On 11/12/21 12:07 PM, Steven Rostedt wrote: > On Fri, 12 Nov 2021 10:08:41 -0500 > Jason Baron <jbaron@xxxxxxxxxx> wrote: > >>> A key difference between that patchset and this patch (besides that >>> small fact that I used +x instead of +T) was that my patchset allowed >>> the dyndbg trace to be emitted to the main buffer and did not force them >>> to be in an instance-specific buffer. >> >> Yes, I agree I'd prefer that we print here to the 'main' buffer - it seems to keep things simpler and easier to combine the output from different >> sources as you mentioned. > > I do not want anything to print to the "main buffer" that can not be > filtered or turned off by the tracing infrastructure itself (aka tracefs > file system). > > Once we allow that, then the trace file will become useless because > everything will write to the main buffer and fill it with noise. > > Events that can be enabled and disabled from tracefs are fine, as they can > be limited. This is why I added that nasty warning if people leave around > trace_printk(), because it does exactly this (write to the main buffer). > It's fine for debugging a custom kernel, but should never be enabled in > something that is shipped, or part of mainline. > > -- Steve > Ok, it looks like Vincent's patch defines a dyndbg event and then uses 'trace_dyndbg()' to output to the 'main' log. So all dynamic output to the 'main' ftrace buffer goes through that event if I understand it correctly. Here's a pointer to it for reference: https://lore.kernel.org/lkml/20200825153338.17061-3-vincent.whitchurch@xxxxxxxx/ Would you be ok with that approach? Thanks, -Jason
