_lock() is disallowed if the CPU is either idle or offline from an RCU perspective, so check for those as well.

int rcu_read_lock_bh_held(void)

might we be in RCU-bh read-side critical section?

Parameters

void
no arguments

Description

Check for bottom half being disabled, which covers both the CONFIG_PROVE_RCU and not cases. Note that if someone uses rcu_read_lock_bh(), but then later enables BH, lockdep (if enabled) will show the situation. This is useful for debug checks in functions that require that they be called within an RCU read-side critical section.

Check debug_lockdep_rcu_enabled() to prevent false positives during boot.

Note that rcu_read_lock_bh() is disallowed if the CPU is either idle or offline from an RCU perspective, so check for those as well.

void wakeme_after_rcu(struct rcu_head * head)

Callback function to awaken a task after grace period

Parameters

struct rcu_head * head
Pointer to rcu_head member within rcu_synchronize structure

Description

Awaken the corresponding task now that a grace period has elapsed.

void init_rcu_head_on_stack(struct rcu_head * head)

initialize on-stack rcu_head for debugobjects

Parameters

struct rcu_head * head
pointer to rcu_head structure to be initialized

Description

This function informs debugobjects of a new rcu_head structure that has been allocated as an auto variable on the stack. This function is not required for rcu_head structures that are statically defined or that are dynamically allocated on the heap. This function has no effect for !CONFIG_DEBUG_OBJECTS_RCU_HEAD kernel builds.

void destroy_rcu_head_on_stack(struct rcu_head * head)

destroy on-stack rcu_head for debugobjects

Parameters

struct rcu_head * head
pointer to rcu_head structure to be initialized

Description

This function informs debugobjects that an on-stack rcu_head structure is about to go out of scope. As with init_rcu_head_on_stack(), this function is not required for rcu_head structures that are statically defined or that are dynamically allocated on the heap. Also as with init_rcu_head_on_stack(), this function has no effect for !CONFIG_DEBUG_OBJECTS_RCU_HEAD kernel builds.

void call_rcu_tasks(struct rcu_head * rhp, rcu_callback_t func)

Queue an RCU for invocation task-based grace period

Parameters

struct rcu_head * rhp
structure to be used for queueing the RCU updates.
rcu_callback_t func
actual callback function to be invoked after the grace period

Description

The callback function will be invoked some time after a full grace period elapses, in other words after all currently executing RCU read-side critical sections have completed. call_rcu_tasks() assumes that the read-side critical sections end at a voluntary context switch (not a preemption!), cond_resched_rcu_qs(), entry into idle, or transition to usermode execution. As such, there are no read-side primitives analogous to rcu_read_lock() and rcu_read_unlock() because this primitive is intended to determine that all tasks have passed through a safe state, not so much for data-strcuture synchronization.

See the description of call_rcu() for more detailed information on memory ordering guarantees.

void synchronize_rcu_tasks(void)

wait until an rcu-tasks grace period has elapsed.

Parameters

void
no arguments

Description

Control will return to the caller some time after a full rcu-tasks grace period has elapsed, in other words after all currently executing rcu-tasks read-side critical sections have elapsed. These read-side critical sections are delimited by calls to schedule(), cond_resched_tasks_rcu_qs(), idle execution, userspace execution, calls to synchronize_rcu_tasks(), and (in theory, anyway) cond_resched().

This is a very specialized primitive, intended only for a few uses in tracing and other situations requiring manipulation of function preambles and profiling hooks. The synchronize_rcu_tasks() function is not (yet) intended for heavy use from multiple CPUs.

Note that this guarantee implies further memory-ordering guarantees. On systems with more than one CPU, when synchronize_rcu_tasks() returns, each CPU is guaranteed to have executed a full memory barrier since the end of its last RCU-tasks read-side critical section whose beginning preceded the call to synchronize_rcu_tasks(). In addition, each CPU having an RCU-tasks read-side critical section that extends beyond the return from synchronize_rcu_tasks() is guaranteed to have executed a full memory barrier after the beginning of synchronize_rcu_tasks() and before the beginning of that RCU-tasks read-side critical section. Note that these guarantees include CPUs that are offline, idle, or executing in user mode, as well as CPUs that are executing in the kernel.

