LinuxLists.cc - [PATCH v1 0/1] add mseal to /proc/pid/smaps

2024-06-14 18:24:29

Subject: [PATCH v1 0/1] add mseal to /proc/pid/smaps

From: Jeff Xu <[email protected]>

Add mseal information in /proc/pid/smaps to indicate the VMA is sealed.

I appreicate Adhemerval Zanella Netto to bring this to my attention. [1]

[1] https://public-inbox.org/libc-alpha/[email protected]/#t

Jeff Xu (1):
/proc/pid/smaps: add mseal info for vma

Documentation/filesystems/proc.rst | 1 +
fs/proc/task_mmu.c | 3 +++
include/linux/mm.h | 5 +++++
mm/internal.h | 5 -----
4 files changed, 9 insertions(+), 5 deletions(-)

--
2.45.2.627.g7a2c4fd464-goog

2024-06-14 18:24:35

by Jeff Xu

[permalink] [raw]

Subject: [PATCH v1 1/1] /proc/pid/smaps: add mseal info for vma

From: Jeff Xu <[email protected]>

Add sp in /proc/pid/smaps to indicate vma is sealed

Signed-off-by: Jeff Xu <[email protected]>
---
Documentation/filesystems/proc.rst | 1 +
fs/proc/task_mmu.c | 3 +++
include/linux/mm.h | 5 +++++
mm/internal.h | 5 -----
4 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst
index 7c3a565ffbef..400217a1589c 100644
--- a/Documentation/filesystems/proc.rst
+++ b/Documentation/filesystems/proc.rst
@@ -571,6 +571,7 @@ encoded manner. The codes are the following:
um userfaultfd missing tracking
uw userfaultfd wr-protect tracking
ss shadow stack page
+ sp sealed page
== =======================================

Note that there is no guarantee that every flag and associated mnemonic will
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 6ed1f56b32b4..ba2db75ca22a 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -711,6 +711,9 @@ static void show_smap_vma_flags(struct seq_file *m, struct vm_area_struct *vma)
#endif /* CONFIG_HAVE_ARCH_USERFAULTFD_MINOR */
#ifdef CONFIG_X86_USER_SHADOW_STACK
[ilog2(VM_SHADOW_STACK)] = "ss",
+#endif
+#ifdef CONFIG_64BIT
+ [ilog2(VM_SEALED)] = "sp",
#endif
};
size_t i;
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 587d34879865..8600564898fa 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -314,6 +314,11 @@ extern unsigned int kobjsize(const void *objp);
#define VM_NOHUGEPAGE 0x40000000 /* MADV_NOHUGEPAGE marked this vma */
#define VM_MERGEABLE 0x80000000 /* KSM may merge identical pages */

+#ifdef CONFIG_64BIT
+/* VM is sealed, in vm_flags */
+#define VM_SEALED _BITUL(63)
+#endif
+
#ifdef CONFIG_ARCH_USES_HIGH_VMA_FLAGS
#define VM_HIGH_ARCH_BIT_0 32 /* bit only usable on 64-bit architectures */
#define VM_HIGH_ARCH_BIT_1 33 /* bit only usable on 64-bit architectures */
diff --git a/mm/internal.h b/mm/internal.h
index fd68c43664d5..72f7c110d563 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -1509,11 +1509,6 @@ void __meminit __init_single_page(struct page *page, unsigned long pfn,
unsigned long shrink_slab(gfp_t gfp_mask, int nid, struct mem_cgroup *memcg,
int priority);

-#ifdef CONFIG_64BIT
-/* VM is sealed, in vm_flags */
-#define VM_SEALED _BITUL(63)
-#endif
-
#ifdef CONFIG_64BIT
static inline int can_do_mseal(unsigned long flags)
{
--
2.45.2.627.g7a2c4fd464-goog

2024-06-14 19:01:14

by Jeff Xu

[permalink] [raw]

