Re: [PATCH] tracing: Do not allow mmap() of persistent ring buffer

From: Steven Rostedt
Date: Thu Feb 13 2025 - 21:21:52 EST


On Fri, 14 Feb 2025 11:07:22 +0900
Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx> wrote:

> This may be good for fixing crash short term but may not allow us to read the
> subbufs which is read right before crash. Can we also add an option to reset
> the read pointer in the previous boot for the persistent ring buffer?

I'm not sure what you mean here.

Note, this is just for mmapping the buffers. Currently we never tried it as
if we did, it would have crashed. But this does not affect normal reads.

-- Steve