Daniel Xu

I've split out some of the commits in this PR into #3368. I'll deal with changing the default string size in a final PR. It'll include the string clamping.

Another idea is to extend string scratch buffer concept to all places where we previously alloca'd on stack to read a string into. This is rather straightforward, as we can...

For custom memset(), we should feature probe for loops and fall back to builtin memset if loops not available. We should also do u64 strides as much as possible, with...

Big memset() support: https://github.com/bpftrace/bpftrace/pull/3237 It uses `bpf_probe_read_kernel(dst, 0, NULL)` trick as suggested by Alexei. This is a nice way to take advantage of optimized kernel memset().

It looks like already have c++: https://github.com/bpftrace/bpftrace/blob/master/tests/testprogs/uprobe_test_cxx.cpp Do we have a good idea of the test cases we want to add? In general I'd prefer to not get too fancy...

> Not really, other than yeah, tedious :) > > Tooling to dynamically generate the stripped-down header would be pretty neat (by recursively expanding dependent types from a list of...

Hey @jaredp-openai ! FWIW we considered implementing this, but ended up going with a kubectl plugin (https://kubernetes.io/docs/tasks/extend-kubectl/kubectl-plugins/). It was too messy to try and reconcile security policies with true remote...

Similar to https://github.com/bpftrace/bpftrace/issues/3028 but this has less concerns about overload the ring buffers. Fairly straightforward and sensible to add an optional named parameter to `clear()`. I asked @4ast and unfortunately...

Will take a closer look at your writeup soon, but I can comment about LBR stuff. I've spent some time hacking on lbr+bpf in the kernel (have a patch on...

> Does it need to happen in the kernel though? If we have the stack and the mapping it would be possible to symbolize user-space stacks in user-space Right, so...