On Fri, Mar 30, 2018 at 08:37:45PM +0300, Michael S. Tsirkin wrote:
> get_user_pages_fast is supposed to be a faster drop-in equivalent of
> get_user_pages. As such, callers expect it to return a negative return
> code when passed an invalid address, and never expect it to
> return 0 when passed a positive number of pages, since
> its documentation says:
>
> * Returns number of pages pinned. This may be fewer than the number
> * requested. If nr_pages is 0 or negative, returns 0. If no pages
> * were pinned, returns -errno.
>
> Unfortunately this is not what the implementation does: it returns 0 if
> passed a kernel address, confusing callers: for example, the following
> is pretty common but does not appear to do the right thing with a kernel
> address:
>
> ret = get_user_pages_fast(addr, 1, writeable, &page);
> if (ret < 0)
> return ret;
>
> Change get_user_pages_fast to return -EFAULT when supplied a
> kernel address to make it match expectations.
>
> __get_user_pages_fast does not seem to be used like this, but let's
> change __get_user_pages_fast as well for consistency and to match
> documentation.
>
> Lightly tested.
>
> Cc: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Huang Ying <ying.huang@xxxxxxxxx>
> Cc: Jonathan Corbet <corbet@xxxxxxx>
> Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Thorsten Leemhuis <regressions@xxxxxxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> Fixes: 5b65c4677a57 ("mm, x86/mm: Fix performance regression in get_user_pages_fast()")
> Reported-by: syzbot+6304bf97ef436580fede@xxxxxxxxxxxxxxxxxxxxxxxxx
> Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx>
Any feedback on this? As this fixes a bug in vhost, I'll merge
through the vhost tree unless someone objects.
> ---
> mm/gup.c | 10 ++++++++--
> 1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/mm/gup.c b/mm/gup.c
> index 6afae32..5642521 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -1749,6 +1749,9 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
> unsigned long flags;
> int nr = 0;
>
> + if (nr_pages <= 0)
> + return 0;
> +
> start &= PAGE_MASK;
> addr = start;
> len = (unsigned long) nr_pages << PAGE_SHIFT;
> @@ -1756,7 +1759,7 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
>
> if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
> (void __user *)start, len)))
> - return 0;
> + return -EFAULT;
>
> /*
> * Disable interrupts. We use the nested form as we can already have
> @@ -1806,9 +1809,12 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
> len = (unsigned long) nr_pages << PAGE_SHIFT;
> end = start + len;
>
> + if (nr_pages <= 0)
> + return 0;
> +
> if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
> (void __user *)start, len)))
> - return 0;
> + return -EFAULT;
>
> if (gup_fast_permitted(start, nr_pages, write)) {
> local_irq_disable();
> --
> MST