Why does `kill -l` not list signal numbers of 32 and 33?

Question

Executing kill -l on linux gives:

 1) SIGHUP       2) SIGINT       3) SIGQUIT      4) SIGILL       5) SIGTRAP
 6) SIGABRT      7) SIGBUS       8) SIGFPE       9) SIGKILL     10) SIGUSR1
11) SIGSEGV     12) SIGUSR2     13) SIGPIPE     14) SIGALRM     15) SIGTERM
16) SIGSTKFLT   17) SIGCHLD     18) SIGCONT     19) SIGSTOP     20) SIGTSTP
21) SIGTTIN     22) SIGTTOU     23) SIGURG      24) SIGXCPU     25) SIGXFSZ
26) SIGVTALRM   27) SIGPROF     28) SIGWINCH    29) SIGIO       30) SIGPWR
31) SIGSYS      34) SIGRTMIN    35) SIGRTMIN+1  36) SIGRTMIN+2  37) SIGRTMIN+3
38) SIGRTMIN+4  39) SIGRTMIN+5  40) SIGRTMIN+6  41) SIGRTMIN+7  42) SIGRTMIN+8
43) SIGRTMIN+9  44) SIGRTMIN+10 45) SIGRTMIN+11 46) SIGRTMIN+12 47) SIGRTMIN+13
48) SIGRTMIN+14 49) SIGRTMIN+15 50) SIGRTMAX-14 51) SIGRTMAX-13 52) SIGRTMAX-12
53) SIGRTMAX-11 54) SIGRTMAX-10 55) SIGRTMAX-9  56) SIGRTMAX-8  57) SIGRTMAX-7
58) SIGRTMAX-6  59) SIGRTMAX-5  60) SIGRTMAX-4  61) SIGRTMAX-3  62) SIGRTMAX-2
63) SIGRTMAX-1  64) SIGRTMAX

What happened to 32 and 33? Why is it not listed? They could have started at 1 and ended at 62 instead of skipping 2 in the middle?

Solution 1

It is because of NPTL. Since it is part of the GNU C library nearly every modern linux distribution don't uses the first two real time signals anymore. NPTL is an implementation of the POSIX Threads. NPTL makes internal use of the first two real-time signals.

This part of the signal manpage is very interesting:

The Linux kernel supports a range of 32 different real-time signals, numbered 33 to 64. However, the glibc POSIX threads implementation internally uses two (for NPTL) or three (for LinuxThreads) real-time signals (see pthreads(7)), and adjusts the value of SIGRTMIN suitably (to 34 or 35). Because the range of available real-time signals varies according to the glibc threading implementation (and this variation can occur at run time according to the available kernel and glibc), and indeed the range of real-time signals varies across UNIX systems, programs should never refer to real-time signals using hard-coded numbers, but instead should always refer to real-time signals using the notation SIGRTMIN+n, and include suitable (run-time) checks that SIGRTMIN+n does not exceed SIGRTMAX.

I also checked the source code for glibc; see line 22. __SIGRTMIN is increased +2, so the first two real time signals are excluded from the range of real time signals.

Solution 2

Because the signals are:

SIGWAITING 32 Ignore All LWPs blocked 
    SIGLWP 33 Ignore Virtual Interprocessor Interrupt for Threads Library 

Neither of which are supported in Linux. (LWP stands for LightWeight Process)

Source: IBM DeveloperWorks Solaris to Linux Porting Guide

Comments

[G-Man Says 'Reinstate Monica']

P.S. There are not "error codes" -- they are signal numbers.

[cuonglm]

Linux had signal 32 and 33. See: unixhelp.ed.ac.uk/CGI/man-cgi?signal+7, Real-time Signals section.

[garethTheRed]

@Gnouc - according to kill -l RTMIN starts at 34, not 32 as per you referenced document. This one says it starts at 33, but goes on to say that you shouldn't reference them by numbers as the numbers may vary depending on glibc threading implementation.

[cuonglm]

Of course it varies base on system, but the term Linux does not supported is incorrect. You can refer this: cse.psu.edu/~dheller/cmpsc311/Lectures/Process-Control-Signa‌​ls/…. Maybe we need a vereran who working with Linux long time ago to let us know what happen with signal 32, 33. Why aren't they documented.

[SlySven]

They are used internally by the RT pthread library - and in the past there would be three not two because a different system used that many for internal purposes. From a coding point of view you are not supposed to use compile-time constants for the signals above SIGSYS e.g. a SIGRTMIN set with a #define line because that number is subject to possible change later when the executable is run. This would have shown up a few years ago when both pthread libs were in use if an application compiled against the one pthread library was run on a system that had been rebooted with the other!

[SlySven]

So how does one get the current value of SIGRTMIN at run-time in a piece of executing code?