mm: mremap: downgrade mmap_sem to read when shrinking

Other than munmap, mremap might be used to shrink memory mapping too.
So, it may hold write mmap_sem for long time when shrinking large
mapping, as what commit ("mm: mmap: zap pages with read mmap_sem in
munmap") described.

The mremap() will not manipulate vmas anymore after __do_munmap() call for
the mapping shrink use case, so it is safe to downgrade to read mmap_sem.

So, the same optimization, which downgrades mmap_sem to read for zapping
pages, is also feasible and reasonable to this case.

The period of holding exclusive mmap_sem for shrinking large mapping
would be reduced significantly with this optimization.

MREMAP_FIXED and MREMAP_MAYMOVE are more complicated to adopt this
optimization since they need manipulate vmas after do_munmap(),
downgrading mmap_sem may create race window.

Simple mapping shrink is the low hanging fruit, and it may cover the
most cases of unmap with munmap together.

[[email protected]: tweak comment]
[[email protected]: fix unsigned compare against 0 issue]
  Link: http://lkml.kernel.org/r/[email protected]
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Yang Shi <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Acked-by: Kirill A. Shutemov <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Matthew Wilcox <[email protected]>
Cc: Laurent Dufour <[email protected]>
Cc: Colin Ian King <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

include/linux/mm.h

@@ -2306,6 +2306,8 @@ extern unsigned long do_mmap(struct file *file, unsigned long addr,
 	unsigned long len, unsigned long prot, unsigned long flags,
 	vm_flags_t vm_flags, unsigned long pgoff, unsigned long *populate,
 	struct list_head *uf);
+extern int __do_munmap(struct mm_struct *, unsigned long, size_t,
+		       struct list_head *uf, bool downgrade);
 extern int do_munmap(struct mm_struct *, unsigned long, size_t,
 		     struct list_head *uf);
 

mm/mmap.c

@@ -2687,8 +2687,8 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
  * work.  This now handles partial unmappings.
  * Jeremy Fitzhardinge <[email protected]>
  */
-static int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len,
-		       struct list_head *uf, bool downgrade)
+int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len,
+		struct list_head *uf, bool downgrade)
 {
 	unsigned long end;
 	struct vm_area_struct *vma, *prev, *last;

mm/mremap.c

@@ -521,6 +521,7 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len,
 	unsigned long ret = -EINVAL;
 	unsigned long charged = 0;
 	bool locked = false;
+	bool downgraded = false;
 	struct vm_userfaultfd_ctx uf = NULL_VM_UFFD_CTX;
 	LIST_HEAD(uf_unmap_early);
 	LIST_HEAD(uf_unmap);
@@ -557,12 +558,20 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len,
 	/*
 	 * Always allow a shrinking remap: that just unmaps
 	 * the unnecessary pages..
-	 * do_munmap does all the needed commit accounting
+	 * __do_munmap does all the needed commit accounting, and
+	 * downgrades mmap_sem to read if so directed.
 	 */
 	if (old_len >= new_len) {
-		ret = do_munmap(mm, addr+new_len, old_len - new_len, &uf_unmap);
-		if (ret && old_len != new_len)
+		int retval;
+
+		retval = __do_munmap(mm, addr+new_len, old_len - new_len,
+				  &uf_unmap, true);
+		if (retval < 0 && old_len != new_len) {
+			ret = retval;
 			goto out;
+		/* Returning 1 indicates mmap_sem is downgraded to read. */
+		} else if (retval == 1)
+			downgraded = true;
 		ret = addr;
 		goto out;
 	}
@@ -627,7 +636,10 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len,
 		vm_unacct_memory(charged);
 		locked = 0;
 	}
-	up_write(&current->mm->mmap_sem);
+	if (downgraded)
+		up_read(&current->mm->mmap_sem);
+	else
+		up_write(&current->mm->mmap_sem);
 	if (locked && new_len > old_len)
 		mm_populate(new_addr + old_len, new_len - old_len);
 	userfaultfd_unmap_complete(mm, &uf_unmap_early);