From: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Mon, 26 Aug 2019 22:14:22 +0200
Subject: mm/mmu_notifiers: prime lockdep
Patch-mainline: v5.4-rc1
Git-commit: 66204f1d2d1b42962033dfa867442f3dfd898d5f
References: jsc#SLE-15176, jsc#SLE-16387

We want to teach lockdep that mmu notifiers can be called from direct
reclaim paths, since on many CI systems load might never reach that
level (e.g. when just running fuzzer or small functional tests).

I've put the annotation into mmu_notifier_register since only when we have
mmu notifiers registered is there any point in teaching lockdep about
them. Also, we already have a kmalloc(, GFP_KERNEL), so this is safe.

Link: https://lore.kernel.org/r/20190826201425.17547-3-daniel.vetter@ffwll.ch
Suggested-by: Jason Gunthorpe <jgg@mellanox.com>
Reviewed-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Acked-by: Thomas Bogendoerfer <tbogendoerfer@suse.de>
---
 mm/mmu_notifier.c |    7 +++++++
 1 file changed, 7 insertions(+)

--- a/mm/mmu_notifier.c
+++ b/mm/mmu_notifier.c
@@ -243,6 +243,13 @@ int __mmu_notifier_register(struct mmu_n
 	lockdep_assert_held_write(&mm->mmap_sem);
 	BUG_ON(atomic_read(&mm->mm_users) <= 0);
 
+	if (IS_ENABLED(CONFIG_LOCKDEP)) {
+		fs_reclaim_acquire(GFP_KERNEL);
+		lock_map_acquire(&__mmu_notifier_invalidate_range_start_map);
+		lock_map_release(&__mmu_notifier_invalidate_range_start_map);
+		fs_reclaim_release(GFP_KERNEL);
+	}
+
 	mn->mm = mm;
 	mn->users = 1;