Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@node Bit Manipulation, Date and Time, Arithmetic, Top
|
|
|
|
|
@c %MENU% Bit manipulation
|
|
|
|
|
@chapter Bit Manipulation
|
|
|
|
|
|
|
|
|
|
This chapter contains information about functions and macros for
|
|
|
|
|
determining the endianness of integer types and manipulating the bits
|
2024-02-01 11:02:01 +00:00
|
|
|
of unsigned integers. These functions and macros are from ISO C23 and
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
are declared in the header file @file{stdbit.h}.
|
|
|
|
|
|
|
|
|
|
The following macros describe the endianness of integer types. They
|
|
|
|
|
have values that are integer constant expressions.
|
|
|
|
|
|
|
|
|
|
@defmac __STDC_ENDIAN_LITTLE__
|
|
|
|
|
This macro represents little-endian storage.
|
|
|
|
|
@end defmac
|
|
|
|
|
|
|
|
|
|
@defmac __STDC_ENDIAN_BIG__
|
|
|
|
|
This macro represents big-endian storage.
|
|
|
|
|
@end defmac
|
|
|
|
|
|
|
|
|
|
@defmac __STDC_ENDIAN_NATIVE__
|
|
|
|
|
This macro equals @code{__STDC_ENDIAN_LITTLE__} if integer types are
|
|
|
|
|
stored in memory in little-endian format, and equals
|
|
|
|
|
@code{__STDC_ENDIAN_BIG__} if integer types are stored in memory in
|
|
|
|
|
big-endian format.
|
|
|
|
|
@end defmac
|
|
|
|
|
|
|
|
|
|
The following functions manipulate the bits of unsigned integers.
|
|
|
|
|
Each function family has functions for the types @code{unsigned char},
|
|
|
|
|
@code{unsigned short}, @code{unsigned int}, @code{unsigned long int}
|
|
|
|
|
and @code{unsigned long long int}. In addition, there is a
|
|
|
|
|
corresponding type-generic macro (not listed below), named the same as
|
|
|
|
|
the functions but without any suffix such as @samp{_uc}. The
|
|
|
|
|
type-generic macro can only be used with an argument of an unsigned
|
2024-01-31 18:17:27 +00:00
|
|
|
integer type with a width of 8, 16, 32 or 64 bits, or when using
|
|
|
|
|
a compiler with support for
|
2024-02-01 15:36:55 +00:00
|
|
|
@uref{https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html,@code{__builtin_stdc_bit_ceil}},
|
2024-01-31 18:17:27 +00:00
|
|
|
etc.@:, built-in functions such as GCC 14.1 or later
|
|
|
|
|
any unsigned integer type those built-in functions support.
|
|
|
|
|
In GCC 14.1 that includes support for @code{unsigned __int128} and
|
|
|
|
|
@code{unsigned _BitInt(@var{n})} if supported by the target.
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_leading_zeros_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_zeros_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_zeros_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_zeros_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_zeros_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_leading_zeros} functions count the number of leading
|
|
|
|
|
(most significant) zero bits in @var{x}, starting from the most
|
|
|
|
|
significant bit of the argument type. If @var{x} is zero, they return
|
|
|
|
|
the width of @var{x} in bits.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_leading_ones_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_ones_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_ones_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_ones_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_leading_ones_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_leading_ones} functions count the number of leading
|
|
|
|
|
(most significant) one bits in @var{x}, starting from the most
|
|
|
|
|
significant bit of the argument type.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_trailing_zeros_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_zeros_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_zeros_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_zeros_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_zeros_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_trailing_zeros} functions count the number of trailing
|
|
|
|
|
(least significant) zero bits in @var{x}, starting from the least
|
|
|
|
|
significant bit of the argument type. If @var{x} is zero, they return
|
|
|
|
|
the width of @var{x} in bits.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_trailing_ones_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_ones_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_ones_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_ones_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_trailing_ones_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_trailing_ones} functions count the number of trailing
|
|
|
|
|
(least significant) one bits in @var{x}, starting from the least
|
|
|
|
|
significant bit of the argument type.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_first_leading_zero_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_zero_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_zero_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_zero_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_zero_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_first_leading_zero} functions return the position of
|
|
|
|
|
the most significant zero bit in @var{x}, counting from the most
|
|
|
|
|
significant bit of @var{x} as 1, or zero if there is no zero bit in
|
|
|
|
|
@var{x}.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_first_leading_one_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_one_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_one_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_one_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_leading_one_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_first_leading_one} functions return the position of the
|
|
|
|
|
most significant one bit in @var{x}, counting from the most
|
|
|
|
|
significant bit of @var{x} as 1, or zero if there is no one bit in
|
|
|
|
|
@var{x}.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_first_trailing_zero_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_zero_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_zero_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_zero_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_zero_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_first_trailing_zero} functions return the position of
|
|
|
|
|
the least significant zero bit in @var{x}, counting from the least
|
|
|
|
|
significant bit of @var{x} as 1, or zero if there is no zero bit in
|
|
|
|
|
@var{x}.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_first_trailing_one_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_one_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_one_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_one_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_first_trailing_one_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_first_trailing_one} functions return the position of
|
|
|
|
|
the least significant one bit in @var{x}, counting from the least
|
|
|
|
|
significant bit of @var{x} as 1, or zero if there is no one bit in
|
|
|
|
|
@var{x}.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_count_zeros_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_zeros_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_zeros_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_zeros_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_zeros_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_count_zeros} functions count the number of zero bits in
|
|
|
|
|
@var{x}.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_count_ones_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_ones_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_ones_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_ones_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_count_ones_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_count_ones} functions count the number of one bits in
|
|
|
|
|
@var{x}.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {_Bool} stdc_has_single_bit_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {_Bool} stdc_has_single_bit_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {_Bool} stdc_has_single_bit_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {_Bool} stdc_has_single_bit_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {_Bool} stdc_has_single_bit_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_has_single_bit} functions return whether @var{x} has
|
|
|
|
|
exactly one bit set to one.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned int} stdc_bit_width_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_bit_width_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_bit_width_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_bit_width_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_bit_width_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_bit_width} functions return the minimum number of bits
|
|
|
|
|
needed to store @var{x}, not counting leading zero bits. If @var{x}
|
|
|
|
|
is zero, they return zero.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned char} stdc_bit_floor_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned short} stdc_bit_floor_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_bit_floor_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned long int} stdc_bit_floor_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned long long int} stdc_bit_floor_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_bit_floor} functions return the largest integer power
|
|
|
|
|
of two that is less than or equal to @var{x}. If @var{x} is zero,
|
|
|
|
|
they return zero.
|
|
|
|
|
@end deftypefun
|
|
|
|
|
|
|
|
|
|
@deftypefun {unsigned char} stdc_bit_ceil_uc (unsigned char @var{x})
|
|
|
|
|
@deftypefunx {unsigned short} stdc_bit_ceil_us (unsigned short @var{x})
|
|
|
|
|
@deftypefunx {unsigned int} stdc_bit_ceil_ui (unsigned int @var{x})
|
|
|
|
|
@deftypefunx {unsigned long int} stdc_bit_ceil_ul (unsigned long int @var{x})
|
|
|
|
|
@deftypefunx {unsigned long long int} stdc_bit_ceil_ull (unsigned long long int @var{x})
|
2024-02-01 11:02:01 +00:00
|
|
|
@standards{C23, stdbit.h}
|
Implement C23 <stdbit.h>
C23 adds a header <stdbit.h> with various functions and type-generic
macros for bit-manipulation of unsigned integers (plus macro defines
related to endianness). Implement this header for glibc.
The functions have both inline definitions in the header (referenced
by macros defined in the header) and copies with external linkage in
the library (which are implemented in terms of those macros to avoid
duplication). They are documented in the glibc manual. Tests, as
well as verifying results for various inputs (of both the macros and
the out-of-line functions), verify the types of those results (which
showed up a bug in an earlier version with the type-generic macro
stdc_has_single_bit wrongly returning a promoted type), that the
macros can be used at top level in a source file (so don't use ({})),
that they evaluate their arguments exactly once, and that the macros
for the type-specific functions have the expected implicit conversions
to the relevant argument type.
Jakub previously referred to -Wconversion warnings in type-generic
macros, so I've included a test with -Wconversion (but the only
warnings I saw and fixed from that test were actually in inline
functions in the <stdbit.h> header - not anything coming from use of
the type-generic macros themselves).
This implementation of the type-generic macros does not handle
unsigned __int128, or unsigned _BitInt types with a width other than
that of a standard integer type (and C23 doesn't require the header to
handle such types either). Support for those types, using the new
type-generic built-in functions Jakub's added for GCC 14, can
reasonably be added in a followup (along of course with associated
tests).
This implementation doesn't do anything special to handle C++, or have
any tests of functionality in C++ beyond the existing tests that all
headers can be compiled in C++ code; it's not clear exactly what form
this header should take in C++, but probably not one using macros.
DIS ballot comment AT-107 asks for the word "count" to be added to the
names of the stdc_leading_zeros, stdc_leading_ones,
stdc_trailing_zeros and stdc_trailing_ones functions and macros. I
don't think it's likely to be accepted (accepting any technical
comments would mean having an FDIS ballot), but if it is accepted at
the WG14 meeting (22-26 January in Strasbourg, starting with DIS
ballot comment handling) then there would still be time to update
glibc for the renaming before the 2.39 release.
The new functions and header are placed in the stdlib/ directory in
glibc, rather than creating a new toplevel stdbit/ or putting them in
string/ alongside ffs.
Tested for x86_64 and x86.
2024-01-03 12:07:14 +00:00
|
|
|
@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
|
|
|
|
|
The @code{stdc_bit_ceil} functions return the smallest integer power
|
|
|
|
|
of two that is greater than or equal to @var{x}. If this cannot be
|
|
|
|
|
represented in the return type, they return zero.
|
|
|
|
|
@end deftypefun
|
