From 268a53de823a6750d6256ee1fb1e7707b4b45740 Mon Sep 17 00:00:00 2001 From: Xavier ASUS Date: Fri, 18 Oct 2019 00:31:54 +0200 Subject: sdcc-3.9.0 fork implementing GNU assembler syntax This fork aims to provide better support for stm8-binutils --- sim/ucsim/doc/analyzer.html | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) create mode 100644 sim/ucsim/doc/analyzer.html (limited to 'sim/ucsim/doc/analyzer.html') diff --git a/sim/ucsim/doc/analyzer.html b/sim/ucsim/doc/analyzer.html new file mode 100644 index 0000000..7692fe8 --- /dev/null +++ b/sim/ucsim/doc/analyzer.html @@ -0,0 +1,32 @@ + + + + μCsim: Code Analyzer + + +

Code analyser

+ Code + analyser analyses the code of the simulated program and tries to + figure out which address contains valid instruction.
+
+ Analyser starts + analysing of the code with address 0 which always must contain an + instruction because reset of the device sets zero into program + counter. So, analyser marks address 0 that it contains an + instruction. Then it marks the address of the next instruction and so + on. If the analyser finds a jump instruction it continues with + destination address of the jump instruction. If it finds a + call instruction, it recursively analyses destination address + as well as next instruction follows the call one.
+
+ Analysator + stops if it finds an address which is already marked or if it finds a + return instruction or an indirect jump which is + impossible to follow. All these mean that it is impossible to discover + all instructions of the program. +

This problem is solved in very simple way. If the + execution reaches an unmarked address, the analyser is called to + analyse the code starting at actual address pointed by the PC.

+ + -- cgit v1.2.3