Re: [RFC PATCH 0/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

Re: [RFC PATCH 0/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

[Gerald Schaefer]

On Tue, 24 Jun 2025 13:20:42 +0530
Anshuman Khandual <anshuman.khandual@arm.com> wrote:

> Hello Gerald,
> 
> On 24/06/25 12:13 AM, Gerald Schaefer wrote:
> > Hi,
> > 
> > currently working on enabling THP_SWAP and THP_MIGRATION support for s390,
> > and stumbling over the WARN_ON(args->fixed_pmd_pfn != pmd_pfn(pmd)) in
> > debug_vm_pgtable pmd_swap_tests(). The problem is that pmd_pfn() on s390
> > will use different shift values for leaf (large) and non-leaf PMDs. And
> > when used on swapped PMDs, for which pmd_leaf() will always return false
> > because !pmd_present(), the result is not really well defined.  
> 
> Just curious - pmd_pfn() would have otherwise worked on leaf PMD entries ?
> Because the PMD swap entries are not leaf entries as pmd_present() returns
> negative, pmd_pfn() does not work on those ?

Yes, but there are actually two problems with this. The initial pmd that
is created with pfn_pmd() is already not leaf/large, but present, so
pmd_pfn() would already not work correctly on s390.

Later, after the __pmd_to_swp_entry() / __swp_entry_to_pmd() cycle, the
present bit got removed because of how those helpers will be implemented
for s390. Now it is neither large nor present, and pmd_pfn() will be
extra confused.

IOW, even if we could implement those helpers as simple no-ops similar
to other archs, the check would still not work, even though the PMD would
have the present bit set, but it still wouldn't be leaf/large.

I guess my description was a bit confusing, since the !pmd_present()
case would only show on s390, but it is not the only problem here.
I think the point is that those helpers should only be used on "proper"
swap PTE/PMD entries, which already cannot be present. And of course
that pte/pmd_pfn() is not meant to be used on such entries at all, as
David explained.

> 
> > 
> > I think that pmd_pfn() is not safe or ever meant to be called on swapped
> > PMD entries, and it doesn't seem to be used in that way anywhere else but
> > debug_vm_pgtable. Also, the whole logic to test the various swap helpers  
> 
> But is not the pmd_pfn() called on pmd which is derived from the swap entry
> first.
> 
> 	pmd = pfn_pmd(args->fixed_pmd_pfn, args->page_prot);
> 	swp = __pmd_to_swp_entry(pmd);
> 	pmd = __swp_entry_to_pmd(swp);
> 	WARN_ON(args->fixed_pmd_pfn != pmd_pfn(pmd));

Yes, but this logic is not really testing swap entries. It only works
because on other archs the __pmd_to_swp_entry() / __swp_entry_to_pmd() are
no-ops, and because pmd_pfn() does not care about leaf/large.

> 
> > on normal PTE/PMD entries seems wrong to me. It just works by chance,
> > because e.g. __pmd_to_swp_entry() and __swp_entry_to_pmd() are just no-ops
> > on other architectures (also on s390, but only for PTEs), and also  
> 
> Hmm, basically it just tests pfn_pmd() and pmd_pfn() conversions ?

Correct, but with the extra quirk that the initial PMD created by pfn_pmd()
is not leaf/large, which is apparently not a problem on other archs for
the pmd_pfn() conversion.

Actually, I now wonder why pfn_pmd() would not implicitly mark it as
leaf/large already, as it seems that this should only be used for leaf
PMDs. But maybe there are some special cases where it could also be
used for non-leaf PMDs.

> 
> > pmd_pfn() does not have any dependency on leaf/non-leaf entries there.
> Could you please elaborate on that ?

As explained above, the initial PMD created by pfn_pmd() is not leaf/large.
Well, conceptually it is more or less, but it is not marked as such. This
would lead to incorrect pmd_pfn() result (only) on s390.

> 
> > 
> > So, I started with a small patch to make pmd_swap_tests() use a proper
> > swapped PMD entry as input value, similar to how it is already done in
> > pte_swap_exclusive_tests(), and not use pmd_pfn() for compare but rather
> > compare the whole entries, again similar to pte_swap_exclusive_tests().  
> 
> Agreed, that will make sense as well.
> 
> > 
> > But then I noticed that such a change would probably also make sense for
> > the other swap tests, and also a small inconsistency in Documentation,
> > where it says e.g.
> > 
> > __pte_to_swp_entry        | Creates a swapped entry (arch) from a mapped PTE
> > 
> > I think this is wrong, those helpers should never operate on present and
> > mapped PTEs, and they certainly don't create any swapped entry from a
> > mapped entry, given that they are just no-ops on most architectures.
> > Instead, in this example, it just returns the arch-dependent
> > representation of a swp_entry_t, which happens to be just the entry
> > itself on most architectures. See also pte_to_swp_entry() /
> > swp_entry_to_pte() in include/linux/swapops.h.  
> 
> Alright.
> 
> > 
> > Now it became a larger clean-up, and I hope it makes sense. This is all
> > rather new common code for me, so maybe I got things wrong, feedback is
> > welcome.  
> 
> A quick ran on arm64 looks just fine, will keep looking into this.

Thanks!

Re: [RFC PATCH 1/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

[Gerald Schaefer]

On Mon, 23 Jun 2025 21:10:54 +0200
David Hildenbrand <david@redhat.com> wrote:

> On 23.06.25 20:43, Gerald Schaefer wrote:
> > The various __pte/pmd_to_swp_entry and __swp_entry_to_pte/pmd helper
> > functions are expected to operate on swapped PTE/PMD entries, not on
> > present and mapped entries.
> > 
> > Reflect this in the swap tests by using a swp_entry_t as input value,
> > similar to how it is already done in pte_swap_exclusive_tests().
> > Move the swap entry creation to init_args() and store it in args, so
> > it can also be used in other functions.
> > 
> > The pte/pmd_swap_tests() are also changed to compare entries instead of
> > pfn values, because pte/pmd_pfn() helpers are not expected to operate on
> > swapped entries. E.g. on s390, pmd_pfn() needs different shifts for leaf
> > (large) and non-leaf PMDs.
> > 
> > Also update documentation, to reflect that the helpers operate on
> > swapped and not mapped entries, and use correct names, i.e.
> > __swp_to_pte/pmd_entry -> __swp_entry_to_pte/pmd.
> > 
> > For consistency, also change pte/pmd_swap_soft_dirty_tests() to use
> > args->swp_entry instead of a present and mapped PTE/PMD.
> > 
> > Signed-off-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
> > ---
> >   Documentation/mm/arch_pgtable_helpers.rst |  8 ++--
> >   mm/debug_vm_pgtable.c                     | 55 ++++++++++++++---------
> >   2 files changed, 38 insertions(+), 25 deletions(-)
> > 
> > diff --git a/Documentation/mm/arch_pgtable_helpers.rst b/Documentation/mm/arch_pgtable_helpers.rst
> > index af245161d8e7..e2ac76202a85 100644
> > --- a/Documentation/mm/arch_pgtable_helpers.rst
> > +++ b/Documentation/mm/arch_pgtable_helpers.rst
> > @@ -242,13 +242,13 @@ SWAP Page Table Helpers
> >   ========================
> >   
> >   +---------------------------+--------------------------------------------------+
> > -| __pte_to_swp_entry        | Creates a swapped entry (arch) from a mapped PTE |
> > +| __pte_to_swp_entry        | Creates a swap entry (arch) from a swapped PTE   |  
> 
> Maybe something like:
> 
> "from a swap (!none && !present) PTE"
> 
> or short
> 
> "swap PTE".
> 
> "swapped" might be misleading.
> 
> Same for the other cases below.

Right, it already felt awkward when I wrote it, not sure why I only changed
it for "swapped entry (arch)". I think I like "swap PTE/PMD", naming the actual
entries in the page table, vs. "swap entry (arch)", naming the (arch-dependent)
representation of the swap PTE/PMD as swp_entry_t.

Will change, and also adjust my description, where I also used possibly
misleading "swapped".

> 
> >   +---------------------------+--------------------------------------------------+
> > -| __swp_to_pte_entry        | Creates a mapped PTE from a swapped entry (arch) |
> > +| __swp_entry_to_pte        | Creates a swapped PTE from a swap entry (arch)   |
> >   +---------------------------+--------------------------------------------------+
> > -| __pmd_to_swp_entry        | Creates a swapped entry (arch) from a mapped PMD |
> > +| __pmd_to_swp_entry        | Creates a swap entry (arch) from a swapped PMD   |
> >   +---------------------------+--------------------------------------------------+
> > -| __swp_to_pmd_entry        | Creates a mapped PMD from a swapped entry (arch) |
> > +| __swp_entry_to_pmd        | Creates a swapped PMD from a swap entry (arch)   |
> >   +---------------------------+--------------------------------------------------+
> >   | is_migration_entry        | Tests a migration (read or write) swapped entry  |
> >   +-------------------------------+----------------------------------------------+
> > diff --git a/mm/debug_vm_pgtable.c b/mm/debug_vm_pgtable.c
> > index 7731b238b534..3b0f83ed6c2e 100644
> > --- a/mm/debug_vm_pgtable.c
> > +++ b/mm/debug_vm_pgtable.c
> > @@ -73,6 +73,8 @@ struct pgtable_debug_args {
> >   	unsigned long		fixed_pud_pfn;
> >   	unsigned long		fixed_pmd_pfn;
> >   	unsigned long		fixed_pte_pfn;
> > +
> > +	swp_entry_t		swp_entry;
> >   };
> >     
> 
> Nothing else jumped at me, so LGTM.
> 

Thanks!

Re: [RFC PATCH 1/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

[Gerald Schaefer]

On Wed, 25 Jun 2025 09:58:31 +0530
Anshuman Khandual <anshuman.khandual@arm.com> wrote:

[...]
> > diff --git a/Documentation/mm/arch_pgtable_helpers.rst b/Documentation/mm/arch_pgtable_helpers.rst
> > index af245161d8e7..e2ac76202a85 100644
> > --- a/Documentation/mm/arch_pgtable_helpers.rst
> > +++ b/Documentation/mm/arch_pgtable_helpers.rst
> > @@ -242,13 +242,13 @@ SWAP Page Table Helpers
> >  ========================
> >  
> >  +---------------------------+--------------------------------------------------+
> > -| __pte_to_swp_entry        | Creates a swapped entry (arch) from a mapped PTE |
> > +| __pte_to_swp_entry        | Creates a swap entry (arch) from a swapped PTE   |
> >  +---------------------------+--------------------------------------------------+
> > -| __swp_to_pte_entry        | Creates a mapped PTE from a swapped entry (arch) |
> > +| __swp_entry_to_pte        | Creates a swapped PTE from a swap entry (arch)   |
> >  +---------------------------+--------------------------------------------------+
> > -| __pmd_to_swp_entry        | Creates a swapped entry (arch) from a mapped PMD |
> > +| __pmd_to_swp_entry        | Creates a swap entry (arch) from a swapped PMD   |
> >  +---------------------------+--------------------------------------------------+
> > -| __swp_to_pmd_entry        | Creates a mapped PMD from a swapped entry (arch) |
> > +| __swp_entry_to_pmd        | Creates a swapped PMD from a swap entry (arch)   |
> >  +---------------------------+--------------------------------------------------+
> >  | is_migration_entry        | Tests a migration (read or write) swapped entry  |
> >  +-------------------------------+----------------------------------------------+  
> 
> __pte_to_swp_entry() and __pmd_to_swp_entry() are still being used (and tested)
> even after applying this patch. Should not their entries be preserved ?

