2.6.33-rc4 i686 clone function looping (?)

Jean-Marc Pigeon jmp at safe.ca
Mon Jan 18 14:04:03 PST 2010


Hello,

	Same physical system and configuration
	(small system in 32 Bits i686 mode).
	2.6.32.3: 	working A_one.
	2.6.33-rc4:	clone call, looping and
		    	taking ALL CPU cycle.

Here is code used.
;======================================
cloneflg=SIGCHLD|__WCLONE;
cloneflg|=CLONE_NEWNET|CLONE_NEWIPC|CLONE_NEWNS|CLONE_NEWPID|
CLONE_NEWUTS;
clonestk=(char *)malloc(STKSIZE)
......

;--------------------------------------
  if (argv[0]!=(char *)0) { /*always        */
        static ARGTYP args;

        args.mspid=getpid();
        args.arch=cntarch;
        args.argc=argc;
        args.argv=argv;
        if ((cpid=clone(&doboot,(void *)(clonestk
+STKSIZE),cloneflg,(void *)&args))<0) {
          (void) apl_alert(0,"%s, Unable to clone %s container process
(error=<%s>)",
                              appname,argv[0],strerror(errno));
;======================================

There is some difference within .config file between
both kernel I regenerated, but I do not believe they
are meaningful for the clone/cgroup/veth/bridge functions.
I can provide config diff if needed.

Do we have a real problem with 2.6.33-rc4?? or did I missed
something? could someone confirm trouble with clone call?

Thanks.
-- 
A bientôt
==========================================================================
Jean-Marc Pigeon                                   Internet: jmp at safe.ca
SAFE Inc.                                          Phone: (514) 493-4280
                                                   Fax:   (514) 493-1946
        Clement, 'a kiss solution' to get rid of SPAM (at last)
           Clement' Home base <"http://www.clement.safe.ca">
==========================================================================



More information about the Containers mailing list