Re: Resident memory set size
Michael Kelly, le lun. 12 mai 2025 19:34:13 +0100, a ecrit:
> Some of these things I had thought of as possibilities and some I hadn't
> although I should have (for example that the vm_page_copy would set the
> hardware dirty flag).
It wasn't completely immediate to me either:)
> It is seems misleading to report the RSS as being those just mapped
> when the memory is really still in use (although not mapped) to
> preserve the dirty information but I suppose it doesn't affect much.
I don't think the two are related.
I guess not remapping was just simpler (you'd otherwise have to remember
which process what mapping what), while preserving the dirty information
is just about copying the page flag and clearing the hardware flag if
needed.
Samuel
Reply to: