Repository navigation
Fix stack buffer overflow in get_view_as_real_copy_out_target_size - #15625
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15625
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 2 Unrelated FailuresAs of commit b8250c5 with merge base 3dbc15b ( NEW FAILURE - The following job has failed:
BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This PR needs a
|
…ytorch#15625) Summary: Check that self.dim() is smaller than kTensorDimensionLimit, otherwise we may write past the end of the out tensor. Reviewed By: JacobSzwejbka Differential Revision: D86356264
31eea79 to
0bed02f
Compare
…ytorch#15625) Summary: Check that self.dim() is smaller than kTensorDimensionLimit, otherwise we may write past the end of the out tensor. Reviewed By: JacobSzwejbka Differential Revision: D86356264
0bed02f to
b8250c5
Compare
|
@pytorchbot cherry-pick --onto release/1.0 -c critical |
Cherry picking #15625The cherry pick PR is at #15718 and it is recommended to link a critical cherry pick PR with an issue. The following tracker issues are updated: Details for Dev Infra teamRaised by workflow job |
Differential Revision: D86356264 Pull Request resolved: pytorch#15625
Summary: Check that self.dim() is smaller than kTensorDimensionLimit, otherwise we may write past the end of the out tensor.
Differential Revision: D86356264