Closed expend20 closed 2 years ago
Hello @ifratric,
I've encountered a small problem in cmp coverage mechanics. It looks like the order of parameters does count. For example:
cmp rcx, [rsp+0x78] ; this won't work
vs:
cmp [rsp+0x78], rcx ; this is working
But in the first case some samples are found, so just as an assumptoin, the displacement may be calculated not accurately.
Project attached: jack_cmp.zip
UPD: just in case the parameters for the fuzzer:
fuzzer.exe -in c:\temp\in_1\ -out out -t 5000 -instrument_module test.exe -target_module test.exe -target_method FuzzIteration -nargs 1 -iterations 10000 -persist -loop -cmp_coverage -- test.exe @@
This should be fixed by https://github.com/googleprojectzero/TinyInst/commit/022421abb6e027d15d11bef835c4cb4b52b2d45f
I ran your test target and confirmed it's now working correctly. Cheers!
Hello @ifratric,
I've encountered a small problem in cmp coverage mechanics. It looks like the order of parameters does count. For example:
vs:
But in the first case some samples are found, so just as an assumptoin, the displacement may be calculated not accurately.
Project attached: jack_cmp.zip
UPD: just in case the parameters for the fuzzer: