Skip to contentSkip to conversation

The price of SELinux (CPU)

DISCUSSION John Richard Moser 6 participants 24 messages Oct 4, 2005 – Oct 5, 2005

lore
Share & tools
lore
#0 lore raw

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've heard that SELinux has produced benchmarks such as 7% increased CPU load. Is this true and current? Is it dependent on policy? What is the policy lookup complexity ( O(1), O(n), O(nlogn)...)? Are there other places where a bottleneck may exist aside from gruffing with the policy? Isn't the policy actually in xattrs so it's O(1)? Where else would an overhead that big come from aside from a lookup in a table?

....

Why is the sky blue? Why do you have a mustach? Why doesn't mommy have one? Does she shave it?

At any rate, my personal end goal is a secure high-performance operating system, as user friendly as Ubuntu, Mandriva, or Win----. To this end, I'm (still; a lot of you have seen me before) evaluating the performance hit of various user and kernel security enhancements like PaX, ProPolice, various OpenWall/GrSecurity niceness that needs to be divided out, and of course LSM/SELinux. Also wondering about that PHKMalloc thing on openbsd; is it really all that, is it junk, how's it compare to the recent ptmalloc work, and can it run on Linux for direct benching . . . but that's off topic.

- --
All content of all messages exchanged herein are left in the Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited resource. They shouldn't be wasted on re-inventing the wheel when there are so many fascinating new problems waiting out there.
-- Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQgT4hDd4aOud5P8RAoWBAJ0foEe4JcqDDlb7mMXQ6Z6FjCFjLACfdmJz +j2lCH7DpTlZK6zUztldEGI=
=RzhA
-----END PGP SIGNATURE-----

#1 lore raw

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

daniel

Quoted 45 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've heard that SELinux has produced benchmarks such as 7% increased CPU
load. Is this true and current? Is it dependent on policy? What is
the policy lookup complexity ( O(1), O(n), O(nlogn)...)? Are there
other places where a bottleneck may exist aside from gruffing with the
policy? Isn't the policy actually in xattrs so it's O(1)? Where else
would an overhead that big come from aside from a lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

At any rate, my personal end goal is a secure high-performance operating
system, as user friendly as Ubuntu, Mandriva, or Win----. To this end,
I'm (still; a lot of you have seen me before) evaluating the performance
hit of various user and kernel security enhancements like PaX,
ProPolice, various OpenWall/GrSecurity niceness that needs to be divided
out, and of course LSM/SELinux. Also wondering about that PHKMalloc
thing on openbsd; is it really all that, is it junk, how's it compare to
the recent ptmalloc work, and can it run on Linux for direct benching .
. . but that's off topic.

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 13 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird -
http://enigmail.mozdev.org

iD8DBQFDQgT4hDd4aOud5P8RAoWBAJ0foEe4JcqDDlb7mMXQ6Z6FjCFjLACfdmJz

Quoted 11 lines

+j2lCH7DpTlZK6zUztldEGI=
=RzhA
-----END PGP SIGNATURE-----
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#2 lore raw

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an abortionist; if I hear something has an ugly side, I try to find out if it can be fixed, and if the trade-off is worth getting rid of it. SELinux and LSM are quite useful you know; the overhead is probably not even that significant on the desktop to gamers (although if you TELL them about it they'll piss themselves), from a practical viewpoint considering their excessive hardware.

Quoted 55 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

daniel

--- John Richard Moser <[email protected]> wrote:

I've heard that SELinux has produced benchmarks such as 7% increased CPU
load. Is this true and current? Is it dependent on policy? What is
the policy lookup complexity ( O(1), O(n), O(nlogn)...)? Are there
other places where a bottleneck may exist aside from gruffing with the
policy? Isn't the policy actually in xattrs so it's O(1)? Where else
would an overhead that big come from aside from a lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

At any rate, my personal end goal is a secure high-performance operating
system, as user friendly as Ubuntu, Mandriva, or Win----. To this end,
I'm (still; a lot of you have seen me before) evaluating the performance
hit of various user and kernel security enhancements like PaX,
ProPolice, various OpenWall/GrSecurity niceness that needs to be divided
out, and of course LSM/SELinux. Also wondering about that PHKMalloc
thing on openbsd; is it really all that, is it junk, how's it compare to
the recent ptmalloc work, and can it run on Linux for direct benching .
. . but that's off topic.

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 7 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited resource. They shouldn't be wasted on re-inventing the wheel when there are so many fascinating new problems waiting out there.
-- Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQgwahDd4aOud5P8RArHEAJ9GFTpKPX3BbAR9vF/UCxeqbXO8DQCgi3sC R8bKVy1wxP2SiGJyc0MB4Xw=
=vvMx
-----END PGP SIGNATURE-----

#3 lore raw

i suggested you to disable selinux in order to have something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

Quoted 15 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off is worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop to gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 7 lines
Quoted 5 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

Quoted 85 lines
Quoted 6 lines

daniel

--- John Richard Moser <[email protected]>

Quoted 3 lines · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 3 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 13 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird -
http://enigmail.mozdev.org

iD8DBQFDQgwahDd4aOud5P8RArHEAJ9GFTpKPX3BbAR9vF/UCxeqbXO8DQCgi3sC

Quoted 11 lines

R8bKVy1wxP2SiGJyc0MB4Xw=
=vvMx
-----END PGP SIGNATURE-----
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#4 lore raw

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an expert in this kind of stuff. I wonder where the numbers come from; i.e. is 7% from policy? A O(1) policy lookup would be immune to big policies; a O(n) would probably not have that much impact from a typical policy lookup. Still perhaps interpreting the policy is a chore in itself, which still says bigger policy means bigger hit. Or is 7% constant?

