1 |
> The problem is that a PAX_FLAGS program header is not |
2 |
> standard while user defined xattrs are. eg. EI_PAX used to put the |
3 |
> markers in the ELF header (in a non-standard way) until that got |
4 |
> clobbered by a commit in glibc; similarly pax markings in the program |
5 |
> header cause issue in cases like the above. XT_PAX has the advantage |
6 |
> of not violating a standard while the disadvantage of needing |
7 |
> end-to-end xattr support. |
8 |
|
9 |
I think the issue here is not, that a non-standard way is used. It is, |
10 |
that the best way to use PAX markings is not standardized. It should be |
11 |
added to the ELF standard (in a generalized way, so that other runtime |
12 |
exploit mitigation suites could use that standard too). |
13 |
|
14 |
I do not know, whether there are valid use cases for using PAX-marked |
15 |
binaries on non-xattr-capable file systems. It just feels like that |
16 |
markings belong more to the binary itself than to the reference to that |
17 |
binary in the file system. |
18 |
|
19 |
> I have no intentions of dropping PT_PAX anytime soon. toolchain did |
20 |
> indicate a desire to do so because the program header causes issues in |
21 |
> binutils' test suite, but dropping PT_AX is a long range plan if it |
22 |
> will ever happen. |
23 |
|
24 |
I am relieved to read that. So i do not have to rebuild the kernels yet |
25 |
(i may have configured Grsec to only use the markings in the ELF header |
26 |
- but i am not sure about that). |
27 |
|
28 |
> The current issue, in my opinion, is how to spead up the install |
29 |
> wrapper which is written in python and slow as hell. |
30 |
|
31 |
Has the bottleneck already been identified? Python should not be much |
32 |
slower than other languages for solving mostly IO-based problems. |
33 |
|
34 |
|
35 |
|
36 |
-- |
37 |
Allan Wegan |
38 |
Jabber: allanwegan@×××××.de |
39 |
ICQ: 209459114 |