[v2,1/4] mm: kmsan: handle alloc failures in kmsan_vmap_pages_range_noflush()

  As reported by Dipanjan Das, when KMSAN is used together with kernel
fault injection (or, generally, even without the latter), calls to
kcalloc() or __vmap_pages_range_noflush() may fail, leaving the
metadata mappings for the virtual mapping in an inconsistent state.
When these metadata mappings are accessed later, the kernel crashes.

To address the problem, we return a non-zero error code from
kmsan_vmap_pages_range_noflush() in the case of any allocation/mapping
failure inside it, and make vmap_pages_range_noflush() return an error
if KMSAN fails to allocate the metadata.

This patch also removes KMSAN_WARN_ON() from vmap_pages_range_noflush(),
as these allocation failures are not fatal anymore.

Reported-by: Dipanjan Das <mail.dipanjan.das@gmail.com>
Link: https://lore.kernel.org/linux-mm/CANX2M5ZRrRA64k0hOif02TjmY9kbbO2aCBPyq79es34RXZ=cAw@mail.gmail.com/
Fixes: b073d7f8aee4 ("mm: kmsan: maintain KMSAN metadata for page operations")
Signed-off-by: Alexander Potapenko <glider@google.com>

---
v2:
 -- return 0 from the inline version of kmsan_vmap_pages_range_noflush()
    (spotted by kernel test robot <lkp@intel.com>)
---
 include/linux/kmsan.h | 20 +++++++++++---------
 mm/kmsan/shadow.c     | 27 ++++++++++++++++++---------
 mm/vmalloc.c          |  6 +++++-
 3 files changed, 34 insertions(+), 19 deletions(-)
  

On Thu, 13 Apr 2023 at 15:12, Alexander Potapenko <glider@google.com> wrote:
>
> As reported by Dipanjan Das, when KMSAN is used together with kernel
> fault injection (or, generally, even without the latter), calls to
> kcalloc() or __vmap_pages_range_noflush() may fail, leaving the
> metadata mappings for the virtual mapping in an inconsistent state.
> When these metadata mappings are accessed later, the kernel crashes.
>
> To address the problem, we return a non-zero error code from
> kmsan_vmap_pages_range_noflush() in the case of any allocation/mapping
> failure inside it, and make vmap_pages_range_noflush() return an error
> if KMSAN fails to allocate the metadata.
>
> This patch also removes KMSAN_WARN_ON() from vmap_pages_range_noflush(),
> as these allocation failures are not fatal anymore.
>
> Reported-by: Dipanjan Das <mail.dipanjan.das@gmail.com>
> Link: https://lore.kernel.org/linux-mm/CANX2M5ZRrRA64k0hOif02TjmY9kbbO2aCBPyq79es34RXZ=cAw@mail.gmail.com/
> Fixes: b073d7f8aee4 ("mm: kmsan: maintain KMSAN metadata for page operations")
> Signed-off-by: Alexander Potapenko <glider@google.com>

Reviewed-by: Marco Elver <elver@google.com>

Looks reasonable, thanks.

> ---
> v2:
>  -- return 0 from the inline version of kmsan_vmap_pages_range_noflush()
>     (spotted by kernel test robot <lkp@intel.com>)
> ---
>  include/linux/kmsan.h | 20 +++++++++++---------
>  mm/kmsan/shadow.c     | 27 ++++++++++++++++++---------
>  mm/vmalloc.c          |  6 +++++-
>  3 files changed, 34 insertions(+), 19 deletions(-)
>
> diff --git a/include/linux/kmsan.h b/include/linux/kmsan.h
> index e38ae3c346184..c7ff3aefc5a13 100644
> --- a/include/linux/kmsan.h
> +++ b/include/linux/kmsan.h
> @@ -134,11 +134,12 @@ void kmsan_kfree_large(const void *ptr);
>   * @page_shift:        page_shift passed to vmap_range_noflush().
>   *
>   * KMSAN maps shadow and origin pages of @pages into contiguous ranges in
> - * vmalloc metadata address range.
> + * vmalloc metadata address range. Returns 0 on success, callers must check
> + * for non-zero return value.
>   */
> -void kmsan_vmap_pages_range_noflush(unsigned long start, unsigned long end,
> -                                   pgprot_t prot, struct page **pages,
> -                                   unsigned int page_shift);
> +int kmsan_vmap_pages_range_noflush(unsigned long start, unsigned long end,
> +                                  pgprot_t prot, struct page **pages,
> +                                  unsigned int page_shift);
>
>  /**
>   * kmsan_vunmap_kernel_range_noflush() - Notify KMSAN about a vunmap.
> @@ -281,12 +282,13 @@ static inline void kmsan_kfree_large(const void *ptr)
>  {
>  }
>
> -static inline void kmsan_vmap_pages_range_noflush(unsigned long start,
> -                                                 unsigned long end,
> -                                                 pgprot_t prot,
> -                                                 struct page **pages,
> -                                                 unsigned int page_shift)
> +static inline int kmsan_vmap_pages_range_noflush(unsigned long start,
> +                                                unsigned long end,
> +                                                pgprot_t prot,
> +                                                struct page **pages,
> +                                                unsigned int page_shift)
>  {
> +       return 0;
>  }
>
>  static inline void kmsan_vunmap_range_noflush(unsigned long start,
> diff --git a/mm/kmsan/shadow.c b/mm/kmsan/shadow.c
> index a787c04e9583c..b8bb95eea5e3d 100644
> --- a/mm/kmsan/shadow.c
> +++ b/mm/kmsan/shadow.c
> @@ -216,27 +216,29 @@ void kmsan_free_page(struct page *page, unsigned int order)
>         kmsan_leave_runtime();
>  }
>
> -void kmsan_vmap_pages_range_noflush(unsigned long start, unsigned long end,
> -                                   pgprot_t prot, struct page **pages,
> -                                   unsigned int page_shift)
> +int kmsan_vmap_pages_range_noflush(unsigned long start, unsigned long end,
> +                                  pgprot_t prot, struct page **pages,
> +                                  unsigned int page_shift)
>  {
>         unsigned long shadow_start, origin_start, shadow_end, origin_end;
>         struct page **s_pages, **o_pages;
> -       int nr, mapped;
> +       int nr, mapped, err = 0;
>
>         if (!kmsan_enabled)
> -               return;
> +               return 0;
>
>         shadow_start = vmalloc_meta((void *)start, KMSAN_META_SHADOW);
>         shadow_end = vmalloc_meta((void *)end, KMSAN_META_SHADOW);
>         if (!shadow_start)
> -               return;
> +               return 0;
>
>         nr = (end - start) / PAGE_SIZE;
>         s_pages = kcalloc(nr, sizeof(*s_pages), GFP_KERNEL);
>         o_pages = kcalloc(nr, sizeof(*o_pages), GFP_KERNEL);
> -       if (!s_pages || !o_pages)
> +       if (!s_pages || !o_pages) {
> +               err = -ENOMEM;
>                 goto ret;
> +       }
>         for (int i = 0; i < nr; i++) {
>                 s_pages[i] = shadow_page_for(pages[i]);
>                 o_pages[i] = origin_page_for(pages[i]);
> @@ -249,10 +251,16 @@ void kmsan_vmap_pages_range_noflush(unsigned long start, unsigned long end,
>         kmsan_enter_runtime();
>         mapped = __vmap_pages_range_noflush(shadow_start, shadow_end, prot,
>                                             s_pages, page_shift);
> -       KMSAN_WARN_ON(mapped);
> +       if (mapped) {
> +               err = mapped;
> +               goto ret;
> +       }
>         mapped = __vmap_pages_range_noflush(origin_start, origin_end, prot,
>                                             o_pages, page_shift);
> -       KMSAN_WARN_ON(mapped);
> +       if (mapped) {
> +               err = mapped;
> +               goto ret;
> +       }
>         kmsan_leave_runtime();
>         flush_tlb_kernel_range(shadow_start, shadow_end);
>         flush_tlb_kernel_range(origin_start, origin_end);
> @@ -262,6 +270,7 @@ void kmsan_vmap_pages_range_noflush(unsigned long start, unsigned long end,
>  ret:
>         kfree(s_pages);
>         kfree(o_pages);
> +       return err;
>  }
>
>  /* Allocate metadata for pages allocated at boot time. */
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index a50072066221a..1355d95cce1ca 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -605,7 +605,11 @@ int __vmap_pages_range_noflush(unsigned long addr, unsigned long end,
>  int vmap_pages_range_noflush(unsigned long addr, unsigned long end,
>                 pgprot_t prot, struct page **pages, unsigned int page_shift)
>  {
> -       kmsan_vmap_pages_range_noflush(addr, end, prot, pages, page_shift);
> +       int ret = kmsan_vmap_pages_range_noflush(addr, end, prot, pages,
> +                                                page_shift);
> +
> +       if (ret)
> +               return ret;
>         return __vmap_pages_range_noflush(addr, end, prot, pages, page_shift);
>  }
>
> --
> 2.40.0.577.gac1e443424-goog
>