Question
My question is how fast is mprotect
. What will be the difference between mprotecting say 1 MB of contiguous memory as compared to 1 GB of contiguous memory? Of course I can measure the time, but I want to know what goes under the hood.
La solution
A quick check on the source seems to indicate that it iterates over the process mappings in the selected region and change their flags. If you protect less than a whole mapping it'll split it into two or three.
So in short it's O(n)
where n
is the number of times you've called mmap.
You can see all the current maps in /proc/pid/maps
Autres conseils
It is O(n) on page count in the region too, because it should change a access bits on all PTEs (Page translation Entries, which describes virtual->physical page mappings in PageTable). Calltree:
mprotect
->
mprotect_fixup
->
change_pte_range
http://lxr.free-electrons.com/source/mm/mprotect.c#L32
47 do {
48 oldpte = *pte;
49 if (pte_present(oldpte)) {
50 pte_t ptent;
51
52 ptent = ptep_modify_prot_start(mm, addr, pte);
53 ptent = pte_modify(ptent, newprot);
54
55 /*
56 * Avoid taking write faults for pages we know to be
57 * dirty.
58 */
59 if (dirty_accountable && pte_dirty(ptent))
60 ptent = pte_mkwrite(ptent);
61
62 ptep_modify_prot_commit(mm, addr, pte, ptent);
63 } else if (PAGE_MIGRATION && !pte_file(oldpte)) {
64 swp_entry_t entry = pte_to_swp_entry(oldpte);
65
66 if (is_write_migration_entry(entry)) {
67 /*
68 * A protection check is difficult so
69 * just be safe and disable write
70 */
71 make_migration_entry_read(&entry);
72 set_pte_at(mm, addr, pte,
73 swp_entry_to_pte(entry));
74 }
75 }
76 } while (pte++, addr += PAGE_SIZE, addr != end);
Note the increment: addr += PAGE_SIZE, addr != end);