[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=wg2Wk9ZgVBDCBHa3-b0fSfByiRJnGA_F8snMy=3HHg_gw@mail.gmail.com>
Date: Tue, 4 Feb 2020 13:19:06 +0000
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: LKML <linux-kernel@...r.kernel.org>,
Ingo Molnar <mingo@...nel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Masami Hiramatsu <mhiramat@...nel.org>,
Tom Zanussi <zanussi@...nel.org>,
Peter Zijlstra <peterz@...radead.org>
Subject: Re: [GIT PULL] tracing: Changes for 5.6
On Tue, Feb 4, 2020 at 12:29 PM Steven Rostedt <rostedt@...dmis.org> wrote:
>
> Would this work?
Yeah, looks workable. At least this wary if you screw up your
bootconfig file somehow (or you have a kernel bug that interacts badly
with a good bootconfig file), you don't need to worry about rewriting
the initrd.
The reason _I_ care is that the initrd creation scripts tend to come
mostly pre-packaged from a distro, and editing the initrd is a big
step.
The one reaction I have is that I wonder if we should just do this the
other way around instead: instead of disabling bootconfig, have a
"enable bootconfig" model instead.
Because it strikes me that the bootconfig should be the special case
(ie "bootconfig does setup for boot-time tracing"), and that you
should explicitly say "I want you to read the extended config" on the
regular kernel command line.
So from looking at this, I do have to say that I'd have a slight
preference for simply making this be an option like
"config=bootconfig" that says "extend cmdline with the data from the
'bootconfig' file".
Would that be horribly painful for your uses?
Linus
Powered by blists - more mailing lists