Using the serialized protobuf definitions from here, the list of basic blocks for function 0x1B380 of linux\libs\clang_O0\libtiff.so.5.4.0 are incorrect.
The ida pro ground truth(linux\libs\clang_O0\Block-idaBlocks-libtiff.so.5.4.0.pb) lists these blocks as children of 0x1B380:
0x1B380
0x1B3B2
0x1B3BE
0x1B3DC
0x1B3E7
0x1B3F2
0x1B42E
0x1B43B
0x1B45C
0x1B468
0x1B479
0x1B48B
The OracleGT ground truth(linux\libs\clang_O0\gtBlock_libtiff.so.5.4.0.pb) lists these blocks as children of 0x1B380:
0x1B479
0x1B48B
0x1B4A0
0x1B4C5
0x1B4DD
0x1B4F1
0x1B517
0x1B53D
0x1B563
0x1B589
0x1B5AF
0x1B5D5
0x1B5FB
0x1B61A
I cross referenced with both IDA & Binary Ninja, and the IDA pro ground truth is definitely correct - OracleGT is listing blocks which are not owned by 0x1B380. Is this a bug?
Using the serialized protobuf definitions from here, the list of basic blocks for function
0x1B380
oflinux\libs\clang_O0\libtiff.so.5.4.0
are incorrect.The ida pro ground truth(
linux\libs\clang_O0\Block-idaBlocks-libtiff.so.5.4.0.pb
) lists these blocks as children of0x1B380
:0x1B48B
The OracleGT ground truth(
linux\libs\clang_O0\gtBlock_libtiff.so.5.4.0.pb
) lists these blocks as children of0x1B380
:I cross referenced with both IDA & Binary Ninja, and the IDA pro ground truth is definitely correct - OracleGT is listing blocks which are not owned by
0x1B380
. Is this a bug?