[PATCH 0/8] QI: make phase2.c easier to change

Andy Green andy at openmoko.com
Fri Jan 9 08:36:21 CET 2009

Hash: SHA1

Somebody in the thread at some point said:
| Andy Green wrote:
|> No, I prefer that fruitless discussion does not replace doing work.
| Actually, most of the so-called "fruitless" discussions here have been
| fruitless because you've made up your mind ahead of time, and stubbornly
| refuse to see anyone else's point-of-view.

I'm afraid I made my mind up ahead of time that I don't see your point
of view :-)

|> It's not uncommon that evangelists turn out to have feet of clay
|> themselves when it comes to what they're trying to sell to others.
| And then, after the discussion has gone on and on and on, and you've run
| out of technical arguments, or just grown weary, you resort to
| dismissing those who argue with you by means of a generalized
| "put-down", exactly as the line above.

Actually this was a specific put-down, I did not see any kernel stuff go
upstream during the time Werner was looking after it.  I mean I know
why, it's a freaking nightmare to move pretty much any of the bitrotted
or unsuitable code and still move forward.  Nothing happened until we
got extra resource with Balaji and now Nelson is trying for it too,
they're largely dedicated to upstream effort.  Everything else is way
more complicated trying to deal with that effort at the same time as new
kernel and moving other stuff forward, literally we pay for the upstream
effort with delay and delay is money.

I wonder what would happen if the upstream-before-working-product people
were locked in a steel cage with the combined subscribers to community
and support.

| In any case, this whole discussion is indeed silly, because we can have
| OE generate and package kernel images in any format we need,
| simultaneously -- those wishing to use kexec need only install the
| correct ipk package, and no modifications to your Qi need be made at
| all.  Easy - and no need for put-downs or smart remarks (whether it's
| technically better or not is completely irrelevant, of course).

For someone who doesn't approve of put-downs, you seem to like them :-)

Well Werner told that kexec and uImage are not compatible, I don't know
anything about it (I think he does though) but it's not the whole point.

Despite the 11 patches the actual impact on Qi is a couple of dozen
bytes and as Werner noted, it would mean we have the option to ditch the
uImage stuff that still causes confusion and doesn't add much.  So
whatever the reasoning that led to it, the cleanups and zImage support
seems worth having.

Is that stubbornness or a logical reasoning process that isn't accounted
for by your calling it "silly"?

- -Andy
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org


More information about the openmoko-kernel mailing list