I don't know what the frame of reference is or was. I'm sure with selinux with no policy it's rather 0ish; what I don't know is what I'm supposed to be looking at for benchmarking. Just randomly turning SELinux on and off and looking might give me an invalid measure.

Quoted 136 lines · Dan C Marinescu

i suggested you to disable selinux in order to have something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

--- John Richard Moser <[email protected]> wrote:

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off is worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop to gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 13 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

daniel

--- John Richard Moser <[email protected]>

Quoted 1 line · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 2 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 7 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited resource. They shouldn't be wasted on re-inventing the wheel when there are so many fascinating new problems waiting out there.
-- Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQh9FhDd4aOud5P8RAt30AJ9Tj2VZJwWh8EfzPocOcTkAIY/kOACfe03m wwtaci0G/aXWXok9NiWJR8E=
=78tr
-----END PGP SIGNATURE-----

#5 lore raw

for me, that 7% is about the price of tea in china... :-)
have no clue...

d

Quoted 39 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an expert in this kind of stuff. I wonder where the numbers
come from; i.e. is 7% from policy? A O(1) policy lookup would be immune
to big policies; a O(n) would probably not have that much impact from a
typical policy lookup. Still perhaps interpreting the policy is a chore
in itself, which still says bigger policy means bigger hit. Or is 7%
constant?

I don't know what the frame of reference is or was. I'm sure with
selinux with no policy it's rather 0ish; what I don't know is what I'm
supposed to be looking at for benchmarking. Just randomly turning
SELinux on and off and looking might give me an invalid measure.

Quoted 1 line · Dan C Marinescu

i suggested you to disable selinux in order to

have

Quoted 6 lines

something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

--- John Richard Moser <[email protected]>

Quoted 5 lines · wrote:

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off

Quoted 21 lines

is

Quoted 4 lines

worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop

to

Quoted 14 lines

gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 7 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

Quoted 133 lines
Quoted 101 lines
Quoted 6 lines

daniel

--- John Richard Moser <[email protected]>

Quoted 1 line · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 2 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 3 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 9 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----

=== message truncated ===

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#6 lore raw
In reply to John Richard Moser

my answer is "does not compute, insuficient data"

this is like saying "light travels fast but there is a 7% trajectory distortion in certain conditions"...

*** it _depends_ on so many factors... the answer _may_ be somewhere in fs, look at the source... etc...

i would rathere doubt the source... (tell me it's
slashdot or msdn or something :-)

d

Quoted 39 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an expert in this kind of stuff. I wonder where the numbers
come from; i.e. is 7% from policy? A O(1) policy lookup would be immune
to big policies; a O(n) would probably not have that much impact from a
typical policy lookup. Still perhaps interpreting the policy is a chore
in itself, which still says bigger policy means bigger hit. Or is 7%
constant?

I don't know what the frame of reference is or was. I'm sure with
selinux with no policy it's rather 0ish; what I don't know is what I'm
supposed to be looking at for benchmarking. Just randomly turning
SELinux on and off and looking might give me an invalid measure.

Quoted 1 line · Dan C Marinescu

i suggested you to disable selinux in order to

have

Quoted 6 lines

something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

--- John Richard Moser <[email protected]>

Quoted 5 lines · wrote:

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off

Quoted 21 lines

is

Quoted 4 lines

worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop

to

Quoted 14 lines

gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 7 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

Quoted 133 lines
Quoted 101 lines
Quoted 6 lines

daniel

--- John Richard Moser <[email protected]>

Quoted 1 line · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 2 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 3 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 9 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----

=== message truncated ===

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#7 lore raw
In reply to John Richard Moser

John,

Try this link:

http://www.nsa.gov/selinux/list-archive/0505/11459.cfm

It's from NSA... Hope this helps somehow...

Daniel

Quoted 39 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an expert in this kind of stuff. I wonder where the numbers
come from; i.e. is 7% from policy? A O(1) policy lookup would be immune
to big policies; a O(n) would probably not have that much impact from a
typical policy lookup. Still perhaps interpreting the policy is a chore
in itself, which still says bigger policy means bigger hit. Or is 7%
constant?

I don't know what the frame of reference is or was. I'm sure with
selinux with no policy it's rather 0ish; what I don't know is what I'm
supposed to be looking at for benchmarking. Just randomly turning
SELinux on and off and looking might give me an invalid measure.

Quoted 1 line · Dan C Marinescu

i suggested you to disable selinux in order to

have

Quoted 6 lines

something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

--- John Richard Moser <[email protected]>

Quoted 5 lines · wrote:

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off

Quoted 21 lines

is

Quoted 4 lines

worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop

to

Quoted 14 lines

gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 7 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

Quoted 133 lines
Quoted 101 lines
Quoted 6 lines

daniel

--- John Richard Moser <[email protected]>

Quoted 1 line · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 2 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 3 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 9 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----

=== message truncated ===

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#8 lore raw
Quoted 7 lines · Dan C Marinescu

John,

Try this link:

http://www.nsa.gov/selinux/list-archive/0505/11459.cfm

It's from NSA... Hope this helps somehow...

