me trying to understand linux display internals: "huh why does the display panel mode timing struct have a 'hskew' property?"

The doc-comment in the kernel:

@cas i’m not sure i’m following - i’ve coded for at least one memory-mapped framebuffer which had each line aligned to %8, and the concept of horizontal skew makes perfect sense to me? is this something else?
@domi that sounds like the right usecase, it's used in the qualcomm drivers so clearly it's used for something. it's just funny that whoever wrote the doc comment shared my bemusement at this being a thing for digital displays
@cas oh! yeah i think i’ve had the same reaction the very first time, now that i think about it
@domi @cas this and a few other things are residual from analog display technology that stuck around. proprietary implementations are usually way worse in this fashion lol
@spinach @cas i wouldn’t call them residual, lots of those values map 1:1 to current concepts. even vblank is still a thing
@domi @cas well yeah there's a good reason they stuck around
@cas If i were to guess this is probably a relic relating to CRT displays
@cas this is for analog CRT overscan where there is extra non-visible sections of each line for synchronization. digital signalling does not use this but it still exists for backwards compatibility