mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH 2/2] iommufd: only hold references to files memfd_pin_folios() can pin
  2026-11-19  2:44 [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper Jiahao Xu
@ 2026-11-19  2:44 ` Jiahao Xu
  2026-11-19  7:48 ` [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper David Hildenbrand (Arm)
  1 sibling, 0 replies; 3+ messages in thread
From: Jiahao Xu @ 2026-11-19  2:44 UTC (permalink / raw)
  To: Andrew Morton, Jason Gunthorpe
  Cc: linux-mm, David Hildenbrand, John Hubbard, Peter Xu,
	Vivek Kasireddy, Kevin Tian, Joerg Roedel (AMD),
	Will Deacon, Robin Murphy, iommu, linux-kernel,
	Samiullah Khawaja, syzbot+7666ed2c3d42196bf7e3, Jiahao Xu

iopt_map_file_pages() stores a reference to the file passed to
IOMMU_IOAS_MAP_FILE in the iopt_pages of the new mapping:

	iopt_map_file_pages()
	  iopt_alloc_file_pages()
	    pages->file = get_file(file)

iopt_release_pages() drops that reference only when the iopt_pages is
destroyed, which for a file mapping happens while the iommufd file is
being torn down.  Nothing restricts which file can be passed, so
userspace can map the iommufd file itself, or a vfio device fd bound to
the same iommufd, both of which own an iommufd_ctx reference.  The
reference graph then contains a cycle and the file, the iopt_pages, the
IOAS and the iommufd_ctx are all leaked; kmemleak reports it as
"memory leak in path_openat" with the holder taken through
iopt_map_file_pages().

Reverse the test instead of trying to enumerate the files that own an
iommufd_ctx: accept only the files that memfd_pin_folios() is able to
pin, which is exactly the set of files the mapping can be built from.
memfd_pin_folios_supported() reports that set, so the same test can run
here before the reference is handed to iopt_pages.  dma-bufs are
unaffected: they are taken by the dma_buf_get() branch above and never
reach the file branch, and iopt_pages holds the dma_buf there, not a
struct file.

Fixes: f4986a72d6e4 ("iommufd: Add IOMMU_IOAS_MAP_FILE")
Reported-by: syzbot+7666ed2c3d42196bf7e3@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=7666ed2c3d42196bf7e3
Signed-off-by: Jiahao Xu <xujiahao@cmss.chinamobile.com>
---
v3:
 - Call the new memfd_pin_folios_supported() helper instead of shmem_file().
   shmem_file() became an out-of-line function in mm/shmem.c that is not
   exported, so calling it from iommufd.ko failed at modpost with
   "symbol 'shmem_file' undefined!".  Reported by Jason Gunthorpe.
v2: https://lore.kernel.org/all/20261117093244.652569-1-xujiahao@cmss.chinamobile.com/
v1: https://lore.kernel.org/all/20261112105945.285526-1-xujiahao@cmss.chinamobile.com/
---
 drivers/iommu/iommufd/io_pagetable.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/iommu/iommufd/io_pagetable.c b/drivers/iommu/iommufd/io_pagetable.c
index 5ac736cf5..ab43e656d 100644
--- a/drivers/iommu/iommufd/io_pagetable.c
+++ b/drivers/iommu/iommufd/io_pagetable.c
@@ -515,6 +515,18 @@ int iopt_map_file_pages(struct iommufd_ctx *ictx, struct io_pagetable *iopt,
 		if (!file)
 			return -EBADF;
 
+		/*
+		 * iopt_pages keeps a reference to the file until the mapping is
+		 * destroyed, so only accept the files memfd_pin_folios() is able
+		 * to pin.  Any other fd, in particular one that itself owns an
+		 * iommufd_ctx (iommufd, vfio, ...), would put the file into a
+		 * reference cycle with this iommufd and never be released.
+		 */
+		if (!memfd_pin_folios_supported(file)) {
+			fput(file);
+			return -EINVAL;
+		}
+
 		pages = iopt_alloc_file_pages(file, start_byte, start, length,
 					      iommu_prot & IOMMU_WRITE);
 		fput(file);
-- 
2.43.0



