Re: [PATCH 2.6.14-rc2] fix incorrect mm->hiwater_vm and mm->hiwater_rss

[Hugh Dickins <hugh@veritas.com>]

On Tue, 27 Sep 2005, Jay Lan wrote:
> 
> While in the work on separating hiwater_vm from hiwater_rss, i noticed
> that __vm_stat_account() was not called in these functions where
> total_vm was updated:
> mm/mmap.c                           do_brk
> mm/nommu.c                          do_mmap_pgoff
> mm/nommu.c                          do_munmap
> arch/ppc64/kernel/vdso.c            arch_setup_additional_pages
> arch/x86_64/ia32/syscall32.c        syscall32_setup_pages
> 
> Frank tried to touch the latter two in his proposed patch.
> Does it make sense we add __vm_stat_account() calls to the above
> routines?

Probably not.  Partly because of the PROCFS issue you noticed after.
And partly because that's a long list, whereas the evidence is that
__vm_stat_account is nowadays being called in the places that need it.

What, you couldn't find the call to __vm_stat_account next to
updating total_vm in do_mmap_pgoff :-?  And I don't see that total_vm
is updated in do_munmap, that and the vm_stat_account have to be done
at a lower, per-vma level.  Haven't looked at the others yet.

I think, better leave hiwater_vm to me for now.  Since I've got that
patchset in the next -mm, it's maybe easier for me to put something
together on top of that.  And it seems quite possible (haven't checked
yet) that actually we should be moving the update of total_vm in to the
core from out all over, rather than splattering hiwater_vm updates about.

I'll take a look in the next couple of days.

Hugh