An interesting thread problem

Typedef

I have two boxes.

One running Red hat 7.0 Linux 2.4.18-14

One running Red at 8.0 Linux 2.2.16-22smp

Struct similar to this

struct typie{
int a;
int b;
char value[200];
}typie_t;

In function x on Red hat 7.0
I go

typie_t v;
v.a=100;

pthread_create(instance,0,function,&v);

in function()
all is well and this executes a happy as pie.

In Red Hat 8.0

I do the same thing and I get
interesting and yet worrying results

[webop@searcher tjulia]$ strace -p 28894
accept(1, {sin_family=AF_UNSPEC, {sa_family=0, sa_data="\0\0\0\0\0\0\0\0\0\0\0\0\0\0"}, [16]) = 2
getrlimit(0x3, 0xbfff9ce0) = 0
pipe([3, 4]) = 0
clone(child_stack=0x80fd958, flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND) = 28940
write(4, "\373\211\0@\5\0\0\0\0\0\0\0\0000\3@H\240\377\277\204\207"..., 148) = 148
rt_sigprocmask(SIG_SETMASK, NULL, [RTMIN], 8) = 0
write(4, "\240\340\6@\0\0\0\0\0\0\0\0p\272\4\10\254\241\377\277\0"..., 148) = 148
rt_sigprocmask(SIG_SETMASK, NULL, [RTMIN], 8) = 0
rt_sigsuspend([] <unfinished ...>
--- SIGRTMIN (Unknown signal 32) ---
<... rt_sigsuspend resumed> ) = -1 EINTR (Interrupted system call)
sigreturn() = ? (mask now [RTMIN])
--- SIGSEGV (Segmentation fault) ---

That's the strace from the failed call on RH8.

Here is the lexically sucessful equivalent on RH 7.

[bodonoghue@converatest tjulia]$ strace -p 29328
accept(1, {sin_family=AF_UNSPEC, {sa_family=0, sa_data="\0\0\0\0\0\0\0\0\0\0\0\0\0\0"}, [16]) = 2
getrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=2147483647}) = 0
setrlimit(RLIMIT_STACK, {rlim_cur=2044*1024, rlim_max=2147483647}) = 0
pipe([3, 4]) = 0
clone(child_stack=0x80fdf00, flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND) = 29332
write(4, "\0\0\0\0\5\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\2\0\0"..., 148) = 148
rt_sigprocmask(SIG_SETMASK, NULL, [RT_0], 8) = 0
write(4, "@:\7@\0\0\0\0\0\0\0\0p\266\4\10\340k\377\277\0\0\0\200"..., 148) = 148
rt_sigprocmask(SIG_SETMASK, NULL, [RT_0], 8) = 0
rt_sigsuspend([] <unfinished ...>
--- SIGRT_0 (Real-time signal 0) ---

Can somebody shed a little light on this error or me? It's looks to me that "interrupted system call" is the culprit, but... I still don't quite fathom how to cicrumvent this error.
Any help appreciated.

Muchos Gracias.

Typedef

void networking::networking_do_loop(){
/*Main loop to accept connections and call the search functions*/
int state,b,num_bytes,fd;
sigset_t sigs;
socklen_t slen;
char incoming[22];
MAX_FD=0;
trans_t tr;

tr.iMaxDocs=iMaxDocs;
tr.iWeight=iWeight;
strncpy(tr.cRWServer,cRWServer,1024);
strncpy(tr.pgstr,pgstr,1024);
#if defined(DEBUG)
syslog(LOG_NOTICE,"Object vales %s %s %d %d\n",
tr.cRWServer,tr.pgstr,tr.iMaxDocs,tr.iWeight);
#endif
while(1){
fd=accept(srv_skfd,(struct sockaddr*)&cli,&slen);
#if defined(DEBUG)
syslog(LOG_NOTICE,"Accepted client connection on %d \n",fd,MAX_FD);
#endif
tr.skfd=fd;
/*Create a separate thread for each separate connection instance!*/
sigemptyset(&sigs);
sigaddset(&sigs,SIGRTMIN);
pthread_sigmask(SIG_BLOCK,&sigs,0);
pthread_create(&thread_dref,0,socket_thread,&tr);
}
return;
};

I'm glad we had this chance to talk...

"oh there'll be a hot time... in the old town ... tonight...."""

Very interesting.

Trojan

Ah, so it's pthreads (same codebase, right?).

What happens in single processor mode on 2.2.16-22smp? Can you switch that on?

Al.

Typedef

I think the problem was that Linux -2.4.x was sending a signal I wasn't handling, while 2.2, wasn't sending any signals, so the proggie didn't crash.