Re: [PATCH v2] tracing: Introduce pipe_cpumask to avoid race on trace_pipes

[Google]

On Fri, 18 Aug 2023 09:41:28 -0400
Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:

> On Fri, 18 Aug 2023 14:03:09 +0900
> Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx> wrote:
> 
> > Looks good to me. 
> > 
> > Reviewed-by: Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>
> > 
> > BTW, this reminds me that splice read support had been introduced for
> > virtio-trace support. I've tested it with/without this patch, but I hit
> > a different issue, which seems like a virtio-serial issue (per-cpu buffers
> > on the guest are spliced correctly, but the data can not be read from host
> > side). Let me investigate it.
> 
> Is virtio-trace using trace_pipe or trace_pipe_raw? Those two are handled
> differently.

It uses trace_pipe_raw. I guess if splice(from trace_pipe_raw to virtio-serial)
returns -1 and errno == EAGAIN, the trace data will be lost?

Thank you,

> 
> -- Steve


-- 
Masami Hiramatsu (Google) <mhiramat@xxxxxxxxxx>