[WIP] Avoid using dladdr with libunwindstack #658
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
fractalwrench
force-pushed
the
termination-handler
branch
from
6c37fb7
to
a6d0807
Compare
fractalwrench
force-pushed
the
avoid-dladdr
branch
from
b3ed3d4
to
6b9bf04
Compare
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## termination-handler #658 +/- ##
=======================================================
- Coverage 80.06% 79.97% -0.09%
=======================================================
Files 416 416
Lines 10934 10934
Branches 1673 1673
=======================================================
- Hits 8754 8745 -9
- Misses 1506 1515 +9
Partials 674 674 |
fnewberg
reviewed
Apr 15, 2024
| @@ -44,6 +44,9 @@ static inline void emb_copy_frame_data(unwindstack::AndroidUnwinderData &android | |||
| data->offset = map_info->offset(); | |||
| data->flags = map_info->flags(); | |||
| emb_strncpy(data->full_name, map_info->GetFullName().c_str(), EMB_FRAME_STR_SIZE); | |||
|
|
|||
| // calculate line number | |||
| data->line_num = data->pc - data->start; | |||
There was a problem hiding this comment.
isn't line_num meant to refer to the symbolicated line number? this looks like it's computing the module address offset
| RETURN_ON_JSON_FAILURE(json_object_set_number(frame_object, kModuleAddrKey, frame->module_addr)); | ||
| RETURN_ON_JSON_FAILURE(json_object_set_number(frame_object, kLineNumKey, frame->line_num)); | ||
|
|
||
| RETURN_ON_JSON_FAILURE(json_object_set_number(frame_object, kFrameAddrKey, frame->pc)); |
There was a problem hiding this comment.
i'm not sure i follow these changes. some of the names in emb_sframe seem a bit confusing and some of the fields seem like they might be redundant. probably worth a chat to go over that to see if we can clean up some stuff there
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Goal
Avoids using
dladdrby using functions available in libunwindstack directly. This avoids the need for extra symbolication code that doesn't necessarily account for edge cases introduced by Android.The SO load address is different by a few thousand bytes when calculated by libunwindstack. I believe this is because libunwindstack accounts for how Android reads elf files whereas dladdr does not.
Testing
I tested 3 representative crashes on an ARM emulator & compared stacktraces (see attached).
The symbolication for libart.so appears to be more complete for libunwindstack (it correctly gets
mterp_op_invoke_static, whereas dladdr doesn't). However, all the embrace frames seem to incorrectly point atnullPointerDereference. I'd assume the different SO load address means our backend symbolication chooses this function.