Note that these benchmarks were done quite some time ago. Among many other changes, selinux's memory overhead was recently significantly reduced which should help these benchmarks quite a bit.

Hopefully we can do another round of benchmarks soon, but we were waiting for particular machine to become available.

-serge

#9 lore raw
In reply to John Richard Moser

the following link:

http://www.nsa.gov/selinux/list-archive/0505/11459.cfm

describes _some_ particular cases... among the over 100,000,000,000 factors involved are:

1. fs type (block-size, etc) // not sure on this one :-(
2. number of processors
3. nature of processes // they are relatively vague on this one
4. kernel configuration (they mentioned about 2.6.12-rc2 or something) but then again, they didn't post their kernel .config (maybe it's a fedora stock .config, maybe not...)
...
...
...
etc

And even then, to draw the line at 7% sounds more like a quick and dirty conclusion (@ least 2 me)...

d

Quoted 39 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an expert in this kind of stuff. I wonder where the numbers
come from; i.e. is 7% from policy? A O(1) policy lookup would be immune
to big policies; a O(n) would probably not have that much impact from a
typical policy lookup. Still perhaps interpreting the policy is a chore
in itself, which still says bigger policy means bigger hit. Or is 7%
constant?

I don't know what the frame of reference is or was. I'm sure with
selinux with no policy it's rather 0ish; what I don't know is what I'm
supposed to be looking at for benchmarking. Just randomly turning
SELinux on and off and looking might give me an invalid measure.

Quoted 1 line · Dan C Marinescu

i suggested you to disable selinux in order to

have

Quoted 6 lines

something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

--- John Richard Moser <[email protected]>

Quoted 5 lines · wrote:

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off

Quoted 21 lines

is

Quoted 4 lines

worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop

to

Quoted 14 lines

gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 7 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

Quoted 133 lines
Quoted 101 lines
Quoted 6 lines

daniel

--- John Richard Moser <[email protected]>

Quoted 1 line · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 2 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 3 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 9 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----

=== message truncated ===

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#10 lore raw
In reply to John Richard Moser

the benchmark "results" _look_ like being authored by some qa engineers... or sysadmins or something...

*** only a deep/intimate knowledge of kernel and fs and acl implementation details and many other areas could suggest a credible conclusion (most likely without even needing any "profiling" at all... on pure theoretical basis, mostly because you would know what goes where and when and how and why and how much it adds here and there, etc, etc, etc)

and i personally have a strong doubt that if the cpu activity was statistically increased with 7% for the very same elementary I/O, linus would have accepted this degradation... my $0.02... :-)

d

Quoted 39 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm not an expert in this kind of stuff. I wonder where the numbers
come from; i.e. is 7% from policy? A O(1) policy lookup would be immune
to big policies; a O(n) would probably not have that much impact from a
typical policy lookup. Still perhaps interpreting the policy is a chore
in itself, which still says bigger policy means bigger hit. Or is 7%
constant?

I don't know what the frame of reference is or was. I'm sure with
selinux with no policy it's rather 0ish; what I don't know is what I'm
supposed to be looking at for benchmarking. Just randomly turning
SELinux on and off and looking might give me an invalid measure.

Quoted 1 line · Dan C Marinescu

i suggested you to disable selinux in order to

have

Quoted 6 lines

something to compare to... (engineers compare, measure, instead of believing in rummors...)

d

--- John Richard Moser <[email protected]>

Quoted 5 lines · wrote:

I'm not an abortionist; if I hear something has an ugly side, I try to
find out if it can be fixed, and if the trade-off

Quoted 21 lines

is

Quoted 4 lines

worth getting rid
of it. SELinux and LSM are quite useful you know; the overhead is
probably not even that significant on the desktop

to

Quoted 14 lines

gamers (although if
you TELL them about it they'll piss themselves), from a practical
viewpoint considering their excessive hardware.

Quoted 7 lines · Dan C Marinescu

try selinux=0, _if u feel that way :-)

about big o:

http://www.maththinking.com/boat/compsciBooksIndex.html

Quoted 133 lines
Quoted 101 lines
Quoted 6 lines

daniel

--- John Richard Moser <[email protected]>

Quoted 1 line · wrote:

I've heard that SELinux has produced benchmarks

such

Quoted 2 lines

as 7% increased CPU
load. Is this true and current? Is it dependent

on

Quoted 4 lines

policy? What is
the policy lookup complexity ( O(1), O(n),
O(nlogn)...)? Are there
other places where a bottleneck may exist aside

from

Quoted 2 lines

gruffing with the
policy? Isn't the policy actually in xattrs so

it's

Quoted 10 lines

O(1)? Where else
would an overhead that big come from aside from a
lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

Quoted 7 lines

At any rate, my personal end goal is a secure
high-performance operating
system, as user friendly as Ubuntu, Mandriva, or
Win----. To this end,
I'm (still; a lot of you have seen me before)
evaluating the performance
hit of various user and kernel security

enhancements

Quoted 2 lines

like PaX,
ProPolice, various OpenWall/GrSecurity niceness

that

Quoted 4 lines

needs to be divided
out, and of course LSM/SELinux. Also wondering
about that PHKMalloc
thing on openbsd; is it really all that, is it

junk,

Quoted 5 lines

how's it compare to
the recent ptmalloc work, and can it run on Linux
for direct benching .
. . but that's off topic.

