Re: [PATCH v4] mm/memory: use pmdp_get_lockless() for the pmd read in do_fault_around()

[Hui Zhu <hui.zhu@linux.dev>]

2026/11/6 20:08, Greg Ungerer:

Hi Greg,

> Hi Hui,
>
> On 6/11/26 13:03, Hui Zhu wrote:
>> 2026/11/5 20:00, David Hildenbrand (Arm) :
>>> On 11/4/26 10:10, Hui Zhu wrote:
>>>> From: Hui Zhu <zhuhui@kylinos.cn>
>>>>
>>>> KCSAN reports a data-race in do_read_fault / pmd_install; the same code
>>>> is in mainline:
>>>>
>>>>     write (marked) to 0xffff8d52a75fb810 of 8 bytes by task 27487 on cpu 5:
>>>>      native_set_pmd arch/x86/include/asm/pgtable_64.h:83 [inline]
>>>>      pmd_populate arch/x86/include/asm/pgalloc.h:84 [inline]
>>>>      pmd_install+0xe0/0x120 mm/memory.c:444
>>>>      finish_fault+0x5bc/0xb00 mm/memory.c:5204
>>>>      do_cow_fault mm/memory.c:5432 [inline]
>>>>      do_fault+0x5d0/0x6e0 mm/memory.c:5526
>>>>      handle_mm_fault+0x211/0x4c0 mm/memory.c:6218
>>>>      __mm_populate+0x144/0x270 mm/gup.c:2207
>>>>      __do_sys_mlockall+0x2cd/0x380 mm/mlock.c:764
>>>>
>>>>     read to 0xffff8d52a75fb810 of 8 bytes by task 27527 on cpu 9:
>>>>      do_fault_around mm/memory.c:5341 [inline]
>>>>      do_read_fault+0x162/0x670 mm/memory.c:5381
>>>>      do_fault+0x31a/0x6e0 mm/memory.c:5524
>>>>      handle_mm_fault+0x211/0x4c0 mm/memory.c:6218
>>>>      faultin_page_range+0x227/0x530 mm/gup.c:2156
>>>>      __x64_sys_madvise+0x67/0x80 mm/madvise.c:1479
>>>>
>>>>     value changed: 0x0000000000000000 -> 0x000000001386c067
>>>>
>>>> Both paths only hold mmap_lock for read, so do_fault_around() tests
>>>> vmf->pmd while pmd_install() populates it under the pmd lock.  The race is
>>>> benign: a stale pmd_none() only wastes the prealloc_pte, which do_fault()
>>>> frees again, and a pmd seen as populated only makes ->map_pages() find no
>>>> PTE and do_read_fault() fall back to __do_fault().
>>>>
>>>> The load still has to be lockless-safe.  The store side already is: on
>>>> i386-PAE, pmd_install() publishes both halves of the pmd with a single
>>>> cmpxchg8b.  The load side is not.  pmdp_get() is a plain READ_ONCE() over
>>>> a pmd_t that is wider than a word there, so it becomes two 32-bit loads
>>>> and can return the low half of one pmd together with the high half of
>>>> another.  Use pmdp_get_lockless(), the accessor for pmd reads that do not
>>>> hold the pmd lock: it re-reads until both halves come from the same pmd.
>>>> It is also what __pte_offset_map() already uses on this very pmd a few
>>>> instructions later.
>>>>
>>>> Only X86_32+X86_PAE and SH+X2TLB compile the retry loop, these being the
>>>> GUP_GET_PXX_LOW_HIGH selectors that have more than two page table levels.
>>>> mips32 with 64-bit physical addresses selects GUP_GET_PXX_LOW_HIGH too,
>>>> but is 2-level and so falls back to pmdp_get().  On every other
>>>> configuration, x86-64 included, pmdp_get_lockless() is pmdp_get(): the
>>>> pmd is loaded in one go, the READ_ONCE() silences the report the same
>>>> way, and there is no functional change.
>>>>
>>>> Fixes: 7267ec008b5c ("mm: postpone page table allocation until we have page to map")
>>>> Cc: stable+noautosel@kernel.org # KCSAN annotation only, no functional change
>>>> Signed-off-by: Hui Zhu <zhuhui@kylinos.cn>
>>>> Acked-by: Lorenzo Stoakes <ljs@kernel.org>
>>>> ---
>>>> Changelog:
>>>> v4:
>>>> Commit message only, the diff is identical to v3.  Spell out the affected
>>>> arches and add the lore links for v1-v3 (Lorenzo Stoakes).  Reword the
>>>> cmpxchg8b sentence: the store is already atomic, it is the load that tears.
>>>> Add the Fixes: tag (David Hildenbrand).
>>>> v3:
>>>> https://lore.kernel.org/all/20261103014749.96764-1-hui.zhu@linux.dev/
>>>> Shorten the comment to one line (David Hildenbrand).
>>>> v2:
>>>> https://lore.kernel.org/all/20261102093120.92654-1-hui.zhu@linux.dev/
>>>> Use pmdp_get_lockless() instead of pmdp_get() (David Hildenbrand).
>>>> v1:
>>>> https://lore.kernel.org/all/20261102031308.66177-1-hui.zhu@linux.dev/
>>>>
>>>>    mm/memory.c | 3 ++-
>>>>    1 file changed, 2 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/mm/memory.c b/mm/memory.c
>>>> index 0fb676dd2bd63..f25c2ce0ef153 100644
>>>> --- a/mm/memory.c
>>>> +++ b/mm/memory.c
>>>> @@ -6017,7 +6017,8 @@ static vm_fault_t do_fault_around(struct vm_fault *vmf)
>>>>        to_pte = min3(from_pte + nr_pages, (pgoff_t)PTRS_PER_PTE,
>>>>                  pte_off + vma_pages(vmf->vma) - vma_off) - 1;
>>>>    -    if (pmd_none(*vmf->pmd)) {
>>>> +    /* pmd_install() rechecks pmd_none() under the pmd lock. */
>>>> +    if (pmd_none(pmdp_get_lockless(vmf->pmd))) {
>>>>            vmf->prealloc_pte = pte_alloc_one(vmf->vma->vm_mm);
>>>>            if (!vmf->prealloc_pte)
>>>>                return VM_FAULT_OOM;
>>> Fun, this breaks some odd m68k configs because they do some legacy weird stuff with pmd_none().
>>>
>>> I'll throw in the following cleanup. CCing m68k maintainers: I would drag
>>> this along with this fix here unless you disagree.
>>>
>>>   From e3c739933a04743e3923f6f39a603e93f2700304 Mon Sep 17 00:00:00 2001
>>> From: "David Hildenbrand (Arm)" <david@kernel.org>
>>> Date: Thu, 5 Nov 2026 12:49:38 +0100
>>> Subject: [PATCH] m68k/mm: simplify
>>>    pmd_none()/pmd_val()/pmd_present()/pmd_bad()
>>>
>>> pmd_t on m68k used to contain 16 words. Commit ef22d8abd876 ("m68k:
>>> mm: Restructure Motorola MMU page-table layout") reduced Motorola PMDs
>>> to one word; Sun3 and ColdFire use word-sized folded PMDs.
>>>
>>> Use direct member access in pmd_val() and replace the Sun3 and ColdFire
>>> pointer-taking predicate wrappers with inline functions taking values.
>>>
>>> This allows us to write something like:
>>>
>>>     if (pmd_none(pmdp_get(pmdp)))
>>>
>>> While at it remove unused pmd_present2() and commented-out pmd_bad()
>>> on sun3.
>>>
>>> Signed-off-by: David Hildenbrand (Arm) <david@kernel.org>
>>> ---
>>>    arch/m68k/include/asm/mcf_pgtable.h  | 20 +++++++++++++++-----
>>>    arch/m68k/include/asm/page.h         |  2 +-
>>>    arch/m68k/include/asm/sun3_pgtable.h | 23 +++++++++++++++--------
>>>    3 files changed, 31 insertions(+), 14 deletions(-)
>>>
>>> diff --git a/arch/m68k/include/asm/mcf_pgtable.h b/arch/m68k/include/asm/mcf_pgtable.h
>>> index f45a882238dbb..4efa8e3e1e98a 100644
>>> --- a/arch/m68k/include/asm/mcf_pgtable.h
>>> +++ b/arch/m68k/include/asm/mcf_pgtable.h
>>> @@ -130,11 +130,21 @@ static inline void pte_clear(struct mm_struct *mm, unsigned long addr,
>>>      #define pte_page(pte)    virt_to_page(__pte_page(pte))
>>>    -static inline int pmd_none2(pmd_t *pmd) { return !pmd_val(*pmd); }
>>> -#define pmd_none(pmd) pmd_none2(&(pmd))
>>> -static inline int pmd_bad2(pmd_t *pmd) { return 0; }
>>> -#define pmd_bad(pmd) pmd_bad2(&(pmd))
>>> -#define pmd_present(pmd) (!pmd_none2(&(pmd)))
>>> +static inline int pmd_none(pmd_t pmd)
>>> +{
>>> +    return !pmd_val(pmd);
>>> +}
>>> +
>>> +static inline int pmd_bad(pmd_t pmd)
>>> +{
>>> +    return 0;
>>> +}
>>> +
>>> +static inline int pmd_present(pmd_t pmd)
>>> +{
>>> +    return !pmd_none(pmd);
>>> +}
>>> +
>>>    static inline void pmd_clear(pmd_t *pmdp) { pmd_val(*pmdp) = 0; }
>>>      diff --git a/arch/m68k/include/asm/page.h b/arch/m68k/include/asm/page.h
>>> index d30f8b2f15927..09a2c020ec992 100644
>>> --- a/arch/m68k/include/asm/page.h
>>> +++ b/arch/m68k/include/asm/page.h
>>> @@ -17,7 +17,7 @@
>>>     */
>>>    #if !defined(CONFIG_MMU) || CONFIG_PGTABLE_LEVELS == 3
>>>    typedef struct { unsigned long pmd; } pmd_t;
>>> -#define pmd_val(x)    ((&x)->pmd)
>>> +#define pmd_val(x)    ((x).pmd)
>>>    #define __pmd(x)    ((pmd_t) { (x) } )
>>>    #endif
>>>    diff --git a/arch/m68k/include/asm/sun3_pgtable.h b/arch/m68k/include/asm/sun3_pgtable.h
>>> index 704442a391fd7..5867d24032655 100644
>>> --- a/arch/m68k/include/asm/sun3_pgtable.h
>>> +++ b/arch/m68k/include/asm/sun3_pgtable.h
>>> @@ -109,14 +109,21 @@ static inline void pte_clear (struct mm_struct *mm, unsigned long addr, pte_t *p
>>>    #define pmd_page(pmd)        virt_to_page((void *)pmd_page_vaddr(pmd))
>>>      -static inline int pmd_none2 (pmd_t *pmd) { return !pmd_val (*pmd); }
>>> -#define pmd_none(pmd) pmd_none2(&(pmd))
>>> -//static inline int pmd_bad (pmd_t pmd) { return (pmd_val (pmd) & SUN3_PMD_MASK) != SUN3_PMD_MAGIC; }
>>> -static inline int pmd_bad2 (pmd_t *pmd) { return 0; }
>>> -#define pmd_bad(pmd) pmd_bad2(&(pmd))
>>> -static inline int pmd_present2 (pmd_t *pmd) { return pmd_val (*pmd) & SUN3_PMD_VALID; }
>>> -/* #define pmd_present(pmd) pmd_present2(&(pmd)) */
>>> -#define pmd_present(pmd) (!pmd_none2(&(pmd)))
>>> +static inline int pmd_none(pmd_t pmd)
>>> +{
>>> +    return !pmd_val(pmd);
>>> +}
>>> +
>>> +static inline int pmd_bad(pmd_t pmd)
>>> +{
>>> +    return 0;
>>> +}
>>> +
>>> +static inline int pmd_present(pmd_t pmd)
>>> +{
>>> +    return !pmd_none(pmd);
>>> +}
>>> +
>>>    static inline void pmd_clear (pmd_t *pmdp) { pmd_val (*pmdp) = 0; }
>>>      /*
>> Hi David,
>>
>> Thanks for picking this up, and for the m68k cleanup!
>>
>> Tested-by: Hui Zhu <zhuhui@kylinos.cn>
>>
>> Tested with gcc-m68k-linux-gnu 13.3.0 (Ubuntu 13.3.0-6ubuntu2~24.04), on
>> three configs that between them cover all three m68k families:
>>
>>    amiga_defconfig                    Motorola MMU, PGTABLE_LEVELS=3
>>    sun3_defconfig                     SUN3,  PGTABLE_LEVELS=2
>>    m5208evb_defconfig + CONFIG_MMU=y  ColdFire,  PGTABLE_LEVELS=2
>>
>> Without your patch none of the three can build mm/memory.o, and each fails
>> at a different site:
>>
>>    arch/m68k/include/asm/page.h:20:27: error: lvalue required as unary '&' operand
>>    arch/m68k/include/asm/sun3_pgtable.h:113:33: error: lvalue required as unary '&' operand
>>    arch/m68k/include/asm/mcf_pgtable.h:134:33: error: lvalue required as unary '&' operand
>>
>> With it, all three build a full vmlinux with zero warnings.
>>
>> Two things that might save you some time:
>>
>> - m68k looks like the only arch with this pattern. Grepping arch/ and
>>    include/ for '(&x)->pmd' and '(&(pmd))' finds nothing outside arch/m68k/,
