I've summarized my work on Userspace, Statically Defined Tracing support for Bitcoin Core in my Coinbase grant half-time report.

· · Web · 1 · 3 · 2

I'm looking for topics, mini projects to do online Socratics on. Would you be up for it with this USDT project in next few months? BPF Performance Tools (Gregg) has been sitting on my reading list since the PR review club and this might give me the kick to read it :)

@michaelfolkson Yes! I've been thinking about some kind of USDT workshop for CoreDev too.

A Socratic might be better once we have more tracepoints merged and people building tools/using the tracepoints? Maybe sometime between CoreDev and 23.0. release?

@0xb10c Absolutely, sounds good. I will have to read it before CoreDev then if you're doing a workshop. It still hasn't clicked with me when it makes sense to use tracing over debugging (or vice versa). I guess if I was a more experienced debugger I'd encounter more problems where I hit the limits of what a debugger can do.

@michaelfolkson I'm not sure if I'd see reading "BPF Performance Tools" as a requirement. The tools are very specific for tracepoints in the Linux kernel. We don't use any of these tools or tracepoints.

We just implement our own tracepoints in our userspace application (Bitcoin Core) and write custom scripts to get insights into the parts we find important.

Additionally, I think 'tracing' via USDT and debugging are two very different things. You can use tracing via USDT as part of the search for a bug, but you can't replace debugging with tracing.

Sometimes debugging, as in stepping through code and looking at state with e.g. GDB, is something you do manually.

The tracing done with USDT allows you to use internal state programatically. You can automatically react to some state.

@0xb10c Thanks for this. This still sounds to me like USDT tracing is something you do once you hit the limits of what a debugger can do? Like deeper, richer debugging? The stop, start nature of breakpoints versus allowing the code to run continuously and still manipulate it? Or maybe you locate the bug using debugging and then you want to explore the impact of the bug with USDT tracing?

@michaelfolkson Yes, debugging is definitely one of the USDT applications. Also developers can be use the tracepoints to evaluate/review changes like I do with the erlay vs master nodes dashboard. Or for general monitoring, anomaly detection, benchmarking, and testing.

Sign in to participate in the conversation
unidentified instance

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!