As I stated give it a try or ask Canon. That might get to the bottom of it. What you quoted is that 3rd party software will ignore the setting. It doesn't state that the camera has not already monkeyed with your image. That is a carefully worded quote including the part about "most" software.
Regarding HTP, yes it does bring up the noise by about one stop...and in some cases it's better to go the other way and overexpose with a higher ISO to push the noise down by a stop or two in post.
In the case of this aircraft I would have over exposed the aircraft by a 1 or 1.5 stop (depending on full sun or overcast) and blown out the sky then pulled the aircraft back down with DPP or ACR. The sky in this image appears too dark. Could very well be HTP was on and was trying to appease the highlights. Would also be nice to see the metadata tagged to the image as reported by DPP.
As far as HTP and HINR is concerned, I am the source and now turn both off due to the effects it produces on RAW images. There are a couple of options in the FW that affect raw images and those are two if them.
In that case, I'd like to respectfully ask 'the source' for some evidence. I believe that the only in-camera setting that affects the RAW image data is long exposure NR, since the dark frame is subtracted before the RAW data are written.
FYI, 'the real source' (aka Canon) disagrees with you, and states that while some Canon software will try to replicate the effects of in-camera HINR (but DPP, not so much), "Virtually all third-party RAW file software programs, such as Adobe’s Camera Raw™ software, will ignore in-camera settings such as High ISO Noise Reduction....Don’t expect the High ISO Noise Reduction you may have set in-camera to have any effect with most third-party software programs."
As for HTP, that also doesn't affect the RAW image data directly, although I suppose you can reasonably say it does, since you're getting a RAW exposure that's a stop underexposed relative to the ISO you picked. Like HINR, it also sets a flag in the metadata, but HTP also 'lies' in the metadata, recording ISO that's one stop higher than was actually used to expose the shot. For example, if you set ISO 400 with HTP, the camera really exposes at ISO 200, but records ISO 400 in the metadata and then applies a tone curve to the JPG file to boost the shadows by a stop - that underexposure is how it preserves highlights (and of course, you also get an extra stop of shadow noise as a consequence). Most RAW converters recognize the flag, and try to duplicate Canon's HTP tone curve with their own version (DPP uses Canon's version, obviously, but unlike HINR, 3rd party converters don't just ignore the metadata flag). RAW apps like rawnalyze just show you the 1-stop underexposed file. Point is, IMO rather than setting HTP, you're better off just exposing to preserve the highlights yourself, then boosting shadows in post yourself, where you have more control over the process.