fadd / faddl / fsub / fsubl / fmul / fmull / fdiv / fdivl / fsqrt / fsqrtl (3) - glibc

Split from the chapter glibc: arithmetic functions into the per-function layout; see the method page and the glibc ledger for coverage status.

Declaration and gates

#include <math.h>

float fadd(double x, double y);            /* (double, double) -> float */
float faddl(long double x, long double y);
float fsub(double x, double y);
float fsubl(long double x, long double y);
float fmul(double x, double y);
float fmull(long double x, long double y);
float fdiv(double x, double y);
float fdivl(long double x, long double y);
float fsqrt(double x);
float fsqrtl(long double x);

All are inside #if __GLIBC_USE (IEC_60559_BFP_EXT_C23) (/usr/include/math.h line 530): the f##name block lines 532-535 and the f##name##l block lines 540-556 both include bits/mathcalls-narrow.h (lines 23-39, __MATHCALL_NARROW of add/div/fma/mul/sqrt/sub). The gate is defined in bits/libc-header-start.h lines 67-77: on with _GNU_SOURCE or __STDC_WANT_IEC_60559_BFP_EXT__, or in C23 mode. Preprocessing <math.h> on this host shows no fadd under plain gnu17, but the declaration appears with _GNU_SOURCE or -std=c23 (verified with gcc -E).

Semantics

Per the manual (node Misc FP Arithmetic, arith.texi lines 2318-2410): these functions return x + y (respectively x - y, x * y, x / y, sqrt(x)), rounded once to the return type of the function without any intermediate rounding to the type of the arguments. The _l names take long double arguments and still return float. This gives a wider-precision computation with a single rounding into float.

Gotchas and portability

• Narrow operations are not aliases of the plain operators: plain (float)(x + y) on two floats rounds twice, once per operation.
• tgmath.h provides a generic macro fadd under the same gate (/usr/include/tgmath.h line 1124).
• The related TS 18661-3 mixed-type functions (fMaddfN, fMxaddfNx, …) are separate ledger rows and separate symbols (f32addf64, …).
• No man page exists for these names in the cached man-pages set.
• musl 1.2.5: absent from musl include/math.h (verified, 0 matches).

References

• glibc manual: arith.texi, node Misc FP Arithmetic (lines 2091-2432)
• headers: /usr/include/math.h, bits/mathcalls-narrow.h, bits/libc-header-start.h, /usr/include/tgmath.h
• symbols: nm -D --with-symbol-versions libm.so.6