diff mbox series

[v4,10/25] mm/mm_init: Move p2pdma page refcount initialisation to p2pdma

Message ID aaa23e6f315a2d9b30a422c3769100cdfa42e85a.1734407924.git-series.apopple@nvidia.com
State New
Headers show
Series fs/dax: Fix ZONE_DEVICE page reference counts | expand

Commit Message

Alistair Popple Dec. 17, 2024, 5:12 a.m. UTC
Currently ZONE_DEVICE page reference counts are initialised by core
memory management code in __init_zone_device_page() as part of the
memremap() call which driver modules make to obtain ZONE_DEVICE
pages. This initialises page refcounts to 1 before returning them to
the driver.

This was presumably done because it drivers had a reference of sorts
on the page. It also ensured the page could always be mapped with
vm_insert_page() for example and would never get freed (ie. have a
zero refcount), freeing drivers of manipulating page reference counts.

However it complicates figuring out whether or not a page is free from
the mm perspective because it is no longer possible to just look at
the refcount. Instead the page type must be known and if GUP is used a
secondary pgmap reference is also sometimes needed.

To simplify this it is desirable to remove the page reference count
for the driver, so core mm can just use the refcount without having to
account for page type or do other types of tracking. This is possible
because drivers can always assume the page is valid as core kernel
will never offline or remove the struct page.

This means it is now up to drivers to initialise the page refcount as
required. P2PDMA uses vm_insert_page() to map the page, and that
requires a non-zero reference count when initialising the page so set
that when the page is first mapped.

Signed-off-by: Alistair Popple <apopple@nvidia.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>

---

Changes since v2:

 - Initialise the page refcount for all pages covered by the kaddr
---
 drivers/pci/p2pdma.c | 13 +++++++++++--
 mm/memremap.c        | 17 +++++++++++++----
 mm/mm_init.c         | 22 ++++++++++++++++++----
 3 files changed, 42 insertions(+), 10 deletions(-)

Comments

David Hildenbrand Dec. 17, 2024, 10:14 p.m. UTC | #1
On 17.12.24 06:12, Alistair Popple wrote:
> Currently ZONE_DEVICE page reference counts are initialised by core
> memory management code in __init_zone_device_page() as part of the
> memremap() call which driver modules make to obtain ZONE_DEVICE
> pages. This initialises page refcounts to 1 before returning them to
> the driver.
> 
> This was presumably done because it drivers had a reference of sorts
> on the page. It also ensured the page could always be mapped with
> vm_insert_page() for example and would never get freed (ie. have a
> zero refcount), freeing drivers of manipulating page reference counts.

It probably dates back to copying that code from other zone-init code 
where we
(a) Treat all available-at-boot memory as allocated before we release it 
to the buddy
(b) Treat all hotplugged memory as allocated until we release it to the 
buddy

As a side note, I'm working on converting (b) -- PageOffline pages -- to 
have a refcount of 0 ("frozen").

> 
> However it complicates figuring out whether or not a page is free from
> the mm perspective because it is no longer possible to just look at
> the refcount. Instead the page type must be known and if GUP is used a
> secondary pgmap reference is also sometimes needed.
> 
> To simplify this it is desirable to remove the page reference count
> for the driver, so core mm can just use the refcount without having to
> account for page type or do other types of tracking. This is possible
> because drivers can always assume the page is valid as core kernel
> will never offline or remove the struct page.
> 
> This means it is now up to drivers to initialise the page refcount as
> required. P2PDMA uses vm_insert_page() to map the page, and that
> requires a non-zero reference count when initialising the page so set
> that when the page is first mapped.
> 
> Signed-off-by: Alistair Popple <apopple@nvidia.com>
> Reviewed-by: Dan Williams <dan.j.williams@intel.com>
> 
> ---
> 
> Changes since v2:
> 
>   - Initialise the page refcount for all pages covered by the kaddr
> ---
>   drivers/pci/p2pdma.c | 13 +++++++++++--
>   mm/memremap.c        | 17 +++++++++++++----
>   mm/mm_init.c         | 22 ++++++++++++++++++----
>   3 files changed, 42 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/pci/p2pdma.c b/drivers/pci/p2pdma.c
> index 0cb7e0a..04773a8 100644
> --- a/drivers/pci/p2pdma.c
> +++ b/drivers/pci/p2pdma.c
> @@ -140,13 +140,22 @@ static int p2pmem_alloc_mmap(struct file *filp, struct kobject *kobj,
>   	rcu_read_unlock();
>   
>   	for (vaddr = vma->vm_start; vaddr < vma->vm_end; vaddr += PAGE_SIZE) {
> -		ret = vm_insert_page(vma, vaddr, virt_to_page(kaddr));
> +		struct page *page = virt_to_page(kaddr);
> +
> +		/*
> +		 * Initialise the refcount for the freshly allocated page. As
> +		 * we have just allocated the page no one else should be
> +		 * using it.
> +		 */
> +		VM_WARN_ON_ONCE_PAGE(!page_ref_count(page), page);
> +		set_page_count(page, 1);
> +		ret = vm_insert_page(vma, vaddr, page);
>   		if (ret) {
>   			gen_pool_free(p2pdma->pool, (uintptr_t)kaddr, len);
>   			return ret;
>   		}
>   		percpu_ref_get(ref);
> -		put_page(virt_to_page(kaddr));
> +		put_page(page);
>   		kaddr += PAGE_SIZE;
>   		len -= PAGE_SIZE;
>   	}
> diff --git a/mm/memremap.c b/mm/memremap.c
> index 40d4547..07bbe0e 100644
> --- a/mm/memremap.c
> +++ b/mm/memremap.c
> @@ -488,15 +488,24 @@ void free_zone_device_folio(struct folio *folio)
>   	folio->mapping = NULL;
>   	folio->page.pgmap->ops->page_free(folio_page(folio, 0));
>   
> -	if (folio->page.pgmap->type != MEMORY_DEVICE_PRIVATE &&
> -	    folio->page.pgmap->type != MEMORY_DEVICE_COHERENT)
> +	switch (folio->page.pgmap->type) {
> +	case MEMORY_DEVICE_PRIVATE:
> +	case MEMORY_DEVICE_COHERENT:
> +		put_dev_pagemap(folio->page.pgmap);
> +		break;
> +
> +	case MEMORY_DEVICE_FS_DAX:
> +	case MEMORY_DEVICE_GENERIC:
>   		/*
>   		 * Reset the refcount to 1 to prepare for handing out the page
>   		 * again.
>   		 */
>   		folio_set_count(folio, 1);
> -	else
> -		put_dev_pagemap(folio->page.pgmap);
> +		break;
> +
> +	case MEMORY_DEVICE_PCI_P2PDMA:
> +		break;
> +	}
>   }
>   
>   void zone_device_page_init(struct page *page)
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 24b68b4..f021e63 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -1017,12 +1017,26 @@ static void __ref __init_zone_device_page(struct page *page, unsigned long pfn,
>   	}
>   
>   	/*
> -	 * ZONE_DEVICE pages are released directly to the driver page allocator
> -	 * which will set the page count to 1 when allocating the page.
> +	 * ZONE_DEVICE pages other than MEMORY_TYPE_GENERIC and
> +	 * MEMORY_TYPE_FS_DAX pages are released directly to the driver page
> +	 * allocator which will set the page count to 1 when allocating the
> +	 * page.
> +	 *
> +	 * MEMORY_TYPE_GENERIC and MEMORY_TYPE_FS_DAX pages automatically have
> +	 * their refcount reset to one whenever they are freed (ie. after
> +	 * their refcount drops to 0).
>   	 */
> -	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
> -	    pgmap->type == MEMORY_DEVICE_COHERENT)
> +	switch (pgmap->type) {
> +	case MEMORY_DEVICE_PRIVATE:
> +	case MEMORY_DEVICE_COHERENT:
> +	case MEMORY_DEVICE_PCI_P2PDMA:
>   		set_page_count(page, 0);
> +		break;
> +
> +	case MEMORY_DEVICE_FS_DAX:
> +	case MEMORY_DEVICE_GENERIC:
> +		break;
> +	}
>   }
>   
>   /*


But that's a bit weird: we call __init_single_page()->init_page_count() 
to initialize it to 1, to then set it back to 0.


Maybe we can just pass to __init_single_page() the refcount we want to 
have directly? Can be a patch on top of course.

Apart from that

Acked-by: David Hildenbrand <david@redhat.com>
Alistair Popple Dec. 18, 2024, 10:49 p.m. UTC | #2
On Tue, Dec 17, 2024 at 11:14:42PM +0100, David Hildenbrand wrote:
> On 17.12.24 06:12, Alistair Popple wrote:
> > Currently ZONE_DEVICE page reference counts are initialised by core
> > memory management code in __init_zone_device_page() as part of the
> > memremap() call which driver modules make to obtain ZONE_DEVICE
> > pages. This initialises page refcounts to 1 before returning them to
> > the driver.
> > 
> > This was presumably done because it drivers had a reference of sorts
> > on the page. It also ensured the page could always be mapped with
> > vm_insert_page() for example and would never get freed (ie. have a
> > zero refcount), freeing drivers of manipulating page reference counts.
> 
> It probably dates back to copying that code from other zone-init code where
> we
> (a) Treat all available-at-boot memory as allocated before we release it to
> the buddy
> (b) Treat all hotplugged memory as allocated until we release it to the
> buddy
 
Argh, thanks for the background.

> As a side note, I'm working on converting (b) -- PageOffline pages -- to
> have a refcount of 0 ("frozen").

[...]

> > diff --git a/mm/mm_init.c b/mm/mm_init.c
> > index 24b68b4..f021e63 100644
> > --- a/mm/mm_init.c
> > +++ b/mm/mm_init.c
> > @@ -1017,12 +1017,26 @@ static void __ref __init_zone_device_page(struct page *page, unsigned long pfn,
> >   	}
> >   	/*
> > -	 * ZONE_DEVICE pages are released directly to the driver page allocator
> > -	 * which will set the page count to 1 when allocating the page.
> > +	 * ZONE_DEVICE pages other than MEMORY_TYPE_GENERIC and
> > +	 * MEMORY_TYPE_FS_DAX pages are released directly to the driver page
> > +	 * allocator which will set the page count to 1 when allocating the
> > +	 * page.
> > +	 *
> > +	 * MEMORY_TYPE_GENERIC and MEMORY_TYPE_FS_DAX pages automatically have
> > +	 * their refcount reset to one whenever they are freed (ie. after
> > +	 * their refcount drops to 0).
> >   	 */
> > -	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
> > -	    pgmap->type == MEMORY_DEVICE_COHERENT)
> > +	switch (pgmap->type) {
> > +	case MEMORY_DEVICE_PRIVATE:
> > +	case MEMORY_DEVICE_COHERENT:
> > +	case MEMORY_DEVICE_PCI_P2PDMA:
> >   		set_page_count(page, 0);
> > +		break;
> > +
> > +	case MEMORY_DEVICE_FS_DAX:
> > +	case MEMORY_DEVICE_GENERIC:
> > +		break;
> > +	}
> >   }
> >   /*
> 
> 
> But that's a bit weird: we call __init_single_page()->init_page_count() to
> initialize it to 1, to then set it back to 0.
> 
> 
> Maybe we can just pass to __init_single_page() the refcount we want to have
> directly? Can be a patch on top of course.

Once the dust settles on this series we won't need the pgmap->type check at
all because all ZONE_DEVICE pages will get an initial count of 0. I have some
follow up clean-ups for after this series is applied (particularly with regards
to pgmap refcounts), so if it's ok I'd rather do this as a follow-up.

> Apart from that
> 
> Acked-by: David Hildenbrand <david@redhat.com>
> 
> -- 
> Cheers,
> 
> David / dhildenb
>
David Hildenbrand Dec. 20, 2024, 6:29 p.m. UTC | #3
>>
>>
>> But that's a bit weird: we call __init_single_page()->init_page_count() to
>> initialize it to 1, to then set it back to 0.
>>
>>
>> Maybe we can just pass to __init_single_page() the refcount we want to have
>> directly? Can be a patch on top of course.
> 
> Once the dust settles on this series we won't need the pgmap->type check at
> all because all ZONE_DEVICE pages will get an initial count of 0. I have some
> follow up clean-ups for after this series is applied (particularly with regards
> to pgmap refcounts), so if it's ok I'd rather do this as a follow-up.