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 2 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

Quoted 8 lines

--
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there

are

Quoted 3 lines

so many fascinating
new problems waiting out there.

--

Quoted 1 line

Eric Steven Raymond

- -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 3 lines

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 9 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----

=== message truncated ===

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#11 lore raw
Quoted 10 lines · Dan C Marinescu

the benchmark "results" _look_ like being authored by some qa engineers... or sysadmins or something...

*** only a deep/intimate knowledge of kernel and fs and acl implementation details and many other areas could suggest a credible conclusion (most likely without even needing any "profiling" at all... on pure theoretical basis, mostly because you would know what goes where and when and how and why and how much it adds here and there, etc, etc, etc)

Any results not based on actual measurement are called "guesses" rather than "data." Such deep knowlege is useful to determine what to measure, not what you would measure if you thought it were necessary.

The measurements are very useful, in that they show the magnitude of the performance impact using a benchmark which was constructed to emulate certain real world loads. Since no one number or even series of numbers can fully describe what *will* happen, but these numbers show what *could* happen.

Quoted 5 lines

and i personally have a strong doubt that if the cpu activity was statistically increased with 7% for the very same elementary I/O, linus would have accepted this degradation... my $0.02... :-)

For some applications the issue isn't how fast the O/S runs, but if it is secure enough to be run at all. Given the speed of even commodity computers, it's probable that even a 2:1 slowdown would still result in useful operation, compared to doing the work without a computer.

I can't speak for Linus' thinking of course, but I have worked in secure environments before, both DOD and DOE, and information control is vital.

Signature · 3 lines
   -bill davidsen ([email protected])
"The secret to procrastination is to put things off until the
 last possible moment - but no longer"  -me
#12 lore raw
Quoted 6 lines

Any results not based on actual measurement are
called "guesses" rather
than "data." Such deep knowlege is useful to
determine what to measure,
not what you would measure if you thought it were
necessary.

in my world, we design stuff, calculate bigO, then implement and finally measure. we don't write code at inspiration and then measure and if kinda s*x we apply patches then measure again, etc etc etc, that's _all_ i meant... (keep measuring something not really designed for this O or that O may be a huge waste of time). even in qa, the best engineers find short-cuts... even in black box testing, but then you get gray box, and white box testing and finally us (r&d).

Quoted 9 lines

The measurements are very useful, in that they show the magnitude of the
performance impact using a benchmark which was constructed to emulate
certain real world loads. Since no one number or even series of numbers
can fully describe what *will* happen, but these numbers show what
*could* happen.

correct, but you should measure something which you first designed then implemented... not the other way around...

Quoted 7 lines
Quoted 1 line

very same elementary I/O, linus would have

accepted

Quoted 1 line

this degradation... my $0.02... :-)

For some applications the issue isn't how fast the O/S runs, but if it
is secure enough to be run at all. Given the speed

hey, absolutely... i was about to add that too (last night) but it was kinda late... but how would you define "secure"... it's kinda like huge, eh? in my $0.02, secure means secure enough for the purpose (whatever that may be...) and it's way over the scope of this email... when this is equivalated with "oh, it's kinda slow, but it's worth cause it's safer"... well, i kinda have doubts on that... (check for a second oppinion in your sec strategy, etc...)

Quoted 5 lines

of even commodity
computers, it's probable that even a 2:1 slowdown would still result in
useful operation, compared to doing the work without a computer.

well, yes and now... it's a long story...

Quoted 4 lines

I can't speak for Linus' thinking of course, but I have worked in secure
environments before, both DOD and DOE, and information control is vital.

yeah... remember the old days (running around with floppies because networking was "unsafe" blah blah blah...)

nice talking 2u bill,
d

Quoted 7 lines
Signature · 5 lines
-- 
    -bill davidsen ([email protected])
"The secret to procrastination is to put things off
until the
  last possible moment - but no longer"  -me

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#13 lore raw
In reply to John Richard Moser

Hi John,

Don't buy that 7% increased CPU (can easily verify it...) start the kernel with selinux=0 (totally disable selinux) and compare the results for yourself...

Now about big_o... In two words, big O is a way of describing the performance of an algorithm. If a system has 2 deal with n steps it is said to be:
*constant O(1) if n doesn't affect the total run_time of that system (eats the same amount of time regardless n). O(n) is called linear (total computation time is a linear dependency of n, that is if it took 3 secs when n = 3, it would take 11 secs when n = 11. And so on... (detail: in case of a polinomial, only the highest power matters!) of course, the lower the better! I have __big__ doubts that NSA implemented something higher than linear... (I suspect that their folks go from O(1) to O(ln(n)) // quality work... Anyway, if O(n) is somehow acceptable for certain algorithms, O(n to power 2, 3,
n) are to be avoided at all cost! (see the widowz “kernel” live example of quadratic micro-kernels ;-)

*** The perfect case is not (yet) defined in general theory of computation. That would be O(0) when a system performs an infinite number of elementary computations in ZERO seconds :-)
*** Same about O(infinite) // obviously the worst case, when a (super lazy) system needs an eternity to do... nothing! :-) // see frequent blue screens on costco purchased personal computers :-)

The smartest the author, the lower the O! In user_land O(n) is considered acceptable in most cases...

