Re: [PATCH v2 04/19] net: Use trace_call__##name() at guarded tracepoint call sites

Steven Rostedt <[email protected]> Tue, 24 Mar 2026 11:37:31 -0400
Newsgroups org.kernel.vger.linux-sctp,org.kernel.vger.bpf,org.kernel.vger.linux-kernel,org.kernel.vger.linux-trace-kernel,org.kernel.vger.netdev
Message-ID <[email protected]>
On Tue, 24 Mar 2026 11:33:28 -0400
Aaron Conole <[email protected]> wrote:

> "Vineeth Pillai (Google)" <[email protected]> writes:
> 
> > Replace trace_foo() with the new trace_call__foo() at sites already
> > guarded by trace_foo_enabled(), avoiding a redundant
> > static_branch_unlikely() re-evaluation inside the tracepoint.
> > trace_call__foo() calls the tracepoint callbacks directly without
> > utilizing the static branch again.
> >
> > Suggested-by: Steven Rostedt <[email protected]>
> > Suggested-by: Peter Zijlstra <[email protected]>
> > Signed-off-by: Vineeth Pillai (Google) <[email protected]>
> > Assisted-by: Claude:claude-sonnet-4-6
> > ---  
> 
> As noted before, I think we can just eliminate the guard in the
> openvswitch module (and that would probably be a more understandable
> fix) rather than replacing the call-in.

That should probably be a separate patch, as it actually changes the
code beyond the purpose of this patch.

Removing the guard first, and then have a patch like this to modify the
rest.

-- Steve