After a fair amount of trial and error, it seems to me that the I/O modules are identical, and there is some key in the shelf hardware (backplane perhaps) that sets the personality to be either a D2600 or M6612.
As you've seen, a Proliant Smart Array won't even detect an M6612 in hpsum, presumambly because they are not a supported combination. What I have done, however, is move I/O modules from an M6612 shelf to a "genuine" D2600 shelf, and then I can detect and update firmware to the latest version (150). I have done this with many modules that originally had M6612 firmware versions on them.
I have successfully run the module (now) with D2600 firmware in the M6612, and have experienced no issues in a few months deployment. Hpsum still won't detect the shelf, but it runs fine in normal operation. I didn't test it beyond simple drive failures and rebuilds, power supply fails, and the SPP diagnostic routine, so something esoteric could always show up.
I'd prefer to use the "real" D2600, but most people selling them don't even know the difference. The M6612 just had an AJ832A sticker stuck over the AJ940A sticker. That's easily removed, and a it often is, since the D2600 is worth more on the secondary market than is an M6612. A serial number search (HPE warranty check) should reveal what it is, but the last one I looked up said it was a power supply. Oh well. I've also compared the Smart Array diagnostic report with different combinations of hardware, but many of the values vary, and I didn't really find characteristics that definitely established a particular "gold" combination.
know for sure the EVA disk chassis stores a value like this in the midplane:
Enclosure
Vendor ID: HP
Product ID: M6412 AG638A
and I assume a D2000 stores an approriate product number - that's the difference.
Thank you very much for your post and revelations - this was the only useful post/information I could find on the topic and it helped me tremendously to get my "journey" started. What I found out in addition after following your lead (and I hope that this might help others in turn) is the following:
1) You can query the type (and firmware rev. etc.) of the enclosure with "ssacli" using the command:
controller slot=1 enclosure all show detail
This (and the steps below) works also for multipath and cascaded enclosures (NB: my p812 controller was at the most recent firmware version of 6.64). Here the M6612 clearly show as "AJ832A" instead of "AJ940A" in a pure "D2600"
2) I didn't have the luxury of a "real" D2600 being available for flashing, so I dug somewhat deeper. And while the M6612 wasn't detected by any update mechanism (not even the ccissflash utility in discovery mode, I guess that's the reason why not even patching the XML file helped), there is a trick to flash the latest firmware (0150 in this case) anyway: Just get the .rpm file from the latest (publicly available) SPP ISO (file is named hp-firmware-d2600-d2700-0150-2.1.x86_64.rpm) install it (in my case I did it on ubuntu via alien), then go to the directory with the installed files (here it was /usr/lib/x86_64-linux-gnu/hp-firmware-d2600-d2700-0150-2.1) and issue the command according to the data you see from the ssacli-output. In my case:
./ccissflash -iCAMSPR0150_6G_app_4MB_ai.pmc -t"D2600 SAS AJ832A" -v0150
(you can see that the AJ940A simply needs to be replaced with AJ832A)
After confirming (drives need to be offline), I had 4 enclosures successfully flashed in no time.
NB: Depending on your setup (i.e. with singlepath) you need to repeat the steps - after powering down and switching connections - to make sure you flash all SAS-modules in the enclosure
Thanks again @MidOfa , this wouldn't have been possible without your research!
P.S. My sticker with "AJ832A" was on the back of the enclosure (quite small) and nowhere near the "AJ940A" label