From: Muchun Song <muchun.song@linux.dev>
To: Qi Zheng <zhengqi.arch@bytedance.com>
Cc: david@redhat.com, hughd@google.com, willy@infradead.org,
vbabka@kernel.org, akpm@linux-foundation.org, rppt@kernel.org,
vishal.moola@gmail.com, peterx@redhat.com, ryan.roberts@arm.com,
christophe.leroy2@cs-soprasteria.com,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
linux-arm-kernel@lists.infradead.org,
linuxppc-dev@lists.ozlabs.org
Subject: Re: [PATCH v4 07/13] mm: khugepaged: collapse_pte_mapped_thp() use pte_offset_map_rw_nolock()
Date: Tue, 24 Sep 2024 15:14:25 +0800 [thread overview]
Message-ID: <79699B24-0D99-4051-91F3-5695D32D62AC@linux.dev> (raw)
In-Reply-To: <07d975c50fe09c246e087303b39998430b1a66bd.1727148662.git.zhengqi.arch@bytedance.com>
> On Sep 24, 2024, at 14:11, Qi Zheng <zhengqi.arch@bytedance.com> wrote:
> In collapse_pte_mapped_thp(), we may modify the pte and pmd entry after
> acquring the ptl, so convert it to using pte_offset_map_rw_nolock(). At
> this time, the pte_same() check is not performed after the PTL held. So we
> should get pgt_pmd and do pmd_same() check after the ptl held.
>
> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
> ---
> mm/khugepaged.c | 14 +++++++++++---
> 1 file changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index 6498721d4783a..8ab79c13d077f 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -1605,7 +1605,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
> if (userfaultfd_armed(vma) && !(vma->vm_flags & VM_SHARED))
> pml = pmd_lock(mm, pmd);
>
> - start_pte = pte_offset_map_nolock(mm, pmd, haddr, &ptl);
> + start_pte = pte_offset_map_rw_nolock(mm, pmd, haddr, &pgt_pmd, &ptl);
> if (!start_pte) /* mmap_lock + page lock should prevent this */
> goto abort;
> if (!pml)
> @@ -1613,6 +1613,9 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
> else if (ptl != pml)
> spin_lock_nested(ptl, SINGLE_DEPTH_NESTING);
>
> + if (unlikely(!pmd_same(pgt_pmd, pmdp_get_lockless(pmd))))
> + goto abort;
> +
> /* step 2: clear page table and adjust rmap */
> for (i = 0, addr = haddr, pte = start_pte;
> i < HPAGE_PMD_NR; i++, addr += PAGE_SIZE, pte++) {
> @@ -1645,7 +1648,6 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
> nr_ptes++;
> }
>
> - pte_unmap(start_pte);
> if (!pml)
> spin_unlock(ptl);
>
> @@ -1658,13 +1660,19 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr,
> /* step 4: remove empty page table */
> if (!pml) {
> pml = pmd_lock(mm, pmd);
> - if (ptl != pml)
> + if (ptl != pml) {
> spin_lock_nested(ptl, SINGLE_DEPTH_NESTING);
> + if (unlikely(!pmd_same(pgt_pmd, pmdp_get_lockless(pmd)))) {
> + spin_unlock(pml);
> + goto abort;
Drop the reference of folio and the mm counter twice at the label of abort and the step 3.
> + }
> + }
> }
> pgt_pmd = pmdp_collapse_flush(vma, haddr, pmd);
> pmdp_get_lockless_sync();
> if (ptl != pml)
> spin_unlock(ptl);
> + pte_unmap(start_pte);
> spin_unlock(pml);
Why not?
pte_unmap_unlock(start_pte, ptl);
if (pml != ptl)
spin_unlock(pml);
>
> mmu_notifier_invalidate_range_end(&range);
> --
> 2.20.1
next prev parent reply other threads:[~2024-09-24 7:15 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-24 6:09 [PATCH v4 00/13] introduce pte_offset_map_{ro|rw}_nolock() Qi Zheng
2024-09-24 6:09 ` [PATCH v4 01/13] mm: pgtable: " Qi Zheng
2024-09-24 6:24 ` Muchun Song
2024-09-24 12:58 ` David Hildenbrand
2024-09-24 6:09 ` [PATCH v4 02/13] powerpc: assert_pte_locked() use pte_offset_map_ro_nolock() Qi Zheng
2024-09-24 6:09 ` [PATCH v4 03/13] mm: filemap: filemap_fault_recheck_pte_none() " Qi Zheng
2024-09-24 6:09 ` [PATCH v4 04/13] mm: khugepaged: __collapse_huge_page_swapin() " Qi Zheng
2024-09-24 6:09 ` [PATCH v4 05/13] arm: adjust_pte() use pte_offset_map_rw_nolock() Qi Zheng
2024-09-24 6:09 ` [PATCH v4 06/13] mm: handle_pte_fault() " Qi Zheng
2024-09-24 6:09 ` [PATCH v4 07/13] mm: khugepaged: collapse_pte_mapped_thp() " Qi Zheng
2024-09-24 7:14 ` Muchun Song [this message]
2024-09-24 7:29 ` Qi Zheng
2024-09-24 8:52 ` Muchun Song
2024-09-24 8:57 ` Qi Zheng
2024-09-24 9:03 ` Muchun Song
2024-09-24 6:10 ` [PATCH v4 08/13] mm: copy_pte_range() " Qi Zheng
2024-09-24 7:35 ` Muchun Song
2024-09-24 6:10 ` [PATCH v4 09/13] mm: mremap: move_ptes() " Qi Zheng
2024-09-24 7:39 ` Muchun Song
2024-09-24 6:10 ` [PATCH v4 10/13] mm: page_vma_mapped_walk: map_pte() " Qi Zheng
2024-09-24 8:25 ` Muchun Song
2024-09-24 8:33 ` Qi Zheng
2024-09-24 8:39 ` Muchun Song
2024-09-24 8:45 ` Qi Zheng
2024-09-24 6:10 ` [PATCH v4 11/13] mm: userfaultfd: move_pages_pte() " Qi Zheng
2024-09-24 6:10 ` [PATCH v4 12/13] mm: multi-gen LRU: walk_pte_range() " Qi Zheng
2024-09-24 14:39 ` David Hildenbrand
2024-09-24 6:10 ` [PATCH v4 13/13] mm: pgtable: remove pte_offset_map_nolock() Qi Zheng
2024-09-24 13:02 ` David Hildenbrand
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=79699B24-0D99-4051-91F3-5695D32D62AC@linux.dev \
--to=muchun.song@linux.dev \
--cc=akpm@linux-foundation.org \
--cc=christophe.leroy2@cs-soprasteria.com \
--cc=david@redhat.com \
--cc=hughd@google.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=peterx@redhat.com \
--cc=rppt@kernel.org \
--cc=ryan.roberts@arm.com \
--cc=vbabka@kernel.org \
--cc=vishal.moola@gmail.com \
--cc=willy@infradead.org \
--cc=zhengqi.arch@bytedance.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox