linux-kernel - [PATCH mm-unstable v2] mm/truncate: batch-clear shadow entries

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Message-ID: <20240710060933.3979380-1-yuzhao@google.com>
Date: Wed, 10 Jul 2024 00:09:33 -0600
From: Yu Zhao <yuzhao@...gle.com>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: Bharata B Rao <bharata@....com>, "Matthew Wilcox (Oracle)" <willy@...radead.org>, linux-mm@...ck.org, 
	linux-kernel@...r.kernel.org, Yu Zhao <yuzhao@...gle.com>
Subject: [PATCH mm-unstable v2] mm/truncate: batch-clear shadow entries

Make clear_shadow_entry() clear shadow entries in `struct folio_batch`
so that it can reduce contention on i_lock and i_pages locks, e.g.,

  watchdog: BUG: soft lockup - CPU#29 stuck for 11s! [fio:2701649]
    clear_shadow_entry+0x3d/0x100
    mapping_try_invalidate+0x117/0x1d0
    invalidate_mapping_pages+0x10/0x20
    invalidate_bdev+0x3c/0x50
    blkdev_common_ioctl+0x5f7/0xa90
    blkdev_ioctl+0x109/0x270

Also, rename clear_shadow_entry() to clear_shadow_entries() accordingly.

Reported-by: Bharata B Rao <bharata@....com>
Closes: https://lore.kernel.org/d2841226-e27b-4d3d-a578-63587a3aa4f3@amd.com/
Tested-by: Bharata B Rao <bharata@....com>
Signed-off-by: Yu Zhao <yuzhao@...gle.com>
---
 mm/truncate.c | 68 +++++++++++++++++++++++----------------------------
 1 file changed, 31 insertions(+), 37 deletions(-)

diff --git a/mm/truncate.c b/mm/truncate.c
index 5ce62a939e55..dfb3d1f4d456 100644
--- a/mm/truncate.c
+++ b/mm/truncate.c
@@ -39,12 +39,25 @@ static inline void __clear_shadow_entry(struct address_space *mapping,
 	xas_store(&xas, NULL);
 }
 
-static void clear_shadow_entry(struct address_space *mapping, pgoff_t index,
-			       void *entry)
+static void clear_shadow_entries(struct address_space *mapping,
+				 struct folio_batch *fbatch, pgoff_t *indices)
 {
+	int i;
+
+	/* Handled by shmem itself, or for DAX we do nothing. */
+	if (shmem_mapping(mapping) || dax_mapping(mapping))
+		return;
+
 	spin_lock(&mapping->host->i_lock);
 	xa_lock_irq(&mapping->i_pages);
-	__clear_shadow_entry(mapping, index, entry);
+
+	for (i = 0; i < folio_batch_count(fbatch); i++) {
+		struct folio *folio = fbatch->folios[i];
+
+		if (xa_is_value(folio))
+			__clear_shadow_entry(mapping, indices[i], folio);
+	}
+
 	xa_unlock_irq(&mapping->i_pages);
 	if (mapping_shrinkable(mapping))
 		inode_add_lru(mapping->host);
@@ -105,36 +118,6 @@ static void truncate_folio_batch_exceptionals(struct address_space *mapping,
 	fbatch->nr = j;
 }
 
-/*
- * Invalidate exceptional entry if easily possible. This handles exceptional
- * entries for invalidate_inode_pages().
- */
-static int invalidate_exceptional_entry(struct address_space *mapping,
-					pgoff_t index, void *entry)
-{
-	/* Handled by shmem itself, or for DAX we do nothing. */
-	if (shmem_mapping(mapping) || dax_mapping(mapping))
-		return 1;
-	clear_shadow_entry(mapping, index, entry);
-	return 1;
-}
-
-/*
- * Invalidate exceptional entry if clean. This handles exceptional entries for
- * invalidate_inode_pages2() so for DAX it evicts only clean entries.
- */
-static int invalidate_exceptional_entry2(struct address_space *mapping,
-					 pgoff_t index, void *entry)
-{
-	/* Handled by shmem itself */
-	if (shmem_mapping(mapping))
-		return 1;
-	if (dax_mapping(mapping))
-		return dax_invalidate_mapping_entry_sync(mapping, index);
-	clear_shadow_entry(mapping, index, entry);
-	return 1;
-}
-
 /**
  * folio_invalidate - Invalidate part or all of a folio.
  * @folio: The folio which is affected.
@@ -494,6 +477,7 @@ unsigned long mapping_try_invalidate(struct address_space *mapping,
 	unsigned long ret;
 	unsigned long count = 0;
 	int i;
+	bool xa_has_values = false;
 
 	folio_batch_init(&fbatch);
 	while (find_lock_entries(mapping, &index, end, &fbatch, indices)) {
@@ -503,8 +487,8 @@ unsigned long mapping_try_invalidate(struct address_space *mapping,
 			/* We rely upon deletion not changing folio->index */
 
 			if (xa_is_value(folio)) {
-				count += invalidate_exceptional_entry(mapping,
-							     indices[i], folio);
+				xa_has_values = true;
+				count++;
 				continue;
 			}
 
@@ -522,6 +506,10 @@ unsigned long mapping_try_invalidate(struct address_space *mapping,
 			}
 			count += ret;
 		}
+
+		if (xa_has_values)
+			clear_shadow_entries(mapping, &fbatch, indices);
+
 		folio_batch_remove_exceptionals(&fbatch);
 		folio_batch_release(&fbatch);
 		cond_resched();
@@ -616,6 +604,7 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
 	int ret = 0;
 	int ret2 = 0;
 	int did_range_unmap = 0;
+	bool xa_has_values = false;
 
 	if (mapping_empty(mapping))
 		return 0;
@@ -629,8 +618,9 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
 			/* We rely upon deletion not changing folio->index */
 
 			if (xa_is_value(folio)) {
-				if (!invalidate_exceptional_entry2(mapping,
-						indices[i], folio))
+				xa_has_values = true;
+				if (dax_mapping(mapping) &&
+				    !dax_invalidate_mapping_entry_sync(mapping, indices[i]))
 					ret = -EBUSY;
 				continue;
 			}
@@ -666,6 +656,10 @@ int invalidate_inode_pages2_range(struct address_space *mapping,
 				ret = ret2;
 			folio_unlock(folio);
 		}
+
+		if (xa_has_values)
+			clear_shadow_entries(mapping, &fbatch, indices);
+
 		folio_batch_remove_exceptionals(&fbatch);
 		folio_batch_release(&fbatch);
 		cond_resched();
-- 
2.45.2.803.g4e1b14247a-goog