Re: [PATCH] net: dsa: mv88e6xxx: don't dispose of Global2 IRQ mappings from mdiobus code

[Vladimir Oltean]

On Wed, Mar 15, 2023 at 07:07:57AM +0100, Klaus Kudielka wrote:
> On Tue, 2023-03-14 at 22:01 +0200, Vladimir Oltean wrote:
> > 
> > I'm a bit puzzled as to how you managed to get just this one patch to
> > have a different subject-prefix from the others?
> 
> A long story, don't laugh at me.
> 
> I imported your patch with "git am", but I imported the "mbox" of the
> complete message. That was the start of the disaster.
> 
> The whole E-mail was in the commit message (also the notes before the
> patch), but that was easy to fix.
> 
> After git format-patch, checkpatch complained that your "From" E-mail
> != "Signed-off-by" E-mail. Obviously git has taken the "From" from the
> first E-mail header.
> 
> I looked again at your patch, there it was right, and there was also
> a different date (again same root cause).
> 
> So I took the shortcut: Just copy/pasted the whole patch header into
> the generated patch file, without thinking further -> Boom.
> 
> (a) Don't use "git am" blindly
> (b) Don't take shortcuts in the process

Ok, so you need to go through the submission process again, to get it right.
We don't want to accept patches which were edited in-place for anything
other than the change log (the portion between "---" and the short
diffstat, which gets discarded by git anyway). The patches that are
accepted should exactly match the patches from your working git tree.
Also, netdev maintainers extremely rarely edit the patches that they
apply, to avoid introducing traceability issues.