but I also know, that if previously there was no sound at all, or it changes from clicking to squeaking, it's usually the result of updating camera firmware.
Where did you get the idea a firmware change can initiate squeaking in the sensor cleaning operation?
I can never understand why people fall for the 'send your body with your lenses' for calibration by third-party lens makers. The whole idea of interchangable lens cameras is that all lenses work on all bodies.
PaulB, you don't understand tolerancing and the effect of tolerance stackup which occurs when multiple devices, each with a tolerance, are "married" together and the stackup of the two individual tolerances affect the outcome. In the case of a camera/lens combination each has a tolerance. Lets just refer to the tolerances for sake of simplicity as a (+) plus and (-) minus situation and each device can have a plus or minus 5 units tolerance and still be within manufacturer's tolerance. Further (+) and (-) will refer to the same type (front or back focus) of tolerance in both the camera and lens. So to the specifics; Canon assembles a camera and the tolerance in the lens mount falls at +4. It is acceptable by their standards since the specification is some quantity ±5 units. Sigma assembles a lens and puts it on the test fixture and finds it is at -4 units, again within the specs of ±5 so it is deemed acceptable. That body/lens combination is essentially a perfect match since a +4 units and a -4 units cancel out. You can pixel peep and be very happy with that combination, so happy that you go out an buy another Sigma lens. The bad thing is that when that 2nd lens was assembled, it tested out at +4 units. Sigma is happy with the lens which meets the ±5 unit specification. However when that lens is married up to the +4 Canon body, the combination is now out +8 units. The +8 represents the tolerance stackup which is out of the ±5 unit spec. Pixel peep on that combination and you might not be happy. The important thing here is each device, the camera body and the lens, separately meet their respective specification but the combination can fail.
As to the 1st part of your statement "If (for eg.) Sigma calibrate a lens to a particular body then the odds are that it will not work properly with a different body". That part is fundamentally true but Mag-1981 is happy with the combination, should that individual chose to pixel peep.
To the 2nd part of your statement, "Sigma should calibrate the lens to be within specification and if the body is out of specification then Sigma will not be able to adjust it anyway". Sigma won't adjust the body but may, note I said may, adjust the lens to match the body and still be within the lens specification. In our example the body was at +4 units; the lens upon initial examination by Sigma was found to also be at +4 units, so they adjust the lens to -4 units. The lens is in spec as is the body and a perfect match is the result.