| 1 |
måndag 17 maj 2010 22.28.05 skrev Ed W: |
| 2 |
> On 16/05/2010 21:20, Magnus Granberg wrote: |
| 3 |
> > Hi |
| 4 |
> > |
| 5 |
> > Here is the summary of the meeting 2010-05-16 |
| 6 |
> > |
| 7 |
> > 1,0 Toolchain |
| 8 |
> > We have an open bug #318171 for the merge of SSP and GCC>=4.4.3 |
| 9 |
> > support. http://bugs.gentoo.org/show_bug.cgi?id=318171 |
| 10 |
> > We are waiting for toolchain to approve the changes to toolchain.eclass |
| 11 |
> > and glibc that we need. Then we will have GCC 4.4.3 and 4.5.0 with full |
| 12 |
> > hardened (PIE/SSP) support in the tree. Grub need to be bumped to the new |
| 13 |
> > patchset. |
| 14 |
> > We have no time line on it for we are waiting on toolchain. |
| 15 |
> |
| 16 |
> I see a comment in there: "Cleaned some code and removed SSP support for |
| 17 |
> gcc 4.3.X " - I think this might need some watching and perhaps a |
| 18 |
> warning here? Sounds like if you now update say a "stable" hardened |
| 19 |
> amd64 machine pulling in stable gcc 4.3.X then you might be suddenly |
| 20 |
> loosing your hardened compiler? |
| 21 |
> |
| 22 |
> I understand this is avoided if using your overlay, but it seems like a |
| 23 |
> potential pitfall for anyone using the "stable" hardened tree? |
| 24 |
> |
| 25 |
> Can anyone comment if this is the case or I'm worrying over nothing? |
| 26 |
> |
| 27 |
> Ta |
| 28 |
> |
| 29 |
> Ed W |
| 30 |
> |
| 31 |
I only removed the code for default enable option for SSP. GCC 4.3.X still |
| 32 |
support SSP if you add -fstack-protector. The GCC 4.4.3 is on the way to get |
| 33 |
stable in 1-4 weeks i hope. Is up to the archs teams now to mark it stablel. |
| 34 |
|
| 35 |
Hardened at gentoo.org |
| 36 |
Magnus Granberg (Zorry) |
Archives