Furthermore, if CPU A invoked synchronize_rcu_tasks(), which returned to its caller on CPU B, then both CPU A and CPU B are guaranteed to have executed a full memory barrier during the execution of synchronize_rcu_tasks() – even if CPU A and CPU B are the same CPU (but again only if the system has more than one CPU).

void rcu_barrier_tasks(void)

Wait for in-flight call_rcu_tasks() callbacks.

Parameters

void
no arguments

Description

Although the current implementation is guaranteed to wait, it is not obligated to, for example, if there are no pending callbacks.

int srcu_read_lock_held(const struct srcu_struct * ssp)

might we be in SRCU read-side critical section?

Parameters

const struct srcu_struct * ssp
The srcu_struct structure to check

Description

If CONFIG_DEBUG_LOCK_ALLOC is selected, returns nonzero iff in an SRCU read-side critical section. In absence of CONFIG_DEBUG_LOCK_ALLOC, this assumes we are in an SRCU read-side critical section unless it can prove otherwise.

Checks debug_lockdep_rcu_enabled() to prevent false positives during boot and while lockdep is disabled.

Note that SRCU is based on its own statemachine and it doesn’t relies on normal RCU, it can be called from the CPU which is in the idle loop from an RCU point of view or offline.

srcu_dereference_check(p, ssp, c)

fetch SRCU-protected pointer for later dereferencing

Parameters

p
the pointer to fetch and protect for later dereferencing
ssp
pointer to the srcu_struct, which is used to check that we really are in an SRCU read-side critical section.
c
condition to check for update-side use

Description

If PROVE_RCU is enabled, invoking this outside of an RCU read-side critical section will result in an RCU-lockdep splat, unless c evaluates to 1. The c argument will normally be a logical expression containing lockdep_is_held() calls.

srcu_dereference(p, ssp)

fetch SRCU-protected pointer for later dereferencing

Parameters

p
the pointer to fetch and protect for later dereferencing
ssp
pointer to the srcu_struct, which is used to check that we really are in an SRCU read-side critical section.

Description

Makes rcu_dereference_check() do the dirty work. If PROVE_RCU is enabled, invoking this outside of an RCU read-side critical section will result in an RCU-lockdep splat.

srcu_dereference_notrace(p, ssp)

no tracing and no lockdep calls from here

Parameters

p
the pointer to fetch and protect for later dereferencing
ssp
pointer to the srcu_struct, which is used to check that we really are in an SRCU read-side critical section.
int srcu_read_lock(struct srcu_struct * ssp)

register a new reader for an SRCU-protected structure.

Parameters

struct srcu_struct * ssp
srcu_struct in which to register the new reader.

Description

Enter an SRCU read-side critical section. Note that SRCU read-side critical sections may be nested. However, it is illegal to call anything that waits on an SRCU grace period for the same srcu_struct, whether directly or indirectly. Please note that one way to indirectly wait on an SRCU grace period is to acquire a mutex that is held elsewhere while calling synchronize_srcu() or synchronize_srcu_expedited().

Note that srcu_read_lock() and the matching srcu_read_unlock() must occur in the same context, for example, it is illegal to invoke srcu_read_unlock() in an irq handler if the matching srcu_read_lock() was invoked in process context.

void srcu_read_unlock(struct srcu_struct * ssp, int idx)

unregister a old reader from an SRCU-protected structure.

Parameters

struct srcu_struct * ssp
srcu_struct in which to unregister the old reader.
int idx
return value from corresponding srcu_read_lock().

Description

Exit an SRCU read-side critical section.

void smp_mb__after_srcu_read_unlock(void)

ensure full ordering after srcu_read_unlock

Parameters

void
no arguments

Description

Converts the preceding srcu_read_unlock into a two-way memory barrier.

