Closed bxatnarf closed 4 years ago
The strange thing about stack traces printed in the log is that none of them are for the migrated process -- they are all for sshd or something with PID 51. Nevertheless, I think this is popcorn-related.
Going to close this after 1f12a34a25a122f6b0e512b0326b5b199daf215c. Thanks!
branch: merge arch: x86-64 triggering example: https://github.com/ssrg-vt/popcorn-kernel-lib/blob/master/src/stack.c
~NB:
stack
does not successfully execute on native popcorn (master branch) either (it segfaults), but no BUG statement are triggered in this version.~ NB:stack
works for me if I compile it on one of the VMs. I'm guessing that it has something to do with its use of__builtin_frame_address
. Therefore, to trigger this bug you should buildstack
on a different enough x86-64 host .After
stack
exits (unsuccessfully), the remote kernel encounters bad state. Remote kernel log (truncated): notice that many of these stack traces say "BUG: Bad page state in process bash pfn:1374cd
" There are also a few that reference thesshd
process.