The patch titled
Subject: mm: handle shared faults under the VMA lock
has been added to the -mm mm-unstable branch. Its filename is
mm-handle-shared-faults-under-the-vma-lock.patch
This patch will shortly appear at
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-handle-shared-faults-under-the-vma-lock.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: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx>
Subject: mm: handle shared faults under the VMA lock
Date: Fri, 6 Oct 2023 20:53:15 +0100
There are many implementations of ->fault and some of them depend on
mmap_lock being held. All vm_ops that implement ->map_pages() end up
calling filemap_fault(), which I have audited to be sure it does not rely
on mmap_lock. So (for now) key off ->map_pages existing as a flag to
indicate that it's safe to call ->fault while only holding the vma lock.
Link: https://lkml.kernel.org/r/20231006195318.4087158-4-willy@xxxxxxxxxxxxx
Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Cc: Suren Baghdasaryan <surenb@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/memory.c | 22 ++++++++++++++++++----
1 file changed, 18 insertions(+), 4 deletions(-)
--- a/mm/memory.c~mm-handle-shared-faults-under-the-vma-lock
+++ a/mm/memory.c
@@ -3045,6 +3045,21 @@ static inline void wp_page_reuse(struct
count_vm_event(PGREUSE);
}
+/*
+ * We could add a bitflag somewhere, but for now, we know that all
+ * vm_ops that have a ->map_pages have been audited and don't need
+ * the mmap_lock to be held.
+ */
+static inline vm_fault_t vmf_can_call_fault(const struct vm_fault *vmf)
+{
+ struct vm_area_struct *vma = vmf->vma;
+
+ if (vma->vm_ops->map_pages || !(vmf->flags & FAULT_FLAG_VMA_LOCK))
+ return 0;
+ vma_end_read(vma);
+ return VM_FAULT_RETRY;
+}
+
static vm_fault_t vmf_anon_prepare(struct vm_fault *vmf)
{
struct vm_area_struct *vma = vmf->vma;
@@ -4672,10 +4687,9 @@ static vm_fault_t do_shared_fault(struct
vm_fault_t ret, tmp;
struct folio *folio;
- if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
- vma_end_read(vma);
- return VM_FAULT_RETRY;
- }
+ ret = vmf_can_call_fault(vmf);
+ if (ret)
+ return ret;
ret = __do_fault(vmf);
if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY)))
_
Patches currently in -mm which might be from willy@xxxxxxxxxxxxx are
mm-make-lock_folio_maybe_drop_mmap-vma-lock-aware.patch
mm-call-wp_page_copy-under-the-vma-lock.patch
mm-handle-shared-faults-under-the-vma-lock.patch
mm-handle-cow-faults-under-the-vma-lock.patch
mm-handle-read-faults-under-the-vma-lock.patch
mm-handle-write-faults-to-ro-pages-under-the-vma-lock.patch
iomap-hold-state_lock-over-call-to-ifs_set_range_uptodate.patch
iomap-protect-read_bytes_pending-with-the-state_lock.patch
mm-add-folio_end_read.patch
ext4-use-folio_end_read.patch
buffer-use-folio_end_read.patch
iomap-use-folio_end_read.patch
bitops-add-xor_unlock_is_negative_byte.patch
alpha-implement-xor_unlock_is_negative_byte.patch
m68k-implement-xor_unlock_is_negative_byte.patch
mips-implement-xor_unlock_is_negative_byte.patch
powerpc-implement-arch_xor_unlock_is_negative_byte-on-32-bit.patch
riscv-implement-xor_unlock_is_negative_byte.patch
s390-implement-arch_xor_unlock_is_negative_byte.patch
mm-delete-checks-for-xor_unlock_is_negative_byte.patch
mm-add-folio_xor_flags_has_waiters.patch
mm-make-__end_folio_writeback-return-void.patch
mm-use-folio_xor_flags_has_waiters-in-folio_end_writeback.patch
