Bug 846790

Summary: missing information about threads in kill(2) man page
Product: Red Hat Enterprise Linux 6 Reporter: David Tonhofer <bughunt>
Component: util-linux-ngAssignee: Karel Zak <kzak>
Status: CLOSED ERRATA QA Contact: qe-baseos-daemons
Severity: low Docs Contact:
Priority: low    
Version: 6.3CC: azelinka, ovasik, pschiffe, psklenar
Target Milestone: rcKeywords: ManPageChange
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Fixed In Version: Doc Type: Enhancement
Doc Text:
Feature: The kill(1) man page has been improved to explicitly explain interaction between kill syscall and threads. Reason: It wasn't obvious to users how kill(1) works with threads. Result (if any):
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-11-21 20:43:37 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Description David Tonhofer 2012-08-08 17:54:16 UTC
Description of problem:

A little question on the usage of kill(2) I have run into:

I have scripts that check whether a process is alive by getting a known pidfile,
then issuing a "kill -o PID" to the PID contained in that pidfile. This is a standard trick of course.

However, it seems that nowadays "kill -o PID" also sends signals to threads, i.e. if there is no PID, the value is assumed to be a TID and the thread is signalled (I suppose, it probably just signals whatever is in the process table at the given ID)

This does not conform to the specification. Consider /bin/kill:

- An strace of kill(1) shows a call to the kill(2) syscall but:
- The kill(2) manpage doesn't mention threads at all.
- Indeed, there is a specially designed tgkill(2) to signal threads.
- The manpage of tgkill says; "By contrast, kill(2) can only be used to send a
  signal to a process (i.e., thread group) as a whole, and the signal will be
  delivered to an arbitrary thread within that process.)"


Should kill(1) not check whether the pid it is given belongs to a thread or a process and, in case of a thread id, simply refuse to send a signal ("kill can only be used to send a signal to a process")

Shouldn't there be a tgkill(1) ?

Or else, shouldn't the manpage be changed?

So either the kill(2) manpage and the tgkill(2) are wrong / need to be
completed or there is an implementation problem with the kill syscall.

Comment 11 errata-xmlrpc 2013-11-21 20:43:37 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.