stefanberger
commented
4 months ago Not easily reproduced given the spurious nature of the problem.
Do you have a hint which type of commands are timing out?
No timeouts happen or a parameter to increase the timeout is available which we can use.
There aren't any knobs to turn. We're dealing with hardcoded timeouts: https://elixir.bootlin.com/linux/v6.8/source/include/linux/tpm.h#L182 Are you seeing the timeouts on a particularly busy host where not enough cycles are available for TPM command processing maybe?
Describe the bug
We're starting to run the systemd integration tests in virtual machines with an emulated TPM attached (backed by swtpm). We've been seeing spurious "tpm0: Operation timed out" failures in the virtual machines, causing tests to fail.
Required: To Reproduce (without these steps your issue may be deleted)
Not easily reproduced given the spurious nature of the problem.
Expected behavior
No timeouts happen or a parameter to increase the timeout is available which we can use.
Desktop (please complete the following information):
Versions of relevant components
Log files
Not available due to spurious nature of the issue