This code was using get_user_pages*(), in a "Case 1" scenario
(Direct IO), using the categorization from [1]. That means that it's
time to convert the get_user_pages*() + put_page() calls to
pin_user_pages*() + unpin_user_pages() calls.
There is some helpful background in [2]: basically, this is a small
part of fixing a long-standing disconnect between pinning pages, and
file systems' use of those pages.
[1] Documentation/core-api/pin_user_pages.rst
[2] "Explicit pinning of user-space pages":
    https://lwn.net/Articles/807108/
Cc: Mike Marshall <hubcap@omnibond.com>
Cc: Martin Brandenburg <martin@omnibond.com>
Cc: devel@lists.orangefs.org
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
 static void
 orangefs_bufmap_unmap(struct orangefs_bufmap *bufmap)
 {
-       int i;
-
-       for (i = 0; i < bufmap->page_count; i++)
-               put_page(bufmap->page_array[i]);
+       unpin_user_pages(bufmap->page_array, bufmap->page_count);
 }
 
 static void
        int offset = 0, ret, i;
 
        /* map the pages */
-       ret = get_user_pages_fast((unsigned long)user_desc->ptr,
+       ret = pin_user_pages_fast((unsigned long)user_desc->ptr,
                             bufmap->page_count, FOLL_WRITE, bufmap->page_array);
 
        if (ret < 0)
 
                for (i = 0; i < ret; i++) {
                        SetPageError(bufmap->page_array[i]);
-                       put_page(bufmap->page_array[i]);
+                       unpin_user_page(bufmap->page_array[i]);
                }
                return -ENOMEM;
        }