linux-kernel - Re: [PATCH] tracing: Allow the top level trace

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: <20250409122853.7e5a29e6@gandalf.local.home>
Date: Wed, 9 Apr 2025 12:28:53 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
Cc: LKML <linux-kernel@...r.kernel.org>, Linux Trace Kernel
 <linux-trace-kernel@...r.kernel.org>, Masami Hiramatsu
 <mhiramat@...nel.org>
Subject: Re: [PATCH] tracing: Allow the top level trace_marker to write into
 another instance

On Wed, 9 Apr 2025 12:18:22 -0400
Mathieu Desnoyers <mathieu.desnoyers@...icios.com> wrote:

> > The top level instance can set this option but it can not clear it, and
> > it will bring the writing of the trace_marker back to its buffer.  
> 
> Why choose to redirect the top level trace marker from the global
> buffer to the specific one rather than allow it to iterate on a
> set of "interested" buffers ?

Hmm, interesting idea. I guess we could do it that way instead, although it
may cause a bit more overhead needing to keep track of the buffers that are
interested and also make sure they do not disappear.

I guess this could be done only in the write system call and the open
doesn't get affected. I guess the list of "interested buffers" can be
maintained by RCU, so that when a write happens, it writes to each
interested buffer via a list_for_each_entry_rcu(), under a RCU read lock so
that the buffers do not disappear when writing to them.

Masami, what do you think?

-- Steve