wallace added inline comments.
================
Comment at: lldb/source/Plugins/Trace/intel-pt/DecodedThread.h:296
std::vector<TraceItemStorage> m_item_data;
+ std::vector<TraceInstructionExtraInfo> m_insn_extra_info;
/// The TraceItemKind for each trace item encoded as uint8_t. We don't
include
----------------
jj10306 wrote:
> do we need to store this or can this information be calculated on demand?
> just thinking about massive traces and want to make sure we're intentional
> about only storing what we absolutely need to for each instruction
I thought about it and it will help us have more correctness for now and a
little bit more speed with little memory cost. The total size per instruction
went from 9 bytes to 11, which is very decent.
Now some details: the Disassembler fails with some instructions. I was tracing
dynolog and found many instances of the disassembler failing, but libipt
doesn't fail. In other words, we can't get the control flow kind from LLDB but
we can from libipt. Same for the byte size. That means that we need to fix the
disassembler, but that will require a lot of work that I can hardly prioritize,
so I'd rather use libipt for now.
================
Comment at: lldb/source/Plugins/Trace/intel-pt/TraceIntelPT.cpp:317
events_stats.total_count);
- for (const auto &event_to_count : events_stats.events_counts) {
- s.Format(" {0}: {1}\n",
- TraceCursor::EventKindToString(event_to_count.first),
- event_to_count.second);
+ for (auto it = events_stats.events_counts.begin();
+ it != events_stats.events_counts.end(); ++it) {
----------------
jj10306 wrote:
> what's the reason for this change?
wtf, i think this is from a previous commit i never intended to publish. I was
playing with some stuff. I need to revert this
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D137614/new/
https://reviews.llvm.org/D137614
_______________________________________________
lldb-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
