On Sat, 9 Oct 2021 20:07:10 +0800 Hou Tao <houtao1@xxxxxxxxxx> wrote: > Not tried yet, but considering that VFS maintainer refused to have tracepoint in > VFS layer, I'm not sure it is worth trying. The reason they refuse to is because it shows data that can become an API. But if that data is just a pointer, then it's not possible to become an API no more than a RAW tracepoint that BPF can hook to. Try asking. > > > > That is, it only gives you a pointer to what is passed in, but does not > > give you anything else to form any API against it. > > This way, not only does BPF have access to this information, so do the > > other tracers, through the new eprobe interface: > Or in a opposite way can eprobe add support for bare tracepoint ? If there's a way to expose the parameters of a bare tracepoint, then it should not be difficult to do so. Heck, we can just have "bare tracepoints" be events, that do nothing but repeat their pointer parameters, as that's basically all they do anyway. There's nothing fundamentally different between a "bare tracepoint" and a trace event that just defines the fields as the same as the parameters to a tracepoint, except now you have another way to know what the parameters are. -- Steve
