Closed 10110111 closed 8 years ago
Agreed, I actually have something in mind which will revamp how reads work. Basically I want to have all read operations done in terms of DebuggerCore::read_pages
as you suggest. But to make it even more efficient, I want to have a caching system (which is expired whenever the debugger leaves a stopped state). That way, we can do lots of small reads without hurting performance :-).
BTW, I've currently done read_bytes
in terms of reading /proc/$PID/mem
, and now the running process can have its memory read (which addresses your issue 191).
In fact, this particular issue has been addressed in my bitness
branch (commit ddfd07c727045b96f0ff52a27abbe9176d05e1f8 and its neighborhood).
PlatformProcess::read_pages()
should callDebuggerCore::read_pages()
by definition. AndPlatformProcess::read_bytes()
should also useDebuggerCore::read_pages()
, otherwise it's very slow on 32-bit EDB with 64-bit debuggee on reading e.g. stack contents.