40-150 f2.8 bricked by EM1mkii

Photo Pete wrote:

radsaq wrote:

As a software developer, I'm surprised that it's even possible for this to happen. What theyshould be doing, and what it appears like given the appearance of the update process, is:

1. Download update file from to PC
2. Verify checksum/signature of firmware file on PC
3. Copy update file to camera
4. Verify checksum/signature of firmware file on camera
5. Erase existing firmware
6. Apply new firmware from file

If this is how it worked, then the only failure case should be if power is removed after step 5 completes and before step 6 completes. If the initial download from the internet doesn't get a complete file, or the camera doesn't get a complete file from the PC, it is trivial to detect this. And certainly the camera shouldn't erase anything before it knows it has a valid update file.

Of course, I've never verified that these are in fact the steps that the Olympus update does take. But it certainly looks like it goes through a download, copy to camera, then flash step. Maybe they're not properly looking at signatures/checksums? This is fairly basic stuff, though. It shouldn't be hard to make it mostly bullet-proof (aside from potential loss of power to the camera).

Exactly my thoughts. Communication errors should not be a risk to the update process if the update process is sufficiently robust. If the file is installed to the camera and checked before the update starts then there should be no need to be tethered to a computer at all when the firmware update is actually being applied.

However, the Olympus instructions clearly state that you should not disconnect your camera from the computer whilst it is being updated (until 'OK' is displayed on the camera rear screen)... so I guess the connection is required for something and we can assume that they are not installing in the way that you have indicated?

Regardless of the technicalities, I have found out the hard way that the update process is not particularly robust and that a lens can be rendered useless should a communication error occur, even when the user is following Olympus's instructions precisely.

I always assumed that the "don't disconnect until it's done" is there simply so that people don't get careless. But I've never been willing to experiment and find out.