1 |
On Saturday 17 February 2007, Brian Harring wrote: |
2 |
> On Sat, Feb 17, 2007 at 09:39:58AM -0500, Mike Frysinger wrote: |
3 |
> > On Saturday 17 February 2007, Brian Harring wrote: |
4 |
> > > Security impact is from a pkg potentially dragging along old libs; if |
5 |
> > > you've got a stable pkg that gets an update once every blue moon, it |
6 |
> > > can hold onto the lib for a *long* time while still using the lib; |
7 |
> > > thus if a vuln. in the lib, said pkg still is screwed. |
8 |
> > |
9 |
> > umm, no ... the package itself is updated against the new copy while the |
10 |
> > old copy exists for other packages that have already been built |
11 |
> |
12 |
> Suspect you're ignoring soname changes, which is about all this patch |
13 |
> would address- for example, ssl's old form of breakage. In that case, |
14 |
> *yes* the package gets updated, anything recompiled should get the |
15 |
> correct lib |
16 |
|
17 |
i'm not ignoring soname changes, those are exactly what i'm talking about |
18 |
|
19 |
> (assuming the code knows the appropriate linker args) |
20 |
|
21 |
there is no such thing ... it's always "-lfoo" |
22 |
|
23 |
> the old vuln. lib still will hang around as long as anything refs it. |
24 |
|
25 |
of course and this is the desired behavior ... people need to run |
26 |
revdep-rebuild, there's no two ways about it |
27 |
|
28 |
> > > Other angle is someone intentionally forcing usage of a known bad |
29 |
> > > library that is still dangling. Corner case, but doable. |
30 |
> > |
31 |
> > as i said, this is the "invalid" syntax: |
32 |
> > ... /usr/lib/libfoo.so.3 ... |
33 |
> |
34 |
> Not to LD_PRELOAD :) |
35 |
> |
36 |
> Haven't tried anything crazy, but suspect it can be abused to override |
37 |
> to the old. |
38 |
|
39 |
again, not in any scenario that actually matters ... so this too is a |
40 |
pointless line of thought to pursue as it has no real security impact |
41 |
|
42 |
> > > Bit curious how this is going to behave if via linked in libs, new loc |
43 |
> > > and old get loaded alongside... |
44 |
> > |
45 |
> > this would require multiple libraries to be involved in the equation and |
46 |
> > the answer is undefined behavior which most certainly result in runtime |
47 |
> > failures ... |
48 |
> |
49 |
> Point there was that instead of just bailing with "lib is missing", |
50 |
> suspect it'll manage to run, then segfault at potentially crappy |
51 |
> times. |
52 |
|
53 |
this is really an outside case and not worth worrying over |
54 |
-mike |