| 1 |
On man, 2006-01-09 at 17:20 +0100, Peter S. Mazinger wrote: |
| 2 |
> On Sat, 7 Jan 2006, Mike Frysinger wrote: |
| 3 |
> |
| 4 |
> > On Saturday 07 January 2006 02:30, Natanael Copa wrote: |
| 5 |
> > > On lør, 2006-01-07 at 00:44 -0500, Mike Frysinger wrote: |
| 6 |
> > > > On Wednesday 04 January 2006 03:52, Natanael Copa wrote: |
| 7 |
> > > > > I got a report from someone who experienced lockups when using |
| 8 |
> > > > > gentoo-embedded built system (2.6.14-hardened-r1) |
| 9 |
|
| 10 |
> > > When I disabled "Configure standard kernel features (for small |
| 11 |
> > > systems)" (gcc -Os) the problem magically went away. |
| 12 |
> > |
| 13 |
> > the kernel can be pretty sensitive to optimizations |
| 14 |
> |
| 15 |
> if kernel is built w/ gcc-3.4.4 and -Os (at least I had that for 2.4.x |
| 16 |
> kernels) it "optimizes away" memcmp and the kernels have gotten a patch |
| 17 |
> to overcome this, the same problem does not occur w/ gcc-3.3.x. |
| 18 |
> |
| 19 |
> > > So I think its a bug in gcc version 3.4.4 (Gentoo Hardened 3.4.4-r1, |
| 20 |
> > > ssp-3.4.4-1.0, pie-8.7.8) |
| 21 |
> > > |
| 22 |
> > > Makes me nervous because I have built everything with -Os. |
| 23 |
> |
| 24 |
> I have built everything w/ -Os and do not have trouble w/ it |
| 25 |
|
| 26 |
I have researched this and found out that it is not the -Os itself that |
| 27 |
causes this, but it is something else under "Configure standard kernel |
| 28 |
features (for small systems)". I'm not sure of exactly what but I know |
| 29 |
at least that it is not "Optimize for size". I have turned all options |
| 30 |
under "Configure standard kernel features (for small systems)" except |
| 31 |
"Load all symbols for debugging/kksymoops" (including optimize for size |
| 32 |
(-Os)) and done the netload test all night. This morning the computer |
| 33 |
was still up. When I have an unstable kernel it hangs after an hour or |
| 34 |
two. |
| 35 |
|
| 36 |
I have tried to flip the "futex support" and that did not make any |
| 37 |
difference (anyone who know how this affects uclibc based apps btw? Do I |
| 38 |
need futex support?) |
| 39 |
|
| 40 |
I think I have narrowed down the problem to one of those: |
| 41 |
* BUG() support |
| 42 |
* Enable full-sized data structures for core |
| 43 |
* Enable eventpoll support |
| 44 |
|
| 45 |
But exactly what, I dont know. If anyone here have had any problems with |
| 46 |
disabling one of those, I'm very interested to hear. |
| 47 |
|
| 48 |
Lots of bla bla here, I just thought it could be good for you guys to |
| 49 |
know that turning off of those might make your kernel unstable, while |
| 50 |
-Os should be ok with current gcc. |
| 51 |
|
| 52 |
|
| 53 |
-- |
| 54 |
Natanael Copa |
| 55 |
|
| 56 |
-- |
| 57 |
gentoo-embedded@g.o mailing list |
Archives