1 |
Kevin: this is the build failures, i would have made them end up adding |
2 |
into bugzilla but one of us both wasn't awake. |
3 |
|
4 |
PSM: if you're reading up: i don't have a clue. |
5 |
|
6 |
In the meantime, working with bugs.g.o. it's getting a real hassle, so |
7 |
can someone please kick that pos in the shins so we can work on bugs |
8 |
again? TIA |
9 |
|
10 |
I think we should have a talk or two and maybe a test case for building |
11 |
that libiberty on all arches with either PIC or PIE code. |
12 |
I honestly don't care howe we build it, i just want it to work on all |
13 |
Gentoo supported arches for hardened. |
14 |
|
15 |
For those of you having issues with the file: here is the log for pitr |
16 |
and miranda: |
17 |
|
18 |
http://dev.gentoo.org/~pappy/tmp/amd64-pie-libiberty-failure.txt |
19 |
|
20 |
Here is what it says on the #gcc channel- no big deal anyway. |
21 |
|
22 |
07:24 < pappy-_> what is the role of libiberty, for what is it used |
23 |
during building gcc |
24 |
07:24 < pinskia> zwol: a million dollar question |
25 |
07:25 < pinskia> pappy-_: to support OS's that don't have some system |
26 |
functions and also to add some |
27 |
functions which are abstractions like launching programs |
28 |
07:25 < pinskia> and also demangler |
29 |
07:25 < pappy-_> does the code inside libiberty has to be PIC |
30 |
07:25 < pappy-_> on all arches? |
31 |
07:25 < DrNick> most importantly, it's linker command line is -liberty |
32 |
07:25 < pinskia> DrNick: :) |
33 |
07:25 < DrNick> which I think was the deciding factor in it's creation |
34 |
07:26 < pinskia> pappy-_: it only builds that for the target side |
35 |
07:26 < pinskia> and not for the host or build sides |
36 |
07:27 < pappy-_> yes, but it also uses it for generating stage2 stuff |
37 |
from stage1 |
38 |
07:27 < pappy-_> the host cc is used for building it in stage1 |
39 |
07:27 < pappy-_> and the stage1 libiberty (correct me if i'm wrong) is |
40 |
used for stage2 creation |
41 |
07:27 -!- don2718 [~chatzilla@×××××××××××××××××××××××××.net] has quit |
42 |
[Ping timeout: 480 seconds] |
43 |
07:28 -!- don2718 [~chatzilla@×××××××××××××××××××××××××.net] has joined #gcc |
44 |
07:29 < pinskia> pappy-_: but that is still a host library |
45 |
07:32 -!- Lightkey [~jonas@×××××××××××××××××××××××××××××××××××.net] has |
46 |
joined #gcc |
47 |
07:33 -!- gabor [~ggreif@××××××××××××××××××××××.net] has joined #gcc |
48 |
07:35 < pappy-_> /usr/x86_64-pc-linux-gnu/bin/ld: |
49 |
|
50 |
../build-x86_64-pc-linux-gnu/libiberty/libiberty.a(hashtab.o): |
51 |
relocation R_X86_64_32S |
52 |
against `a local symbol' can not be used when making a shared |
53 |
object; recompile with -fPIC |
54 |
07:35 < pappy-_> ../build-x86_64-pc-linux-gnu/libiberty/libiberty.a: |
55 |
could not read symbols: Bad value |
56 |
07:35 < pinskia> %tell dnovillo I found the fix for the latent bug, it |
57 |
was in phicprop pass and it was |
58 |
because EXECUTE_IF_SET_IN_BITMAP does not like the bitmap to |
59 |
change from underneath itself |
60 |
07:36 < gccbot> pinskia: The operation succeeded. |
61 |
07:36 -!- kazu [~chatzilla@×××××××××××××××××××××××××××××××.net] has quit |
62 |
[Ping timeout: 480 seconds] |
63 |
07:36 -!- rsandifo [~user@××××××××××××××××××××××.uk] has joined #gcc |
64 |
07:36 < pappy-_> this happens when i compile the object code for |
65 |
libiberty.a with -fPIE |
66 |
07:36 < pappy-_> to use it in stage2 fails with this relocation error |
67 |
07:36 < pappy-_> which is creating PIE executables |
68 |
07:37 < pappy-_> the 'problem' is that i thought i know a statical |
69 |
library contains object code for |
70 |
the code segment of executables. |
71 |
07:37 < pappy-_> and that executables can contain code of that library |
72 |
07:38 < pappy-_> so the object code inside that statical library should |
73 |
be pie in my understanding |
74 |
07:38 < pappy-_> however, it doesn't seem to be the reality |
75 |
07:42 < pappy-_> ? |
76 |
|
77 |
|
78 |
I'm catching up on some REM phase now, cya |
79 |
|
80 |
|
81 |
Alex |