[PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

The events xdp_cpumap_kthread, xdp_cpumap_enqueue and xdp_devmap_xmit are
only called when CONFIG_BPF_SYSCALL is defined.  As each event can take up
to 5K regardless if they are used or not, it's best not to define them
when they are not used. Add #ifdef around these events when they are not
used.

Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
Changes since v1: https://lore.kernel.org/20250612101612.3d4509cc@batman.local.home

- Rebased on top of bpf-next

 include/trace/events/xdp.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/trace/events/xdp.h b/include/trace/events/xdp.h
index d3ef86c97ae3..746a9e95a52a 100644
--- a/include/trace/events/xdp.h
+++ b/include/trace/events/xdp.h
@@ -187,6 +187,7 @@ DEFINE_EVENT(xdp_redirect_template, xdp_redirect_map_err,
 	TP_ARGS(dev, xdp, tgt, err, map_type, map_id, index)
 );
 
+#ifdef CONFIG_BPF_SYSCALL
 TRACE_EVENT(xdp_cpumap_kthread,
 
 	TP_PROTO(int map_id, unsigned int processed,  unsigned int drops,
@@ -300,6 +301,7 @@ TRACE_EVENT(xdp_devmap_xmit,
 		  __entry->sent, __entry->drops,
 		  __entry->err)
 );
+#endif /* CONFIG_BPF_SYSCALL */
 
 /* Expect users already include <net/xdp.h>, but not xdp_priv.h */
 #include <net/xdp_priv.h>
-- 
2.47.2

Re: [PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[Alexei Starovoitov]

On Thu, Jun 12, 2025 at 3:20 PM Steven Rostedt <rostedt@goodmis.org> wrote:
>
> From: Steven Rostedt <rostedt@goodmis.org>
>
> The events xdp_cpumap_kthread, xdp_cpumap_enqueue and xdp_devmap_xmit are
> only called when CONFIG_BPF_SYSCALL is defined.  As each event can take up
> to 5K regardless if they are used or not, it's best not to define them
> when they are not used. Add #ifdef around these events when they are not
> used.
>
> Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
> ---
> Changes since v1: https://lore.kernel.org/20250612101612.3d4509cc@batman.local.home
>
> - Rebased on top of bpf-next

We can certainly take it, but you mentioned you're working
on some patches that will warn when tracepoint is not used.
So do you need this to land sooner than the next merge window ?

Re: [PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[Steven Rostedt]

On Thu, 12 Jun 2025 19:16:33 -0700
Alexei Starovoitov <alexei.starovoitov@gmail.com> wrote:

> On Thu, Jun 12, 2025 at 3:20 PM Steven Rostedt <rostedt@goodmis.org> wrote:
> >
> > From: Steven Rostedt <rostedt@goodmis.org>
> >
> > The events xdp_cpumap_kthread, xdp_cpumap_enqueue and xdp_devmap_xmit are
> > only called when CONFIG_BPF_SYSCALL is defined.  As each event can take up
> > to 5K regardless if they are used or not, it's best not to define them
> > when they are not used. Add #ifdef around these events when they are not
> > used.
> >
> > Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
> > Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
> > ---
> > Changes since v1: https://lore.kernel.org/20250612101612.3d4509cc@batman.local.home
> >
> > - Rebased on top of bpf-next  
> 
> We can certainly take it, but you mentioned you're working
> on some patches that will warn when tracepoint is not used.
> So do you need this to land sooner than the next merge window ?

No. I plan on sending that code near the end of the next merge window
to let these patches get in before they start to add warnings.

I'm also going to wait till near the end of this cycle before I add
that code to linux-next to keep the warnings happening there too soon.

So, please take it. That way there's less likelihood of another
conflict.

Thanks,

-- Steve

Re: [PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[Steven Rostedt]

On Thu, 12 Jun 2025 22:26:52 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:
> 
> > We can certainly take it, but you mentioned you're working
> > on some patches that will warn when tracepoint is not used.
> > So do you need this to land sooner than the next merge window ?  
> 

BTW, if you are curious about those patches, I just posted them:

https://lore.kernel.org/linux-trace-kernel/20250612235827.011358765@goodmis.org/

-- Steve

Re: [PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[Toke Høiland-Jørgensen]

Steven Rostedt <rostedt@goodmis.org> writes:

> From: Steven Rostedt <rostedt@goodmis.org>
>
> The events xdp_cpumap_kthread, xdp_cpumap_enqueue and xdp_devmap_xmit are
> only called when CONFIG_BPF_SYSCALL is defined.  As each event can take up
> to 5K regardless if they are used or not, it's best not to define them
> when they are not used. Add #ifdef around these events when they are not
> used.
>
> Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>

Reviewed-by: Toke Høiland-Jørgensen <toke@kernel.org>

Re: [PATCH v2] xdp: tracing: Hide some xdp events under CONFIG_BPF_SYSCALL

[Alexei Starovoitov]

On Thu, Jun 12, 2025 at 7:27 PM Steven Rostedt <rostedt@goodmis.org> wrote:
>
> On Thu, 12 Jun 2025 19:16:33 -0700
> Alexei Starovoitov <alexei.starovoitov@gmail.com> wrote:
>
> > On Thu, Jun 12, 2025 at 3:20 PM Steven Rostedt <rostedt@goodmis.org> wrote:
> > >
> > > From: Steven Rostedt <rostedt@goodmis.org>
> > >
> > > The events xdp_cpumap_kthread, xdp_cpumap_enqueue and xdp_devmap_xmit are
> > > only called when CONFIG_BPF_SYSCALL is defined.  As each event can take up
> > > to 5K regardless if they are used or not, it's best not to define them
> > > when they are not used. Add #ifdef around these events when they are not
> > > used.
> > >
> > > Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
> > > Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
> > > ---
> > > Changes since v1: https://lore.kernel.org/20250612101612.3d4509cc@batman.local.home
> > >
> > > - Rebased on top of bpf-next
> >
> > We can certainly take it, but you mentioned you're working
> > on some patches that will warn when tracepoint is not used.
> > So do you need this to land sooner than the next merge window ?
>
> No. I plan on sending that code near the end of the next merge window
> to let these patches get in before they start to add warnings.
>
> I'm also going to wait till near the end of this cycle before I add
> that code to linux-next to keep the warnings happening there too soon.
>
> So, please take it. That way there's less likelihood of another
> conflict.

It was applied to bpf-next/net yesterday.
pw-bot didn't notice.