Intel em receive hang and possible pr #72970 + some offtop

[Gleb Smirnoff]

  Rob,

  to compile the driver taken from RELENG_6 branch on the
6.0-RELEASE (or 6.1-RELEASE) system you need to also merge
this commit into your system:

  http://lists.freebsd.org/pipermail/cvs-src/2006-July/065867.html

After this driver from RELENG_6 will be buildable. I wonder, why
noone in the thread had mention this?

....

And I will also add my opinion on this thread, since many people
has expressed theirs.

If one experiences a bug in a FreeBSD release, and he/she asserts
that he/she is not going to install STABLE branch snapshop on
their system, because he/she has a very mission critical installation
and STABLE branch is a crap, then he/she has three options:

- Merge the necessary bugfix and its dependencies theirselves. If
  one has such mission critical installation, then he/she should
  have an employee experienced enough to run few cvs(1) commands.

- Find a person who can do this. Pay him.

- Find a person who can do this for free. Ask him kindly, and he
  will help. It took me 15 seconds to find the URL above. It will
  take me 5 - 10 minutes to prepare ready patch for 6.1-RELEASE.
  I'll probably help one, if he/she asks me. However, I'd prefer to
  point at actual CVS revisions and ask to do the merge theirselves.

Well, I've already moved this post to offtopic, and I will continue.
Rob, the below isn't said directly to you. This is what I'd like
to say to many FreeBSD users.

I'm always disappointed when I read about somebody strongly refusing
to install STABLE. You know, the releases you trust in are cut from
this branch! The more people install STABLE, the better the release
would be, because more bugs will be uncovered. This is the community's
contribution to the FreeBSD, the price of the free software.

One may say that his installation is too commercially important to
test STABLE on it, and refuse to do so. Well, apart from the fact
that this man is using a free thingie without giving anything back,
it may happen that he didn't avoid the bugs he tried to. It may
happen, that the bug he would found if installed STABLE, was not
found during release process and leaked into next FreeBSD release.
Then he installs release and experiences this bug. Due to his
"release only" policy, he needs to wait for the next one, again
not testing it. Believe me, I've seen several such cases.

At my job, I install STABLE on important servers about two or
three times between releases. And this servers aren't less
important than someone's else. There are a number of ways to test
new software (and hardware as well) in a safe manner, so that
if anything goes wrong, you have ability either switch to
backup, or quickly rollback to previous version. This is a day
to day sysadmin job - test forthcoming changes. For example, this
week we have upgraded few servers to STABLE, so that we test it
and 6.2-RELEASE won't give us any surprises. If we find any bugs,
we would report about them or fix them ourselves. And when
6.2-RELEASE comes out, we will upgrade the rest of the servers
being sure in this release.

-- 
Totus tuus, Glebius.
GLEBIUS-RIPN GLEB-RIPE