Nothing is removed here. Only adjusted description, where David already
posted some improvement. And renamed __swp_to_pte/pmd_entry() to the
correct names __swp_entry_to_pte/pmd().

[...]
> > @@ -804,17 +811,11 @@ static void __init pmd_swap_soft_dirty_tests(struct pgtable_debug_args *args) {
> >  
> >  static void __init pte_swap_exclusive_tests(struct pgtable_debug_args *args)
> >  {
> > -	unsigned long max_swap_offset;
> >  	swp_entry_t entry, entry2;
> >  	pte_t pte;
> >  
> >  	pr_debug("Validating PTE swap exclusive\n");
> > -
> > -	/* See generic_max_swapfile_size(): probe the maximum offset */
> > -	max_swap_offset = swp_offset(pte_to_swp_entry(swp_entry_to_pte(swp_entry(0, ~0UL))));
> > -
> > -	/* Create a swp entry with all possible bits set */
> > -	entry = swp_entry((1 << MAX_SWAPFILES_SHIFT) - 1, max_swap_offset);
> > +	entry = args->swp_entry;  
> args->swp_entry should be reused here as well.

Yes, and it is. I just moved the swap entry creation logic from here to
init_args(), and instead use args->swp_entry here.

> 
> >  
> >  	pte = swp_entry_to_pte(entry);
> >  	WARN_ON(pte_swp_exclusive(pte));
> > @@ -838,30 +839,36 @@ static void __init pte_swap_exclusive_tests(struct pgtable_debug_args *args)
> >  
> >  static void __init pte_swap_tests(struct pgtable_debug_args *args)
> >  {
> > -	swp_entry_t swp;
> > -	pte_t pte;
> > +	swp_entry_t entry, arch_entry;
> > +	pte_t pte, pte2;  
> A very small nit - s/pte2/pte as the first one is pmd not pte or
> make it pte1, pte2 if preferred.

Sure, pte1/2 looks better. Same for pmd1/2 in pmd_swap_tests().

> 
> >  
> >  	pr_debug("Validating PTE swap\n");
> > -	pte = pfn_pte(args->fixed_pte_pfn, args->page_prot);
> > -	swp = __pte_to_swp_entry(pte);
> > -	pte = __swp_entry_to_pte(swp);
> > -	WARN_ON(args->fixed_pte_pfn != pte_pfn(pte));
> > +	entry = args->swp_entry;  
> 
> Should args->swp_entry be used directly here and 'entry' local variable
> be dropped ?

Right, should be possible, also in pmd_swap_tests().

[...]
> > @@ -1166,6 +1173,7 @@ static void __init init_fixed_pfns(struct pgtable_debug_args *args)
> >  
> >  static int __init init_args(struct pgtable_debug_args *args)
> >  {
> > +	unsigned long max_swap_offset;
> >  	struct page *page = NULL;
> >  	int ret = 0;
> >  
> > @@ -1248,6 +1256,11 @@ static int __init init_args(struct pgtable_debug_args *args)
> >  
> >  	init_fixed_pfns(args);
> >  
> > +	/* See generic_max_swapfile_size(): probe the maximum offset */
> > +	max_swap_offset = swp_offset(pte_to_swp_entry(swp_entry_to_pte(swp_entry(0, ~0UL))));  
> Why not directly use generic_max_swapfile_size() which is doing exact same thing.
> 
> unsigned long generic_max_swapfile_size(void)
> {
> 	return swp_offset(pte_to_swp_entry(
> 			swp_entry_to_pte(swp_entry(0, ~0UL)))) + 1;
> }

Good question. I just moved this code here from pte_swap_exclusive_tests(),
see above, and did not think about that. Now I also wonder why
generic_max_swapfile_size() wasn't used before.

But it is not exactly the same thing, there is an extra "+ 1" there.
Maybe that is the reason, but I don't really understand the details /
difference, and therefore would not want to change it.

David, do you remember why you didn't use generic_max_swapfile_size()
in your pte_swap_exclusive_tests()?

> 
> > +	/* Create a swp entry with all possible bits set */
> > +	args->swp_entry = swp_entry((1 << MAX_SWAPFILES_SHIFT) - 1, max_swap_offset);
> > +  
> 
> Makes sense to use maximum possible bits while creating the swap entry for testing.
> 
> >  	/*
> >  	 * Allocate (huge) pages because some of the tests need to access
> >  	 * the data in the pages. The corresponding tests will be skipped  

Re: [RFC PATCH 1/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

[David Hildenbrand]

> [...]
>>> @@ -1166,6 +1173,7 @@ static void __init init_fixed_pfns(struct pgtable_debug_args *args)
>>>   
>>>   static int __init init_args(struct pgtable_debug_args *args)
>>>   {
>>> +	unsigned long max_swap_offset;
>>>   	struct page *page = NULL;
>>>   	int ret = 0;
>>>   
>>> @@ -1248,6 +1256,11 @@ static int __init init_args(struct pgtable_debug_args *args)
>>>   
>>>   	init_fixed_pfns(args);
>>>   
>>> +	/* See generic_max_swapfile_size(): probe the maximum offset */
>>> +	max_swap_offset = swp_offset(pte_to_swp_entry(swp_entry_to_pte(swp_entry(0, ~0UL))));
>> Why not directly use generic_max_swapfile_size() which is doing exact same thing.
>>
>> unsigned long generic_max_swapfile_size(void)
>> {
>> 	return swp_offset(pte_to_swp_entry(
>> 			swp_entry_to_pte(swp_entry(0, ~0UL)))) + 1;
>> }
> 
> Good question. I just moved this code here from pte_swap_exclusive_tests(),
> see above, and did not think about that. Now I also wonder why
> generic_max_swapfile_size() wasn't used before.
> 
> But it is not exactly the same thing, there is an extra "+ 1" there.
> Maybe that is the reason, but I don't really understand the details /
> difference, and therefore would not want to change it.
> 
> David, do you remember why you didn't use generic_max_swapfile_size()
> in your pte_swap_exclusive_tests()?

Excellent question. If only I would remember :)

generic_max_swapfile_size() resides in mm/swapfile.c, which is only 
around with CONFIG_SWAP.

It makes sense to have that function only if there are ... actual swapfiles.

These checks here are independent of CONFIG_SWAP (at least in theory -- 
for migration entries etc we don't need CONFIG_SWAP), and we simply want 
to construct a swap PTE with all possible bits set.

-- 
Cheers,

David / dhildenb

Re: [RFC PATCH 1/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

[Anshuman Khandual]

On 25/06/25 10:17 PM, David Hildenbrand wrote:
>> [...]
>>>> @@ -1166,6 +1173,7 @@ static void __init init_fixed_pfns(struct pgtable_debug_args *args)
>>>>     static int __init init_args(struct pgtable_debug_args *args)
>>>>   {
>>>> +    unsigned long max_swap_offset;
>>>>       struct page *page = NULL;
>>>>       int ret = 0;
>>>>   @@ -1248,6 +1256,11 @@ static int __init init_args(struct pgtable_debug_args *args)
>>>>         init_fixed_pfns(args);
>>>>   +    /* See generic_max_swapfile_size(): probe the maximum offset */
>>>> +    max_swap_offset = swp_offset(pte_to_swp_entry(swp_entry_to_pte(swp_entry(0, ~0UL))));
>>> Why not directly use generic_max_swapfile_size() which is doing exact same thing.
>>>
>>> unsigned long generic_max_swapfile_size(void)
>>> {
>>>     return swp_offset(pte_to_swp_entry(
>>>             swp_entry_to_pte(swp_entry(0, ~0UL)))) + 1;
>>> }
>>
>> Good question. I just moved this code here from pte_swap_exclusive_tests(),
>> see above, and did not think about that. Now I also wonder why
>> generic_max_swapfile_size() wasn't used before.
>>
>> But it is not exactly the same thing, there is an extra "+ 1" there.
>> Maybe that is the reason, but I don't really understand the details /
>> difference, and therefore would not want to change it.
>>
>> David, do you remember why you didn't use generic_max_swapfile_size()
>> in your pte_swap_exclusive_tests()?
> 
> Excellent question. If only I would remember :)
> 
> generic_max_swapfile_size() resides in mm/swapfile.c, which is only around with CONFIG_SWAP.
> 
> It makes sense to have that function only if there are ... actual swapfiles.
> 
> These checks here are independent of CONFIG_SWAP (at least in theory -- for migration entries etc we don't need CONFIG_SWAP), and we simply want to construct a swap PTE with all possible bits set.

After this modification of PMD based swap test - there will be now
two uses for generic_max_swapfile_size(). Rather than refactoring
these into a similar helper in mm/debug_vm_pgtable.c - should the
existing helper just be moved outside of CONFIG_SWAP, thus making
it available in general ?

Re: [RFC PATCH 1/1] mm/debug_vm_pgtable: Use a swp_entry_t input value for swap tests

[David Hildenbrand]

On 30.06.25 06:18, Anshuman Khandual wrote:
> 
> 
> On 25/06/25 10:17 PM, David Hildenbrand wrote:
>>> [...]
>>>>> @@ -1166,6 +1173,7 @@ static void __init init_fixed_pfns(struct pgtable_debug_args *args)
>>>>>      static int __init init_args(struct pgtable_debug_args *args)
>>>>>    {
>>>>> +    unsigned long max_swap_offset;
>>>>>        struct page *page = NULL;
>>>>>        int ret = 0;
>>>>>    @@ -1248,6 +1256,11 @@ static int __init init_args(struct pgtable_debug_args *args)
>>>>>          init_fixed_pfns(args);
>>>>>    +    /* See generic_max_swapfile_size(): probe the maximum offset */
>>>>> +    max_swap_offset = swp_offset(pte_to_swp_entry(swp_entry_to_pte(swp_entry(0, ~0UL))));
>>>> Why not directly use generic_max_swapfile_size() which is doing exact same thing.
>>>>
>>>> unsigned long generic_max_swapfile_size(void)
>>>> {
>>>>      return swp_offset(pte_to_swp_entry(
>>>>              swp_entry_to_pte(swp_entry(0, ~0UL)))) + 1;
>>>> }
>>>
>>> Good question. I just moved this code here from pte_swap_exclusive_tests(),
>>> see above, and did not think about that. Now I also wonder why
>>> generic_max_swapfile_size() wasn't used before.
>>>
>>> But it is not exactly the same thing, there is an extra "+ 1" there.
>>> Maybe that is the reason, but I don't really understand the details /
>>> difference, and therefore would not want to change it.
>>>
>>> David, do you remember why you didn't use generic_max_swapfile_size()
>>> in your pte_swap_exclusive_tests()?
>>
>> Excellent question. If only I would remember :)
>>
>> generic_max_swapfile_size() resides in mm/swapfile.c, which is only around with CONFIG_SWAP.
>>
>> It makes sense to have that function only if there are ... actual swapfiles.
>>
>> These checks here are independent of CONFIG_SWAP (at least in theory -- for migration entries etc we don't need CONFIG_SWAP), and we simply want to construct a swap PTE with all possible bits set.
> 
> After this modification of PMD based swap test - there will be now
> two uses for generic_max_swapfile_size(). Rather than refactoring
> these into a similar helper in mm/debug_vm_pgtable.c - should the
> existing helper just be moved outside of CONFIG_SWAP, thus making
> it available in general ?

As I said, having generic_max_swapfile_size() that returns something != 
0 with !CONFIG_SWAP feels rather odd.

No swap -> no swapfile -> no swapfile max size.

No strong opinion though,, but desperately trying to de-duplicate 2 LOC 
might not be worth the churn at all.

-- 
Cheers,

David / dhildenb