So, in two words, simply put, fast is good :-) && slow is bad :-(

Daniel

Quoted 45 lines · --- John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've heard that SELinux has produced benchmarks such as 7% increased CPU
load. Is this true and current? Is it dependent on policy? What is
the policy lookup complexity ( O(1), O(n), O(nlogn)...)? Are there
other places where a bottleneck may exist aside from gruffing with the
policy? Isn't the policy actually in xattrs so it's O(1)? Where else
would an overhead that big come from aside from a lookup in a table?

....

Why is the sky blue? Why do you have a mustach?
Why doesn't mommy have
one? Does she shave it?

At any rate, my personal end goal is a secure high-performance operating
system, as user friendly as Ubuntu, Mandriva, or Win----. To this end,
I'm (still; a lot of you have seen me before) evaluating the performance
hit of various user and kernel security enhancements like PaX,
ProPolice, various OpenWall/GrSecurity niceness that needs to be divided
out, and of course LSM/SELinux. Also wondering about that PHKMalloc
thing on openbsd; is it really all that, is it junk, how's it compare to
the recent ptmalloc work, and can it run on Linux for direct benching .
. . but that's off topic.

- --
All content of all messages exchanged herein are left in the
Public Domain, unless otherwise explicitly stated.

Quoted 13 lines

Creative brains are a valuable, limited
resource. They shouldn't be
wasted on re-inventing the wheel when there are
so many fascinating
new problems waiting out there.
--
Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird -
http://enigmail.mozdev.org

iD8DBQFDQgT4hDd4aOud5P8RAoWBAJ0foEe4JcqDDlb7mMXQ6Z6FjCFjLACfdmJz

Quoted 11 lines

+j2lCH7DpTlZK6zUztldEGI=
=RzhA
-----END PGP SIGNATURE-----
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected] More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#14 lore raw
In reply to John Richard Moser
Quoted 5 lines · John Richard Moser

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've heard that SELinux has produced benchmarks such as 7% increased CPU load.

The overall performance hit across several micro and macro benchmarks, when last measured last year sometime, was around 7%, depending on workload and what you were testing. It's a very rough figure and any serious benchmarking needs to be done for the intended workload.

The AVC is now linearly scalable (measured up to 32 processors) thanks to RCU and work by NEC.

Quoted 5 lines

Is this true and current? Is it dependent on policy? What is
the policy lookup complexity ( O(1), O(n), O(nlogn)...)? Are there other places where a bottleneck may exist aside from gruffing with the policy? Isn't the policy actually in xattrs so it's O(1)? Where else would an overhead that big come from aside from a lookup in a table?

The overhead is generally independent of policy size, as policy is cached in the AVC and most workloads use a trivial number of policy rules in a steady state (often less than 20).

So, generally, you'll only have a very small number of AVC entries active, although you could have some longish hash chains if policy has not been reloaded since boot.

Look in /selinux/avc for stats.

Googling for "selinux performance" will guide you to:
http://www.livejournal.com/users/james_morris/2153.html

- James

Signature · 2 lines
James Morris
<[email protected]>
#15 lore raw
In reply to John Richard Moser
Quoted 2 lines · John Richard Moser

At any rate, my personal end goal is a secure high-performance operating system, as user friendly as

Step 0: Sooner or later, "secure" and "user friendly" *will* come into conflict. At that point, you have to make a choice. Note that in many cases, we *made* the choice years or even decades ago, and we've gotten used to the choices made. For instance, you'd certainly get better performance and user friendliness if you just stubbed out permission() in fs/namei.c and capable(), and just had them return "let the guy do it". But somehow, I don't think anybody would find that very palatable.

Similarly, the stuff that comes out of Redmond, in general, has security issues precisely because they chose "user friendly" when they got to Step 0. Being able to put Javascript and/or executable binaries in e-mail for automatic execution is certainly user-friendly - but it's not secure.

In any case, the overhead isn't 7%. If anything, it's probably closer to 0.7%, and dropping with each kernel release as the code gets tuned and optimized even more. And beware the impact of micro-optimizations and macro-performance - there was recently a code change to reduce the number and size of avtab entries. That slowed down the actual code path slightly, but overall was actually a performance win, especially on smaller memory-constrained machines, due to the drastic drop in overall slab consumption.

And remember - the first time that a security system prevents (for example) an exploit against an Apache bug from being used to take over a system, it's paid for itself. When the FBI faxes you that "Hold Evidence" order, it means you may not be seeing that server again for weeks, if ever.....

#16 lore raw

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 8 lines · [email protected] wrote:
Quoted 2 lines · John Richard Moser

At any rate, my personal end goal is a secure high-performance operating system, as user friendly as

Step 0: Sooner or later, "secure" and "user friendly" *will* come into conflict.

It's a lot later than you think. A home desktop OS isn't a server OS; and a server OS isn't a home desktop OS. That being said, something doesn't have to be as wide-open as the goatse guy's ass to be suitable for every tom dick and moron.

