Re: [PATCH v2 5/6] mm: introduce mod_vm_flags_nolock and use it in untrack_pfn
On Wed, Jan 25, 2023 at 12:38:50AM -0800, Suren Baghdasaryan wrote:
> In cases when VMA flags are modified after VMA was isolated and mmap_lock
> was downgraded, flags modifications would result in an assertion because
> mmap write lock is not held.
> Introduce mod_vm_flags_nolock to be used in such situation.
vm_flags_mod_nolock?
> Pass a hint to untrack_pfn to conditionally use mod_vm_flags_nolock for
> flags modification and to avoid assertion.
>
> Signed-off-by: Suren Baghdasaryan
> ---
> arch/x86/mm/pat/memtype.c | 10 +++---
> include/linux/mm.h| 12 +---
> include/linux/pgtable.h | 5 +++--
> mm/memory.c | 13 +++--
> mm/memremap.c | 4 ++--
> mm/mmap.c | 16 ++--
> 6 files changed, 38 insertions(+), 22 deletions(-)
>
> diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
> index ae9645c900fa..d8adc0b42cf2 100644
> --- a/arch/x86/mm/pat/memtype.c
> +++ b/arch/x86/mm/pat/memtype.c
> @@ -1046,7 +1046,7 @@ void track_pfn_insert(struct vm_area_struct *vma,
> pgprot_t *prot, pfn_t pfn)
> * can be for the entire vma (in which case pfn, size are zero).
> */
> void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
> - unsigned long size)
> + unsigned long size, bool mm_wr_locked)
> {
> resource_size_t paddr;
> unsigned long prot;
> @@ -1065,8 +1065,12 @@ void untrack_pfn(struct vm_area_struct *vma, unsigned
> long pfn,
> size = vma->vm_end - vma->vm_start;
> }
> free_pfn_range(paddr, size);
> - if (vma)
> - clear_vm_flags(vma, VM_PAT);
> + if (vma) {
> + if (mm_wr_locked)
> + clear_vm_flags(vma, VM_PAT);
> + else
> + mod_vm_flags_nolock(vma, 0, VM_PAT);
> + }
> }
>
> /*
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 55335edd1373..48d49930c411 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -656,12 +656,18 @@ static inline void clear_vm_flags(struct vm_area_struct
> *vma,
> vma->vm_flags &= ~flags;
> }
>
> +static inline void mod_vm_flags_nolock(struct vm_area_struct *vma,
> +unsigned long set, unsigned long clear)
> +{
> + vma->vm_flags |= set;
> + vma->vm_flags &= ~clear;
> +}
> +
> static inline void mod_vm_flags(struct vm_area_struct *vma,
> unsigned long set, unsigned long clear)
> {
> mmap_assert_write_locked(vma->vm_mm);
> - vma->vm_flags |= set;
> - vma->vm_flags &= ~clear;
> + mod_vm_flags_nolock(vma, set, clear);
> }
>
> static inline void vma_set_anonymous(struct vm_area_struct *vma)
> @@ -2087,7 +2093,7 @@ static inline void zap_vma_pages(struct vm_area_struct
> *vma)
> }
> void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt,
> struct vm_area_struct *start_vma, unsigned long start,
> - unsigned long end);
> + unsigned long end, bool mm_wr_locked);
>
> struct mmu_notifier_range;
>
> diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
> index 5fd45454c073..c63cd44777ec 100644
> --- a/include/linux/pgtable.h
> +++ b/include/linux/pgtable.h
> @@ -1185,7 +1185,8 @@ static inline int track_pfn_copy(struct vm_area_struct
> *vma)
> * can be for the entire vma (in which case pfn, size are zero).
