1 |
Are there any numbers (benchmarks) about the performance penalty of |
2 |
pageexec and/or segmexec on intel x86 machines? |
3 |
|
4 |
The idea that I have is that page-exec on x86 involves a page-fault |
5 |
for every (execute) access to a new page that will be treated by |
6 |
pax... and that is performance-wise .. bad.. |
7 |
|
8 |
And that segmexec is a diferent approach that involves, mirroring the |
9 |
process address space on two segments with diferent "write" |
10 |
permissions, and compairing those two, to check if there was any |
11 |
overwrite of the code segment. |
12 |
This would mean doubling the mem-usage, at least for the code-segment |
13 |
in segmexec mode. |
14 |
|
15 |
I have the idea that segmexec is advised for being faster (on x86), |
16 |
but I don't have any numbers, and I was trying to understand the |
17 |
performance-wise consequences of each implementation. |
18 |
|
19 |
And in arches that suport no-exec pages (has sparc or amd64), what are |
20 |
the performance penalties? Anyone can give me some pointers? |
21 |
|
22 |
stuff like: kernel compiles, mysql benches, or... any other benchmark |
23 |
is good for me.. just to "grasp" a idea.. |
24 |
|
25 |
|
26 |
Best regards, |
27 |
|
28 |
|
29 |
-- |
30 |
Miguel Sousa Filipe |
31 |
|
32 |
-- |
33 |
gentoo-hardened@g.o mailing list |