Quoted 7 lines

At that point, you have to make a choice. Note that in many cases, we *made* the choice years or even decades ago, and we've gotten used to the choices made. For instance, you'd certainly get better performance and user friendliness if you just stubbed out permission() in fs/namei.c and capable(), and just had them return "let the guy do it". But somehow, I don't think anybody would find that very palatable.

Would you now? The performance gain would be negligible, even if it were there; the user friendly factor would be pretty nil. I mean the user now can administrate his system without entering a password before hitting the configuration center -- which he does every several weeks if that.

Aside from this, viruses and spyware and worms can now run rampant and do what they want to his system, and other users' idiotic actions on a multi-user system affect him. This is more user friendly? No, I think it's going in the opposite direction. . . .

The choice was made at Windows with "Let the desktop user run as administrator;" in Linux it's typically made at "We've designed an OS that runs very, very well with your account limited." They're both roughly equivalent in terms of user friendly (I think linux with Gnome or KDE is actually easier, so does my mom).

Quoted 5 lines

Similarly, the stuff that comes out of Redmond, in general, has security issues precisely because they chose "user friendly" when they got to Step 0. Being able to put Javascript and/or executable binaries in e-mail for automatic execution is certainly user-friendly - but it's not secure.

There's "user friendly," and then there's just ass. Switzerland gives each and every child a rifle and trains them to use it at age 12 IIRC; this would be "user friendly." Now, if you want to be just ass, hand every 4 year old a gun with live ammunition and wait for them to put a bullet in someone's brain and learn on their own that you shouldn't shoot people unless you really mean it.

Open source programs achieve "user friendly" in a responsible manner. Firefox doesn't have a local machine zone with javascript able to write to files directly; thunderbird doesn't auto-run certain scripts; the file browser isn't integrated into the web browser in anything but KDE (which I personally dislike for other reasons).

Quoted 7 lines

In any case, the overhead isn't 7%. If anything, it's probably closer to 0.7%, and dropping with each kernel release as the code gets tuned and optimized even more. And beware the impact of micro-optimizations and macro-performance - there was recently a code change to reduce the number and size of avtab entries. That slowed down the actual code path slightly, but overall was actually a performance win, especially on smaller memory-constrained machines, due to the drastic drop in overall slab consumption.

Nice. Some IBM guys said they're gonna rebench soon so I'm looking forward to that, but this is reassuring.

Quoted 5 lines

And remember - the first time that a security system prevents (for example) an exploit against an Apache bug from being used to take over a system, it's paid for itself. When the FBI faxes you that "Hold Evidence" order, it means you may not be seeing that server again for weeks, if ever.....

- --
All content of all messages exchanged herein are left in the Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited resource. They shouldn't be wasted on re-inventing the wheel when there are so many fascinating new problems waiting out there.
-- Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQsnwhDd4aOud5P8RAmmWAJ9JJquzIPzjVlm5w0OxrBAwOJP6gwCeOHYv sVpFxYCDZvKbhUOq86dqog4=
=i9Fd
-----END PGP SIGNATURE-----

#17 lore raw
Quoted 4 lines · John Richard Moser

Aside from this, viruses and spyware and worms can now run rampant and do what they want to his system, and other users' idiotic actions on a multi-user system affect him. This is more user friendly? No, I think it's going in the opposite direction. . . .

Virus writers are users too, you know. :)

And the other users are users as well - what if the other user's "idiotic action" is to nuke your 500Mbyte archive of alt.binaries.pictures.llama.sex that's taking up the disk space that is keeping him from running the payroll software? In your world, rather than him being able to fix the problem, he has to go find a sysadmin with the root password to fix it, causing delays and being less friendly....

You seem to be intentionally trying to miss the basic point, which is that any additional security ends up trading off against other things.

Non-execute stack is a Good Thing security-wise - but it breaks some code, forcing upgrades and/or having to track down binaries and flag them as "don't enforce NX stack". And then those binaries are still vulnerable....

SELinux is, in general, also a Good Thing. However, the fact that the policy restricts what stuff can happen in the security context associated with mail delivery (after all, you *don't* want arbitrary binaries running then, right?) did some serious damage to the way I use procmail, which in some cases ended up running other binaries. OK, so my .procmailrc *is* a 600-line monster that does a lot of odd stuff - the point was that I had to add even *more* contortions to the way it works, which is even less user-friendly....

#18 lore raw

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 9 lines · [email protected] wrote:
Quoted 4 lines · John Richard Moser

Aside from this, viruses and spyware and worms can now run rampant and do what they want to his system, and other users' idiotic actions on a multi-user system affect him. This is more user friendly? No, I think it's going in the opposite direction. . . .

Quoted 9 lines

Virus writers are users too, you know. :)

And the other users are users as well - what if the other user's "idiotic action" is to nuke your 500Mbyte archive of alt.binaries.pictures.llama.sex that's taking up the disk space that is keeping him from running the payroll software? In your world, rather than him being able to fix the problem, he has to go find a sysadmin with the root password to fix it, causing delays and being less friendly....

Oh sure, except that. . .

1)  You shouldn't be screwing with the payroll system
2)  You're quota'd on any good setup
3)  You're fired
Quoted 3 lines

You seem to be intentionally trying to miss the basic point, which is that any additional security ends up trading off against other things.

It does, but put the degree to which it trades off in perspective.

Utilizing ProPolice, functions which contain a local character pointer (char[]) are guarded. These functions experience a minimal performance hit; practically it can top-out at around 8% theoretical, although from my perspective you could produce an empty function where the guard code is 100% of its body. The 8% theoretical is along the lines of:

void foo() {
char a[6];
strcpy(a, "hello");
}

Obviously strcpy()'s complexity decreases the overhead; this is some strange hybrid of "theoretical maximum" with "practical maximum," which comes out to be "theoretical practical maximum" which is nonsense, but we'll just for the sake of discussion let that slide.

Anyway, a finite number of functions have such guard code; and the overhead can be shown as a function y=1/x, where y is the overhead and x is the number of units of code run between entering and exiting the function, assuming the code added by propolice is 1 unit. In the end, the overhead is pretty much nil for practical applications.

For your trade-off of some practical 0.1% increase in CPU load, your applications do two things when a stack buffer is overflowed. First, they tell you what function produced the buffer that was overflowed, in what source file; second, they immediately terminate (complain loudly).

This means that things that are fuzz may crash the program; but so will attacks. It also means that these crashes are easy to report in valuable detail, and thus the bugs are fixed rather quickly. In the end, this produces cleaner, more stable code due to attention being brought to bugs more directly. Additionally, such bugs would cause intermittent odd behavior ranging from things not working to program freezes to data corruption; a sudden crash in place of these things is hardly much of a trade-off. Finally, deploying such a protection doesn't suddenly unearth massive amounts of buggy code, because if there were such bugs in force then there'd be massive amounts of odd behavior already.

So you've traded nearly nothing in the practical sense for a great enhancement in security in this example. In some special applications, a program with its own bug will trigger this, which may be a more major trade-off; but you need access to the source code to add this protection to anything anyway, so in this case it's just going to tell you how to fix the bug anyway.

Quoted 4 lines

Non-execute stack is a Good Thing security-wise - but it breaks some code, forcing upgrades and/or having to track down binaries and flag them as "don't enforce NX stack". And then those binaries are still vulnerable....

Right. It breaks some (very little, typically) code. Perspective please.

Quoted 8 lines

SELinux is, in general, also a Good Thing. However, the fact that the policy restricts what stuff can happen in the security context associated with mail delivery (after all, you *don't* want arbitrary binaries running then, right?) did some serious damage to the way I use procmail, which in some cases ended up running other binaries. OK, so my .procmailrc *is* a 600-line monster that does a lot of odd stuff - the point was that I had to add even *more* contortions to the way it works, which is even less user-friendly....

Special case that would have likely never arisen if you had such restrictions when you started; or in the very least would have been solved more elegantly.

Quoted 1 line

In the end, massive, intrusive security is not exactly the best thing for security's sake; but anything you can get away with significantly cleanly (i.e. you don't break 99% of the applications on 99% of home users' desktops) is worth immediate focus for those who are so inclined.

- --
All content of all messages exchanged herein are left in the Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited resource. They shouldn't be wasted on re-inventing the wheel when there are so many fascinating new problems waiting out there.
-- Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQuGhhDd4aOud5P8RAtbhAJ9p22xB3KhPQ9iywk7ug6VbAgKFlQCeN9Yp si7fx6ngk4UU/H8KTNgeR0U=
=soXe
-----END PGP SIGNATURE-----

#19 lore raw
Quoted 11 lines · John Richard Moser
Quoted 6 lines

And the other users are users as well - what if the other user's "idiotic action" is to nuke your 500Mbyte archive of alt.binaries.pictures.llama.sex that's taking up the disk space that is keeping him from running the payroll software? In your world, rather than him being able to fix the problem, he has to go find a sysadmin with the root password to fix it, causing delays and being less friendly....

Oh sure, except that. . .

1) You shouldn't be screwing with the payroll system
2) You're quota'd on any good setup

Ahem. You're adding in more "user unfriendly" constraints again. :)

Quoted 4 lines

In the end, massive, intrusive security is not exactly the best thing for security's sake; but anything you can get away with significantly cleanly (i.e. you don't break 99% of the applications on 99% of home users' desktops) is worth immediate focus for those who are so inclined.

Good. Now hand me that crystal ball that lets us know for sure which of those two categories any given security measure falls into. How often do we see "this shouldn't break anything" patches on this list that do, in fact, manage to break something anyhow?

#20 lore raw
Quoted 7 lines

Good. Now hand me that crystal ball that lets us
know for sure which of
those two categories any given security measure
falls into. How often do
we see "this shouldn't break anything" patches on
this list that do, in fact,
manage to break something anyhow?

it seams to me that the crystal ball is that there is no crystal ball at all... and there shouldn't be... :-) // @ least in sec...

d

__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com

#21 lore raw
In reply to [email protected]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 26 lines · [email protected] wrote:
Quoted 8 lines · John Richard Moser
Quoted 6 lines

And the other users are users as well - what if the other user's "idiotic action" is to nuke your 500Mbyte archive of alt.binaries.pictures.llama.sex that's taking up the disk space that is keeping him from running the payroll software? In your world, rather than him being able to fix the problem, he has to go find a sysadmin with the root password to fix it, causing delays and being less friendly....

Quoted 4 lines

Oh sure, except that. . .

