From: Suren Baghdasaryan <surenb@google.com>
To: Laurent Dufour <ldufour@linux.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Michel Lespinasse <michel@lespinasse.org>,
Jerome Glisse <jglisse@google.com>,
Michal Hocko <mhocko@suse.com>, Vlastimil Babka <vbabka@suse.cz>,
Johannes Weiner <hannes@cmpxchg.org>,
Mel Gorman <mgorman@suse.de>,
Davidlohr Bueso <dave@stgolabs.net>,
Matthew Wilcox <willy@infradead.org>,
"Liam R. Howlett" <liam.howlett@oracle.com>,
Peter Zijlstra <peterz@infradead.org>,
Laurent Dufour <laurent.dufour@fr.ibm.com>,
"Paul E . McKenney" <paulmck@kernel.org>,
Andy Lutomirski <luto@kernel.org>,
Song Liu <songliubraving@fb.com>, Peter Xu <peterx@redhat.com>,
David Hildenbrand <david@redhat.com>,
dhowells@redhat.com, Hugh Dickins <hughd@google.com>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Kent Overstreet <kent.overstreet@linux.dev>,
David Rientjes <rientjes@google.com>,
Axel Rasmussen <axelrasmussen@google.com>,
Joel Fernandes <joelaf@google.com>,
Minchan Kim <minchan@google.com>,
kernel-team <kernel-team@android.com>,
linux-mm <linux-mm@kvack.org>,
linux-arm-kernel@lists.infradead.org,
linuxppc-dev@lists.ozlabs.org, x86@kernel.org,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [RFC PATCH RESEND 06/28] mm: mark VMA as locked whenever vma->vm_flags are modified
Date: Tue, 6 Sep 2022 12:00:47 -0700 [thread overview]
Message-ID: <CAJuCfpFi07ZpAuSLwADGO0mfu2Hon9jWb0hsCytorchQ=cgK3Q@mail.gmail.com> (raw)
In-Reply-To: <1624be86-4c17-46e5-fafc-eb8afb7b9b4a@linux.ibm.com>
On Tue, Sep 6, 2022 at 7:27 AM Laurent Dufour <ldufour@linux.ibm.com> wrote:
>
> Le 01/09/2022 à 19:34, Suren Baghdasaryan a écrit :
> > VMA flag modifications should be done under VMA lock to prevent concurrent
> > page fault handling in that area.
> >
> > Signed-off-by: Suren Baghdasaryan <surenb@google.com>
> > ---
> > fs/proc/task_mmu.c | 1 +
> > fs/userfaultfd.c | 6 ++++++
> > mm/madvise.c | 1 +
> > mm/mlock.c | 2 ++
> > mm/mmap.c | 1 +
> > mm/mprotect.c | 1 +
> > 6 files changed, 12 insertions(+)
>
> There are few changes also done in the driver's space, for instance:
>
> *** arch/x86/kernel/cpu/sgx/driver.c:
> sgx_mmap[98] vma->vm_flags |= VM_PFNMAP | VM_DONTEXPAND |
> VM_DONTDUMP | VM_IO;
> *** arch/x86/kernel/cpu/sgx/virt.c:
> sgx_vepc_mmap[108] vma->vm_flags |= VM_PFNMAP | VM_IO |
> VM_DONTDUMP | VM_DONTCOPY;
> *** drivers/dax/device.c:
> dax_mmap[311] vma->vm_flags |= VM_HUGEPAGE;
>
> I guess these changes to vm_flags should be protected as well, or to be
> checked one by one.
Thanks for noting these! I'll add necessary locking here and will look
for other places I might have missed.
>
> >
> > diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
> > index 4e0023643f8b..ceffa5c2c650 100644
> > --- a/fs/proc/task_mmu.c
> > +++ b/fs/proc/task_mmu.c
> > @@ -1285,6 +1285,7 @@ static ssize_t clear_refs_write(struct file *file, const char __user *buf,
> > for (vma = mm->mmap; vma; vma = vma->vm_next) {
> > if (!(vma->vm_flags & VM_SOFTDIRTY))
> > continue;
> > + vma_mark_locked(vma);
> > vma->vm_flags &= ~VM_SOFTDIRTY;
> > vma_set_page_prot(vma);
> > }
> > diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
> > index 175de70e3adf..fe557b3d1c07 100644
> > --- a/fs/userfaultfd.c
> > +++ b/fs/userfaultfd.c
> > @@ -620,6 +620,7 @@ static void userfaultfd_event_wait_completion(struct userfaultfd_ctx *ctx,
> > mmap_write_lock(mm);
> > for (vma = mm->mmap; vma; vma = vma->vm_next)
> > if (vma->vm_userfaultfd_ctx.ctx == release_new_ctx) {
> > + vma_mark_locked(vma);
> > vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
> > vma->vm_flags &= ~__VM_UFFD_FLAGS;
> > }
> > @@ -653,6 +654,7 @@ int dup_userfaultfd(struct vm_area_struct *vma, struct list_head *fcs)
> >
> > octx = vma->vm_userfaultfd_ctx.ctx;
> > if (!octx || !(octx->features & UFFD_FEATURE_EVENT_FORK)) {
> > + vma_mark_locked(vma);
> > vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
> > vma->vm_flags &= ~__VM_UFFD_FLAGS;
> > return 0;
> > @@ -734,6 +736,7 @@ void mremap_userfaultfd_prep(struct vm_area_struct *vma,
> > atomic_inc(&ctx->mmap_changing);
> > } else {
> > /* Drop uffd context if remap feature not enabled */
> > + vma_mark_locked(vma);
> > vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
> > vma->vm_flags &= ~__VM_UFFD_FLAGS;
> > }
> > @@ -891,6 +894,7 @@ static int userfaultfd_release(struct inode *inode, struct file *file)
> > vma = prev;
> > else
> > prev = vma;
> > + vma_mark_locked(vma);
> > vma->vm_flags = new_flags;
> > vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
> > }
> > @@ -1449,6 +1453,7 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx,
> > * the next vma was merged into the current one and
> > * the current one has not been updated yet.
> > */
> > + vma_mark_locked(vma);
> > vma->vm_flags = new_flags;
> > vma->vm_userfaultfd_ctx.ctx = ctx;
> >
> > @@ -1630,6 +1635,7 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx,
> > * the next vma was merged into the current one and
> > * the current one has not been updated yet.
> > */
> > + vma_mark_locked(vma);
> > vma->vm_flags = new_flags;
> > vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
> >
> > diff --git a/mm/madvise.c b/mm/madvise.c
> > index 5f0f0948a50e..a173f0025abd 100644
> > --- a/mm/madvise.c
> > +++ b/mm/madvise.c
> > @@ -181,6 +181,7 @@ static int madvise_update_vma(struct vm_area_struct *vma,
> > /*
> > * vm_flags is protected by the mmap_lock held in write mode.
> > */
> > + vma_mark_locked(vma);
> > vma->vm_flags = new_flags;
> > if (!vma->vm_file) {
> > error = replace_anon_vma_name(vma, anon_name);
> > diff --git a/mm/mlock.c b/mm/mlock.c
> > index b14e929084cc..f62e1a4d05f2 100644
> > --- a/mm/mlock.c
> > +++ b/mm/mlock.c
> > @@ -380,6 +380,7 @@ static void mlock_vma_pages_range(struct vm_area_struct *vma,
> > */
> > if (newflags & VM_LOCKED)
> > newflags |= VM_IO;
> > + vma_mark_locked(vma);
> > WRITE_ONCE(vma->vm_flags, newflags);
> >
> > lru_add_drain();
> > @@ -456,6 +457,7 @@ static int mlock_fixup(struct vm_area_struct *vma, struct vm_area_struct **prev,
> >
> > if ((newflags & VM_LOCKED) && (oldflags & VM_LOCKED)) {
> > /* No work to do, and mlocking twice would be wrong */
> > + vma_mark_locked(vma);
> > vma->vm_flags = newflags;
> > } else {
> > mlock_vma_pages_range(vma, start, end, newflags);
> > diff --git a/mm/mmap.c b/mm/mmap.c
> > index 693e6776be39..f89c9b058105 100644
> > --- a/mm/mmap.c
> > +++ b/mm/mmap.c
> > @@ -1818,6 +1818,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
> > out:
> > perf_event_mmap(vma);
> >
> > + vma_mark_locked(vma);
> > vm_stat_account(mm, vm_flags, len >> PAGE_SHIFT);
> > if (vm_flags & VM_LOCKED) {
> > if ((vm_flags & VM_SPECIAL) || vma_is_dax(vma) ||
>
> I guess, this doesn't really impact, but the call to vma_mark_locked(vma)
> may be done only in the case the vm_flags field is touched.
