Re: SIGSTOP signal

2000-06-13 Thread Bolan Meek
Nuno Almeida wrote:
 
 I would like to know if this is simply a singularity of Debian, or if it's a
 bug of mine.
 
 When I'm programming in C/C++ to other linux distr. and I make a signal trap
 I can't, and that's absolutly normal, trap the signals 9 and 17, for SIGKILL
 and SIGSTOP.
 On Debian I can catch the SIGSTOP signal is this normal? Why the diference?

Nothing can trap SIGKILL.  Elsewise, how could anything be killed for
sure?
I'm not really familiar with SIGSTOP.  But let me check... (snips from
man
below).

Looks like it is _not_ normal to trap the SIGSTOP, not even on Linux,
including
Debian.  If it _can_ it should be a bug to report to the maintainer of
kernel-image*.

My HP-UX box says (from `man kill`):
  0 SIGNULL   NullCheck access to pid
  1 SIGHUPHangup  Terminate; can be trapped
  2 SIGINTInterrupt   Terminate; can be trapped
  3 SIGQUIT   QuitTerminate with core dump; can be
trapped
  9 SIGKILL   KillForced termination; cannot be
trapped
 15 SIGTERM   Terminate   Terminate; can be trapped
 24 SIGSTOP   StopPause the process; cannot be
trapped  ==
 25 SIGTSTP   Terminal stop   Pause the process; can be trapped
 26 SIGCONT   ContinueRun a stopped process

My debian-sparc box says:

Linux   November 21, 1999   1

() ()

   ALRM 14   exitHUP  1exitINT  2exit
   KILL 9exit this   signal   maynotbeblocked
   PIPE 13   exitPOLL  exitPROF  exit
   TERM 15   exitUSR1  exitUSR2  exit
   VTALRM exit  STKFLT exit may not be imple-
   mented  PWR   ignoremay  exit  on   some   systems
   WINCH  ignore   CHLD  ignore  URG   ignore
   TSTP  stop may interact with the shell
   TTIN  stop may interact with the shell
   TTOU  stop may interact with the shell
   STOP  stop thissignalmaynot   be   blocked  
==
   CONT  restart   continue if stopped, otherwise  ignore
   ABRT 6coreFPE  8coreILL  4core
   QUIT 3coreSEGV 11   coreTRAP 5core
   SYS   core may  not  be implemented EMT   core may
   not  be  implemented  BUS   core core  dump  may  fail
   XCPU  core core dump may fail XFSZ  core core dump
   may fail

The man page on Solaris 7 says:

...
 The signal() and sigset() functions modify  signal  disposi-
 tions.  The  sig argument specifies the signal, which may be
 any signal except SIGKILL and  SIGSTOP.  The  disp  argument  

 specifies  the  signal's  disposition, which may be SIG_DFL,
...



SIGSTOP signal

2000-06-12 Thread Nuno Almeida
I would like to know if this is simply a singularity of Debian, or if it's a
bug of mine.

When I'm programming in C/C++ to other linux distr. and I make a signal trap
I can't, and that's absolutly normal, trap the signals 9 and 17, for SIGKILL
and SIGSTOP.
On Debian I can catch the SIGSTOP signal is this normal? Why the diference?