Re: [PATCH 5/7] mm: free user PTE page table pages

[Qi Zheng]

On 7/19/21 6:01 AM, Kirill A. Shutemov wrote:
> On Sun, Jul 18, 2021 at 12:30:31PM +0800, Qi Zheng wrote:
> > Some malloc libraries(e.g. jemalloc or tcmalloc) usually
> > allocate the amount of VAs by mmap() and do not unmap
> > those VAs. They will use madvise(MADV_DONTNEED) to free
> > physical memory if they want. But the page tables do not
> > be freed by madvise(), so it can produce many page tables
> > when the process touches an enormous virtual address space.
> >
> > The following figures are a memory usage snapshot of one
> > process which actually happened on our server:
> >
> >          VIRT:  55t
> >          RES:   590g
> >          VmPTE: 110g
> >
> > As we can see, the PTE page tables size is 110g, while the
> > RES is 590g. In theory, the process only need 1.2g PTE page
> > tables to map those physical memory. The reason why PTE page
> > tables occupy a lot of memory is that madvise(MADV_DONTNEED)
> > only empty the PTE and free physical memory but doesn't free
> > the PTE page table pages. So we can free those empty PTE page
> > tables to save memory. In the above cases, we can save memory
> > about 108g(best case). And the larger the difference between
> > the size of VIRT and RES, the more memory we save.
> >
> > In this patch series, we add a pte_refcount field to the
> > struct page of page table to track how many users of PTE page
> > table. Similar to the mechanism of page refcount, the user of
> > PTE page table should hold a refcount to it before accessing.
> > The PTE page table page will be freed when the last refcount
> > is dropped.
>
> The patch is very hard to review.
>
> Could you split up introduction of the new API in the separate patch? With
> a proper documentation of the API.

Good idea, i will do it.

> Why pte_refcount is atomic? Looks like you do everything under pmd_lock().
> Do I miss something?

When we do pte_get_unless_zero(), we hold pmd_lock to protect against
free_pte_table(). But we don't need to hold the pmd lock when we do
pte_get()/pte_put() in mapping/unmapping routine.

> And performance numbers should be included. I don't expect pmd_lock() in
> all hotpaths to scale well.

Yeah, so we use rcu lock to replace the pmd lock in some routines in the
subsequent patch (mm: defer freeing PTE page table for a grace period).

Thanks,

Qi