logb(3)

✨ logb — get exponent of a floating-point value ✨

okay bestie, so this is logb(3) from the Library Functions Manual, and it's literally just "girl, what's your exponent?" but for floats 💅

---

🌸 NAME 🌸

logb, logbf, logbl — get exponent of a floating-point value

📚 LIBRARY 📚

Math library (that's libm, so link with -lm bestie, don't forget or the linker will be SO dramatic)

🎀 SYNOPSIS 🎀

#include <math.h>

double logb(double x);
float logbf(float x);
long double logbl(long double x);

💖 Feature Test Macro Requirements for glibc 💖

(yes, peep feature_test_macros(7) if you're confused, she explains everything)

logb():

_ISOC99_SOURCE || _POSIX_C_SOURCE >= 200112L
    || _XOPEN_SOURCE >= 500
    || /* Since glibc 2.19: */ _DEFAULT_SOURCE
    || /* glibc <= 2.19: */ _BSD_SOURCE || _SVID_SOURCE

logbf(), logbl():

_ISOC99_SOURCE || _POSIX_C_SOURCE >= 200112L
    || /* Since glibc 2.19: */ _DEFAULT_SOURCE
    || /* glibc <= 2.19: */ _BSD_SOURCE || _SVID_SOURCE

---

🦋 DESCRIPTION 🦋

okay so the way this works 👑 these functions extract the exponent straight out of x's internal floating-point representation and hand it back to you as a floating-point value. no cap, that's it, that's the whole job.

the integer constant FLT_RADIX, which lives in <float.h>, tells you the radix your system's floating-point representation is using. if FLT_RADIX is 2, then logb(x) is similar to floor(log2(fabs(x))) — but bestie, hear me: the latter can give you a straight-up incorrect integer because of intermediate rounding 😭 so don't be out here swapping them like they're the same thing, because they are NOT.

and the iconic part: if x is subnormal, logb() returns the exponent x would have if it were normalized. it's giving "the version of me that got the glow-up" ✨

---

⭐ RETURN VALUE ⭐

On success, these functions return the exponent of x.

• If x is a NaN, a NaN is returned. (garbage in, garbage out, we love consistency 💀)
• If x is zero, then a pole error occurs, and the functions return -HUGE_VAL, -HUGE_VALF, or -HUGE_VALL, respectively.
• If x is negative infinity or positive infinity, then positive infinity is returned.

---

🔥 ERRORS 🔥

peep math_error(7) for the tea on how to figure out whether an error actually happened when you called these functions, bestie.

here's what can go down:

• Pole error: x is 0 — a divide-by-zero floating-point exception (FE_DIVBYZERO) is raised.

🫶 IMPORTANT: These functions do not set errno. not a single one. so don't go checking it, she's not talking.

---

💅 ATTRIBUTES 💅

for an explanation of the terms used in this section, see attributes(7).

• logb(), logbf(), logbl() — Thread safety: MT-Safe (we stan a thread-safe queen 👑)

---

🌈 STANDARDS 🌈

C11, POSIX.1-2008.

🕰️ HISTORY 🕰️

C99, POSIX.1-2001.

logb() is out here serving since 4.3BSD — see IEEE.3 in the 4.3BSD manual if you wanna see the original slay.

---

👀 SEE ALSO 👀

• ilogb(3)
• log(3)

---

okay that's the whole page, you ate that 🍓 this was Linux man-pages version 6.20, dated 2026-02-08, and it's still logb(3) at the end of the day. go take the exponent of something, bestie 💖✨🎀🦋