The patch titled
Subject: mm: userfaultfd: don't pass around both mm and vma
has been added to the -mm mm-unstable branch. Its filename is
mm-userfaultfd-dont-pass-around-both-mm-and-vma.patch
This patch will shortly appear at
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-userfaultfd-dont-pass-around-both-mm-and-vma.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: Axel Rasmussen <axelrasmussen@xxxxxxxxxx>
Subject: mm: userfaultfd: don't pass around both mm and vma
Date: Tue, 14 Mar 2023 15:12:48 -0700
Quite a few userfaultfd functions took both mm and vma pointers as
arguments. Since the mm is trivially accessible via vma->vm_mm, there's
no reason to pass both; it just needlessly extends the already long
argument list.
Get rid of the mm pointer, where possible, to shorten the argument list.
Link: https://lkml.kernel.org/r/20230314221250.682452-3-axelrasmussen@xxxxxxxxxx
Signed-off-by: Axel Rasmussen <axelrasmussen@xxxxxxxxxx>
Acked-by: Peter Xu <peterx@xxxxxxxxxx>
Acked-by: Mike Rapoport (IBM) <rppt@xxxxxxxxxx>
Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Cc: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: James Houghton <jthoughton@xxxxxxxxxx>
Cc: Jan Kara <jack@xxxxxxx>
Cc: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx>
Cc: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Cc: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
Cc: Muchun Song <muchun.song@xxxxxxxxx>
Cc: Nadav Amit <namit@xxxxxxxxxx>
Cc: Shuah Khan <shuah@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
fs/userfaultfd.c | 2 -
include/linux/hugetlb.h | 5 +-
include/linux/shmem_fs.h | 4 +-
include/linux/userfaultfd_k.h | 4 +-
mm/hugetlb.c | 4 +-
mm/shmem.c | 7 +--
mm/userfaultfd.c | 61 +++++++++++++++-----------------
7 files changed, 41 insertions(+), 46 deletions(-)
--- a/fs/userfaultfd.c~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/fs/userfaultfd.c
@@ -1644,7 +1644,7 @@ static int userfaultfd_unregister(struct
/* Reset ptes for the whole vma range if wr-protected */
if (userfaultfd_wp(vma))
- uffd_wp_range(mm, vma, start, vma_end - start, false);
+ uffd_wp_range(vma, start, vma_end - start, false);
new_flags = vma->vm_flags & ~__VM_UFFD_FLAGS;
prev = vma_merge(&vmi, mm, prev, start, vma_end, new_flags,
--- a/include/linux/hugetlb.h~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/include/linux/hugetlb.h
@@ -158,7 +158,7 @@ unsigned long hugetlb_total_pages(void);
vm_fault_t hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma,
unsigned long address, unsigned int flags);
#ifdef CONFIG_USERFAULTFD
-int hugetlb_mfill_atomic_pte(struct mm_struct *dst_mm, pte_t *dst_pte,
+int hugetlb_mfill_atomic_pte(pte_t *dst_pte,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
@@ -393,8 +393,7 @@ static inline void hugetlb_free_pgd_rang
}
#ifdef CONFIG_USERFAULTFD
-static inline int hugetlb_mfill_atomic_pte(struct mm_struct *dst_mm,
- pte_t *dst_pte,
+static inline int hugetlb_mfill_atomic_pte(pte_t *dst_pte,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
--- a/include/linux/shmem_fs.h~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/include/linux/shmem_fs.h
@@ -153,14 +153,14 @@ extern void shmem_uncharge(struct inode
#ifdef CONFIG_USERFAULTFD
#ifdef CONFIG_SHMEM
-extern int shmem_mfill_atomic_pte(struct mm_struct *dst_mm, pmd_t *dst_pmd,
+extern int shmem_mfill_atomic_pte(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
bool zeropage, bool wp_copy,
struct page **pagep);
#else /* !CONFIG_SHMEM */
-#define shmem_mfill_atomic_pte(dst_mm, dst_pmd, dst_vma, dst_addr, \
+#define shmem_mfill_atomic_pte(dst_pmd, dst_vma, dst_addr, \
src_addr, zeropage, wp_copy, pagep) ({ BUG(); 0; })
#endif /* CONFIG_SHMEM */
#endif /* CONFIG_USERFAULTFD */
--- a/include/linux/userfaultfd_k.h~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/include/linux/userfaultfd_k.h
@@ -56,7 +56,7 @@ enum mcopy_atomic_mode {
MCOPY_ATOMIC_CONTINUE,
};
-extern int mfill_atomic_install_pte(struct mm_struct *dst_mm, pmd_t *dst_pmd,
+extern int mfill_atomic_install_pte(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr, struct page *page,
bool newly_allocated, bool wp_copy);
@@ -73,7 +73,7 @@ extern ssize_t mfill_atomic_continue(str
extern int mwriteprotect_range(struct mm_struct *dst_mm,
unsigned long start, unsigned long len,
bool enable_wp, atomic_t *mmap_changing);
-extern long uffd_wp_range(struct mm_struct *dst_mm, struct vm_area_struct *vma,
+extern long uffd_wp_range(struct vm_area_struct *vma,
unsigned long start, unsigned long len, bool enable_wp);
/* mm helpers */
--- a/mm/hugetlb.c~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/mm/hugetlb.c
@@ -6159,8 +6159,7 @@ out_mutex:
* Used by userfaultfd UFFDIO_* ioctls. Based on userfaultfd's mfill_atomic_pte
* with modifications for hugetlb pages.
*/
-int hugetlb_mfill_atomic_pte(struct mm_struct *dst_mm,
- pte_t *dst_pte,
+int hugetlb_mfill_atomic_pte(pte_t *dst_pte,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
@@ -6168,6 +6167,7 @@ int hugetlb_mfill_atomic_pte(struct mm_s
struct page **pagep,
bool wp_copy)
{
+ struct mm_struct *dst_mm = dst_vma->vm_mm;
bool is_continue = (mode == MCOPY_ATOMIC_CONTINUE);
struct hstate *h = hstate_vma(dst_vma);
struct address_space *mapping = dst_vma->vm_file->f_mapping;
--- a/mm/shmem.c~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/mm/shmem.c
@@ -2519,8 +2519,7 @@ static struct inode *shmem_get_inode(str
}
#ifdef CONFIG_USERFAULTFD
-int shmem_mfill_atomic_pte(struct mm_struct *dst_mm,
- pmd_t *dst_pmd,
+int shmem_mfill_atomic_pte(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
@@ -2610,11 +2609,11 @@ int shmem_mfill_atomic_pte(struct mm_str
goto out_release;
ret = shmem_add_to_page_cache(folio, mapping, pgoff, NULL,
- gfp & GFP_RECLAIM_MASK, dst_mm);
+ gfp & GFP_RECLAIM_MASK, dst_vma->vm_mm);
if (ret)
goto out_release;
- ret = mfill_atomic_install_pte(dst_mm, dst_pmd, dst_vma, dst_addr,
+ ret = mfill_atomic_install_pte(dst_pmd, dst_vma, dst_addr,
&folio->page, true, wp_copy);
if (ret)
goto out_delete_from_cache;
--- a/mm/userfaultfd.c~mm-userfaultfd-dont-pass-around-both-mm-and-vma
+++ a/mm/userfaultfd.c
@@ -55,12 +55,13 @@ struct vm_area_struct *find_dst_vma(stru
* This function handles both MCOPY_ATOMIC_NORMAL and _CONTINUE for both shmem
* and anon, and for both shared and private VMAs.