> */
> static inline void untrack_pfn(struct vm_area_struct *vma,
> -unsigned long pfn, unsigned long size)
> +unsigned long pfn, unsigned long size,
> +bool mm_wr_locked)
> {
> }
>
> @@ -1203,7 +1204,7 @@ extern void track_pfn_insert(struct vm_area_struct
> *vma, pgprot_t *prot,
>pfn_t pfn);
> extern int track_pfn_copy(struct vm_area_struct *vma);
> extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
> - unsigned long size);
> + unsigned long size, bool mm_wr_locked);
> extern void untrack_pfn_moved(struct vm_area_struct *vma);
> #endif
>
> diff --git a/mm/memory.c b/mm/memory.c
> index d6902065e558..5b11b50e2c4a 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -1613,7 +1613,7 @@ void unmap_page_range(struct mmu_gather *tlb,
> static void unmap_single_vma(struct mmu_gather *tlb,
> struct vm_area_struct *vma, unsigned long start_addr,
> unsigned long end_addr,
> - struct zap_details *details)
> + struct zap_details *details, bool mm_wr_locked)
> {
> unsigned long start = max(vma->vm_start, start_addr);
> unsigned long end;
> @@ -1628,7 +1628,7 @@ static void unmap_single_vma(struct mmu_gather *tlb,
> uprobe_munmap(vma, start, end);
>
> if (unlikely(vma->vm_flags & VM_PFNMAP))
> - untrack_pfn(vma, 0, 0);
> + untrack_pfn(vma, 0, 0,
Re: [PATCH v2 5/6] mm: introduce mod_vm_flags_nolock and use it in untrack_pfn
On Wed, Jan 25, 2023 at 1:42 AM Michal Hocko wrote:
>
> On Wed 25-01-23 00:38:50, Suren Baghdasaryan wrote:
> > In cases when VMA flags are modified after VMA was isolated and mmap_lock
> > was downgraded, flags modifications would result in an assertion because
> > mmap write lock is not held.
> > Introduce mod_vm_flags_nolock to be used in such situation.
> > Pass a hint to untrack_pfn to conditionally use mod_vm_flags_nolock for
> > flags modification and to avoid assertion.
>
> The changelog nor the documentation of mod_vm_flags_nolock
> really explain when it is safe to use it. This is really important for
> future potential users.
True. I'll add clarification in the comments and in the changelog. Thanks!
>
> > Signed-off-by: Suren Baghdasaryan
> > ---
> > arch/x86/mm/pat/memtype.c | 10 +++---
> > include/linux/mm.h| 12 +---
> > include/linux/pgtable.h | 5 +++--
> > mm/memory.c | 13 +++--
> > mm/memremap.c | 4 ++--
> > mm/mmap.c | 16 ++--
> > 6 files changed, 38 insertions(+), 22 deletions(-)
> >
> > diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
> > index ae9645c900fa..d8adc0b42cf2 100644
> > --- a/arch/x86/mm/pat/memtype.c
> > +++ b/arch/x86/mm/pat/memtype.c
> > @@ -1046,7 +1046,7 @@ void track_pfn_insert(struct vm_area_struct *vma,
> > pgprot_t *prot, pfn_t pfn)
> > * can be for the entire vma (in which case pfn, size are zero).