>>    and pmd_none2()/pmd_bad2()/pmd_present2() only exist in mcf_pgtable.h and
>>    sun3_pgtable.h. So your three hunks look complete to me.
>>
>> - No m52xx/m54xx defconfig enables CONFIG_MMU, so the ColdFire path through
>>    mcf_pgtable.h has no defconfig coverage at all; you have to flip it on by
>>    hand (scripts/config -e MMU && make olddefconfig) to reach it. That is
>>    probably how this went unnoticed for so long - my patch has been breaking
>>    m68k since v1, because pmdp_get() returns by value just like
>>    pmdp_get_lockless() does.
> The m5475evb_defconfig is supposed to have MMU enabled:
>
>       $ ARCH=m68k CROSS_COMPILE=m68k-linux-uclibc- make m5475evb_defconfig
>         HOSTCC  scripts/basic/fixdep
>         [snip]
>         HOSTLD  scripts/kconfig/conf
>       #
>       # configuration written to .config
>       #
>       $
>       $ grep CONFIG_MMU .config
>       CONFIG_MMU=y
>       CONFIG_MMU_COLDFIRE=y
>       CONFIG_MMU_GATHER_NO_RANGE=y
>       CONFIG_MMU_GATHER_MERGE_VMAS=y
>       CONFIG_MMU_LAZY_TLB_REFCOUNT=y
>
> But something ends up disabling it when you compile.
> I'll fix this.
>
>
>> I also checked whether your patch changes code generation, by building the
>> same tree with and without it and comparing every object with objdump -d
>> (plain cmp is useless here, timestamps and link layout shift everything).
>> Counting only objects whose disassembly actually differs:
>>
>>    amiga      3733 objects, 0 changed
>>    sun3       3553 objects, 3 changed: mm/memory.o, mm/pgtable-generic.o,
>>                                        mm/vmalloc.o
>>    m5208evb    716 objects, 1 changed: mm/pgtable-generic.o
>>
>> amiga is the interesting one. It is the only config where pmd_val() itself
>> changed, and it comes out bit-for-bit identical, which also says the build
>> is reproducible enough for this comparison to mean anything. sun3 and
>> ColdFire differ because pmd_none() stops taking the address of its
>> argument, and the direction is an improvement. In mm/vmalloc.o on sun3 one
>> function loses a forced spill:
>>
>>    -  subql #4,%sp
>>    -  movel %d0,%sp@(12)     /* spill the pmd value */
>>    -  lea %sp@(12),%a1       /* &(pmd) */
>>    -  moveal %a1@,%a0        /* read it back */
>>
>> and turns into a 13 instruction function with no stack slot at all.
>> Instruction counts are otherwise flat (sun3 mm/memory.o 7373 -> 7373,
>> mm/vmalloc.o 5514 -> 5515), so it is not a size regression.
>>
>> Dragging the cleanup along with the fix sounds right to me. It is a hard
>> prerequisite, and sent on its own it would just look like a cleanup with no
>> urgency. I assume it goes in ahead of 6ab55cfd in core-7.5.misc, so that
>> linux-next never sees a broken m68k in between.
>>
>> One red herring in case anyone else pokes at this: m5208evb with its stock
>> defconfig (CONFIG_MMU not set) fails to build mm/memory.o for reasons that
>> have nothing to do with any of this - implicit declarations of
>> pte_is_uffd_wp_marker() and pmd_pgtable(), plus a parse error in
>> include/linux/mm.h. Pre-existing, and it goes away once MMU is on.
> You should not be able to set MMU=y if configured for the M5208.
> It doesn't have an MMU. You want the 5475, 5485 or 54418 to test
> valid MMU=y configurations.
>
> Regards
> Greg
>
>
>

You are right on both counts, and that mail of mine had three things
wrong.

1. "No m52xx/m54xx defconfig enables CONFIG_MMU" - wrong.  I grepped the
defconfig fragments for CONFIG_MMU=y, but MMU is 'default y', so an
enabled MMU never appears in the fragment.  m5475evb and m54418evb
both give CONFIG_MMU=y, as you show.

2. My "m5208evb + CONFIG_MMU=y" test config was not a valid config.
M520x is 'depends on !MMU', so forcing MMU on leaves the SoC choice
unsatisfiable and olddefconfig silently drops it:

before:  # CONFIG_MMU is not set / CONFIG_M520x=y
after:   CONFIG_MMU=y / CONFIG_MMU_COLDFIRE=y / CONFIG_M520x gone

Please disregard the m5208evb row in my codegen numbers.

3. The "red herring" I flagged was my own doing, not a pre-existing
problem.  mm/Makefile has 'mmu-$(CONFIG_MMU) := ... memory.o', so a
NOMMU config never builds mm/memory.o at all; I asked for a target
that config does not have.  Sorry for the noise on that one.

Retested with the configs you named (7.4-rc1 plus the do_fault_around
change plus David's cleanup, gcc-m68k-linux-gnu 13.3.0):

m5475evb_defconfig   without David's patch: mm/memory.o fails at
mcf_pgtable.h:134 with it: full -j24 build ok, vmlinux 4215344
bytes, 0 warnings

m54418evb_defconfig  without David's patch: same mcf_pgtable.h:134
failure

So the Tested-by I sent covers amiga, sun3 and m5475evb.  It is
build-tested only - I have no m68k hardware here to boot.

On MMU being disabled when you compile: I cannot reproduce that. Here
m5475evb_defconfig gives CONFIG_MMU=y, olddefconfig reports no change,
two consecutive defconfig runs produce an identical .config, and a full
build finishes with CONFIG_MMU=y and CONFIG_M547x=y still set and
.config untouched.  Two differences from your setup that might matter:
I am on a glibc toolchain rather than uclibc, and building out of tree
with O= rather than in tree.

One observation, which may well be unrelated.  m5475evb_defconfig never
sets CONFIG_M547x - it only ends up with it because the "ColdFire SoC
type" choice defaults to M520x, which is unavailable under MMU=y, so
kconfig falls back to the first usable member. m54418evb_defconfig sets
CONFIG_M5441x=y explicitly.  If that fallback is what breaks on your
side, making m5475evb explicit would sidestep it.

Best,
Hui


>