linux-kernel - Re: [PATCH] x86/tracing: introduce enter/exit tracepoint pairs for page faults

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: <20250418213945.GJaALGod4-j86Squ43@fat_crate.local>
Date: Fri, 18 Apr 2025 23:39:45 +0200
From: Borislav Petkov <bp@...en8.de>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: Junxuan Liao <ljx@...wisc.edu>, linux-trace-kernel@...r.kernel.org,
	linux-kernel@...r.kernel.org, x86@...nel.org,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	"H. Peter Anvin" <hpa@...or.com>,
	Masami Hiramatsu <mhiramat@...nel.org>,
	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
	"Paul E. McKenney" <paulmck@...nel.org>
Subject: Re: [PATCH] x86/tracing: introduce enter/exit tracepoint pairs for
 page faults

On Wed, Apr 16, 2025 at 02:01:15PM -0400, Steven Rostedt wrote:
> The above was created by:
> 
>  # trace-cmd sqlhist -e -n user_faults SELECT TIMESTAMP_DELTA_USECS as delta FROM page_fault_user_enter as start JOIN \
>      page_fault_user_exit as end ON start.common_pid = end.common_pid

Pfff, that's SQL.

You're too old fashioned - you need an AI thing now :-P

So tell me: I as a silly user, how do I figure out how to use trace-cmd?
I guess it has docs somewhere...

And apparently I need trace-cmd now - I can't type all that into sysfs... Or
I guess I can use trace-cmd to generate the commands and then I can echo them
into the target machine.

I.e., ftrace still doesn't need a special tool to be used...

> This was just a simple example. I rather see where in the kernel it happens.
> I can still use the synthetic events and user stack trace to find where the
> big faults occur.

Right, so it would be great to have the actual use case in a patch's commit
message... i.e., why is a patch important.

> > Now if only that info were in the commit message along with the usage scenario
> > so that people can *actually* do what you guys are bragging about...
> 
> I plan on adding things like this to Documentation/trace/debugging.rst
> 
> I need to get time out to add a bunch of helpful tricks there.

Yap, please.

I have to admit I was able to catch a trace for tglx using only the ftrace
documentation we have so it must be good. But moar is better. :-P

Thx man!

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette