Document Signed-off-by, DCO 1.1. Maybe require Signed-off-by.

Three potential options here:

1. Squash the 2nd and 3rd patch to require Signed-off-by in Mesa. (my preference)
2. Drop the 3rd patch to only document how Signed-off-by is used.
3. Drop all patches, and those using git commit -s can continue to interpret it however they want. :)

From previous discussions, my impression is that:

• @eric and @evelikov also prefer option #1 (closed). Possibly @kusma as well.
• @idr is ok with option #1 (closed), but not particular in favor of it.
• @mattst88 may also fall into the same category as @idr, but I don't think he likes option #2 (closed).
• @anholt might prefer #3 (closed).

Please correct me if I got the wrong impression based on our previous discussion on the email list. And, I hope others with an opinion will add to the discussion.