| 1 |
On Sunday, April 5, 2009 08:54:22 Grant wrote: |
| 2 |
> >> I use a program called miro a lot, but since upgrading to gcc-4.3.3, |
| 3 |
> >> it segfaults if I don't issue: |
| 4 |
> > |
| 5 |
> > what's the package name? or what ebuild do you use? |
| 6 |
> > |
| 7 |
> >> grsec: signal 11 sent to /usr/bin/miro.real[miro.real:19177] |
| 8 |
> >> uid/euid:1000/1000 gid/egid:100/100, parent /usr/bin/miro[miro:19176] |
| 9 |
> >> uid/euid:1000/1000 gid/egid:100/100 |
| 10 |
> >> |
| 11 |
> >> and since miro.real is a python script: |
| 12 |
> > |
| 13 |
> > in that case you have to use paxctl on the python interpreter, not |
| 14 |
> > the script. but before that it's better to find out why it crashes, |
| 15 |
> > so you should produce a coredump and analyze it. |
| 16 |
> |
| 17 |
> Thank you, I ran 'paxctl -m /usr/bin/python' and now miro starts |
| 18 |
> without softmode enabled. I use the miro ebuild from this bug: |
| 19 |
> |
| 20 |
> http://bugs.gentoo.org/show_bug.cgi?id=131527 |
| 21 |
> |
| 22 |
> - Gramt |
| 23 |
|
| 24 |
Don't know what miro is and didn't bother to look at the bug, but removing |
| 25 |
MPROTECT on python is not good. One alternative might be to write simple |
| 26 |
wrapper in C for miro and paxctl that instead. |
| 27 |
|
| 28 |
Gordon Malm (gengor) |
Archives