*/
-int mfill_atomic_install_pte(struct mm_struct *dst_mm, pmd_t *dst_pmd,
+int mfill_atomic_install_pte(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr, struct page *page,
bool newly_allocated, bool wp_copy)
{
int ret;
+ struct mm_struct *dst_mm = dst_vma->vm_mm;
pte_t _dst_pte, *dst_pte;
bool writable = dst_vma->vm_flags & VM_WRITE;
bool vm_shared = dst_vma->vm_flags & VM_SHARED;
@@ -127,8 +128,7 @@ out_unlock:
return ret;
}
-static int mfill_atomic_pte_copy(struct mm_struct *dst_mm,
- pmd_t *dst_pmd,
+static int mfill_atomic_pte_copy(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
@@ -190,10 +190,10 @@ static int mfill_atomic_pte_copy(struct
__SetPageUptodate(page);
ret = -ENOMEM;
- if (mem_cgroup_charge(page_folio(page), dst_mm, GFP_KERNEL))
+ if (mem_cgroup_charge(page_folio(page), dst_vma->vm_mm, GFP_KERNEL))
goto out_release;
- ret = mfill_atomic_install_pte(dst_mm, dst_pmd, dst_vma, dst_addr,
+ ret = mfill_atomic_install_pte(dst_pmd, dst_vma, dst_addr,
page, true, wp_copy);
if (ret)
goto out_release;
@@ -204,8 +204,7 @@ out_release:
goto out;
}
-static int mfill_atomic_pte_zeropage(struct mm_struct *dst_mm,
- pmd_t *dst_pmd,
+static int mfill_atomic_pte_zeropage(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr)
{
@@ -217,7 +216,7 @@ static int mfill_atomic_pte_zeropage(str
_dst_pte = pte_mkspecial(pfn_pte(my_zero_pfn(dst_addr),
dst_vma->vm_page_prot));
- dst_pte = pte_offset_map_lock(dst_mm, dst_pmd, dst_addr, &ptl);
+ dst_pte = pte_offset_map_lock(dst_vma->vm_mm, dst_pmd, dst_addr, &ptl);
if (dst_vma->vm_file) {
/* the shmem MAP_PRIVATE case requires checking the i_size */
inode = dst_vma->vm_file->f_inode;
@@ -230,7 +229,7 @@ static int mfill_atomic_pte_zeropage(str
ret = -EEXIST;
if (!pte_none(*dst_pte))
goto out_unlock;
- set_pte_at(dst_mm, dst_addr, dst_pte, _dst_pte);
+ set_pte_at(dst_vma->vm_mm, dst_addr, dst_pte, _dst_pte);
/* No need to invalidate - it was non-present before */
update_mmu_cache(dst_vma, dst_addr, dst_pte);
ret = 0;
@@ -240,8 +239,7 @@ out_unlock:
}
/* Handles UFFDIO_CONTINUE for all shmem VMAs (shared or private). */
-static int mfill_atomic_pte_continue(struct mm_struct *dst_mm,
- pmd_t *dst_pmd,
+static int mfill_atomic_pte_continue(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
bool wp_copy)
@@ -269,7 +267,7 @@ static int mfill_atomic_pte_continue(str
goto out_release;
}
- ret = mfill_atomic_install_pte(dst_mm, dst_pmd, dst_vma, dst_addr,
+ ret = mfill_atomic_install_pte(dst_pmd, dst_vma, dst_addr,
page, false, wp_copy);
if (ret)
goto out_release;
@@ -310,7 +308,7 @@ static pmd_t *mm_alloc_pmd(struct mm_str
* mfill_atomic processing for HUGETLB vmas. Note that this routine is
* called with mmap_lock held, it will release mmap_lock before returning.
*/
-static __always_inline ssize_t mfill_atomic_hugetlb(struct mm_struct *dst_mm,
+static __always_inline ssize_t mfill_atomic_hugetlb(
struct vm_area_struct *dst_vma,
unsigned long dst_start,
unsigned long src_start,
@@ -318,6 +316,7 @@ static __always_inline ssize_t mfill_ato
enum mcopy_atomic_mode mode,
bool wp_copy)
{
+ struct mm_struct *dst_mm = dst_vma->vm_mm;
int vm_shared = dst_vma->vm_flags & VM_SHARED;
ssize_t err;
pte_t *dst_pte;
@@ -411,7 +410,7 @@ retry:
goto out_unlock;
}
- err = hugetlb_mfill_atomic_pte(dst_mm, dst_pte, dst_vma,
+ err = hugetlb_mfill_atomic_pte(dst_pte, dst_vma,
dst_addr, src_addr, mode, &page,
wp_copy);
@@ -463,17 +462,15 @@ out:
}
#else /* !CONFIG_HUGETLB_PAGE */
/* fail at build time if gcc attempts to use this */
-extern ssize_t mfill_atomic_hugetlb(struct mm_struct *dst_mm,
- struct vm_area_struct *dst_vma,
- unsigned long dst_start,
- unsigned long src_start,
- unsigned long len,
- enum mcopy_atomic_mode mode,
- bool wp_copy);
+extern ssize_t mfill_atomic_hugetlb(struct vm_area_struct *dst_vma,
+ unsigned long dst_start,
+ unsigned long src_start,
+ unsigned long len,
+ enum mcopy_atomic_mode mode,
+ bool wp_copy);
#endif /* CONFIG_HUGETLB_PAGE */
-static __always_inline ssize_t mfill_atomic_pte(struct mm_struct *dst_mm,
- pmd_t *dst_pmd,
+static __always_inline ssize_t mfill_atomic_pte(pmd_t *dst_pmd,
struct vm_area_struct *dst_vma,
unsigned long dst_addr,
unsigned long src_addr,
@@ -484,7 +481,7 @@ static __always_inline ssize_t mfill_ato
ssize_t err;
if (mode == MCOPY_ATOMIC_CONTINUE) {
- return mfill_atomic_pte_continue(dst_mm, dst_pmd, dst_vma,
+ return mfill_atomic_pte_continue(dst_pmd, dst_vma,
dst_addr, wp_copy);
}
@@ -500,14 +497,14 @@ static __always_inline ssize_t mfill_ato
*/
if (!(dst_vma->vm_flags & VM_SHARED)) {
if (mode == MCOPY_ATOMIC_NORMAL)
- err = mfill_atomic_pte_copy(dst_mm, dst_pmd, dst_vma,
+ err = mfill_atomic_pte_copy(dst_pmd, dst_vma,
dst_addr, src_addr, page,
wp_copy);
else
- err = mfill_atomic_pte_zeropage(dst_mm, dst_pmd,
+ err = mfill_atomic_pte_zeropage(dst_pmd,
dst_vma, dst_addr);
} else {
- err = shmem_mfill_atomic_pte(dst_mm, dst_pmd, dst_vma,
+ err = shmem_mfill_atomic_pte(dst_pmd, dst_vma,
dst_addr, src_addr,
mode != MCOPY_ATOMIC_NORMAL,
wp_copy, page);
@@ -588,7 +585,7 @@ retry:
* If this is a HUGETLB vma, pass off to appropriate routine
*/
if (is_vm_hugetlb_page(dst_vma))
- return mfill_atomic_hugetlb(dst_mm, dst_vma, dst_start,
+ return mfill_atomic_hugetlb(dst_vma, dst_start,
src_start, len, mcopy_mode,
wp_copy);
@@ -641,7 +638,7 @@ retry:
BUG_ON(pmd_none(*dst_pmd));
BUG_ON(pmd_trans_huge(*dst_pmd));
- err = mfill_atomic_pte(dst_mm, dst_pmd, dst_vma, dst_addr,
+ err = mfill_atomic_pte(dst_pmd, dst_vma, dst_addr,
src_addr, &page, mcopy_mode, wp_copy);
cond_resched();
@@ -710,7 +707,7 @@ ssize_t mfill_atomic_continue(struct mm_
mmap_changing, 0);
}
-long uffd_wp_range(struct mm_struct *dst_mm, struct vm_area_struct *dst_vma,
+long uffd_wp_range(struct vm_area_struct *dst_vma,
unsigned long start, unsigned long len, bool enable_wp)
{
unsigned int mm_cp_flags;
@@ -732,7 +729,7 @@ long uffd_wp_range(struct mm_struct *dst
*/
if (!enable_wp && vma_wants_manual_pte_write_upgrade(dst_vma))
mm_cp_flags |= MM_CP_TRY_CHANGE_WRITABLE;
- tlb_gather_mmu(&tlb, dst_mm);
+ tlb_gather_mmu(&tlb, dst_vma->vm_mm);
ret = change_protection(&tlb, dst_vma, start, start + len, mm_cp_flags);
tlb_finish_mmu(&tlb);
@@ -788,7 +785,7 @@ int mwriteprotect_range(struct mm_struct
_start = max(dst_vma->vm_start, start);
_end = min(dst_vma->vm_end, end);
- err = uffd_wp_range(dst_mm, dst_vma, _start, _end - _start, enable_wp);
+ err = uffd_wp_range(dst_vma, _start, _end - _start, enable_wp);
/* Return 0 on success, <0 on failures */
if (err < 0)
_
Patches currently in -mm which might be from axelrasmussen@xxxxxxxxxx are
mm-userfaultfd-rename-functions-for-clarity-consistency.patch
mm-userfaultfd-dont-pass-around-both-mm-and-vma.patch
mm-userfaultfd-combine-mode-and-wp_copy-arguments.patch
mm-userfaultfd-add-uffdio_continue_mode_wp-to-install-wp-ptes.patch
