well, then lets go back to the very basis of this whole ... box match.
:) It was about a broken networking patch that i stumbled upon
(unwillingly, via testing), which commit was _not_ posted on netdev and
_not_ posted on lkml:
http://lkml.org/lkml/2008/4/19/51
I simply hit a (trivial) regression in the networking code, but i simply
found no existing discussion of the suspect patch (commit 5e8fbe2a).
the development process is an integral part of the source code of this
OS and not a private matter of developers or maintainers. It is not a
religion and it is not taboo to criticise it, it is a crutial part of
our technology. So i will continue to criticise the development process
in the future too when i think it has aspects that hurts us. [and will
try to address all incoming criticism as well.]
the problem is that linux-next alone would not have helped much in this
specific matter. For example the softlockup warnings annoyance you
reported would have triggered immediately had you booted your Sparc64
box with linux-next or -mm even just _once_ :-)
But the same holds for me: had i ran linux-next i could have reported
some of the networking regressions sooner.
So how about making mutual use of linux-next and 'promise' to each other
to at least minimally build/boot the integrated tree, or at least
promise to not complain too loudly about bugs that could have been found
and reported there via reasonable mutual testing of linux-next? Does
that sound reasonable?
Ingo
--