Call this after srcu_read_unlock, to guarantee that all memory operations that occur after smp_mb__after_srcu_read_unlock will appear to happen after the preceding srcu_read_unlock.

int init_srcu_struct(struct srcu_struct * ssp)

initialize a sleep-RCU structure

Parameters

struct srcu_struct * ssp
structure to initialize.

Description

Must invoke this on a given srcu_struct before passing that srcu_struct to any other function. Each srcu_struct represents a separate domain of SRCU protection.

bool srcu_readers_active(struct srcu_struct * ssp)

returns true if there are readers. and false otherwise

Parameters

struct srcu_struct * ssp
which srcu_struct to count active readers (holding srcu_read_lock).

Description

Note that this is not an atomic primitive, and can therefore suffer severe errors when invoked on an active srcu_struct. That said, it can be useful as an error check at cleanup time.

void cleanup_srcu_struct(struct srcu_struct * ssp)

deconstruct a sleep-RCU structure

Parameters

struct srcu_struct * ssp
structure to clean up.

Description

Must invoke this after you are finished using a given srcu_struct that was initialized via init_srcu_struct(), else you leak memory.

void call_srcu(struct srcu_struct * ssp, struct rcu_head * rhp, rcu_callback_t func)

Queue a callback for invocation after an SRCU grace period

Parameters

struct srcu_struct * ssp
srcu_struct in queue the callback
struct rcu_head * rhp
structure to be used for queueing the SRCU callback.
rcu_callback_t func
function to be invoked after the SRCU grace period

Description

The callback function will be invoked some time after a full SRCU grace period elapses, in other words after all pre-existing SRCU read-side critical sections have completed. However, the callback function might well execute concurrently with other SRCU read-side critical sections that started after call_srcu() was invoked. SRCU read-side critical sections are delimited by srcu_read_lock() and srcu_read_unlock(), and may be nested.

The callback will be invoked from process context, but must nevertheless be fast and must not block.

void synchronize_srcu_expedited(struct srcu_struct * ssp)

Brute-force SRCU grace period

Parameters

struct srcu_struct * ssp
srcu_struct with which to synchronize.

Description

Wait for an SRCU grace period to elapse, but be more aggressive about spinning rather than blocking when waiting.

Note that synchronize_srcu_expedited() has the same deadlock and memory-ordering properties as does synchronize_srcu().

void synchronize_srcu(struct srcu_struct * ssp)

wait for prior SRCU read-side critical-section completion

Parameters

struct srcu_struct * ssp
srcu_struct with which to synchronize.

Description

Wait for the count to drain to zero of both indexes. To avoid the possible starvation of synchronize_srcu(), it waits for the count of the index=((->srcu_idx & 1) ^ 1) to drain to zero at first, and then flip the srcu_idx and wait for the count of the other index.

Can block; must be called from process context.

Note that it is illegal to call synchronize_srcu() from the corresponding SRCU read-side critical section; doing so will result in deadlock. However, it is perfectly legal to call synchronize_srcu() on one srcu_struct from some other srcu_struct’s read-side critical section, as long as the resulting graph of srcu_structs is acyclic.

There are memory-ordering constraints implied by synchronize_srcu(). On systems with more than one CPU, when synchronize_srcu() returns, each CPU is guaranteed to have executed a full memory barrier since the end of its last corresponding SRCU read-side critical section whose beginning preceded the call to synchronize_srcu(). In addition, each CPU having an SRCU read-side critical section that extends beyond the return from synchronize_srcu() is guaranteed to have executed a full memory barrier after the beginning of synchronize_srcu() and before the beginning of that SRCU read-side critical section. Note that these guarantees include CPUs that are offline, idle, or executing in user mode, as well as CPUs that are executing in the kernel.

Furthermore, if CPU A invoked synchronize_srcu(), which returned to its caller on CPU B, then both CPU A and CPU B are guaranteed to have executed a full memory barrier during the execution of synchronize_srcu(). This guarantee applies even if CPU A and CPU B are the same CPU, but again only if the system has more than one CPU.

