Here are some things to check if there are problems running syzkaller.
-
Use the
-debug
command line option to make syzkaller print all possible debug output, from both thesyz-manager
top-level program and thesyz-fuzzer
instances. With this option syzkaller will only run one VM instance. -
Use the
-vv N
command line option to increase the amount of logging output, from both thesyz-manager
top-level program and thesyz-fuzzer
instances (which go to the output files in thecrashes
subdirectory of the working directory). Higher values of N give more output. -
If logging indicates problems with the executor program (e.g.
executor failure
), try manually running a short sequence of system calls:- Copy
syz-executor
andsyz-execprog
into a running VM. - In the VM run
./syz-execprog -executor ./syz-executor -debug sampleprog
where sampleprog is a simple system call script (e.g. just containinggetpid()
). - For example, if this reports that
clone
has failed, this probably indicates that the test kernel does not include support for all of the required namespaces. In this case, running thesyz-execprog
test with the-sandbox=setuid
option fixes the problem, so the main configuration needs to be updated to setsandbox
tosetuid
.
- Copy
Also see this for Linux kernel specific troubleshooting advice.
If none of the above helps, file a bug on the bug tracker
or ask us directly on the [email protected] mailing list.
Please include syzkaller commit id that you use and syz-manager
output with -debug
flag enabled if applicable.