angel(2) system call, the quest for immortality, aka kill(2) with SIGSTOP/SIGKILL will *not* work

[Warm White Wolf]

Greetings !

I have developed a new system call, be it named angel(2),
on Linux operating system (this is what I know), which makes
a program invulnerable to kill(2) calls, including SIGKILL and
SIGSTOP.

The uses may involve fork() + angel(), daemon() + angel(),
setsid() + angel(), exec*() + angel().

Use the intellectual property I give you, as a gift to the BSD
operating system, using 4- 3- 2- BSD licence. That's it, name
me in the sources.

Thank you, FreeBSD !
You are a great Unix operating system !

[Alan Somers]

Post by Warm White Wolf
Greetings !
I have developed a new system call, be it named angel(2),
on Linux operating system (this is what I know), which makes
a program invulnerable to kill(2) calls, including SIGKILL and
SIGSTOP.
The uses may involve fork() + angel(), daemon() + angel(),
setsid() + angel(), exec*() + angel().
Use the intellectual property I give you, as a gift to the BSD
operating system, using 4- 3- 2- BSD licence. That's it, name
me in the sources.
Thank you, FreeBSD !
You are a great Unix operating system !

What are the applications? Blocking SIGKILL is pretty extreme.
-Alan

[Warm White Wolf]

Seriously : suppose your well crafted, eventually audited program, is more
important than
your whole informatic system, perhaps the raison d'etre of your business.
You want to live,
more than the 1% errors of your sysadmins.

Ludic : you have an account in the Unix machines at your university. You
wrote your small
HTTPD, and you want that your sysadmins won't kill your power-httpd.

fork(void) was pretty extremy at it's time...

As a power Unix-user, a config option for the kernel, at compile-time, can
be provided...

Post by Alan Somers

Post by Warm White Wolf
Greetings !
I have developed a new system call, be it named angel(2),
on Linux operating system (this is what I know), which makes
a program invulnerable to kill(2) calls, including SIGKILL and
SIGSTOP.
The uses may involve fork() + angel(), daemon() + angel(),
setsid() + angel(), exec*() + angel().
Use the intellectual property I give you, as a gift to the BSD
operating system, using 4- 3- 2- BSD licence. That's it, name
me in the sources.
Thank you, FreeBSD !
You are a great Unix operating system !

What are the applications? Blocking SIGKILL is pretty extreme.
-Alan

[Warm White Wolf]

You know what ? It is *my* work. I do not accept negative comments, may it
be practical my work, or not. Like Einstein's theory on E=mc^2, the bad
guys developed bombs, not wonderful purposes. Take a timeout, and come with
positive uses. Bye.

Interesting, next time someone inject a bitcoin miner in a jail running an
untrusted php system you'll need a reboot to stop it...

Post by Warm White Wolf
Seriously : suppose your well crafted, eventually audited program, is more
important than
your whole informatic system, perhaps the raison d'etre of your business.
You want to live,
more than the 1% errors of your sysadmins.
Ludic : you have an account in the Unix machines at your university. You
wrote your small
HTTPD, and you want that your sysadmins won't kill your power-httpd.
fork(void) was pretty extremy at it's time...
As a power Unix-user, a config option for the kernel, at compile-time, can
be provided...

On Tue, Aug 28, 2018 at 2:59 AM Warm White Wolf <

Post by Warm White Wolf
Greetings !
I have developed a new system call, be it named angel(2),
on Linux operating system (this is what I know), which makes
a program invulnerable to kill(2) calls, including SIGKILL and
SIGSTOP.
The uses may involve fork() + angel(), daemon() + angel(),
setsid() + angel(), exec*() + angel().
Use the intellectual property I give you, as a gift to the BSD
operating system, using 4- 3- 2- BSD licence. That's it, name
me in the sources.
Thank you, FreeBSD !
You are a great Unix operating system !

What are the applications? Blocking SIGKILL is pretty extreme.
-Alan

_______________________________________________
https://lists.freebsd.org/mailman/listinfo/freebsd-hackers
"

[Eugene Grosbein]

Post by Warm White Wolf
Seriously : suppose your well crafted, eventually audited program, is more
important than
your whole informatic system, perhaps the raison d'etre of your business.
You want to live,
more than the 1% errors of your sysadmins.
Ludic : you have an account in the Unix machines at your university. You
wrote your small
HTTPD, and you want that your sysadmins won't kill your power-httpd.
fork(void) was pretty extremy at it's time...
As a power Unix-user, a config option for the kernel, at compile-time, can
be provided...

This does not protect a process from power outage. And this does not (and should not)
protect a process from system shutdown/reboot to install critical updates, for example.

Well crafted programm must be designed to be ready to such events,
be able to recover and continue execution after restart.

And well crafted operating systems do not kill processes with SIGKILL easily,
they use other means to inform processes to terminate execution, so is FreeBSD.

[Ian Lepore]

Post by Eugene Grosbein

Post by Warm White Wolf
Seriously : suppose your well crafted, eventually audited program, is more
important than
your whole informatic system, perhaps the raison d'etre of your business.
You want to live,
more than the 1% errors of your sysadmins.
Ludic : you have an account in the Unix machines at your
university. You
wrote your small
HTTPD, and you want that your sysadmins won't kill your power-
httpd.
fork(void) was pretty extremy at it's time...
As a power Unix-user, a config option for the kernel, at compile-
time, can
be provided...

This does not protect a process from power outage. And this does not (and should not)
protect a process from system shutdown/reboot to install critical updates, for example.
Well crafted programm must be designed to be ready to such events,
be able to recover and continue execution after restart.
And well crafted operating systems do not kill processes with SIGKILL easily,
they use other means to inform processes to terminate execution, so is FreeBSD.

Why are you feeding this obvious troll?

-- Ian

[Achim Patzner]

Post by Warm White Wolf
Ludic : you have an account in the Unix machines at your university. You
wrote your small HTTPD, and you want that your sysadmins won't kill
your power-httpd.

That's exactly their job. If I found soome user on one of my system
running software I cannot stop if I need to the next thing that would be
gone after the reboot was their account...


Achim

[Warm White Wolf]

I was thinking more in the style of Stevens' UNP servers and clients,
that were configurable to run on every > 1024 / < 1024 port.
Exempli gratia : 8888, 8080, etc.

Post by Achim Patzner

Post by Warm White Wolf
Ludic : you have an account in the Unix machines at your university. You
wrote your small HTTPD, and you want that your sysadmins won't kill
your power-httpd.

That's exactly their job. If I found soome user on one of my system
running software I cannot stop if I need to the next thing that would be
gone after the reboot was their account...
Achim
_______________________________________________
https://lists.freebsd.org/mailman/listinfo/freebsd-hackers