You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@edwardalee and I noticed that deadlines are not inherited when a chain of reactors spans across two enclaves. Here is the example we were looking at:
We were expecting Task1 inherit Task3's deadline (49ms) and be scheduled before Task4 (50ms deadline) within Enclave1, but apparently this is not happening. The first part of the output of the program is:
So Task4 with deadline 50ms is executed before Task1, which we expected to inherit the 49ms deadline, because apparently Task1 has no deadline. If we explicitly assign the deadline to Task1:
The execution is as expected:
Thanks for the report. We should definitely get support for propagating deadlines across enclave boundaries. It is probably not a big fix. What is more important is to actually get the c-enclaves merged. I might get some time to work on this the next weeks. I am happy for contributions to that :)
@edwardalee and I noticed that deadlines are not inherited when a chain of reactors spans across two enclaves. Here is the example we were looking at:
We were expecting Task1 inherit Task3's deadline (49ms) and be scheduled before Task4 (50ms deadline) within Enclave1, but apparently this is not happening. The first part of the output of the program is:
So Task4 with deadline 50ms is executed before Task1, which we expected to inherit the 49ms deadline, because apparently Task1 has no deadline. If we explicitly assign the deadline to Task1:
The execution is as expected:
Here is the code of the example:
Example.txt
And here the code with the explicit deadline for Task1:
ExampleExplicitDeadline.txt
The examples were run with the code from the "c-enclaves" branch.
Thank you!
The text was updated successfully, but these errors were encountered: