The patch titled
Subject: mm: pass VMA instead of MM to follow_pte()
has been added to the -mm mm-unstable branch. Its filename is
mm-pass-vma-instead-of-mm-to-follow_pte.patch
This patch will shortly appear at
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-pass-vma-instead-of-mm-to-follow_pte.patch
This patch will later appear in the mm-unstable branch at
git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next via the mm-everything
branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
and is updated there every 2-3 working days
------------------------------------------------------
From: David Hildenbrand <david@xxxxxxxxxx>
Subject: mm: pass VMA instead of MM to follow_pte()
Date: Wed, 10 Apr 2024 17:55:26 +0200
.. and centralize the VM_IO/VM_PFNMAP sanity check in there. We'll
now also perform these sanity checks for direct follow_pte()
invocations.
For generic_access_phys(), we might now check multiple times: nothing to
worry about, really.
Link: https://lkml.kernel.org/r/20240410155527.474777-3-david@xxxxxxxxxx
Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
Acked-by: Sean Christopherson <seanjc@xxxxxxxxxx> [KVM]
Cc: Alex Williamson <alex.williamson@xxxxxxxxxx>
Cc: Christoph Hellwig <hch@xxxxxx>
Cc: Fei Li <fei1.li@xxxxxxxxx>
Cc: Gerald Schaefer <gerald.schaefer@xxxxxxxxxxxxx>
Cc: Heiko Carstens <hca@xxxxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Cc: Yonghua Huang <yonghua.huang@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
arch/s390/pci/pci_mmio.c | 4 ++--
arch/x86/mm/pat/memtype.c | 5 +----
drivers/vfio/vfio_iommu_type1.c | 4 ++--
drivers/virt/acrn/mm.c | 3 +--
include/linux/mm.h | 2 +-
mm/memory.c | 15 ++++++++-------
virt/kvm/kvm_main.c | 4 ++--
7 files changed, 17 insertions(+), 20 deletions(-)
--- a/arch/s390/pci/pci_mmio.c~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/arch/s390/pci/pci_mmio.c
@@ -169,7 +169,7 @@ SYSCALL_DEFINE3(s390_pci_mmio_write, uns
if (!(vma->vm_flags & VM_WRITE))
goto out_unlock_mmap;
- ret = follow_pte(vma->vm_mm, mmio_addr, &ptep, &ptl);
+ ret = follow_pte(vma, mmio_addr, &ptep, &ptl);
if (ret)
goto out_unlock_mmap;
@@ -308,7 +308,7 @@ SYSCALL_DEFINE3(s390_pci_mmio_read, unsi
if (!(vma->vm_flags & VM_WRITE))
goto out_unlock_mmap;
- ret = follow_pte(vma->vm_mm, mmio_addr, &ptep, &ptl);
+ ret = follow_pte(vma, mmio_addr, &ptep, &ptl);
if (ret)
goto out_unlock_mmap;
--- a/arch/x86/mm/pat/memtype.c~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/arch/x86/mm/pat/memtype.c
@@ -954,10 +954,7 @@ static int follow_phys(struct vm_area_st
pte_t *ptep, pte;
spinlock_t *ptl;
- if (!(vma->vm_flags & (VM_IO | VM_PFNMAP)))
- return -EINVAL;
-
- if (follow_pte(vma->vm_mm, vma->vm_start, &ptep, &ptl))
+ if (follow_pte(vma, vma->vm_start, &ptep, &ptl))
return -EINVAL;
pte = ptep_get(ptep);
--- a/drivers/vfio/vfio_iommu_type1.c~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/drivers/vfio/vfio_iommu_type1.c
@@ -518,7 +518,7 @@ static int follow_fault_pfn(struct vm_ar
spinlock_t *ptl;
int ret;
- ret = follow_pte(vma->vm_mm, vaddr, &ptep, &ptl);
+ ret = follow_pte(vma, vaddr, &ptep, &ptl);
if (ret) {
bool unlocked = false;
@@ -532,7 +532,7 @@ static int follow_fault_pfn(struct vm_ar
if (ret)
return ret;
- ret = follow_pte(vma->vm_mm, vaddr, &ptep, &ptl);
+ ret = follow_pte(vma, vaddr, &ptep, &ptl);
if (ret)
return ret;
}
--- a/drivers/virt/acrn/mm.c~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/drivers/virt/acrn/mm.c
@@ -187,8 +187,7 @@ int acrn_vm_ram_map(struct acrn_vm *vm,
}
for (i = 0; i < nr_pages; i++) {
- ret = follow_pte(vma->vm_mm,
- memmap->vma_base + i * PAGE_SIZE,
+ ret = follow_pte(vma, memmap->vma_base + i * PAGE_SIZE,
&ptep, &ptl);
if (ret)
break;
--- a/include/linux/mm.h~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/include/linux/mm.h
@@ -2420,7 +2420,7 @@ void free_pgd_range(struct mmu_gather *t
unsigned long end, unsigned long floor, unsigned long ceiling);
int
copy_page_range(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma);
-int follow_pte(struct mm_struct *mm, unsigned long address,
+int follow_pte(struct vm_area_struct *vma, unsigned long address,
pte_t **ptepp, spinlock_t **ptlp);
int generic_access_phys(struct vm_area_struct *vma, unsigned long addr,
void *buf, int len, int write);
--- a/mm/memory.c~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/mm/memory.c
@@ -5926,7 +5926,7 @@ int __pmd_alloc(struct mm_struct *mm, pu
/**
* follow_pte - look up PTE at a user virtual address
- * @mm: the mm_struct of the target address space
+ * @vma: the memory mapping
* @address: user virtual address
* @ptepp: location to store found PTE
* @ptlp: location to store the lock for the PTE
@@ -5945,15 +5945,19 @@ int __pmd_alloc(struct mm_struct *mm, pu
*
* Return: zero on success, -ve otherwise.
