carl-daniel: I wonder about the recent commit that related to breaking kontron, that we were unsure caused kontron to fail some time ago but were not sure about.
I ask this question because latest svn no longer works at all, I get absolutely no output on kontron. Nothing.
Can we re-open the question? My memory is very fuzzy on the issues.
thanks
ron
On Fri, Feb 20, 2009 at 5:01 PM, ron minnich rminnich@gmail.com wrote:
Can we re-open the question? My memory is very fuzzy on the issues.
no need. pilot error.
and, amazingly, the SMP code is actually kind of working. I'm actually quite shocked.
ron
On 21.02.2009 02:01, ron minnich wrote:
carl-daniel: I wonder about the recent commit that related to breaking kontron, that we were unsure caused kontron to fail some time ago but were not sure about.
I ask this question because latest svn no longer works at all, I get absolutely no output on kontron. Nothing.
Can we re-open the question? My memory is very fuzzy on the issues.
I need to look at the core2duo startup again. If my memory serves correctly, we now have some stack variables outside the stack. That's an age-old once fixed and then reverted core2duo startup bug.
Depending on how lucky you are (and if gcc plays along), even buggy startup may work for you without visible issues. It certainly worked months (years?) for everyone else although it was buggy. Some SimNow versions did notice the bug in odd ways, but that's it.
Short summary: Even if the current code boots, you need to watch out for stack underflow and corner case behaviour for bottom-of-stack during startup.
Regards, Carl-Daniel
On Fri, Feb 20, 2009 at 5:40 PM, Carl-Daniel Hailfinger
I need to look at the core2duo startup again. If my memory serves correctly, we now have some stack variables outside the stack. That's an age-old once fixed and then reverted core2duo startup bug.
I may be missing something but I don't see any variables out of stack.
ron
-
Carl-Daniel Hailfinger -
ron minnich