2011/10/10 Petr Salinger :
> The main problem is that we do not know whether gcc-4.x
> miscompiles kernel or kernel is not properly coded.
Agreed. I'll do more testing and investigate. In the meantime I've
uploaded a gcc-4.5 built kfreebsd to experimental, so that this
combination can be tested
Apparently, there are more problems than -fgcse. "gcc-4.6 -fno-gcse"
still fails whereas "gcc-4.4 -fno-gcse" doesn't.
I've downgraded the dependency to get a working kernel. Before we
upgrade, I think we could use experimental as testing ground to
resolve the problems. It's very bad that this ki
2011/10/7 Robert Millan :
> 2011/10/5 Petr Salinger :
>> Would you mind to rebuild latest kfreebsd-8 with gcc-4.3 from stable ?
>
> This works. In fact I made a backport out of it and uploaded it to
> bpo (in case someone's interested in running kfreebsd 8.2 in Squeeze).
>
>> And by current "gcc-4
2011/10/5 Petr Salinger :
> Would you mind to rebuild latest kfreebsd-8 with gcc-4.3 from stable ?
This works. In fact I made a backport out of it and uploaded it to
bpo (in case someone's interested in running kfreebsd 8.2 in Squeeze).
> And by current "gcc-4.6 -fno-gcse" ?
Will try that next.
On my main workstation, kfreebsd 8.2-1.1 and later versions CPU-faults
early on boot. This affects the i686 version only (not amd64 one).
8.2-1 is not affected.
Probably related to:
* Build by GCC 4.6. (Closes: #594288)
Would you mind to rebuild latest kfreebsd-8 with gcc-4.3 from stable ?
Package: kfreebsd-image-8.2-1-686
Version: 8.2-1.1
Severity: grave
On my main workstation, kfreebsd 8.2-1.1 and later versions CPU-faults
early on boot. This affects the i686 version only (not amd64 one).
8.2-1 is not affected.
-- System Information:
Debian Release: 6.0.2
APT prefers stable-u
6 matches
Mail list logo