Of course, these memory-ordering guarantees apply only when synchronize_srcu(), srcu_read_lock(), and srcu_read_unlock() are passed the same srcu_struct structure.

If SRCU is likely idle, expedite the first request. This semantic was provided by Classic SRCU, and is relied upon by its users, so TREE SRCU must also provide it. Note that detecting idleness is heuristic and subject to both false positives and negatives.

void srcu_barrier(struct srcu_struct * ssp)

Wait until all in-flight call_srcu() callbacks complete.

Parameters

struct srcu_struct * ssp
srcu_struct on which to wait for in-flight callbacks.
unsigned long srcu_batches_completed(struct srcu_struct * ssp)

return batches completed.

Parameters

struct srcu_struct * ssp
srcu_struct on which to report batch completion.

Description

Report the number of batches, correlated with, but not necessarily precisely the same as, the number of grace periods that have elapsed.

void hlist_bl_del_rcu(struct hlist_bl_node * n)

deletes entry from hash list without re-initialization

Parameters

struct hlist_bl_node * n
the element to delete from the hash list.

Note

hlist_bl_unhashed() on entry does not return true after this, the entry is in an undefined state. It is useful for RCU based lockfree traversal.

In particular, it means that we can not poison the forward pointers that may still be used for walking the hash list.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_bl_add_head_rcu() or hlist_bl_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_bl_for_each_entry().

void hlist_bl_add_head_rcu(struct hlist_bl_node * n, struct hlist_bl_head * h)

Parameters

struct hlist_bl_node * n
the element to add to the hash list.
struct hlist_bl_head * h
the list to add to.

Description

Adds the specified element to the specified hlist_bl, while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_bl_add_head_rcu() or hlist_bl_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_bl_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs. Regardless of the type of CPU, the list-traversal primitive must be guarded by rcu_read_lock().

hlist_bl_for_each_entry_rcu(tpos, pos, head, member)

iterate over rcu list of given type

Parameters

tpos
the type * to use as a loop cursor.
pos
the struct hlist_bl_node to use as a loop cursor.
head
the head for your list.
member
the name of the hlist_bl_node within the struct.
list_tail_rcu(head)

returns the prev pointer of the head of the list

Parameters

head
the head of the list

Note

This should only be used with the list header, and even then only if list_del() and similar primitives are not also used on the list header.

void list_add_rcu(struct list_head * new, struct list_head * head)

add a new entry to rcu-protected list

Parameters

struct list_head * new
new entry to be added
struct list_head * head
list head to add it after

Description

Insert a new entry after the specified head. This is good for implementing stacks.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as list_add_rcu() or list_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as list_for_each_entry_rcu().

void list_add_tail_rcu(struct list_head * new, struct list_head * head)

add a new entry to rcu-protected list

Parameters

struct list_head * new
new entry to be added
struct list_head * head
list head to add it before

Description

Insert a new entry before the specified head. This is useful for implementing queues.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as list_add_tail_rcu() or list_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as list_for_each_entry_rcu().

void list_del_rcu(struct list_head * entry)

deletes entry from list without re-initialization

Parameters

struct list_head * entry
the element to delete from the list.

Note

list_empty() on entry does not return true after this, the entry is in an undefined state. It is useful for RCU based lockfree traversal.

In particular, it means that we can not poison the forward pointers that may still be used for walking the list.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as list_del_rcu() or list_add_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as list_for_each_entry_rcu().

Note that the caller is not permitted to immediately free the newly deleted entry. Instead, either synchronize_rcu() or call_rcu() must be used to defer freeing until an RCU grace period has elapsed.

void hlist_del_init_rcu(struct hlist_node * n)

deletes entry from hash list with re-initialization

Parameters

struct hlist_node * n
the element to delete from the hash list.

