Closed shaobo-he closed 3 years ago
Are you sure that it is still %24
in the analyzed code? The code is transformed prior to alias analysis, so it is possible that the name simply has changed.
I will look deeper, but it may take me a little time before I can get into this. Would be nice if this is not a true bug, but only a renaming :)
Thank you for your quick reply. I grepped the file produced using the -oll
option and it still contains %24
,
$ grep "%24 = getelementptr.*%23" real.ll
%24 = getelementptr inbounds %"class.std::__1::__libcpp_compressed_pair_imp.4", %"class.std::__1::__libcpp_compressed_pair_imp.4"* %23, i32 0, i32 0, !dbg !11275, !verifier.code !6341
%24 = getelementptr inbounds %"struct.std::__1::__list_node_base", %"struct.std::__1::__list_node_base"* %23, i32 0, i32 1, !dbg !11820, !verifier.code !6341
@Shaobo: I'll close this issue. Open it if it's not solved yet.
Hello sea-dsa developers,
While I was debugging a soundness issue when using sea-dsa for SMACK, I couldn't find a node that corresponds to a pointer variable. In
demo.ll
,%24
of function_ZN9lru_cacheC2Ev
is used as the pointer argument to a store instruction,I would expect it to show up in the memory graph (i.e.,
sgx_fopen.mem.dot
),I'm using the
dev10
branch and the following command to generate the memory graph (I had to changelib/seadsa/DsaPrinter.cc
to let it find the graph since the entrypoint issgx_fopen
instead of `main).Please let me know if it's expected or I'm missing anything. Thank you.
debug.zip