fix: strengthen lcd buffer protection

[3djc]

Improve LCD buffer protection by checking negative x/y values

This fixes #4544

NB:

• this should be applied to 2.9.x too

[pfeerick]

What was the actual step to reproduce this... as I wasn't able to do it on my X7 or TX12MK2? And can wind it (offset) all the way up to 100, or down to -100 without issue... just wondering what the missing piece of the puzzle is to repo the issue. 😕

[3djc]

2.9.2, ail input, raise offset, crash. It draws at negative y value

ok on simu obviously, but crashes x7

[3djc]

Adding debug in void drawFunction(FnFuncP fn, uint8_t offset) here:

produces negative Y values

[philmoz]

Posted an additional fix in the issue.

[pfeerick]

Well, since I'm talking about hardware... nice to know simu doesn't have any problem with it 🤪 It seems the 2.10 PR I had on the X7 was immune to this. 2.9.2 isn't though, and the channel doesn't matter (as it obviously shouldn't). Still interesting that the TX12MK2 on 2.9.2 doesn't have any issue... maybe something different with the F4, as the TLite is also susceptible to this? 🤯

Anyway, I'll leave this till tomorrow to let you make any further changes... to be merged before nightly unless you mark it otherwise...

[3djc]

well, you are writing outside LCD buffer, but there is not saying where you actually write, its a bit random, can depend on radio, compiler, ..