^ permalink raw reply	[flat|nested] 3+ messages in thread

* [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper
@ 2026-11-19  2:44 Jiahao Xu
  2026-11-19  2:44 ` [PATCH 2/2] iommufd: only hold references to files memfd_pin_folios() can pin Jiahao Xu
  2026-11-19  7:48 ` [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper David Hildenbrand (Arm)
  0 siblings, 2 replies; 3+ messages in thread
From: Jiahao Xu @ 2026-11-19  2:44 UTC (permalink / raw)
  To: Andrew Morton, Jason Gunthorpe
  Cc: linux-mm, David Hildenbrand, John Hubbard, Peter Xu,
	Vivek Kasireddy, Kevin Tian, Joerg Roedel (AMD),
	Will Deacon, Robin Murphy, iommu, linux-kernel,
	Samiullah Khawaja, syzbot+7666ed2c3d42196bf7e3, Jiahao Xu

memfd_pin_folios() accepts only shmem and file-hugetlb files and enforces
that inline with shmem_file() and is_file_hugepages().  A caller that has
to know up front whether a file may be handed to memfd_pin_folios()
cannot use those predicates: since d3710111839d ("mm/shmem: make
shmem_file() check the file's inode") shmem_file() is an out-of-line
function in mm/shmem.c which is not exported, so modules cannot call it,
and is_file_hugepages() on its own is not the same question.

Provide the check as an exported helper next to memfd_pin_folios() and
use it inside memfd_pin_folios() itself, so that a single place decides
which files may be pinned.

Signed-off-by: Jiahao Xu <xujiahao@cmss.chinamobile.com>
---
 include/linux/mm.h |  1 +
 mm/gup.c           | 18 +++++++++++++++++-
 2 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 040cff643..e401edc93 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -3600,6 +3600,7 @@ long pin_user_pages_unlocked(unsigned long start, unsigned long nr_pages,
 long memfd_pin_folios(struct file *memfd, loff_t start, loff_t end,
 		      struct folio **folios, unsigned int max_folios,
 		      pgoff_t *offset);
+bool memfd_pin_folios_supported(struct file *file);
 int folio_add_pins(struct folio *folio, unsigned int pins);
 
 int get_user_pages_fast(unsigned long start, int nr_pages,
diff --git a/mm/gup.c b/mm/gup.c
index 7605da0e0..90ac51207 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -3449,6 +3449,22 @@ long pin_user_pages_unlocked(unsigned long start, unsigned long nr_pages,
 }
 EXPORT_SYMBOL(pin_user_pages_unlocked);
 
+/**
+ * memfd_pin_folios_supported() - check if memfd_pin_folios() can pin a file
+ * @file: the file to check
+ *
+ * memfd_pin_folios() only accepts files backed by shmem or by hugetlb, so
+ * callers that have to validate a file before pinning it, or before taking a
+ * reference to it, can use this helper instead of duplicating the check.
+ *
+ * Return: true if @file may be passed to memfd_pin_folios().
+ */
+bool memfd_pin_folios_supported(struct file *file)
+{
+	return shmem_file(file) || is_file_hugepages(file);
+}
+EXPORT_SYMBOL_GPL(memfd_pin_folios_supported);
+
 /**
  * memfd_pin_folios() - pin folios associated with a memfd
  * @memfd:      the memfd whose folios are to be pinned
@@ -3493,7 +3509,7 @@ long memfd_pin_folios(struct file *memfd, loff_t start, loff_t end,
 	if (!memfd)
 		return -EINVAL;
 
-	if (!shmem_file(memfd) && !is_file_hugepages(memfd))
+	if (!memfd_pin_folios_supported(memfd))
 		return -EINVAL;
 
 	if (end >= i_size_read(file_inode(memfd)))
-- 
2.43.0



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper
  2026-11-19  2:44 [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper Jiahao Xu
  2026-11-19  2:44 ` [PATCH 2/2] iommufd: only hold references to files memfd_pin_folios() can pin Jiahao Xu
@ 2026-11-19  7:48 ` David Hildenbrand (Arm)
  1 sibling, 0 replies; 3+ messages in thread
From: David Hildenbrand (Arm) @ 2026-11-19  7:48 UTC (permalink / raw)
  To: Jiahao Xu, Andrew Morton, Jason Gunthorpe
  Cc: linux-mm, John Hubbard, Peter Xu, Vivek Kasireddy, Kevin Tian,
	Joerg Roedel (AMD),
	Will Deacon, Robin Murphy, iommu, linux-kernel,
	Samiullah Khawaja, syzbot+7666ed2c3d42196bf7e3

On 11/19/26 03:44, Jiahao Xu wrote:
> memfd_pin_folios() accepts only shmem and file-hugetlb files and enforces
> that inline with shmem_file() and is_file_hugepages().  A caller that has
> to know up front whether a file may be handed to memfd_pin_folios()
> cannot use those predicates: since d3710111839d ("mm/shmem: make
> shmem_file() check the file's inode") shmem_file() is an out-of-line
> function in mm/shmem.c which is not exported, so modules cannot call it,
> and is_file_hugepages() on its own is not the same question.
> 
> Provide the check as an exported helper next to memfd_pin_folios() and
> use it inside memfd_pin_folios() itself, so that a single place decides
> which files may be pinned.
> 
> Signed-off-by: Jiahao Xu <xujiahao@cmss.chinamobile.com>
> ---
>  include/linux/mm.h |  1 +
>  mm/gup.c           | 18 +++++++++++++++++-
>  2 files changed, 18 insertions(+), 1 deletion(-)
> 
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 040cff643..e401edc93 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -3600,6 +3600,7 @@ long pin_user_pages_unlocked(unsigned long start, unsigned long nr_pages,
>  long memfd_pin_folios(struct file *memfd, loff_t start, loff_t end,
>  		      struct folio **folios, unsigned int max_folios,
>  		      pgoff_t *offset);
> +bool memfd_pin_folios_supported(struct file *file);
>  int folio_add_pins(struct folio *folio, unsigned int pins);
>  
>  int get_user_pages_fast(unsigned long start, int nr_pages,
> diff --git a/mm/gup.c b/mm/gup.c
> index 7605da0e0..90ac51207 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -3449,6 +3449,22 @@ long pin_user_pages_unlocked(unsigned long start, unsigned long nr_pages,
>  }
>  EXPORT_SYMBOL(pin_user_pages_unlocked);
>  
> +/**
> + * memfd_pin_folios_supported() - check if memfd_pin_folios() can pin a file
> + * @file: the file to check
> + *
> + * memfd_pin_folios() only accepts files backed by shmem or by hugetlb, so
> + * callers that have to validate a file before pinning it, or before taking a
> + * reference to it, can use this helper instead of duplicating the check.
> + *
> + * Return: true if @file may be passed to memfd_pin_folios().
> + */
> +bool memfd_pin_folios_supported(struct file *file)
> +{
> +	return shmem_file(file) || is_file_hugepages(file);

It's odd that you would be asking an interface specific for "memfd" whether some
random file would be supported. It's in the name: memfd.

I think the better question is: "is this a memfd". or "is this compatible with
memfd thingies" ?

file_is_memfd -- but we accept also other files, there are not strictly memfds

Hm.

file_is_memfd_compatible ?

Maybe there are better suggestions.

-- 
Cheers,

David

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-11-19  7:48 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-19  2:44 [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper Jiahao Xu
2026-11-19  2:44 ` [PATCH 2/2] iommufd: only hold references to files memfd_pin_folios() can pin Jiahao Xu
2026-11-19  7:48 ` [PATCH 1/2] mm/gup: add memfd_pin_folios_supported() helper David Hildenbrand (Arm)

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®