tpm0: Operation timed out

[DaanDeMeyer]

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):

• OS: Fedora
• Version: 39

Versions of relevant components

• swtpm: 0.8.1
• libtpms: 0.9.6
• openssl: 3.1.1
• gnutls: 3.8.5
• qemu: 8.1.3

Log files

Not available due to spurious nature of the issue

[stefanberger]

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?