1 |
On Sat, 7 Jan 2006, Mike Frysinger wrote: |
2 |
|
3 |
> On Saturday 07 January 2006 02:30, Natanael Copa wrote: |
4 |
> > On lør, 2006-01-07 at 00:44 -0500, Mike Frysinger wrote: |
5 |
> > > On Wednesday 04 January 2006 03:52, Natanael Copa wrote: |
6 |
> > > > I got a report from someone who experienced lockups when using |
7 |
> > > > gentoo-embedded built system (2.6.14-hardened-r1) |
8 |
> > > |
9 |
> > > this is a hardened kernel issue rather than a userland issue ... you'd be |
10 |
> > > better off e-mailing the gentoo-hardened list |
11 |
> > |
12 |
> > Probably, but I think I found the reason. Whenever I build the kernel |
13 |
> > with "optimize for size" it hangs after an hour with intense net load |
14 |
> > and io load (I made a loop that dd if=/dev/cdrom) |
15 |
> > |
16 |
> > When I disabled "Configure standard kernel features (for small |
17 |
> > systems)" (gcc -Os) the problem magically went away. |
18 |
> |
19 |
> the kernel can be pretty sensitive to optimizations |
20 |
|
21 |
if kernel is built w/ gcc-3.4.4 and -Os (at least I had that for 2.4.x |
22 |
kernels) it "optimizes away" memcmp and the kernels have gotten a patch |
23 |
to overcome this, the same problem does not occur w/ gcc-3.3.x. |
24 |
|
25 |
> > So I think its a bug in gcc version 3.4.4 (Gentoo Hardened 3.4.4-r1, |
26 |
> > ssp-3.4.4-1.0, pie-8.7.8) |
27 |
> > |
28 |
> > Makes me nervous because I have built everything with -Os. |
29 |
|
30 |
I have built everything w/ -Os and do not have trouble w/ it |
31 |
|
32 |
Peter |
33 |
|
34 |
-- |
35 |
Peter S. Mazinger <ps dot m at gmx dot net> ID: 0xA5F059F2 |
36 |
Key fingerprint = 92A4 31E1 56BC 3D5A 2D08 BB6E C389 975E A5F0 59F2 |
37 |
|
38 |
-- |
39 |
gentoo-embedded@g.o mailing list |