Subject: Re: [PATCH v1 1/1] /proc/pid/smaps: add mseal info for vma

On Fri, Jun 14, 2024 at 11:43 AM Jann Horn <[email protected]> wrote:
>
> Hi!
>
> On Fri, Jun 14, 2024 at 8:24 PM <[email protected]> wrote:
> > Add sp in /proc/pid/smaps to indicate vma is sealed
> >
> > Signed-off-by: Jeff Xu <[email protected]>
> > ---
> > Documentation/filesystems/proc.rst | 1 +
> > fs/proc/task_mmu.c | 3 +++
> > include/linux/mm.h | 5 +++++
> > mm/internal.h | 5 -----
> > 4 files changed, 9 insertions(+), 5 deletions(-)
> >
> > diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst
> > index 7c3a565ffbef..400217a1589c 100644
> > --- a/Documentation/filesystems/proc.rst
> > +++ b/Documentation/filesystems/proc.rst
> > @@ -571,6 +571,7 @@ encoded manner. The codes are the following:
> > um userfaultfd missing tracking
> > uw userfaultfd wr-protect tracking
> > ss shadow stack page
> > + sp sealed page
>
> Nit: Why "page"? The sealing is a property of the VMA, not of the
> pages mapped into it. Maybe "sealed area" and an abbreviation like
> "sl" would make sense?
>
> > diff --git a/include/linux/mm.h b/include/linux/mm.h
> > index 587d34879865..8600564898fa 100644
> > --- a/include/linux/mm.h
> > +++ b/include/linux/mm.h
> > @@ -314,6 +314,11 @@ extern unsigned int kobjsize(const void *objp);
> > #define VM_NOHUGEPAGE 0x40000000 /* MADV_NOHUGEPAGE marked this vma */
> > #define VM_MERGEABLE 0x80000000 /* KSM may merge identical pages */
> >
> > +#ifdef CONFIG_64BIT
> > +/* VM is sealed, in vm_flags */
> > +#define VM_SEALED _BITUL(63)
> > +#endif
> > +
> > #ifdef CONFIG_ARCH_USES_HIGH_VMA_FLAGS
> > #define VM_HIGH_ARCH_BIT_0 32 /* bit only usable on 64-bit architectures */
> > #define VM_HIGH_ARCH_BIT_1 33 /* bit only usable on 64-bit architectures */
>
> Other 64-bit flags are defined further down; maybe it would make sense
> to move this definition below the definition of VM_ALLOW_ANY_UNCACHED,
> so that the definitions are sorted by the number of the bit?

Sure. I will update.

Thanks!
-Jeff

2024-06-14 19:14:27

by Jann Horn

[permalink] [raw]

Subject: Re: [PATCH v1 1/1] /proc/pid/smaps: add mseal info for vma

Hi!

On Fri, Jun 14, 2024 at 8:24 PM <[email protected]> wrote:
> Add sp in /proc/pid/smaps to indicate vma is sealed
>
> Signed-off-by: Jeff Xu <[email protected]>
> ---
> Documentation/filesystems/proc.rst | 1 +
> fs/proc/task_mmu.c | 3 +++
> include/linux/mm.h | 5 +++++
> mm/internal.h | 5 -----
> 4 files changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst
> index 7c3a565ffbef..400217a1589c 100644
> --- a/Documentation/filesystems/proc.rst
> +++ b/Documentation/filesystems/proc.rst
> @@ -571,6 +571,7 @@ encoded manner. The codes are the following:
> um userfaultfd missing tracking
> uw userfaultfd wr-protect tracking
> ss shadow stack page
> + sp sealed page

Nit: Why "page"? The sealing is a property of the VMA, not of the
pages mapped into it. Maybe "sealed area" and an abbreviation like
"sl" would make sense?

> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 587d34879865..8600564898fa 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -314,6 +314,11 @@ extern unsigned int kobjsize(const void *objp);
> #define VM_NOHUGEPAGE 0x40000000 /* MADV_NOHUGEPAGE marked this vma */
> #define VM_MERGEABLE 0x80000000 /* KSM may merge identical pages */
>
> +#ifdef CONFIG_64BIT
> +/* VM is sealed, in vm_flags */
> +#define VM_SEALED _BITUL(63)
> +#endif
> +
> #ifdef CONFIG_ARCH_USES_HIGH_VMA_FLAGS
> #define VM_HIGH_ARCH_BIT_0 32 /* bit only usable on 64-bit architectures */
> #define VM_HIGH_ARCH_BIT_1 33 /* bit only usable on 64-bit architectures */

Other 64-bit flags are defined further down; maybe it would make sense
to move this definition below the definition of VM_ALLOW_ANY_UNCACHED,
so that the definitions are sorted by the number of the bit?

2024-06-14 19:29:33

by Jeff Xu

[permalink] [raw]