Create Linux_Memory_Management_Essentials.md

[igor-stoppa]

Description of internal memory management in Linux, from safety perspective.

[paolonig]

Above I would say "if a non-interference claim cannot be made, due to the extremely complex ...."

[igor-stoppa]

I intentionally made a stronger statement.
Also to make it easier to prove me wrong :-)

[igor-stoppa]

time flies ... but positive testing is still insufficient, at least witout evidence of the exhaustiveness of its coverage

[dweingaertner]

My general comment is that we should separate what is knowledge about the kernel memory management system, provide it with reference anchors, and separately discuss the safety implications, as these are not of the same nature.

[dweingaertner]

Same comment as above. We should to separate the premises from these considerations, and explicitly state the premises each conclusion builds on. Take for example the 1 argument below. Uncertainty in the timing might be a problem, but might not, depending on the requirements or on the size of the uncertainty.

[igor-stoppa]

Agreed to reconsider later the overall posture of the document.
My motivation is to try to keep together statements which are kind of obvious with others that might be less obvious, but derivable.
We could consider - in general - to introduce also the notion of which aspects are likely to be unavoidable, and which ones might be relevant only for some use case scenarios.