linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Baolu Lu <baolu.lu@linux.intel.com>
To: Dave Hansen <dave.hansen@intel.com>,
	David Hildenbrand <david@redhat.com>,
	Joerg Roedel <joro@8bytes.org>, Will Deacon <will@kernel.org>,
	Robin Murphy <robin.murphy@arm.com>,
	Kevin Tian <kevin.tian@intel.com>,
	Jason Gunthorpe <jgg@nvidia.com>, Jann Horn <jannh@google.com>,
	Vasant Hegde <vasant.hegde@amd.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
	Alistair Popple <apopple@nvidia.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Uladzislau Rezki <urezki@gmail.com>,
	Jean-Philippe Brucker <jean-philippe@linaro.org>,
	Andy Lutomirski <luto@kernel.org>, Yi Lai <yi1.lai@intel.com>
Cc: iommu@lists.linux.dev, security@kernel.org, x86@kernel.org,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Dave Hansen <dave.hansen@linux.intel.com>
Subject: Re: [PATCH v5 3/8] x86/mm: Use 'ptdesc' when freeing PMD pages
Date: Sat, 11 Oct 2025 14:26:53 +0800	[thread overview]
Message-ID: <ce126bc8-7672-4071-96b5-9652e736175b@linux.intel.com> (raw)
In-Reply-To: <e4c0c578-50e9-4e65-8e7f-7a6d995b22e0@intel.com>

On 10/10/25 03:31, Dave Hansen wrote:
> On 10/9/25 12:25, David Hildenbrand wrote:
>>>
>>> @@ -750,8 +750,8 @@ int pud_free_pmd_page(pud_t *pud, unsigned long addr)
>>>          for (i = 0; i < PTRS_PER_PMD; i++) {
>>>            if (!pmd_none(pmd_sv[i])) {
>>> -            pte = (pte_t *)pmd_page_vaddr(pmd_sv[i]);
>>> -            pte_free_kernel(&init_mm, pte);
>>> +            pt = page_ptdesc(pmd_page(pmd_sv[i]));
>>> +            pagetable_dtor_free(pt);
>>
>> There is pmd_ptdesc() which does
>>
>>      page_ptdesc(pmd_pgtable_page(pmd));
>>
>> It's buried in a
>>
>>      #if defined(CONFIG_SPLIT_PMD_PTLOCKS)
>>
>> Can't we just make that always available so we can use it here?
> 
> Yes, that looks like a good idea. I never noticed pmd_ptdesc() when I
> was writing this for sure.

I updated the patch like this,

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 6a0bb7fc3148..a0850dc6878e 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -3203,8 +3203,6 @@ pte_t *pte_offset_map_rw_nolock(struct mm_struct 
*mm, pmd_t *pmd,
         ((unlikely(pmd_none(*(pmd))) && __pte_alloc_kernel(pmd))? \
                 NULL: pte_offset_kernel(pmd, address))

-#if defined(CONFIG_SPLIT_PMD_PTLOCKS)
-
  static inline struct page *pmd_pgtable_page(pmd_t *pmd)
  {
         unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1);
@@ -3216,6 +3214,8 @@ static inline struct ptdesc *pmd_ptdesc(pmd_t *pmd)
         return page_ptdesc(pmd_pgtable_page(pmd));
  }

+#if defined(CONFIG_SPLIT_PMD_PTLOCKS)
+
  static inline spinlock_t *pmd_lockptr(struct mm_struct *mm, pmd_t *pmd)
  {
         return ptlock_ptr(pmd_ptdesc(pmd));
diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c
index ddf248c3ee7d..c830ccbc2fd8 100644
--- a/arch/x86/mm/pgtable.c
+++ b/arch/x86/mm/pgtable.c
@@ -729,7 +729,7 @@ int pmd_clear_huge(pmd_t *pmd)
  int pud_free_pmd_page(pud_t *pud, unsigned long addr)
  {
         pmd_t *pmd, *pmd_sv;
-       pte_t *pte;
+       struct ptdesc *pt;
         int i;

         pmd = pud_pgtable(*pud);
@@ -750,8 +750,8 @@ int pud_free_pmd_page(pud_t *pud, unsigned long addr)

         for (i = 0; i < PTRS_PER_PMD; i++) {
                 if (!pmd_none(pmd_sv[i])) {
-                       pte = (pte_t *)pmd_page_vaddr(pmd_sv[i]);
-                       pte_free_kernel(&init_mm, pte);
+                       pt = pmd_ptdesc(&pmd_sv[i]);
+                       pagetable_dtor_free(pt);
                 }
         }

@@ -772,15 +772,15 @@ int pud_free_pmd_page(pud_t *pud, unsigned long addr)
   */
  int pmd_free_pte_page(pmd_t *pmd, unsigned long addr)
  {
-       pte_t *pte;
+       struct ptdesc *pt;

-       pte = (pte_t *)pmd_page_vaddr(*pmd);
+       pt = pmd_ptdesc(pmd);
         pmd_clear(pmd);

         /* INVLPG to clear all paging-structure caches */
         flush_tlb_kernel_range(addr, addr + PAGE_SIZE-1);

-       pte_free_kernel(&init_mm, pte);
+       pagetable_dtor_free(pt);

         return 1;
  }

Thanks,
baolu


  reply	other threads:[~2025-10-11  6:30 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-09-19  5:39 [PATCH v5 0/8] Fix stale IOTLB entries for kernel address space Lu Baolu
2025-09-19  5:39 ` [PATCH v5 1/8] mm: Add a ptdesc flag to mark kernel page tables Lu Baolu
2025-10-08 19:56   ` Matthew Wilcox
2025-10-11  6:24     ` Baolu Lu
2025-09-19  5:40 ` [PATCH v5 2/8] mm: Actually mark kernel page table pages Lu Baolu
2025-10-09 19:19   ` David Hildenbrand
2025-10-13  7:17   ` Mike Rapoport
2025-09-19  5:40 ` [PATCH v5 3/8] x86/mm: Use 'ptdesc' when freeing PMD pages Lu Baolu
2025-10-09 19:25   ` David Hildenbrand
2025-10-09 19:31     ` Dave Hansen
2025-10-11  6:26       ` Baolu Lu [this message]
2025-09-19  5:40 ` [PATCH v5 4/8] mm: Introduce pure page table freeing function Lu Baolu
2025-10-09 19:26   ` David Hildenbrand
2025-10-13  7:24   ` Mike Rapoport
2025-09-19  5:40 ` [PATCH v5 5/8] x86/mm: Use pagetable_free() Lu Baolu
2025-09-24 12:40   ` Jason Gunthorpe
2025-10-09 19:26   ` David Hildenbrand
2025-10-13  7:28   ` Mike Rapoport
2025-09-19  5:40 ` [PATCH v5 6/8] mm: Introduce deferred freeing for kernel page tables Lu Baolu
2025-10-09 19:28   ` David Hildenbrand
2025-10-09 19:32     ` Dave Hansen
2025-10-10 15:47   ` David Hildenbrand
2025-10-11  6:30     ` Baolu Lu
2025-09-19  5:40 ` [PATCH v5 7/8] mm: Hook up Kconfig options for async page table freeing Lu Baolu
2025-09-19  5:40 ` [PATCH v5 8/8] iommu/sva: Invalidate stale IOTLB entries for kernel address space Lu Baolu
2025-09-25 20:24 ` [PATCH v5 0/8] Fix " Dave Hansen
2025-10-08 19:42   ` Dave Hansen
2025-10-09 19:16     ` David Hildenbrand
2025-10-14 13:21     ` Baolu Lu

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=ce126bc8-7672-4071-96b5-9652e736175b@linux.intel.com \
    --to=baolu.lu@linux.intel.com \
    --cc=apopple@nvidia.com \
    --cc=bp@alien8.de \
    --cc=dave.hansen@intel.com \
    --cc=dave.hansen@linux.intel.com \
    --cc=david@redhat.com \
    --cc=iommu@lists.linux.dev \
    --cc=jannh@google.com \
    --cc=jean-philippe@linaro.org \
    --cc=jgg@nvidia.com \
    --cc=joro@8bytes.org \
    --cc=kevin.tian@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=luto@kernel.org \
    --cc=mingo@redhat.com \
    --cc=peterz@infradead.org \
    --cc=robin.murphy@arm.com \
    --cc=security@kernel.org \
    --cc=tglx@linutronix.de \
    --cc=urezki@gmail.com \
    --cc=vasant.hegde@amd.com \
    --cc=will@kernel.org \
    --cc=x86@kernel.org \
    --cc=yi1.lai@intel.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