Note

list_unhashed() on the node return true after this. It is useful for RCU based read lockfree traversal if the writer side must know if the list entry is still hashed or already unhashed.

In particular, it means that we can not poison the forward pointers that may still be used for walking the hash list and we can only zero the pprev pointer so list_unhashed() will return true after this.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_add_head_rcu() or hlist_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_for_each_entry_rcu().

void list_replace_rcu(struct list_head * old, struct list_head * new)

replace old entry by new one

Parameters

struct list_head * old
the element to be replaced
struct list_head * new
the new element to insert

Description

The old entry will be replaced with the new entry atomically.

Note

old should not be empty.

void __list_splice_init_rcu(struct list_head * list, struct list_head * prev, struct list_head * next, void (*sync) (void)

join an RCU-protected list into an existing list.

Parameters

struct list_head * list
the RCU-protected list to splice
struct list_head * prev
points to the last element of the existing list
struct list_head * next
points to the first element of the existing list
void (*)(void) sync
synchronize_rcu, synchronize_rcu_expedited, …

Description

The list pointed to by prev and next can be RCU-read traversed concurrently with this function.

Note that this function blocks.

Important note: the caller must take whatever action is necessary to prevent any other updates to the existing list. In principle, it is possible to modify the list as soon as sync() begins execution. If this sort of thing becomes necessary, an alternative version based on call_rcu() could be created. But only if -really- needed – there is no shortage of RCU API members.

void list_splice_init_rcu(struct list_head * list, struct list_head * head, void (*sync) (void)

splice an RCU-protected list into an existing list, designed for stacks.

Parameters

struct list_head * list
the RCU-protected list to splice
struct list_head * head
the place in the existing list to splice the first list into
void (*)(void) sync
synchronize_rcu, synchronize_rcu_expedited, …
void list_splice_tail_init_rcu(struct list_head * list, struct list_head * head, void (*sync) (void)

splice an RCU-protected list into an existing list, designed for queues.

Parameters

struct list_head * list
the RCU-protected list to splice
struct list_head * head
the place in the existing list to splice the first list into
void (*)(void) sync
synchronize_rcu, synchronize_rcu_expedited, …
list_entry_rcu(ptr, type, member)

get the struct for this entry

Parameters

ptr
the struct list_head pointer.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

This primitive may safely run concurrently with the _rcu list-mutation primitives such as list_add_rcu() as long as it’s guarded by rcu_read_lock().

list_first_or_null_rcu(ptr, type, member)

get the first element from a list

Parameters

ptr
the list head to take the element from.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

Note that if the list is empty, it returns NULL.

This primitive may safely run concurrently with the _rcu list-mutation primitives such as list_add_rcu() as long as it’s guarded by rcu_read_lock().

list_next_or_null_rcu(head, ptr, type, member)

get the first element from a list

Parameters

head
the head for the list.
ptr
the list head to take the next element from.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

Note that if the ptr is at the end of the list, NULL is returned.

This primitive may safely run concurrently with the _rcu list-mutation primitives such as list_add_rcu() as long as it’s guarded by rcu_read_lock().

list_for_each_entry_rcu(pos, head, member, cond)

iterate over rcu list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.
cond
optional lockdep expression if called from non-RCU protection.

Description

This list-traversal primitive may safely run concurrently with the _rcu list-mutation primitives such as list_add_rcu() as long as the traversal is guarded by rcu_read_lock().

list_entry_lockless(ptr, type, member)

get the struct for this entry

Parameters

ptr
the struct list_head pointer.
type
the type of the struct this is embedded in.
member
the name of the list_head within the struct.

Description

This primitive may safely run concurrently with the _rcu list-mutation primitives such as list_add_rcu(), but requires some implicit RCU read-side guarding. One example is running within a special exception-time environment where preemption is disabled and where lockdep cannot be invoked. Another example is when items are added to the list, but never deleted.

list_for_each_entry_lockless(pos, head, member)

iterate over rcu list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_struct within the struct.

Description

This primitive may safely run concurrently with the _rcu list-mutation primitives such as list_add_rcu(), but requires some implicit RCU read-side guarding. One example is running within a special exception-time environment where preemption is disabled and where lockdep cannot be invoked. Another example is when items are added to the list, but never deleted.

list_for_each_entry_continue_rcu(pos, head, member)

continue iteration over list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_head within the struct.

Description

Continue to iterate over list of given type, continuing after the current position which must have been in the list when the RCU read lock was taken. This would typically require either that you obtained the node from a previous walk of the list in the same RCU read-side critical section, or that you held some sort of non-RCU reference (such as a reference count) to keep the node alive and in the list.

This iterator is similar to list_for_each_entry_from_rcu() except this starts after the given position and that one starts at the given position.

list_for_each_entry_from_rcu(pos, head, member)

iterate over a list from current point

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the list_node within the struct.

Description

Iterate over the tail of a list starting from a given position, which must have been in the list when the RCU read lock was taken. This would typically require either that you obtained the node from a previous walk of the list in the same RCU read-side critical section, or that you held some sort of non-RCU reference (such as a reference count) to keep the node alive and in the list.

This iterator is similar to list_for_each_entry_continue_rcu() except this starts from the given position and that one starts from the position after the given position.

void hlist_del_rcu(struct hlist_node * n)

deletes entry from hash list without re-initialization

Parameters

struct hlist_node * n
the element to delete from the hash list.

Note

list_unhashed() on entry does not return true after this, the entry is in an undefined state. It is useful for RCU based lockfree traversal.

In particular, it means that we can not poison the forward pointers that may still be used for walking the hash list.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_add_head_rcu() or hlist_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_for_each_entry().

void hlist_replace_rcu(struct hlist_node * old, struct hlist_node * new)

replace old entry by new one

Parameters

struct hlist_node * old
the element to be replaced
struct hlist_node * new
the new element to insert

Description

The old entry will be replaced with the new entry atomically.

void hlist_add_head_rcu(struct hlist_node * n, struct hlist_head * h)

Parameters

struct hlist_node * n
the element to add to the hash list.
struct hlist_head * h
the list to add to.

Description

Adds the specified element to the specified hlist, while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_add_head_rcu() or hlist_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs. Regardless of the type of CPU, the list-traversal primitive must be guarded by rcu_read_lock().

void hlist_add_tail_rcu(struct hlist_node * n, struct hlist_head * h)

Parameters

struct hlist_node * n
the element to add to the hash list.
struct hlist_head * h
the list to add to.

Description

Adds the specified element to the specified hlist, while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_add_head_rcu() or hlist_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs. Regardless of the type of CPU, the list-traversal primitive must be guarded by rcu_read_lock().

void hlist_add_before_rcu(struct hlist_node * n, struct hlist_node * next)

Parameters

struct hlist_node * n
the new element to add to the hash list.
struct hlist_node * next
the existing element to add the new element before.

Description

Adds the specified element to the specified hlist before the specified node while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_add_head_rcu() or hlist_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs.

void hlist_add_behind_rcu(struct hlist_node * n, struct hlist_node * prev)

Parameters

struct hlist_node * n
the new element to add to the hash list.
struct hlist_node * prev
the existing element to add the new element after.

Description

Adds the specified element to the specified hlist after the specified node while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_add_head_rcu() or hlist_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs.

hlist_for_each_entry_rcu(pos, head, member, cond)

iterate over rcu list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the hlist_node within the struct.
cond
optional lockdep expression if called from non-RCU protection.

Description

This list-traversal primitive may safely run concurrently with the _rcu list-mutation primitives such as hlist_add_head_rcu() as long as the traversal is guarded by rcu_read_lock().

hlist_for_each_entry_rcu_notrace(pos, head, member)

iterate over rcu list of given type (for tracing)

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the hlist_node within the struct.

Description

This list-traversal primitive may safely run concurrently with the _rcu list-mutation primitives such as hlist_add_head_rcu() as long as the traversal is guarded by rcu_read_lock().