> > */
> > void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
> > - unsigned long size)
> > + unsigned long size, bool mm_wr_locked)
> > {
> > resource_size_t paddr;
> > unsigned long prot;
> > @@ -1065,8 +1065,12 @@ void untrack_pfn(struct vm_area_struct *vma,
> > unsigned long pfn,
> > size = vma->vm_end - vma->vm_start;
> > }
> > free_pfn_range(paddr, size);
> > - if (vma)
> > - clear_vm_flags(vma, VM_PAT);
> > + if (vma) {
> > + if (mm_wr_locked)
> > + clear_vm_flags(vma, VM_PAT);
> > + else
> > + mod_vm_flags_nolock(vma, 0, VM_PAT);
> > + }
> > }
> >
> > /*
> > diff --git a/include/linux/mm.h b/include/linux/mm.h
> > index 55335edd1373..48d49930c411 100644
> > --- a/include/linux/mm.h
> > +++ b/include/linux/mm.h
> > @@ -656,12 +656,18 @@ static inline void clear_vm_flags(struct
> > vm_area_struct *vma,
> > vma->vm_flags &= ~flags;
> > }
> >
> > +static inline void mod_vm_flags_nolock(struct vm_area_struct *vma,
> > +unsigned long set, unsigned long clear)
> > +{
> > + vma->vm_flags |= set;
> > + vma->vm_flags &= ~clear;
> > +}
> > +
> > static inline void mod_vm_flags(struct vm_area_struct *vma,
> > unsigned long set, unsigned long clear)
> > {
> > mmap_assert_write_locked(vma->vm_mm);
> > - vma->vm_flags |= set;
> > - vma->vm_flags &= ~clear;
> > + mod_vm_flags_nolock(vma, set, clear);
> > }
> >
> > static inline void vma_set_anonymous(struct vm_area_struct *vma)
> > @@ -2087,7 +2093,7 @@ static inline void zap_vma_pages(struct
> > vm_area_struct *vma)
> > }
> > void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt,
> > struct vm_area_struct *start_vma, unsigned long start,
> > - unsigned long end);
> > + unsigned long end, bool mm_wr_locked);
> >
> > struct mmu_notifier_range;
> >
> > diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
> > index 5fd45454c073..c63cd44777ec 100644
> > --- a/include/linux/pgtable.h
> > +++ b/include/linux/pgtable.h
> > @@ -1185,7 +1185,8 @@ static inline int track_pfn_copy(struct
> > vm_area_struct *vma)
> > * can be for the entire vma (in which case pfn, size are zero).
> > */
> > static inline void untrack_pfn(struct vm_area_struct *vma,
> > -unsigned long pfn, unsigned long size)
> > +unsigned long pfn, unsigned long size,
> > +bool mm_wr_locked)
> > {
> > }
> >
> > @@ -1203,7 +1204,7 @@ extern void track_pfn_insert(struct vm_area_struct
> > *vma, pgprot_t *prot,
> >pfn_t pfn);
> > extern int track_pfn_copy(struct vm_area_struct *vma);
> > extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
> > - unsigned long size);
> > + unsigned long size, bool mm_wr_locked);
> > extern void untrack_pfn_moved(struct vm_area_struct *vma);
> > #endif
> >
> > diff --git a/mm/memory.c b/mm/memory.c
> > index d6902065e558..5b11b50e2c4a 100644
> > --- a/mm/memory.c
> > +++ b/mm/memory.c
> > @@ -1613,7 +1613,7 @@ void unmap_page_range(struct mmu_gather *tlb,
> > static void unmap_single_vma(struct mmu_gather *tlb,
> > struct vm_area_struct *vma, unsigned long start_addr,
> > unsigned long
Re: [PATCH v2 5/6] mm: introduce mod_vm_flags_nolock and use it in untrack_pfn
On Wed 25-01-23 00:38:50, Suren Baghdasaryan wrote:
> In cases when VMA flags are modified after VMA was isolated and mmap_lock
> was downgraded, flags modifications would result in an assertion because
> mmap write lock is not held.
> Introduce mod_vm_flags_nolock to be used in such situation.
> Pass a hint to untrack_pfn to conditionally use mod_vm_flags_nolock for
> flags modification and to avoid assertion.
The changelog nor the documentation of mod_vm_flags_nolock
really explain when it is safe to use it. This is really important for
future potential users.
> Signed-off-by: Suren Baghdasaryan
> ---
> arch/x86/mm/pat/memtype.c | 10 +++---
> include/linux/mm.h| 12 +---
> include/linux/pgtable.h | 5 +++--
> mm/memory.c | 13 +++--
> mm/memremap.c | 4 ++--
> mm/mmap.c | 16 ++--
> 6 files changed, 38 insertions(+), 22 deletions(-)
>
> diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
> index ae9645c900fa..d8adc0b42cf2 100644
> --- a/arch/x86/mm/pat/memtype.c
> +++ b/arch/x86/mm/pat/memtype.c
> @@ -1046,7 +1046,7 @@ void track_pfn_insert(struct vm_area_struct *vma,
> pgprot_t *prot, pfn_t pfn)
> * can be for the entire vma (in which case pfn, size are zero).
> */
> void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
> - unsigned long size)
> + unsigned long size, bool mm_wr_locked)
> {
> resource_size_t paddr;
> unsigned long prot;
> @@ -1065,8 +1065,12 @@ void untrack_pfn(struct vm_area_struct *vma, unsigned
> long pfn,
> size = vma->vm_end - vma->vm_start;
> }
> free_pfn_range(paddr, size);
> - if (vma)
> - clear_vm_flags(vma, VM_PAT);
> + if (vma) {
> + if (mm_wr_locked)
> + clear_vm_flags(vma, VM_PAT);
> + else
> + mod_vm_flags_nolock(vma, 0, VM_PAT);
> + }
> }
>
> /*
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 55335edd1373..48d49930c411 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -656,12 +656,18 @@ static inline void clear_vm_flags(struct vm_area_struct
> *vma,
> vma->vm_flags &= ~flags;
> }
>
> +static inline void mod_vm_flags_nolock(struct vm_area_struct *vma,
> +unsigned long set, unsigned long clear)
> +{
> + vma->vm_flags |= set;
> + vma->vm_flags &= ~clear;
> +}
> +
> static inline void mod_vm_flags(struct vm_area_struct *vma,
> unsigned long set, unsigned long clear)
> {
> mmap_assert_write_locked(vma->vm_mm);
> - vma->vm_flags |= set;
> - vma->vm_flags &= ~clear;
> + mod_vm_flags_nolock(vma, set, clear);
> }
>
> static inline void vma_set_anonymous(struct vm_area_struct *vma)
> @@ -2087,7 +2093,7 @@ static inline void zap_vma_pages(struct vm_area_struct
> *vma)
> }
> void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt,
> struct vm_area_struct *start_vma, unsigned long start,
> - unsigned long end);
> + unsigned long end, bool mm_wr_locked);
>
> struct mmu_notifier_range;
>
> diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
> index 5fd45454c073..c63cd44777ec 100644
> --- a/include/linux/pgtable.h
> +++ b/include/linux/pgtable.h
> @@ -1185,7 +1185,8 @@ static inline int track_pfn_copy(struct vm_area_struct
> *vma)
> * can be for the entire vma (in which case pfn, size are zero).
> */
> static inline void untrack_pfn(struct vm_area_struct *vma,
> -unsigned long pfn, unsigned long size)
> +unsigned long pfn, unsigned long size,
> +bool mm_wr_locked)
> {
> }
>
> @@ -1203,7 +1204,7 @@ extern void track_pfn_insert(struct vm_area_struct
> *vma, pgprot_t *prot,
>pfn_t pfn);
> extern int track_pfn_copy(struct vm_area_struct *vma);
> extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
> - unsigned long size);
> + unsigned long size, bool mm_wr_locked);
> extern void untrack_pfn_moved(struct vm_area_struct *vma);
> #endif
>
> diff --git a/mm/memory.c b/mm/memory.c
> index d6902065e558..5b11b50e2c4a 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -1613,7 +1613,7 @@ void unmap_page_range(struct mmu_gather *tlb,
> static void unmap_single_vma(struct mmu_gather *tlb,
> struct vm_area_struct *vma, unsigned long start_addr,
> unsigned long end_addr,
> - struct zap_details *details)
> + struct zap_details *details, bool mm_wr_locked)
> {
> unsigned long start = max(vma->vm_start, start_addr);
> unsigned long end;
> @@ -1628,7 +1628,7 @@ static void unmap_single_vma(struct mmu_gather *tlb,
> uprobe_munmap(vma, start, end);
>
> if