*/
-int follow_pte(struct mm_struct *mm, unsigned long address,
+int follow_pte(struct vm_area_struct *vma, unsigned long address,
pte_t **ptepp, spinlock_t **ptlp)
{
+ struct mm_struct *mm = vma->vm_mm;
pgd_t *pgd;
p4d_t *p4d;
pud_t *pud;
pmd_t *pmd;
pte_t *ptep;
+ if (!(vma->vm_flags & (VM_IO | VM_PFNMAP)))
+ goto out;
+
pgd = pgd_offset(mm, address);
if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
goto out;
@@ -6007,11 +6011,8 @@ int generic_access_phys(struct vm_area_s
int offset = offset_in_page(addr);
int ret = -EINVAL;
- if (!(vma->vm_flags & (VM_IO | VM_PFNMAP)))
- return -EINVAL;
-
retry:
- if (follow_pte(vma->vm_mm, addr, &ptep, &ptl))
+ if (follow_pte(vma, addr, &ptep, &ptl))
return -EINVAL;
pte = ptep_get(ptep);
pte_unmap_unlock(ptep, ptl);
@@ -6026,7 +6027,7 @@ retry:
if (!maddr)
return -ENOMEM;
- if (follow_pte(vma->vm_mm, addr, &ptep, &ptl))
+ if (follow_pte(vma, addr, &ptep, &ptl))
goto out_unmap;
if (!pte_same(pte, ptep_get(ptep))) {
--- a/virt/kvm/kvm_main.c~mm-pass-vma-instead-of-mm-to-follow_pte
+++ a/virt/kvm/kvm_main.c
@@ -2902,7 +2902,7 @@ static int hva_to_pfn_remapped(struct vm
spinlock_t *ptl;
int r;
- r = follow_pte(vma->vm_mm, addr, &ptep, &ptl);
+ r = follow_pte(vma, addr, &ptep, &ptl);
if (r) {
/*
* get_user_pages fails for VM_IO and VM_PFNMAP vmas and does
@@ -2917,7 +2917,7 @@ static int hva_to_pfn_remapped(struct vm
if (r)
return r;
- r = follow_pte(vma->vm_mm, addr, &ptep, &ptl);
+ r = follow_pte(vma, addr, &ptep, &ptl);
if (r)
return r;
}
_
Patches currently in -mm which might be from david@xxxxxxxxxx are
mm-madvise-make-madv_populate_readwrite-handle-vm_fault_retry-properly.patch
mm-madvise-dont-perform-madvise-vma-walk-for-madv_populate_readwrite.patch
mm-userfaultfd-dont-place-zeropages-when-zeropages-are-disallowed.patch
s390-mm-re-enable-the-shared-zeropage-for-pv-and-skeys-kvm-guests.patch
mm-convert-folio_estimated_sharers-to-folio_likely_mapped_shared.patch
mm-convert-folio_estimated_sharers-to-folio_likely_mapped_shared-fix.patch
selftests-memfd_secret-add-vmsplice-test.patch
mm-merge-folio_is_secretmem-and-folio_fast_pin_allowed-into-gup_fast_folio_allowed.patch
mm-optimize-config_per_vma_lock-member-placement-in-vm_area_struct.patch
mm-remove-prot-parameter-from-move_pte.patch
mm-gup-consistently-name-gup-fast-functions.patch
mm-treewide-rename-config_have_fast_gup-to-config_have_gup_fast.patch
mm-use-gup-fast-instead-fast-gup-in-remaining-comments.patch
drivers-virt-acrn-fix-pfnmap-pte-checks-in-acrn_vm_ram_map.patch
mm-pass-vma-instead-of-mm-to-follow_pte.patch
mm-follow_pte-improvements.patch
