Re: [RFC PATCH v2 02/11] KVM: guest_memfd: Handle final folio_put() of guest_memfd pages

[David Hildenbrand <david@redhat.com>]

On 29.01.25 18:23, Fuad Tabba wrote:
> Before transitioning a guest_memfd folio to unshared, thereby
> disallowing access by the host and allowing the hypervisor to
> transition its view of the guest page as private, we need to be
> sure that the host doesn't have any references to the folio.
> 
> This patch introduces a new type for guest_memfd folios, which
> isn't activated in this series but is here as a placeholder and
> to facilitate the code in the next patch. This will be used in
> the future to register a callback that informs the guest_memfd
> subsystem when the last reference is dropped, therefore knowing
> that the host doesn't have any remaining references.
> 
> Signed-off-by: Fuad Tabba <tabba@google.com>
> ---
>   include/linux/page-flags.h | 7 +++++++
>   mm/debug.c                 | 1 +
>   mm/swap.c                  | 5 +++++
>   3 files changed, 13 insertions(+)
> 
> diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
> index 6615f2f59144..bab3cac1f93b 100644
> --- a/include/linux/page-flags.h
> +++ b/include/linux/page-flags.h
> @@ -942,6 +942,7 @@ enum pagetype {
>   	PGTY_slab	= 0xf5,
>   	PGTY_zsmalloc	= 0xf6,
>   	PGTY_unaccepted	= 0xf7,
> +	PGTY_guestmem	= 0xf8,
>   
>   	PGTY_mapcount_underflow = 0xff
>   };
> @@ -1091,6 +1092,12 @@ FOLIO_TYPE_OPS(hugetlb, hugetlb)
>   FOLIO_TEST_FLAG_FALSE(hugetlb)
>   #endif
>   

Some short doc would be nice, to at least hint that this is related to 
guest_memfd, and that these are otherwise folios.


/*
  * guestmem folios are folios that are used to back VM memory as managed
  * guest_memfd. Once the last reference is put, instead of freeing these
  * folios back to the page allocator, they are returned to guest_memfd.
  *
  * For now, guestmem will only be set on these folios as long as they
  * cannot be mapped to user space ("private state"), with the plan of
  * always setting that type once typed folios can be mapped to user
  * space cleanly.
  */

> +#ifdef CONFIG_KVM_GMEM_MAPPABLE
> +FOLIO_TYPE_OPS(guestmem, guestmem)
> +#else
> +FOLIO_TEST_FLAG_FALSE(guestmem)
> +#endif
> +
>   PAGE_TYPE_OPS(Zsmalloc, zsmalloc, zsmalloc)
>   
>   /*
> diff --git a/mm/debug.c b/mm/debug.c
> index 95b6ab809c0e..db93be385ed9 100644
> --- a/mm/debug.c
> +++ b/mm/debug.c
> @@ -56,6 +56,7 @@ static const char *page_type_names[] = {
>   	DEF_PAGETYPE_NAME(table),
>   	DEF_PAGETYPE_NAME(buddy),
>   	DEF_PAGETYPE_NAME(unaccepted),
> +	DEF_PAGETYPE_NAME(guestmem),
 >   };>
>   static const char *page_type_name(unsigned int page_type)
> diff --git a/mm/swap.c b/mm/swap.c
> index 8a66cd9cb9da..73d61c7f8edd 100644
> --- a/mm/swap.c
> +++ b/mm/swap.c
> @@ -37,6 +37,7 @@
>   #include <linux/page_idle.h>
>   #include <linux/local_lock.h>
>   #include <linux/buffer_head.h>
> +#include <linux/kvm_host.h>
>   
>   #include "internal.h"
>   
> @@ -101,6 +102,10 @@ static void free_typed_folio(struct folio *folio)
>   		if (IS_ENABLED(CONFIG_HUGETLBFS))
>   			free_huge_folio(folio);
>   		return;
> +	case PGTY_guestmem:
> +		if (IS_ENABLED(CONFIG_KVM_GMEM_SHARED_MEM))
> +			WARN_ONCE(1, "A placeholder that shouldn't trigger.");


Does it make sense to directly introduce the callback into guest_memfd 
and handle the WARN_ONCE() in there? Then, we don't have tot ouch this 
core code later again.

-- 
Cheers,

David / dhildenb