linux-kernel - Re: [GIT PULL] trivial for 4.9

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: <CA+55aFzXfy3X5-0V=jNEKUVntfq_2e8-ExxCTfcgSr+LhVBv+w@mail.gmail.com>
Date:   Fri, 7 Oct 2016 14:06:27 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Joe Perches <joe@...ches.com>
Cc:     Jiri Kosina <jikos@...nel.org>,
        Colin Ian King <colin.king@...onical.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [GIT PULL] trivial for 4.9

On Fri, Oct 7, 2016 at 1:33 PM, Joe Perches <joe@...ches.com> wrote:
>>
>> And if there are behavioral issues, they should (a) be pointed out and
>> (b) be fixed.
>>
>> In *no* case does it make sense to randomly just add newline
>> characters without even having a reason for it.
>
> It prevents random interleaving from those other
> 12000+ possible printk calls without an explicit
> KERN_<LEVEL>

YOU ARE NOT LISTENING.

Let me do this really clearly and slowly.

I'm not applying patches that

 (a) are random churn

 (b) make the code uglier

 (c) are about purely theoretical problems IN OTHER CODE.

How hard is that to understand? The "\n"-adding patches are ALL of the
above. They don't fix a problem, they actually *hide* problems, and
they hide problems that

 (1) do not matter

 (2) aren't in the code that the "\n"-adding patch even adds.

See?

So stop with the idiotic theoretical arguments about interleaving that
isn't even true.

Instead, start with the *actual* problems. First off, if somebody
actually reports an issue like the above, fix *that*.

And btw, even without an explicit KERN_<level>, you should still not
get any interleaving. Only an _explicit_ KERN_CONT should cause
interleaving, and dammit, if some interrupt does a KERN_CONT without
having had a non-cont printk before it, that code is buggy and should
damn well be fixed.

And if it's not an interrupt, then the "not the same task as last
time" should add the newline.

In other words, all your arguments seem entirely wrong. IF that
interleaving actually happens, we have bugs that should be fixed.

Again: we should not add stupid churn that doesn't even fix the bugs,
just makes code harder to read and adds churn.

Seriously.

If you can send me a patch to *fix* anything, go ahead. But stop this
idiotic "let's add random pointless newline characters" crap already.

               Linus