mm/gup: Provide callback to check if __GUP_fast() is allowed for the range
authorKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Thu, 16 Mar 2017 15:26:54 +0000 (18:26 +0300)
committerIngo Molnar <mingo@kernel.org>
Sat, 18 Mar 2017 08:48:03 +0000 (09:48 +0100)
This is a preparation patch for the transition of x86 to the generic GUP_fast()
implementation.

On x86, get_user_pages_fast() does a couple of sanity checks to see if we can
call __get_user_pages_fast() for the range.

This kind of wrapping protection should be useful for the generic code too.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-7-kirill.shutemov@linux.intel.com
[ Small readability edits. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
mm/gup.c

index e3d1e80424f495a944132aaf635e412b5a35db10..527ec2c6cca3b758eeb9a5e353b4075146fe8634 100644 (file)
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1614,6 +1614,21 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
        return nr;
 }
 
+#ifndef gup_fast_permitted
+/*
+ * Check if it's allowed to use __get_user_pages_fast() for the range, or
+ * we need to fall back to the slow version:
+ */
+bool gup_fast_permitted(unsigned long start, int nr_pages, int write)
+{
+       unsigned long len, end;
+
+       len = (unsigned long) nr_pages << PAGE_SHIFT;
+       end = start + len;
+       return end >= start;
+}
+#endif
+
 /**
  * get_user_pages_fast() - pin user pages in memory
  * @start:     starting user address
@@ -1633,11 +1648,14 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 int get_user_pages_fast(unsigned long start, int nr_pages, int write,
                        struct page **pages)
 {
-       int nr, ret;
+       int nr = 0, ret = 0;
 
        start &= PAGE_MASK;
-       nr = __get_user_pages_fast(start, nr_pages, write, pages);
-       ret = nr;
+
+       if (gup_fast_permitted(start, nr_pages, write)) {
+               nr = __get_user_pages_fast(start, nr_pages, write, pages);
+               ret = nr;
+       }
 
        if (nr < nr_pages) {
                /* Try to get the remaining pages with get_user_pages */