This is the same as hlist_for_each_entry_rcu() except that it does not do any RCU debugging or tracing.

hlist_for_each_entry_rcu_bh(pos, head, member)

iterate over rcu list of given type

Parameters

pos
the type * to use as a loop cursor.
head
the head for your list.
member
the name of the hlist_node within the struct.

Description

This list-traversal primitive may safely run concurrently with the _rcu list-mutation primitives such as hlist_add_head_rcu() as long as the traversal is guarded by rcu_read_lock().

hlist_for_each_entry_continue_rcu(pos, member)

iterate over a hlist continuing after current point

Parameters

pos
the type * to use as a loop cursor.
member
the name of the hlist_node within the struct.
hlist_for_each_entry_continue_rcu_bh(pos, member)

iterate over a hlist continuing after current point

Parameters

pos
the type * to use as a loop cursor.
member
the name of the hlist_node within the struct.
hlist_for_each_entry_from_rcu(pos, member)

iterate over a hlist continuing from current point

Parameters

pos
the type * to use as a loop cursor.
member
the name of the hlist_node within the struct.
void hlist_nulls_del_init_rcu(struct hlist_nulls_node * n)

deletes entry from hash list with re-initialization

Parameters

struct hlist_nulls_node * n
the element to delete from the hash list.

Note

hlist_nulls_unhashed() on the node return true after this. It is useful for RCU based read lockfree traversal if the writer side must know if the list entry is still hashed or already unhashed.

In particular, it means that we can not poison the forward pointers that may still be used for walking the hash list and we can only zero the pprev pointer so list_unhashed() will return true after this.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_nulls_add_head_rcu() or hlist_nulls_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_nulls_for_each_entry_rcu().

hlist_nulls_first_rcu(head)

returns the first element of the hash list.

Parameters

head
the head of the list.
hlist_nulls_next_rcu(node)

returns the element of the list after node.

Parameters

node
element of the list.
void hlist_nulls_del_rcu(struct hlist_nulls_node * n)

deletes entry from hash list without re-initialization

Parameters

struct hlist_nulls_node * n
the element to delete from the hash list.

Note

hlist_nulls_unhashed() on entry does not return true after this, the entry is in an undefined state. It is useful for RCU based lockfree traversal.

In particular, it means that we can not poison the forward pointers that may still be used for walking the hash list.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_nulls_add_head_rcu() or hlist_nulls_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_nulls_for_each_entry().

void hlist_nulls_add_head_rcu(struct hlist_nulls_node * n, struct hlist_nulls_head * h)

Parameters

struct hlist_nulls_node * n
the element to add to the hash list.
struct hlist_nulls_head * h
the list to add to.

Description

Adds the specified element to the specified hlist_nulls, while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_nulls_add_head_rcu() or hlist_nulls_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_nulls_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs. Regardless of the type of CPU, the list-traversal primitive must be guarded by rcu_read_lock().

void hlist_nulls_add_tail_rcu(struct hlist_nulls_node * n, struct hlist_nulls_head * h)

Parameters

struct hlist_nulls_node * n
the element to add to the hash list.
struct hlist_nulls_head * h
the list to add to.

Description

Adds the specified element to the specified hlist_nulls, while permitting racing traversals.

The caller must take whatever precautions are necessary (such as holding appropriate locks) to avoid racing with another list-mutation primitive, such as hlist_nulls_add_head_rcu() or hlist_nulls_del_rcu(), running on this same list. However, it is perfectly legal to run concurrently with the _rcu list-traversal primitives, such as hlist_nulls_for_each_entry_rcu(), used to prevent memory-consistency problems on Alpha CPUs. Regardless of the type of CPU, the list-traversal primitive must be guarded by rcu_read_lock().

hlist_nulls_for_each_entry_rcu(tpos, pos, head, member)

iterate over rcu list of given type

Parameters

tpos
the type * to use as a loop cursor.
pos
the struct hlist_nulls_node to use as a loop cursor.
head
the head of the list.
member
the name of the hlist_nulls_node within the struct.

