| commit | 22cf3d45ee6d1966d6735d11b4c7e93887a82c62 | |
| author | Mark Williams <mwilliams@fb.com> | |
| Mon, 23 Mar 2020 13:00:01 +0000 (23 06:00 -0700) | ||
| committer | Facebook GitHub Bot <facebook-github-bot@users.noreply.github.com> | |
| Mon, 23 Mar 2020 13:04:19 +0000 (23 06:04 -0700) | ||
| tree | 6ad7610e76327fc0b44c5d01061d163a538d1170 | treesnapshot (tar.gz zip) |
| parent | 5c01868edc5bea3fbae185e74e280d09c3d6ef85 | commitdiff |
Fix unwinding through identical frames
Summary:
It was possible for multiple frames in the same nesting to get the
same id, because we use the sp (which doesn't change through a given
nesting) and pc (which can be the same for multiple different
frames in the backtrace).
Add the fp value as the 'special' field in FrameId to differentiate.
Reviewed By: oulgen
Differential Revision: D20590529
fbshipit-source-id: 37293178d91d23b11a8eef16e2b352a24d67e334
Summary:
It was possible for multiple frames in the same nesting to get the
same id, because we use the sp (which doesn't change through a given
nesting) and pc (which can be the same for multiple different
frames in the backtrace).
Add the fp value as the 'special' field in FrameId to differentiate.
Reviewed By: oulgen
Differential Revision: D20590529
fbshipit-source-id: 37293178d91d23b11a8eef16e2b352a24d67e334
| hphp/tools/gdb/unwind.py | diffblobblamehistory |