linux-kernel - Re: [GIT PULL] ftrace: Fixes for v6.13

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] [day] [month] [year] [list]

Message-ID: <20241216133926.1662f2a4@gandalf.local.home>
Date: Mon, 16 Dec 2024 13:39:26 -0500
From: Steven Rostedt <rostedt@...dmis.org>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: LKML <linux-kernel@...r.kernel.org>, Masami Hiramatsu
 <mhiramat@...nel.org>, Mark Rutland <mark.rutland@....com>, Mathieu
 Desnoyers <mathieu.desnoyers@...icios.com>, Al Viro
 <viro@...iv.linux.org.uk>, Michal Simek <monstr@...str.eu>
Subject: Re: [GIT PULL] ftrace: Fixes for v6.13

On Mon, 16 Dec 2024 09:59:35 -0800
Linus Torvalds <torvalds@...ux-foundation.org> wrote:

> That said, honestly, looking at the resulting diff, I really think a
> lot of those %p users are mindless drivel in the first place, and I am
> not convinced that the real address is even wanted or needed. I think
> people have completely mindlessly added "print out address" without
> any actual reason for it.
> 
> I'm seeing things like just random usb_request pointers being printed
> out, and I'm not convinced that is ever really useful or a situation
> where a '%px' is *any* different from a plain '%p' (ie the main thing
> I suspect you can do with that value is just say "it's the same
> request", and the hashed value works just fine for that).
> 
> So I am not convinced that the original reason for the mindless
> (runtime) conversion of '%p' to '%px' was really even well-motivated.
> I think it was a bad idea, and implemented badly.
> 
> End result: instead of doing that automated sed-script (which is
> certainly easy, but noisy), I actually think it might be much better
> to just remove the runtime '%p' -> '%px' conversion entirely, and see
> if anybody even notices - and then when somebody has a good reason for
> actually caring about a random kernel data structure address, change
> it at *THAT* point.

Masami was the one that implemented this for debugging reasons.

Masami, what was the motivation behind using %px? I know you needed that to
debug crashes to compare some event pointers with where they exist in the
crash dump.

But perhaps for those events that you used for crashes, we can just change
the TP_printk() over to use "%px" and remove the full conversion of all
events that you have.

Thoughts?

-- Steve