Masks are rasters that must match the dimensions and pixel size of the image they will be applied to. You would need to create a mask for images that have different dimensions and pixel sizes, or resize the other data sets to match the mask.

If so, how can I make the different mask for 20m and 60 m bands? The F mask tutorial as mentioned above doesn't show any different options for masking. I mean from the f mask tutorial I can only make a single mak. I am a beginner and not getting complicated kinds of stuff. I would appreciate if could be provided with a simple tutorial.

Sorry, in this case you cannot generate additional masks because the spectral bands do not cover the necessary areas.
What you can do is after the 10m mask is created from the Sentinel-2 data, you can use Layer Stacking to stack the 10, 20, and 60m bands into a single file. Be sure to import the 10m data set first so it sets up the dimensions and pixel size. Then import the remaining data sets. The resulting stack will have the same dimensions and pixel size as the mask. You can then apply the mask to the full data set.
https://www.harrisgeospatial.com/docs/ManageRasterData.html#Layer

Thank you. Will try that way.

ENVI doesn't allow me to stack the bands saying the image is not georeferenced. How can I solve this?

What version of ENVI are you using? Are you opening your Sentinel-2 data using the XML metadata file?
If you open with the XMl, ENVI should recognize the georeferencing. If you open the JP2k files directly, the Layer Stacking tool may not recognize the georeferencing even if they show as having spatial reference when open in the ENVI display. In this case, you may need to save the files to ENVI format so the spatial reference is saved into the ENVI header. Then you can layer stack them.