labs: debugging: add hint to properly generate flamegraphs

The current labs setup have some lacks regarding perf data analysis on host
side (some symbols still missing despite provided --symfs and -k), and
Flamegraph generation is affected by the issue.
Suggest as an alternative to export the output of "perf script" onto host
instead of perf.data file when generating flamegraphs.

Also, system wide recording is not very interesting in this case (no
specific load on the system), so suggest to also reproduce the flamegraph
part on some previous labs.

Signed-off-by: Alexis Lothoré <[email protected]>

labs/debugging-system-wide-profiling/debugging-system-wide-profiling.tex

@@ -235,3 +235,11 @@ \section{(Bonus) System profiling with {\em perf} and FlameGraphs}
 \end{bashinput}
 
 Using this flamegraph, analyze the system load.
+
+{\em TIP: if the generated graph is not relevant/missing symbols, you may try to to all symbols
+translation in the target before generating the graph on host, i.e use perf script on the target,
+and bring back the result on host and provide it to Flamegraph scripts}
+
+You can also generate a Flamegraph on previous labs: for example you can re-do a recording on
+\code{png_convert} from the application profiling lab, generate the corresponding Flamegraph and confirm
+your observations about most time-consuming functions.