| 1 |
On Wed, Aug 22, 2018 at 8:26 AM Ben Kohler <bkohler@g.o> wrote: |
| 2 |
> |
| 3 |
> Hi guys, |
| 4 |
> |
| 5 |
> For some time now, we've been shipping broken i486 stage3s that do not |
| 6 |
> run on pre-i686 hardware [1]. Due to a change in catalyst [2], we no |
| 7 |
> longer set CXXFLAGS in the default make.conf, so the x86 profiles' (imho |
| 8 |
> wrong/broken) defaults [3] kick in. |
| 9 |
> |
| 10 |
> I'd like to get this fixed, and I see 3 possible solutions, listed in |
| 11 |
> order of my own preference: |
| 12 |
> |
| 13 |
> 1) Adjust x86 profile defaults to drop the problematic -march=i686. |
| 14 |
> This would be more in line with amd64 profiles (et al), which set no |
| 15 |
> -march value so it can run on any hardware for this arch. |
| 16 |
|
| 17 |
Based on a quick test, running i686-pc-linuc-gnu-gcc without passing |
| 18 |
-march seems to be equivalent to passing "-mtune=generic -march=i686". |
| 19 |
So, you'll get code that will only run on i686, but has "generic" |
| 20 |
tuning (whatever that means). |
| 21 |
|
| 22 |
I don't think that will make a noticeable difference on user systems. |
| 23 |
Most Gentoo people probably override this to some more specific CPU |
| 24 |
model via CFLAGS in make.conf anyway. |
| 25 |
|
| 26 |
So +1 from me on removing -march=i686 from the x86 arch profile. |
Archives