From 32525238238e96ec0ce64a36f70558f76bc90ff5 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Mon, 15 Jul 2019 11:22:54 +0200 Subject: fixing commit fuckup --- llvm_mode/README.llvm | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) (limited to 'llvm_mode/README.llvm') diff --git a/llvm_mode/README.llvm b/llvm_mode/README.llvm index 77c406f8..779ff47c 100644 --- a/llvm_mode/README.llvm +++ b/llvm_mode/README.llvm @@ -88,13 +88,18 @@ which C/C++ files to actually intrument. See README.whitelist For splitting memcmp, strncmp, etc. please see README.laf-intel -As the original afl llvm_mode implementation has been replaced with -then much more effective instrim (https://github.com/csienslab/instrim/) -there is an option for optimizing loops. This optimization shows which -part of the loop has been selected, but not how many time a loop has been -called in a row (unless its a complex loop and a block inside was -instrumented). If you want to enable this set the environment variable -LOOPHEAD=1 +Then there is an optimized instrumentation strategy that uses CFGs and +markers to just instrument what is needed. This increases speed by 20-25% +however has a lower path discovery. +If you want to use this, set AFL_LLVM_INSTRIM=1 +See README.instrim + +Finally if your llvm version is 8 or lower, you can activate a mode that +prevents that a counter overflow result in a 0 value. This is good for +path discovery, but the llvm implementation for intel for this functionality +is not optimal and was only fixed in llvm 9. +You can set this with AFL_LLVM_NOT_ZERO=1 +See README.neverzero 4) Gotchas, feedback, bugs -- cgit 1.4.1