device bus resets (data loss) with perfecly working disks (bus vs. scterc timeouts)

The default hardware bus timeout is so short (30sec) that it causes complete bus resets before consumer/soho storage devices are done with their error correction, and would return with either read success, or a proper single-block reading errors.

That is why one sees only strange bus resets in the logs [and silent data loss], instead of just experiencing harmless drive or OS block re-mapping or single-block reading errors.

And on raid setups the reset typically breaks the complete raid array, often during recovery re-sync.

Proposed v1.02 udev rules can be found as attachment here: https://www.smartmontools.org/ticket/658