1) You shouldn't be screwing with the payroll system
2) You're quota'd on any good setup

Ahem. You're adding in more "user unfriendly" constraints again. :)

Quoted 4 lines

In the end, massive, intrusive security is not exactly the best thing for security's sake; but anything you can get away with significantly cleanly (i.e. you don't break 99% of the applications on 99% of home users' desktops) is worth immediate focus for those who are so inclined.

Quoted 4 lines

Good. Now hand me that crystal ball that lets us know for sure which of those two categories any given security measure falls into. How often do we see "this shouldn't break anything" patches on this list that do, in fact, manage to break something anyhow?

I've seen PaX break Xine, Java, nVidia's retarded binary OpenGL, and mplayer. Aside from that everything else ran rather smooth on my system, including X itself. To fix this of course was a policy adjustment that weakened the protections partially for each breaking thing. For a time I noticed native x86-64 Xine had a bug and would make bad assumptions, and died on a vanilla linux kernel because an area aside from stack/heap that defaulted to !PROT_EXEC was expected to be executable; I actually needed PaX so I could equate PROT_READ to PROT_EXEC for Xine.

When ASLR went into 2.6.13 there was talk about higher ASLR breaking Oracle; and a little bit about Emacs breaking at times. Wine also has iffy issues with it that get addressed by an early hack, which is the Wine devs being -quite- friendly. Linus mentioned something about mmap()ing a 2 gig email box into memory on IA-32. Aside from that, high-order ASLR shouldn't be a problem; and most of those issues vanish on 64-bit systems anyway.

Exec Shield breaks a few things, about as much as PaX; but the compiler was explicitly modified to turn Exec Shield off if the wind blows, so they started with a relatively wide number of binaries with ES disabled while RH tried to narrow it to essentials (a backwards approach if you ask me). You never see ES break stuff because of this.

Neither Spender nor me have seen the grsecurity/openwall "linking restrictions" break anything, although I've heard one person report on an obscure app not liking it for some strange reason. Aside from that, it's effective at keeping ahead of the game when it comes to /tmp file races.

I don't think grsecurity's banning of any write access to /dev/*mem except for the video memory area ever actually broke anything legitimate.

Position independent executables help ASLR such as from PaX or Exec Shield be more effective; if the program isn't going to work with it, it's flat out not going to compile. If it compiles, it works.

ProPolice's protections only break programs with bugs. To mis-fire ProPolice, you have to actually write past the end of a stack based buffer, ticking off the canary. Of course, it immediately tells you where the problem is so you can go fix your code.

I forsee PHKmalloc in OpenBSD, if it works as advertised, as doing the same things ProPolice does, but to the heap. This means slightly-buggy apps will fall victim to sudden death. Good riddance to bad rubbish; may the developers correct all exposed bugs with expedience.

- --
All content of all messages exchanged herein are left in the Public Domain, unless otherwise explicitly stated.

Creative brains are a valuable, limited resource. They shouldn't be wasted on re-inventing the wheel when there are so many fascinating new problems waiting out there.
-- Eric Steven Raymond
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQzQjhDd4aOud5P8RAtt7AJkBrPxss5dNMNrrSXQeQyZSDyr8OQCgjSVH 3aDc4HOcrLYRYLUGSxlPOFE=
=IZeX
-----END PGP SIGNATURE-----

#22 lore raw
In reply to [email protected]
Quoted 9 lines · [email protected] wrote:
Quoted 4 lines · John Richard Moser

Oh sure, except that. . .

1) You shouldn't be screwing with the payroll system
2) You're quota'd on any good setup

Ahem. You're adding in more "user unfriendly" constraints again. :)

I don't think promising a user that s/he can have N MB of disk is unfriendly at all.

Signature · 3 lines
   -bill davidsen ([email protected])
"The secret to procrastination is to put things off until the
 last possible moment - but no longer"  -me
#23 lore raw
In reply to [email protected]
Quoted 25 lines · [email protected] wrote:
Quoted 4 lines · John Richard Moser

Aside from this, viruses and spyware and worms can now run rampant and do what they want to his system, and other users' idiotic actions on a multi-user system affect him. This is more user friendly? No, I think it's going in the opposite direction. . . .

Virus writers are users too, you know. :)

And the other users are users as well - what if the other user's "idiotic action" is to nuke your 500Mbyte archive of alt.binaries.pictures.llama.sex that's taking up the disk space that is keeping him from running the payroll software? In your world, rather than him being able to fix the problem, he has to go find a sysadmin with the root password to fix it, causing delays and being less friendly....

You seem to be intentionally trying to miss the basic point, which is that any additional security ends up trading off against other things.

Non-execute stack is a Good Thing security-wise - but it breaks some code, forcing upgrades and/or having to track down binaries and flag them as "don't enforce NX stack". And then those binaries are still vulnerable....

Quoted 9 lines

SELinux is, in general, also a Good Thing. However, the fact that the policy restricts what stuff can happen in the security context associated with mail delivery (after all, you *don't* want arbitrary binaries running then, right?) did some serious damage to the way I use procmail, which in some cases ended up running other binaries. OK, so my .procmailrc *is* a 600-line monster that does a lot of odd stuff - the point was that I had to add even *more* contortions to the way it works, which is even less user-friendly....

Doesn't everyone have executables in their .procmailrc? Mine starts with a filter which may add one line to the mail header, quantifying exactly how badly it sucks. That's then used to take preemptive action against spam and other stuff I don't wnat or need to see.

That's a lot to give up.