Sure. For ordinary memory hotplug I'll also convert it to start with 
refcount=0 soonish, so there we're also simply pass 0 to __init_single_page.
diff mbox series

Patch

diff --git a/drivers/pci/p2pdma.c b/drivers/pci/p2pdma.c
index 0cb7e0a..04773a8 100644
--- a/drivers/pci/p2pdma.c
+++ b/drivers/pci/p2pdma.c
@@ -140,13 +140,22 @@  static int p2pmem_alloc_mmap(struct file *filp, struct kobject *kobj,
 	rcu_read_unlock();
 
 	for (vaddr = vma->vm_start; vaddr < vma->vm_end; vaddr += PAGE_SIZE) {
-		ret = vm_insert_page(vma, vaddr, virt_to_page(kaddr));
+		struct page *page = virt_to_page(kaddr);
+
+		/*
+		 * Initialise the refcount for the freshly allocated page. As
+		 * we have just allocated the page no one else should be
+		 * using it.
+		 */
+		VM_WARN_ON_ONCE_PAGE(!page_ref_count(page), page);
+		set_page_count(page, 1);
+		ret = vm_insert_page(vma, vaddr, page);
 		if (ret) {
 			gen_pool_free(p2pdma->pool, (uintptr_t)kaddr, len);
 			return ret;
 		}
 		percpu_ref_get(ref);
-		put_page(virt_to_page(kaddr));
+		put_page(page);
 		kaddr += PAGE_SIZE;
 		len -= PAGE_SIZE;
 	}
diff --git a/mm/memremap.c b/mm/memremap.c
index 40d4547..07bbe0e 100644
--- a/mm/memremap.c
+++ b/mm/memremap.c
@@ -488,15 +488,24 @@  void free_zone_device_folio(struct folio *folio)
 	folio->mapping = NULL;
 	folio->page.pgmap->ops->page_free(folio_page(folio, 0));
 
-	if (folio->page.pgmap->type != MEMORY_DEVICE_PRIVATE &&
-	    folio->page.pgmap->type != MEMORY_DEVICE_COHERENT)
+	switch (folio->page.pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_COHERENT:
+		put_dev_pagemap(folio->page.pgmap);
+		break;
+
+	case MEMORY_DEVICE_FS_DAX:
+	case MEMORY_DEVICE_GENERIC:
 		/*
 		 * Reset the refcount to 1 to prepare for handing out the page
 		 * again.
 		 */
 		folio_set_count(folio, 1);
-	else
-		put_dev_pagemap(folio->page.pgmap);
+		break;
+
+	case MEMORY_DEVICE_PCI_P2PDMA:
+		break;
+	}
 }
 
 void zone_device_page_init(struct page *page)
diff --git a/mm/mm_init.c b/mm/mm_init.c
index 24b68b4..f021e63 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -1017,12 +1017,26 @@  static void __ref __init_zone_device_page(struct page *page, unsigned long pfn,
 	}
 
 	/*
-	 * ZONE_DEVICE pages are released directly to the driver page allocator
-	 * which will set the page count to 1 when allocating the page.
+	 * ZONE_DEVICE pages other than MEMORY_TYPE_GENERIC and
+	 * MEMORY_TYPE_FS_DAX pages are released directly to the driver page
+	 * allocator which will set the page count to 1 when allocating the
+	 * page.
+	 *
+	 * MEMORY_TYPE_GENERIC and MEMORY_TYPE_FS_DAX pages automatically have
+	 * their refcount reset to one whenever they are freed (ie. after
+	 * their refcount drops to 0).
 	 */
-	if (pgmap->type == MEMORY_DEVICE_PRIVATE ||
-	    pgmap->type == MEMORY_DEVICE_COHERENT)
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_COHERENT:
+	case MEMORY_DEVICE_PCI_P2PDMA:
 		set_page_count(page, 0);
+		break;
+
+	case MEMORY_DEVICE_FS_DAX:
+	case MEMORY_DEVICE_GENERIC:
+		break;
+	}
 }
 
 /*