linux-kernel - Re: [GIT PULL] GPIO bulk changes for kernel v4.6

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]

Date:	Thu, 17 Mar 2016 23:07:10 -0700
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	Linus Walleij <linus.walleij@...aro.org>
Cc:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"linux-gpio@...r.kernel.org" <linux-gpio@...r.kernel.org>,
	Laxman Dewangan <ldewangan@...dia.com>
Subject: Re: [GIT PULL] GPIO bulk changes for kernel v4.6

On Thu, Mar 17, 2016 at 1:59 AM, Linus Walleij <linus.walleij@...aro.org> wrote:
>
> NOTE: tree was a bit dirty and I realized it too late: Laxmans
> devm_gpiochip_add() branch was based on my for-next branch rather
> than my devel branch, making some commits appear twice and
> a file named README.md "Share upstreaming patches" appear and
> then get reverted out by me.
>
> The end result should be clean but the history is a bit messy.

Gaah. I took the tree, but I didn't realize just *how* messy it was. I
doubt you did either.

Dammit, had I realized just how screwed up that branch was, I'd have
made you re-do it.

Because that branch is crap.

And the real reason is is crap isn't a "README.md" file that comes and
goes. The real reason it is crap is that it has a new root commit, and
Laxman has done something TOTALLY INSANE.

I'm not even sure what insane tool was used to do this, but there's a
new root commit at

  a101ad945113be3d7f283a181810d76897f0a0d6

that has no parenthood, and that is only used for a completely bogus
merge (merge commit e5451c8f8330e03ad3cfa16048b4daf961af434f). That's
where the README.md file comes from.

Git does support the notion of having multiple roots, and it is a
useful thing to have when you merge two different projects with
separate history. In fact, git itself has that, for 'gitk' that was
merged into the main git history.

We have that in the kernel for the initial btrfs merge too, actually.
btrfs started out outside the full kernel, and had a root of its own
with its own development history, and then it was merged into the
kernel tree under fs/btrfs.

But this is *not* that. That commit
a101ad945113be3d7f283a181810d76897f0a0d6 is just pure garbage, and the
merge that introduces it is pure shit.

Dammit, I noticed too late, so now it's out there. How the hell did
that even happen? Why did Laxman do that insane merge? Why did it get
pulled back?

You actually have to *work* at making shit like this, so I wonder what
workflow you guys had to make that bad merge. It's easy enough to do
with git manually, but it's hard to do by _mistake_. Is there some
broken tool that Laxman uses?

I'm very annoyed, because while the multi-root situation can be
useful, it can also be confusing as hell. It can cause bisection
problems, and it can just cause people to go "WTF?"

I need to know what happened and make sure it doesn't happen again. If
this is some intentional workflow by nvidia, it needs to stop *now*.
It's broken shit.

               Linus