1 |
(I originally posted this on gmane.linux.gentoo.amd64, but even though |
2 |
GMane authorized me to post there, my posts don't get through. So I'm |
3 |
reposting it here.) |
4 |
|
5 |
Usually, when prelinking a system, it's recommended to use prelink's |
6 |
"-m" (or "--conserve-memory") option: |
7 |
|
8 |
When assigning addresses to libraries, allow overlap of |
9 |
address space slots provided that the two libraries are |
10 |
not present together in any of the binaries or libraries. |
11 |
This results in a smaller virtual address space range used |
12 |
for libraries. On the other hand, if prelink sees a |
13 |
binary during incremental prelinking which puts together |
14 |
two libraries which were not present together in any other |
15 |
binary and were given the same virtual address space slots, |
16 |
then the binary cannot be prelinked. Without this option, |
17 |
each library is assigned a unique virtual address space |
18 |
slot. |
19 |
|
20 |
But on AMD64, virtual address space is virtually unlimited (or at least |
21 |
"very huge"). Are there any drawbacks in not using "-m" on AMD64? |