Hi all!
I am studing LRU page replacement policy. I read that , kernel maintains
two list (active list and inactive ) per zone .
When there is first reference to page , page is addeed to inactive list
(PG_lru =1 and RG_referenced=0) . When there is second reference
PG_referenced bit get set . And when there is third reference to same page ,
it will get added to active list.
But when there is page fault , handle_mm_fault get invoked. If page
was not previously refered , it calls do_anonymous_page (demand paging
conept). This function allocate new page frame and directly insert that
page in active list .
My question is that , why do_anonymous_page (infact all page fault
pages) directly add pages to active list , eventhough they have first
reference ?
Thanks
--
Shyam
"Born To Learn"
