Re: sigaction, fork, malloc, and futex

Re: sigaction, fork, malloc, and futex

[Pascal Schmidt]

On Wed, 05 May 2004 01:40:09 +0200, you wrote in linux.kernel:
> The signal handler executes on its own signal stack, it
> can't reach anything thats located on the main program stack.

Only when using sigaltstack(2). Otherwise, by default, there is
no seperate signal stack.

-- 
Ciao,
Pascal

sigaction, fork, malloc, and futex

[Steve Beaty]

	anyone have a clue on this one?  we set up a signal handler, create
	a child that sends that signal, and have the signal handler fork
	another child.  if there is a malloc(), the second child gets stuck
	in a futex(); without the malloc(), no problem.  2.4.20-30.9
	kernel.  straces at the end.  any help would be appreciated.
	thanks!

===========================================================================
#include <stdio.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <unistd.h>
#include <stdlib.h>
#include <assert.h>

/* comment out the following to show the problem */
#define WORKS

void sig_handler (int signum)
{
	int     child;

	fprintf (stderr, "Process %d received a SIGALRM signal\n", getpid());
	if ((child = fork ()) == 0)
	{
		fprintf (stderr, "%d exiting\n", getpid());
		exit (0);
	}
	fprintf (stderr, "%d waiting for %d\n", getpid(), child);
	waitpid (child, NULL, 0);
}

int
main (int argc, char **argv)
{
	int parent = getpid();
	int child;

	struct sigaction action;
	sigemptyset (&action.sa_mask);
	action.sa_handler = sig_handler;

#ifndef WORKS
	malloc (sizeof (int));
#endif

	assert (sigaction (SIGALRM, &action, NULL) == 0);

	if ((child = fork ()) == 0)
	{
		fprintf (stderr, "%d sending SIGALRM to %d\n", getpid(), parent);
		if (kill (parent, SIGALRM) == -1) perror ("kill 1");
		fprintf (stderr, "%d exiting\n", getpid());
		exit (0);
	}
	fprintf (stderr, "%d waiting for %d\n", getpid(), child);
	waitpid (child, NULL, 0);
}

/*
** if WORKS defined:

$ ./a.out
9106 sending SIGALRM to 9105
9106 exiting
Process 9105 received a SIGALRM signal
9107 exiting
9105 waiting for 9107
9105 waiting for 9106

** if WORKS undefined:

$ ./a.out
9095 sending SIGALRM to 9094
9095 exiting
Process 9094 received a SIGALRM signal

$ sh /usr/src/linux-2.4.20-30.9/scripts/ver_linux
If some fields are empty or look unusual you may have an old version.
Compare to the current minimal requirements in Documentation/Changes.
  
Linux emess0.mscd.edu 2.4.20-30.9 #1 Wed Feb 4 20:44:26 EST 2004 i686 i686
i386 GNU/Linux
  
Gnu C                  3.2.2
Gnu make               3.79.1
util-linux             2.11y
mount                  2.11y
modutils               2.4.22
e2fsprogs              1.32
jfsutils               1.0.17
reiserfsprogs          3.6.4
pcmcia-cs              3.1.31
PPP                    2.4.1
isdn4k-utils           3.1pre4
Linux C Library        2.3.2
Dynamic linker (ldd)   2.3.2
Procps                 2.0.11
Net-tools              1.60
Kbd                    1.08
Sh-utils               4.5.3
Modules Loaded         ieee1394 ide-cd cdrom i810_audio ac97_codec
soundcore mga agpgart binfmt_misc autofs e100 ipchains loop lvm-mod keybdev
mousedev hid input usb-uhci usbcore ext3 jbd

*/

#if 0

BAD:
---------------------------------------------------------------------------
$ strace ./a.out
execve("./a.out", ["./a.out"], [/* 42 vars */]) = 0
uname({sys="Linux", node="emess.mscd.edu", ...}) = 0
brk(0)                                  = 0x80499fc
open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=105716, ...}) = 0
old_mmap(NULL, 105716, PROT_READ, MAP_PRIVATE, 3, 0) = 0x40016000
close(3)                                = 0
open("/lib/tls/libc.so.6", O_RDONLY)    = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360W\1"..., 512) =
512
fstat64(3, {st_mode=S_IFREG|0755, st_size=1539996, ...}) = 0
old_mmap(0x42000000, 1267276, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) =
0x42000000
old_mmap(0x42130000, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3,
0x130000) = 0x42130000
old_mmap(0x42133000, 9804, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x42133000
close(3)                                = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0x40030000
set_thread_area({entry_number:-1 -> 6, base_addr:0x400304a0, limit:1048575,
seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1,
seg_not_present:0, useable:1}) = 0
munmap(0x40016000, 105716)              = 0
getpid()                                = 9059
brk(0)                                  = 0x80499fc
brk(0x804a9fc)                          = 0x804a9fc
brk(0)                                  = 0x804a9fc
brk(0x804b000)                          = 0x804b000
rt_sigaction(SIGALRM, {0x8048568, [],
SA_RESTORER|SA_NOMASK|SA_SIGINFO|0x21328d0, 0x420277b0}, NULL, 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|0x11,
<ignored>, <ignored>, 0x400304e8) = 9060
9060 sending SIGALRM to 9059
9060 exiting
--- SIGALRM (Alarm clock) @ 0 (0) ---
getpid()                                = 9059
--- SIGCHLD (Child exited) @ 0 (0) ---
write(2, "Process 9059 received a SIGALRM "..., 39Process 9059 received a
SIGALRM signal
) = 39
futex(0x421336d0, FUTEX_WAIT, 2, NULL
[control-c]
---------------------------------------------------------------------------

GOOD

---------------------------------------------------------------------------
$ strace ./a.out
execve("./a.out", ["./a.out"], [/* 42 vars */]) = 0
uname({sys="Linux", node="emess.mscd.edu", ...}) = 0
brk(0)                                  = 0x80499b8
open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=105716, ...}) = 0
old_mmap(NULL, 105716, PROT_READ, MAP_PRIVATE, 3, 0) = 0x40016000
close(3)                                = 0
open("/lib/tls/libc.so.6", O_RDONLY)    = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360W\1"..., 512) =
512
fstat64(3, {st_mode=S_IFREG|0755, st_size=1539996, ...}) = 0
old_mmap(0x42000000, 1267276, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) =
0x42000000
old_mmap(0x42130000, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3,
0x130000) = 0x42130000
old_mmap(0x42133000, 9804, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x42133000
close(3)                                = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0x40030000
set_thread_area({entry_number:-1 -> 6, base_addr:0x400304a0, limit:1048575,
seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1,
seg_not_present:0, useable:1}) = 0
munmap(0x40016000, 105716)              = 0
getpid()                                = 9087
rt_sigaction(SIGALRM, {0x8048534, [],
SA_RESTORER|SA_NOMASK|SA_SIGINFO|0x21328d0, 0x420277b0}, NULL, 8) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|0x11,
<ignored>, <ignored>, 0x400304e8) = 9088
9088 sending SIGALRM to 9087
9088 exiting
--- SIGALRM (Alarm clock) @ 0 (0) ---
getpid()                                = 9087
--- SIGCHLD (Child exited) @ 0 (0) ---
write(2, "Process 9087 received a SIGALRM "..., 39Process 9087 received a
SIGALRM signal
) = 39
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|0x11,
<ignored>, <ignored>, 0x400304e8) = 9089
getpid()                                = 9087
write(2, "9087 waiting for 9089\n", 229087 waiting for 9089
) = 22
wait4(9089, 9089 exiting
NULL, 0, NULL)              = 9089
rt_sigreturn(0x1200011)                 = 9088
getpid()                                = 9087
write(2, "9087 waiting for 9088\n", 229087 waiting for 9088
) = 22
wait4(9088, NULL, 0, NULL)              = 9088
exit_group(0)                           = ?
---------------------------------------------------------------------------
#endif
===========================================================================

-- 
Dr. Steve Beaty (B80)                                 Associate Professor
Metro State College of Denver                        beaty@emess.mscd.edu
VOX: (303) 556-5321                                 Science Building 134C
FAX: (303) 556-5381                         http://clem.mscd.edu/~beatys/

Re: sigaction, fork, malloc, and futex

[chris]

Hi-

On Tue, 4 May 2004, Steve Beaty wrote:

>
> 	anyone have a clue on this one?  we set up a signal handler, create
> 	a child that sends that signal, and have the signal handler fork
> 	another child.  if there is a malloc(), the second child gets stuck
> 	in a futex(); without the malloc(), no problem.  2.4.20-30.9
> 	kernel.  straces at the end.  any help would be appreciated.
> 	thanks!

Your signal handler function is illegally calling non-reentrant functions.
The *printf() family of functions are going to need to call malloc() to
allocate buffers. malloc() cannot be re-entered.

So specifically your deadlock sequence is:

Parent:
fork()
fprintf()
-> malloc()
   -> take a malloc() lock
(Child schedules and sends SIGALRM at this point)
SIGALRM:
fprintf()
-> malloc()
   -> try to take a malloc() lock
      -> deadlock, lock is already taken and will never be released!

Modern glibc / kernel combinations which use futexes in the malloc code
really seem to expose this race.

Cheers
Chris

Re: sigaction, fork, malloc, and futex

[Zan Lynx]

[-- Attachment #1: Type: text/plain, Size: 1714 bytes --]

On Tue, 2004-05-04 at 16:30, chris@scary.beasts.org wrote:
> Hi-
> 
> On Tue, 4 May 2004, Steve Beaty wrote:
> 
> >
> > 	anyone have a clue on this one?  we set up a signal handler, create
> > 	a child that sends that signal, and have the signal handler fork
> > 	another child.  if there is a malloc(), the second child gets stuck
> > 	in a futex(); without the malloc(), no problem.  2.4.20-30.9
> > 	kernel.  straces at the end.  any help would be appreciated.
> > 	thanks!
> 
> Your signal handler function is illegally calling non-reentrant functions.
> The *printf() family of functions are going to need to call malloc() to
> allocate buffers. malloc() cannot be re-entered.
> 
> So specifically your deadlock sequence is:
> 
> Parent:
> fork()
> fprintf()
> -> malloc()
>    -> take a malloc() lock
> (Child schedules and sends SIGALRM at this point)
> SIGALRM:
> fprintf()
> -> malloc()
>    -> try to take a malloc() lock
>       -> deadlock, lock is already taken and will never be released!
> 
> Modern glibc / kernel combinations which use futexes in the malloc code
> really seem to expose this race.
> 
> Cheers
> Chris

No, it's nothing to do with the fprintf.  I tried the program without
any fprints at all and got the same result.

I'm pretty sure the problem is in glibc.  Look at malloc_atfork and
free_atfork in glibc's malloc/arena.c.  I think the reason you are only
seeing it happen when you malloc is that glibc only initializes the
malloc system when you first use it.

I am not sure it is really a problem though.  I don't think you should
be allowed to fork inside a signal handler.  That seems very wrong.
-- 
Zan Lynx <zlynx@acm.org>

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: sigaction, fork, malloc, and futex

[Andreas Schwab]

Zan Lynx <zlynx@acm.org> writes:

> I am not sure it is really a problem though.  I don't think you should
> be allowed to fork inside a signal handler.  That seems very wrong.

fork is marked async-signal-safe, provided that all atfork handlers are
async-signal-safe.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux AG, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: sigaction, fork, malloc, and futex

[Steve Beaty]

> I am not sure it is really a problem though.  I don't think you should
> be allowed to fork inside a signal handler.  That seems very wrong.

	i can't disagree :-)  but fork() is supposed to be reentrant...

-- 
Dr. Steve Beaty (B80)                                 Associate Professor
Metro State College of Denver                        beaty@emess.mscd.edu
VOX: (303) 556-5321                                 Science Building 134C
FAX: (303) 556-5381                         http://clem.mscd.edu/~beatys/

Re: sigaction, fork, malloc, and futex

[Zan Lynx]

[-- Attachment #1: Type: text/plain, Size: 536 bytes --]

On Thu, 2004-05-06 at 15:37, Steve Beaty wrote:
> > I am not sure it is really a problem though.  I don't think you should
> > be allowed to fork inside a signal handler.  That seems very wrong.
> 
> 	i can't disagree :-)  but fork() is supposed to be reentrant...

In your test program, I just tried using sys_clone to fork the process
directly.  It works.  The problem really is inside glibc.  If glibc is
not POSIX compliant in this instance, I suppose you will have to file a
bug with them.
-- 
Zan Lynx <zlynx@acm.org>

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: sigaction, fork, malloc, and futex

[Steve Beaty]

> Your signal handler function is illegally calling non-reentrant functions.
> The *printf() family of functions are going to need to call malloc() to
> allocate buffers. malloc() cannot be re-entered.

	it deadlocks even with no calls in the client...  weird.

	thanks,

-- 
Dr. Steve Beaty (B80)                                 Associate Professor
Metro State College of Denver                        beaty@emess.mscd.edu
VOX: (303) 556-5321                                 Science Building 134C
FAX: (303) 556-5381                         http://clem.mscd.edu/~beatys/

Re: sigaction, fork, malloc, and futex

[Richard B. Johnson]

On Tue, 4 May 2004, Steve Beaty wrote:

>
> 	anyone have a clue on this one?  we set up a signal handler, create
> 	a child that sends that signal, and have the signal handler fork
> 	another child.  if there is a malloc(), the second child gets stuck
> 	in a futex(); without the malloc(), no problem.  2.4.20-30.9
> 	kernel.  straces at the end.  any help would be appreciated.
> 	thanks!

Sure buggy code. There are only a few things you can do from
a signal handler and I'm fairly sure fork() isn't one of them.
Did you just invent this as a way of concatenating system calls
to see what works? Surely you don't have a program that does this
do you? The signal handler executes on its own signal stack, it
can't reach anything thats located on the main program stack. The
malloc() stuff is in the heap, above the main program stack. Once
you do something that requires access to the heap or the main
program stack, in the signal handler, all bets are off.

So, just don't dooooo this!

>
> ===========================================================================
> #include <stdio.h>
> #include <sys/types.h>
> #include <sys/wait.h>
> #include <unistd.h>
> #include <stdlib.h>
> #include <assert.h>
>
> /* comment out the following to show the problem */
> #define WORKS
>
> void sig_handler (int signum)
> {
> 	int     child;
>
> 	fprintf (stderr, "Process %d received a SIGALRM signal\n", getpid());
> 	if ((child = fork ()) == 0)
> 	{
> 		fprintf (stderr, "%d exiting\n", getpid());
> 		exit (0);
> 	}
> 	fprintf (stderr, "%d waiting for %d\n", getpid(), child);
> 	waitpid (child, NULL, 0);
> }
>
> int
> main (int argc, char **argv)
> {
> 	int parent = getpid();
> 	int child;
>
> 	struct sigaction action;
> 	sigemptyset (&action.sa_mask);
> 	action.sa_handler = sig_handler;
>
> #ifndef WORKS
> 	malloc (sizeof (int));
> #endif
>
> 	assert (sigaction (SIGALRM, &action, NULL) == 0);
>
> 	if ((child = fork ()) == 0)
> 	{
> 		fprintf (stderr, "%d sending SIGALRM to %d\n", getpid(), parent);
> 		if (kill (parent, SIGALRM) == -1) perror ("kill 1");
> 		fprintf (stderr, "%d exiting\n", getpid());
> 		exit (0);
> 	}
> 	fprintf (stderr, "%d waiting for %d\n", getpid(), child);
> 	waitpid (child, NULL, 0);
> }
>
> /*
> ** if WORKS defined:
>
> $ ./a.out
> 9106 sending SIGALRM to 9105
> 9106 exiting
> Process 9105 received a SIGALRM signal
> 9107 exiting
> 9105 waiting for 9107
> 9105 waiting for 9106
>
> ** if WORKS undefined:
>
> $ ./a.out
> 9095 sending SIGALRM to 9094
> 9095 exiting
> Process 9094 received a SIGALRM signal
>
> $ sh /usr/src/linux-2.4.20-30.9/scripts/ver_linux
> If some fields are empty or look unusual you may have an old version.
> Compare to the current minimal requirements in Documentation/Changes.
>
> Linux emess0.mscd.edu 2.4.20-30.9 #1 Wed Feb 4 20:44:26 EST 2004 i686 i686
> i386 GNU/Linux
>
> Gnu C                  3.2.2
> Gnu make               3.79.1
> util-linux             2.11y
> mount                  2.11y
> modutils               2.4.22
> e2fsprogs              1.32
> jfsutils               1.0.17
> reiserfsprogs          3.6.4
> pcmcia-cs              3.1.31
> PPP                    2.4.1
> isdn4k-utils           3.1pre4
> Linux C Library        2.3.2
> Dynamic linker (ldd)   2.3.2
> Procps                 2.0.11
> Net-tools              1.60
> Kbd                    1.08
> Sh-utils               4.5.3
> Modules Loaded         ieee1394 ide-cd cdrom i810_audio ac97_codec
> soundcore mga agpgart binfmt_misc autofs e100 ipchains loop lvm-mod keybdev
> mousedev hid input usb-uhci usbcore ext3 jbd
>
> */
>
> #if 0
>
> BAD:
> ---------------------------------------------------------------------------
> $ strace ./a.out
> execve("./a.out", ["./a.out"], [/* 42 vars */]) = 0
> uname({sys="Linux", node="emess.mscd.edu", ...}) = 0
> brk(0)                                  = 0x80499fc
> open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
> directory)
> open("/etc/ld.so.cache", O_RDONLY)      = 3
> fstat64(3, {st_mode=S_IFREG|0644, st_size=105716, ...}) = 0
> old_mmap(NULL, 105716, PROT_READ, MAP_PRIVATE, 3, 0) = 0x40016000
> close(3)                                = 0
> open("/lib/tls/libc.so.6", O_RDONLY)    = 3
> read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360W\1"..., 512) =
> 512
> fstat64(3, {st_mode=S_IFREG|0755, st_size=1539996, ...}) = 0
> old_mmap(0x42000000, 1267276, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) =
> 0x42000000
> old_mmap(0x42130000, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3,
> 0x130000) = 0x42130000
> old_mmap(0x42133000, 9804, PROT_READ|PROT_WRITE,
> MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x42133000
> close(3)                                = 0
> old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
> 0) = 0x40030000
> set_thread_area({entry_number:-1 -> 6, base_addr:0x400304a0, limit:1048575,
> seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1,
> seg_not_present:0, useable:1}) = 0
> munmap(0x40016000, 105716)              = 0
> getpid()                                = 9059
> brk(0)                                  = 0x80499fc
> brk(0x804a9fc)                          = 0x804a9fc
> brk(0)                                  = 0x804a9fc
> brk(0x804b000)                          = 0x804b000
> rt_sigaction(SIGALRM, {0x8048568, [],
> SA_RESTORER|SA_NOMASK|SA_SIGINFO|0x21328d0, 0x420277b0}, NULL, 8) = 0
> clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|0x11,
> <ignored>, <ignored>, 0x400304e8) = 9060
> 9060 sending SIGALRM to 9059
> 9060 exiting
> --- SIGALRM (Alarm clock) @ 0 (0) ---
> getpid()                                = 9059
> --- SIGCHLD (Child exited) @ 0 (0) ---
> write(2, "Process 9059 received a SIGALRM "..., 39Process 9059 received a
> SIGALRM signal
> ) = 39
> futex(0x421336d0, FUTEX_WAIT, 2, NULL
> [control-c]
> ---------------------------------------------------------------------------
>
> GOOD
>
> ---------------------------------------------------------------------------
> $ strace ./a.out
> execve("./a.out", ["./a.out"], [/* 42 vars */]) = 0
> uname({sys="Linux", node="emess.mscd.edu", ...}) = 0
> brk(0)                                  = 0x80499b8
> open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
> directory)
> open("/etc/ld.so.cache", O_RDONLY)      = 3
> fstat64(3, {st_mode=S_IFREG|0644, st_size=105716, ...}) = 0
> old_mmap(NULL, 105716, PROT_READ, MAP_PRIVATE, 3, 0) = 0x40016000
> close(3)                                = 0
> open("/lib/tls/libc.so.6", O_RDONLY)    = 3
> read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360W\1"..., 512) =
> 512
> fstat64(3, {st_mode=S_IFREG|0755, st_size=1539996, ...}) = 0
> old_mmap(0x42000000, 1267276, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) =
> 0x42000000
> old_mmap(0x42130000, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3,
> 0x130000) = 0x42130000
> old_mmap(0x42133000, 9804, PROT_READ|PROT_WRITE,
> MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x42133000
> close(3)                                = 0
> old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
> 0) = 0x40030000
> set_thread_area({entry_number:-1 -> 6, base_addr:0x400304a0, limit:1048575,
> seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1,
> seg_not_present:0, useable:1}) = 0
> munmap(0x40016000, 105716)              = 0
> getpid()                                = 9087
> rt_sigaction(SIGALRM, {0x8048534, [],
> SA_RESTORER|SA_NOMASK|SA_SIGINFO|0x21328d0, 0x420277b0}, NULL, 8) = 0
> clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|0x11,
> <ignored>, <ignored>, 0x400304e8) = 9088
> 9088 sending SIGALRM to 9087
> 9088 exiting
> --- SIGALRM (Alarm clock) @ 0 (0) ---
> getpid()                                = 9087
> --- SIGCHLD (Child exited) @ 0 (0) ---
> write(2, "Process 9087 received a SIGALRM "..., 39Process 9087 received a
> SIGALRM signal
> ) = 39
> clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|0x11,
> <ignored>, <ignored>, 0x400304e8) = 9089
> getpid()                                = 9087
> write(2, "9087 waiting for 9089\n", 229087 waiting for 9089
> ) = 22
> wait4(9089, 9089 exiting
> NULL, 0, NULL)              = 9089
> rt_sigreturn(0x1200011)                 = 9088
> getpid()                                = 9087
> write(2, "9087 waiting for 9088\n", 229087 waiting for 9088
> ) = 22
> wait4(9088, NULL, 0, NULL)              = 9088
> exit_group(0)                           = ?
> ---------------------------------------------------------------------------
> #endif
> ===========================================================================
>
> --
> Dr. Steve Beaty (B80)                                 Associate Professor
> Metro State College of Denver                        beaty@emess.mscd.edu
> VOX: (303) 556-5321                                 Science Building 134C
> FAX: (303) 556-5381                         http://clem.mscd.edu/~beatys/
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Cheers,
Dick Johnson
Penguin : Linux version 2.4.26 on an i686 machine (5557.45 BogoMips).
            Note 96.31% of all statistics are fiction.

Re: sigaction, fork, malloc, and futex

[Steve Beaty]

> Sure buggy code. There are only a few things you can do from
> a signal handler and I'm fairly sure fork() isn't one of them.

	well, one should only set values and deal with them somewhere in
	signal handlers, but...

> Did you just invent this as a way of concatenating system calls
> to see what works? Surely you don't have a program that does this
> do you?

	well, yes, but it is a classroom assignment so not "real" code.
	the assignment is to create a simulated disk drive (a process) via
	a "system call" that signals a "kernel".  perhaps we should really
	fork and exec...

-- 
Dr. Steve Beaty (B80)                                 Associate Professor
Metro State College of Denver                        beaty@emess.mscd.edu
VOX: (303) 556-5321                                 Science Building 134C
FAX: (303) 556-5381                         http://clem.mscd.edu/~beatys/

Re: sigaction, fork, malloc, and futex

[Edgar Toernig]

Steve Beaty wrote:
>
>... the signal handler fork another child.

Uhh... just read that POSIX actually allows that.  Evil!
I guess you're entering undiscovered land...

> void sig_handler (int signum)
> {
> 	int     child;
> 
> 	fprintf (stderr, "Process %d received a SIGALRM signal\n", getpid());
> 	if ((child = fork ()) == 0)
> 	{
> 		fprintf (stderr, "%d exiting\n", getpid());
> 		exit (0);
> 	}
> 	fprintf (stderr, "%d waiting for %d\n", getpid(), child);
> 	waitpid (child, NULL, 0);
> }

No fprintf and exit in a signal handler!  Use write and _exit instead.

> 	if ((child = fork ()) == 0)
> 	{
> 		fprintf (stderr, "%d sending SIGALRM to %d\n", getpid(), parent);
> 		if (kill (parent, SIGALRM) == -1) perror ("kill 1");
> 		fprintf (stderr, "%d exiting\n", getpid());
> 		exit (0);
> 	}

This exit is dubious too.  _exit is better.

Good luck, you may need it - ET.

Re: sigaction, fork, malloc, and futex

[Steve Beaty]

> Uhh... just read that POSIX actually allows that.  Evil!
> I guess you're entering undiscovered land...

	it's funny how often i am there :-) :-)

> No fprintf and exit in a signal handler!  Use write and _exit instead.

	doesn't seem to matter...

> This exit is dubious too.  _exit is better.

	nor does this...  i know better than to use *printf and exit; i was
	eager to get other's views on this...

> Good luck, you may need it - ET.

	i do! :-)  the code seems to work on Ultrix, Solaris, and OSX.
	this doesn't imply that it should :-)

-- 
Dr. Steve Beaty (B80)                                 Associate Professor
Metro State College of Denver                        beaty@emess.mscd.edu
VOX: (303) 556-5321                                 Science Building 134C
FAX: (303) 556-5381                         http://clem.mscd.edu/~beatys/