Description

The barrier() is needed to make sure compiler doesn’t cache first element [1], as this loop can be restarted [2] [1] Documentation/core-api/atomic_ops.rst around line 114 [2] Documentation/RCU/rculist_nulls.txt around line 146

hlist_nulls_for_each_entry_safe(tpos, pos, head, member)

iterate over list of given type safe against removal of list entry

Parameters

tpos
the type * to use as a loop cursor.
pos
the struct hlist_nulls_node to use as a loop cursor.
head
the head of the list.
member
the name of the hlist_nulls_node within the struct.
bool rcu_sync_is_idle(struct rcu_sync * rsp)

Are readers permitted to use their fastpaths?

Parameters

struct rcu_sync * rsp
Pointer to rcu_sync structure to use for synchronization

Description

Returns true if readers are permitted to use their fastpaths. Must be invoked within some flavor of RCU read-side critical section.

void rcu_sync_init(struct rcu_sync * rsp)

Initialize an rcu_sync structure

Parameters

struct rcu_sync * rsp
Pointer to rcu_sync structure to be initialized
void rcu_sync_enter_start(struct rcu_sync * rsp)

Force readers onto slow path for multiple updates

Parameters

struct rcu_sync * rsp
Pointer to rcu_sync structure to use for synchronization

Description

Must be called after rcu_sync_init() and before first use.

Ensures rcu_sync_is_idle() returns false and rcu_sync_{enter,exit}() pairs turn into NO-OPs.

void rcu_sync_func(struct rcu_head * rhp)

Callback function managing reader access to fastpath

Parameters

struct rcu_head * rhp
Pointer to rcu_head in rcu_sync structure to use for synchronization

Description

This function is passed to call_rcu() function by rcu_sync_enter() and rcu_sync_exit(), so that it is invoked after a grace period following the that invocation of enter/exit.

If it is called by rcu_sync_enter() it signals that all the readers were switched onto slow path.

If it is called by rcu_sync_exit() it takes action based on events that have taken place in the meantime, so that closely spaced rcu_sync_enter() and rcu_sync_exit() pairs need not wait for a grace period.

If another rcu_sync_enter() is invoked before the grace period ended, reset state to allow the next rcu_sync_exit() to let the readers back onto their fastpaths (after a grace period). If both another rcu_sync_enter() and its matching rcu_sync_exit() are invoked before the grace period ended, re-invoke call_rcu() on behalf of that rcu_sync_exit(). Otherwise, set all state back to idle so that readers can again use their fastpaths.

void rcu_sync_enter(struct rcu_sync * rsp)

Force readers onto slowpath

Parameters

struct rcu_sync * rsp
Pointer to rcu_sync structure to use for synchronization

Description

This function is used by updaters who need readers to make use of a slowpath during the update. After this function returns, all subsequent calls to rcu_sync_is_idle() will return false, which tells readers to stay off their fastpaths. A later call to rcu_sync_exit() re-enables reader slowpaths.

When called in isolation, rcu_sync_enter() must wait for a grace period, however, closely spaced calls to rcu_sync_enter() can optimize away the grace-period wait via a state machine implemented by rcu_sync_enter(), rcu_sync_exit(), and rcu_sync_func().

void rcu_sync_exit(struct rcu_sync * rsp)

Allow readers back onto fast path after grace period

Parameters

struct rcu_sync * rsp
Pointer to rcu_sync structure to use for synchronization

Description

This function is used by updaters who have completed, and can therefore now allow readers to make use of their fastpaths after a grace period has elapsed. After this grace period has completed, all subsequent calls to rcu_sync_is_idle() will return true, which tells readers that they can once again use their fastpaths.

void rcu_sync_dtor(struct rcu_sync * rsp)

Clean up an rcu_sync structure

Parameters

struct rcu_sync * rsp
Pointer to rcu_sync structure to be cleaned up