Experimental Physics and
Industrial Control System

"Jeff Hill" <[email protected]> · Mon, 11 Jun 2007 14:50:45 -0600

Till,

Completely agree with your assessment of the situation, and your fix. In my
defense, what actually occurs when one calls exit() has been evolving quite
a bit with each new release of R3.14.

Jeff

-----Original Message-----
From: [email protected] [mailto:[email protected]]
On Behalf Of Till Straumann
Sent: Sunday, June 10, 2007 1:38 PM
To: TECHTALK tech-talk
Subject: CA client on unix bug

I found the following problem (tested on
linux, solaris, base-3.14.9, base-3.14.8.2)
with the CA client on unix:

Scenario:

IF
  - no caRepeater is running
  - CA client tries to fork a caRepeater
    but the 'exec' syscall fails (e.g., because the
    caRepeater is not found in PATH)

THEN
 the forked process may end up blocking
 for an event that never happens and hence
 may never exit (see attached test program for
 more details) and therefore never release system
 resources.

FIX: libCom/osi/os/posix/osdProcess.c:osiSpawnDetachedProcess()
  should call _exit() rather than exit() if execle() fails.

REPRODUCE: try attached code (comments contain analysis
of the problem).

-- Till

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System