1 |
On Tue, 31 Jul 2012 15:16:34 -0400 |
2 |
Rich Freeman <rich0@g.o> wrote: |
3 |
|
4 |
> On Tue, Jul 31, 2012 at 10:56 AM, Ian Stakenvicius <axs@g.o> |
5 |
> wrote: |
6 |
> > |
7 |
> > Although that is true, it would be -WAY- too slow to generate said |
8 |
> > list via equery/q* helpers; I think that's where the |
9 |
> > extended-attributes and/or cache idea comes into play. |
10 |
> |
11 |
> I agree. This needs to be high-performance when it comes to |
12 |
> individual file access. If it takes 10 seconds per build to populate |
13 |
> some database or set up a bazillion bind mounts that isn't the end of |
14 |
> the world, but if it takes an extra 0.1 seconds every time a file is |
15 |
> read that could add up VERY fast on a large build. |
16 |
|
17 |
I'd be more afraid about resources, and whether the kernel will be |
18 |
actually able to handle bazillion bind mounts. And if, whether it won't |
19 |
actually cause more overhead than copying the whole system to some kind |
20 |
of tmpfs. |
21 |
|
22 |
> |
23 |
> Ideally I'd like to see the same thing extended to run-time, and short |
24 |
> of writing some entirely new security model into the kernel or taking |
25 |
> namespaces to a whole new level, part of me thinks that |
26 |
> auto-generating SELinux policies might be the solution, so that the |
27 |
> existing mechanism can be extended. |
28 |
> |
29 |
> The mad scientist in me keeps thinking up crazy schemes so that |
30 |
> package collisions can be handled by each package just seeing whatever |
31 |
> it wants to see - maybe the entire filesystem looks different |
32 |
> depending on what app you use. Then I realize that bash is an |
33 |
> application, and how on earth would a human make sense of a system |
34 |
> where no file has any stable identifier other than maybe a |
35 |
> content-hashed key. Then that makes me wonder why we link to |
36 |
> libraries by filename anyway, when we could just give each library a |
37 |
> GUID and version, and maybe a more general identifier for cases where |
38 |
> you have alternate implementations. |
39 |
> |
40 |
> But, as long as we're still just running Gentoo on Unix-like OSes |
41 |
> maybe tweaking the jail is a good place to start... |
42 |
> |
43 |
> Rich |
44 |
> |
45 |
|
46 |
|
47 |
|
48 |
-- |
49 |
Best regards, |
50 |
Michał Górny |