2.2 is what you should be shooting for. Once you have run the calibration, you can go to "verify profile" under the tools menu (assuming you are using dispcalGUI) to get the deltaE values for each test patch.
I found that I got good enough results (average DeltaE of .65) with the medium quality setting using the default testchart with the "curves + matrix" setting. However, given that I am using DispcalGUI to mke the profile AND test it, it could potentially be inaccurate; I kind of just have to trust that the program knows what it is doing.
You want to make sure and do the preliminary RGB hardware adjustments before you start the software calibration; you want to get as close as you can with hardware.
Here is a helpful video I used when I was learning how to calibrate with dispcalGUI and the ArgyllCMS drivers:
Unfortunately the Qnixs and Xstars have no OSD to alter the RGB channels, so I'm stuck depending on the software to fix up any problems.
I do have dispcalGUI but I'm trying to do it all using argyllcms directly. It is more powerful for sure - I created a preliminary profile using 150 points, used targen with this profile to create a 6800 point monster that I left overnight I also want to profile my printer at some point so I thought it best to get used to the argyllcms commands.
DispcalGUI says in it's documentation that stuff like sRGB is intended to be played back at 2.4, even though it is encoded with a gamma of 2.2, which is confusing me. Pretty much everything I do falls into this category, which make things more complicated.
Not that it matters too much to me, but since I've got it I want the best results possible. Maybe I'll just make lots of profiles.