Hi everyone,

I just wanted to let everyone know that I just did a big import into
the ruby code in the m5 tree.  This basically updated the M5 tree with
all of the recent changes that were done in ruby.  This should
hopefully be the last time that we do something like this for ruby as
we're going to start working on ruby out of the M5 tree.  Yay!

There are numerous improvements in the code (perhaps Derek would like
to give a synopsis).  Additionally, there are two known regressions
because the update was done as a re-import not so much as a merge
(because the trees diverged significantly and were maintained using
different revision control systems).  The regression is that we lost
the changes that made ruby handle atomic memory operations.  As a
result Daniel's multithreading code is broken.  (I think that fixing
this should be one of the top priorities)  The second is that there
was a major refactoring of slicc, so many of the coherence protocols
need some tweaking to get them in working shape again.

At this point, I think we're back to a normal (m5 style) development
process with Ruby code.  This means that you should take care not to
break what is currently working.  You should run regressions.  You
should seek code reviews.  With this import, the M5 codebase now has
over 200,000 lines of code.  While this isn't huge when compared to
things like the linux kernel, it's still big enough that everyone
needs to try hard to keep the code quality high enough that stuff
doesn't just start breaking all the time.  We've been pretty
successful for the past couple of years, but as we add more
collaborators, we need to make sure that everyone works hard at this.

  Nate
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to