[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240703174129.7e19a9a6@rorschach.local.home>
Date: Wed, 3 Jul 2024 17:41:29 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Yang Jialong 杨佳龙 <jialong.yang@...ngroup.cn>
Cc: mhiramat@...nel.org, Karol Herbst <karolherbst@...il.com>, Pekka
Paalanen <ppaalanen@...il.com>, mathieu.desnoyers@...icios.com, Michael
Ellerman <mpe@...erman.id.au>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 1/2] powerpc/mmiotrace: Add MMIO Tracing tool for
PowerPC
On Wed, 3 Jul 2024 15:39:09 +0800
Yang Jialong 杨佳龙 <jialong.yang@...ngroup.cn> wrote:
> Hello, mmiotrace maintainers.
>
> Here is a first version patch used to support mmiotrace function on
> powerpc platform. When I copy mmiotrace to powerpc archtecture, we found
> that there are many code are general, such as code in kmmio.c and
> mmio-mod.c in arch/x86/mm/. But we have file trace_mmiotrace.c as a low
> layer to finish the function. Whether we can abstract out one layer
> again or not?
>
The mmiotracer is at a minimal maintenance level. I constantly test it
and make sure it works (on x86) for every pull request I do. But I
haven't done much more than that.
If I ever got the time, I was going to make it work for SMP machines,
as it currently still requires running in UP (does a CPU hotplug to
shutdown all CPUs but one when enabled). But that's just because there
was no effort to make it multiprocessor safe. It is definitely doable.
But it is very low priority in my TODO list. I have no problem if you
want to clean it up and pull out the more generic parts. Hey, if you
have time, make it SMP safe if you can ;-) Although that may remove
one of my biggest bug finders. You'd be amazed at what breaks when you
aggressively disable all but one CPU and then enable all of them back
up again.
-- Steve
Powered by blists - more mailing lists