> Something like this:
>
> vm_stat_account(mm, vm_flags, len >> PAGE_SHIFT);
> if (vm_flags & VM_LOCKED) {
> if ((vm_flags & VM_SPECIAL) || vma_is_dax(vma) ||
> is_vm_hugetlb_page(vma) ||
> - vma == get_gate_vma(current->mm))
> + vma == get_gate_vma(current->mm)) {
> + vma_mark_locked(vma);
> vma->vm_flags &= VM_LOCKED_CLEAR_MASK;
> - else
> + } else
> mm->locked_vm += (len >> PAGE_SHIFT);
> }
>
>
> > diff --git a/mm/mprotect.c b/mm/mprotect.c
> > index bc6bddd156ca..df47fc21b0e4 100644
> > --- a/mm/mprotect.c
> > +++ b/mm/mprotect.c
> > @@ -621,6 +621,7 @@ mprotect_fixup(struct mmu_gather *tlb, struct vm_area_struct *vma,
> > * vm_flags and vm_page_prot are protected by the mmap_lock
> > * held in write mode.
> > */
> > + vma_mark_locked(vma);
> > vma->vm_flags = newflags;
> > /*
> > * We want to check manually if we can change individual PTEs writable
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2022-09-06 19:02 UTC|newest]
Thread overview: 91+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-01 17:34 [RFC PATCH RESEND 00/28] per-VMA locks proposal Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 01/28] mm: introduce CONFIG_PER_VMA_LOCK Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 02/28] mm: rcu safe VMA freeing Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 03/28] mm: introduce __find_vma to be used without mmap_lock protection Suren Baghdasaryan
2022-09-01 20:22 ` Kent Overstreet
2022-09-01 23:18 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 04/28] mm: move mmap_lock assert function definitions Suren Baghdasaryan
2022-09-01 20:24 ` Kent Overstreet
2022-09-01 20:51 ` Liam Howlett
2022-09-01 23:21 ` Suren Baghdasaryan
2022-09-02 6:23 ` Sebastian Andrzej Siewior
2022-09-02 17:46 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 05/28] mm: add per-VMA lock and helper functions to control it Suren Baghdasaryan
2022-09-06 13:46 ` Laurent Dufour
2022-09-06 17:24 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 06/28] mm: mark VMA as locked whenever vma->vm_flags are modified Suren Baghdasaryan
2022-09-06 14:26 ` Laurent Dufour
2022-09-06 19:00 ` Suren Baghdasaryan [this message]
2022-09-06 20:00 ` Liam Howlett
2022-09-06 20:13 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 07/28] kernel/fork: mark VMAs as locked before copying pages during fork Suren Baghdasaryan
2022-09-06 14:37 ` Laurent Dufour
2022-09-08 23:57 ` Suren Baghdasaryan
2022-09-09 13:27 ` Laurent Dufour
2022-09-09 16:29 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 08/28] mm/khugepaged: mark VMA as locked while collapsing a hugepage Suren Baghdasaryan
2022-09-06 14:43 ` Laurent Dufour
2022-09-09 0:15 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 09/28] mm/mempolicy: mark VMA as locked when changing protection policy Suren Baghdasaryan
2022-09-06 14:47 ` Laurent Dufour
2022-09-09 0:27 ` Suren Baghdasaryan
2022-09-01 17:34 ` [RFC PATCH RESEND 10/28] mm/mmap: mark VMAs as locked in vma_adjust Suren Baghdasaryan
2022-09-06 15:35 ` Laurent Dufour
2022-09-09 0:51 ` Suren Baghdasaryan
2022-09-09 15:52 ` Laurent Dufour
2022-09-01 17:34 ` [RFC PATCH RESEND 11/28] mm/mmap: mark VMAs as locked before merging or splitting them Suren Baghdasaryan
2022-09-06 15:44 ` Laurent Dufour
2022-09-01 17:35 ` [RFC PATCH RESEND 12/28] mm/mremap: mark VMA as locked while remapping it to a new address range Suren Baghdasaryan
2022-09-06 16:09 ` Laurent Dufour
2022-09-01 17:35 ` [RFC PATCH RESEND 13/28] mm: conditionally mark VMA as locked in free_pgtables and unmap_page_range Suren Baghdasaryan
2022-09-09 10:33 ` Laurent Dufour
2022-09-09 16:43 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 14/28] mm: mark VMAs as locked before isolating them Suren Baghdasaryan
2022-09-09 13:35 ` Laurent Dufour
2022-09-09 16:28 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 15/28] mm/mmap: mark adjacent VMAs as locked if they can grow into unmapped area Suren Baghdasaryan
2022-09-09 13:43 ` Laurent Dufour
2022-09-09 16:25 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 16/28] kernel/fork: assert no VMA readers during its destruction Suren Baghdasaryan
2022-09-09 13:56 ` Laurent Dufour
2022-09-09 16:19 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 17/28] mm/mmap: prevent pagefault handler from racing with mmu_notifier registration Suren Baghdasaryan
2022-09-09 14:20 ` Laurent Dufour
2022-09-09 16:12 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 18/28] mm: add FAULT_FLAG_VMA_LOCK flag Suren Baghdasaryan
2022-09-09 14:26 ` Laurent Dufour
2022-09-01 17:35 ` [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock Suren Baghdasaryan
2022-09-06 19:39 ` Peter Xu
2022-09-06 20:08 ` Suren Baghdasaryan
2022-09-06 20:22 ` Peter Xu
2022-09-07 0:58 ` Suren Baghdasaryan
2022-09-09 14:26 ` Laurent Dufour
2022-09-01 17:35 ` [RFC PATCH RESEND 20/28] mm: introduce per-VMA lock statistics Suren Baghdasaryan
2022-09-09 14:28 ` Laurent Dufour
2022-09-09 16:11 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 21/28] mm: introduce find_and_lock_anon_vma to be used from arch-specific code Suren Baghdasaryan
2022-09-09 14:38 ` Laurent Dufour
2022-09-09 16:10 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 22/28] x86/mm: try VMA lock-based page fault handling first Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 23/28] x86/mm: define ARCH_SUPPORTS_PER_VMA_LOCK Suren Baghdasaryan
2022-09-01 20:20 ` Kent Overstreet
2022-09-01 23:17 ` Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 24/28] arm64/mm: try VMA lock-based page fault handling first Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 25/28] arm64/mm: define ARCH_SUPPORTS_PER_VMA_LOCK Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 26/28] powerc/mm: try VMA lock-based page fault handling first Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 27/28] powerpc/mm: define ARCH_SUPPORTS_PER_VMA_LOCK Suren Baghdasaryan
2022-09-01 17:35 ` [RFC PATCH RESEND 28/28] kernel/fork: throttle call_rcu() calls in vm_area_free Suren Baghdasaryan
2022-09-09 15:19 ` Laurent Dufour
2022-09-09 16:02 ` Suren Baghdasaryan
2022-09-09 16:14 ` Laurent Dufour
2022-09-01 20:58 ` [RFC PATCH RESEND 00/28] per-VMA locks proposal Kent Overstreet
2022-09-01 23:26 ` Suren Baghdasaryan
2022-09-11 9:35 ` Vlastimil Babka
2022-09-28 2:28 ` Suren Baghdasaryan
2022-09-29 11:18 ` Vlastimil Babka
2022-09-02 7:42 ` Peter Zijlstra
2022-09-02 14:45 ` Suren Baghdasaryan
2022-09-05 12:32 ` Michal Hocko
2022-09-05 18:32 ` Suren Baghdasaryan
2022-09-05 20:35 ` Kent Overstreet
2022-09-06 15:46 ` Suren Baghdasaryan
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='CAJuCfpFi07ZpAuSLwADGO0mfu2Hon9jWb0hsCytorchQ=cgK3Q@mail.gmail.com' \
--to=surenb@google.com \
--cc=akpm@linux-foundation.org \
--cc=axelrasmussen@google.com \
--cc=bigeasy@linutronix.de \
--cc=dave@stgolabs.net \
--cc=david@redhat.com \
--cc=dhowells@redhat.com \
--cc=hannes@cmpxchg.org \
--cc=hughd@google.com \
--cc=jglisse@google.com \
--cc=joelaf@google.com \
--cc=kent.overstreet@linux.dev \
--cc=kernel-team@android.com \
--cc=laurent.dufour@fr.ibm.com \
--cc=ldufour@linux.ibm.com \
--cc=liam.howlett@oracle.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=luto@kernel.org \
--cc=mgorman@suse.de \
--cc=mhocko@suse.com \
--cc=michel@lespinasse.org \
--cc=minchan@google.com \
--cc=paulmck@kernel.org \
--cc=peterx@redhat.com \
--cc=peterz@infradead.org \
--cc=rientjes@google.com \
--cc=songliubraving@fb.com \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
--cc=x86@kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).