Derek Bruening

One way to solve this is to use the new encodings file for blocks with annotations, if we can detect them during live tracing: maybe by the gap in the...

Function markers between an instruction and its memref was already filed as #5620 and that was already fixed, right? Is your second example after that fix went in, or before?...

This is on the large side -- seeing all the steps listed it maybe it could have been split up? Not requesting now as it is already posted, but in...

This means we need an analysis-tool-visible extended-payload marker type, or maybe make the reader hide that from the user.

Having the actual register size in the operand in the IR at runtime does seem necessary but maybe not sufficient. Tools like DrMemory which are shadowing registers may want to...

> > What about `dr_simd_t` which it looks like is currently just 128 bits? Does it need to be increased to handle 2048 bits... > > Yes. The low 128...

> > > > What about `dr_simd_t` which it looks like is currently just 128 bits? Does it need to be increased to handle 2048 bits... > > > >...

This problem occurs on attach as well: the staggered attach means that in an app with thousands of threads the threads taken over early start producing trace data well before...

From https://github.com/DynamoRIO/dynamorio/pull/5704#discussion_r1007187555: we should add an invariant check that timestamps always increase monotonically through each trace.

Happened on `code_api|tool.drcachesim.threads-with-config-file` at http://139.178.82.61:8080/job/DynamoRIO-AArch64-Precommit/2003/consoleFull