linux-kernel - Re: commit tag order vs. "b4 am"

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20240130-solid-seahorse-from-camelot-e0ecdb@lemur>
Date: Tue, 30 Jan 2024 20:57:44 -0500
From: Konstantin Ryabitsev <konstantin@...uxfoundation.org>
To: Randy Dunlap <rdunlap@...radead.org>
Cc: Tony Luck <tony.luck@...el.com>, x86@...nel.org, 
	Reinette Chatre <reinette.chatre@...el.com>, linux-kernel@...r.kernel.org
Subject: Re: commit tag order vs. "b4 am"

On Tue, Jan 30, 2024 at 04:47:26PM -0800, Randy Dunlap wrote:
> > Reinette noticed that v14 of my resctrl/SNC patch series[1] did not adhere
> > to the tag order proscribed in Documentation/process/maintainer-tip.rst
> > Specifically my "Signed-off-by:" was now the last tag, instead of
> > appearing before the "Reviewed-by:" and "Tested-by" tags as it had in
> > v13.
> > 
> > A little digging showed that my tag had been moved to the end by "b4 am"
> > when I used it to pick up some additonal tags.
> > 
> > An e-mail discussion with Konstantin ensued to determine if this was
> > a bug. Konstantin said:
> > 
> >    This is the intended behaviour, because b4 follows the chain-of-custody
> >    procedure. If we encounter a Signed-off-by trailer matching the identity of
> >    the user preparing the series, we move it to the bottom to indicate that the
> >    chain-of-custody boundary has moved to include the code review trailers
> >    received after the initial submission.
> > 
> >    https://lore.kernel.org/tools/20221031165842.vxr4kp6h7qnkc53l@meerkat.local/
> > 
> >    Basically, the "Signed-off-by" trailer is special because it indicates that
> >    everything above it is the responsibility of the person doing the sign-off. If
> >    we kept your Signed-off-by in the original spot, then it wouldn't be clear who
> >    collected and applied the trailers.
> 
> I can't find "chain of custody" anywhere in Documentation/process/, nor a
> specification or example of this ordering.
> 
> Where did this b4 requirement come from?

Many discussions at the Maintainer Summit and on the tools/users lists. E.g.:
https://lore.kernel.org/tools/20221031165842.vxr4kp6h7qnkc53l@meerkat.local/

The chain of custody approach is the only one that makes sense because it
allows keeping track of who applied which trailers. I know that most people
don't think about it twice, but it matters from the perspective of process.
If we need to put it into the official documentation, I'm happy to submit the
patch.

-K