Skip to contentSkip to conversation

ext4: don't BUG if someone dirty pages without asking ext4 first

PATCH New Sasha Levin 2 participants 3 messages Mar 28, 2022 Stable branch

lore
Share & tools
lore
syzbot
#1 lore raw
In reply to Sasha Levin
I see the command but can't find the corresponding bug. The email is sent to [email protected] address but the HASH does not correspond to any known bug. Please double check the…
Quoted 79 lines

From: Theodore Ts'o <[email protected]>

[ Upstream commit cc5095747edfb054ca2068d01af20be3fcc3634f ]

[un]pin_user_pages_remote is dirtying pages without properly warning the file system in advance. A related race was noted by Jan Kara in 2018[1]; however, more recently instead of it being a very hard-to-hit race, it could be reliably triggered by process_vm_writev(2) which was discovered by Syzbot[2].

This is technically a bug in mm/gup.c, but arguably ext4 is fragile in that if some other kernel subsystem dirty pages without properly notifying the file system using page_mkwrite(), ext4 will BUG, while other file systems will not BUG (although data will still be lost).

So instead of crashing with a BUG, issue a warning (since there may be potential data loss) and just mark the page as clean to avoid unprivileged denial of service attacks until the problem can be properly fixed. More discussion and background can be found in the thread starting at [2].

[1] https://lore.kernel.org/linux-mm/[email protected]
[2] https://lore.kernel.org/r/[email protected]

---
 fs/ext4/inode.c | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
fs/ext4/inode.c +25 -0
index 7959aae4857e..96cf0f57ca95 100644
@@ -2148,6 +2148,15 @@ static int ext4_writepage(struct page *page,
 	else
 		len = PAGE_SIZE;
 
+	/* Should never happen but for bugs in other kernel subsystems */
+	if (!page_has_buffers(page)) {
+		ext4_warning_inode(inode,
+		   "page %lu does not have buffers attached", page->index);
+		ClearPageDirty(page);
+		unlock_page(page);
+		return 0;
+	}
+
 	page_bufs = page_buffers(page);
 	/*
 	 * We cannot do block allocation or other extent handling in this
@@ -2697,6 +2706,22 @@ static int mpage_prepare_extent_to_map(struct mpage_da_data *mpd)
 			wait_on_page_writeback(page);
 			BUG_ON(PageWriteback(page));
 
+			/*
+			 * Should never happen but for buggy code in
+			 * other subsystems that call
+			 * set_page_dirty() without properly warning
+			 * the file system first.  See [1] for more
+			 * information.
+			 *
+			 * [1] https://lore.kernel.org/linux-mm/[email protected]
+			 */
+			if (!page_has_buffers(page)) {
+				ext4_warning_inode(mpd->inode, "page %lu does not have buffers attached", page->index);
+				ClearPageDirty(page);
+				unlock_page(page);
+				continue;
+			}
+
 			if (mpd->map.m_len == 0)
 				mpd->first_page = page->index;
 			mpd->next_page = page->index + 1;
Signature · 1 line
2.34.1

I see the command but can't find the corresponding bug. The email is sent to [email protected] address but the HASH does not correspond to any known bug. Please double check the address.

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/2 ext4: don't BUG if someone dirty pages without asking ext4 first

From: Theodore Ts'o <[email protected]>

[ Upstream commit cc5095747edfb054ca2068d01af20be3fcc3634f ]

[un]pin_user_pages_remote is dirtying pages without properly warning the file system in advance. A related race was noted by Jan Kara in 2018[1]; however, more recently instead of it being a very hard-to-hit race, it could be reliably triggered by process_vm_writev(2) which was discovered by Syzbot[2].

This is technically a bug in mm/gup.c, but arguably ext4 is fragile in that if some other kernel subsystem dirty pages without properly notifying the file system using page_mkwrite(), ext4 will BUG, while other file systems will not BUG (although data will still be lost).

So instead of crashing with a BUG, issue a warning (since there may be potential data loss) and just mark the page as clean to avoid unprivileged denial of service attacks until the problem can be properly fixed. More discussion and background can be found in the thread starting at [2].

[1] https://lore.kernel.org/linux-mm/[email protected]
[2] https://lore.kernel.org/r/[email protected]

---
 fs/ext4/inode.c | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
fs/ext4/inode.c +25 -0
index 7959aae4857e..96cf0f57ca95 100644
@@ -2148,6 +2148,15 @@ static int ext4_writepage(struct page *page,
 	else
 		len = PAGE_SIZE;
 
+	/* Should never happen but for bugs in other kernel subsystems */
+	if (!page_has_buffers(page)) {
+		ext4_warning_inode(inode,
+		   "page %lu does not have buffers attached", page->index);
+		ClearPageDirty(page);
+		unlock_page(page);
+		return 0;
+	}
+
 	page_bufs = page_buffers(page);
 	/*
 	 * We cannot do block allocation or other extent handling in this
@@ -2697,6 +2706,22 @@ static int mpage_prepare_extent_to_map(struct mpage_da_data *mpd)
 			wait_on_page_writeback(page);
 			BUG_ON(PageWriteback(page));
 
+			/*
+			 * Should never happen but for buggy code in
+			 * other subsystems that call
+			 * set_page_dirty() without properly warning
+			 * the file system first.  See [1] for more
+			 * information.
+			 *
+			 * [1] https://lore.kernel.org/linux-mm/[email protected]
+			 */
+			if (!page_has_buffers(page)) {
+				ext4_warning_inode(mpd->inode, "page %lu does not have buffers attached", page->index);
+				ClearPageDirty(page);
+				unlock_page(page);
+				continue;
+			}
+
 			if (mpd->map.m_len == 0)
 				mpd->first_page = page->index;
 			mpd->next_page = page->index + 1;
Signature · 1 line
2.34.1
author maintainer
#2 lore raw

PATCH 2/2 ntfs: add sanity check on allocation size

From: Dongliang Mu <[email protected]>

[ Upstream commit 714fbf2647b1a33d914edd695d4da92029c7e7c0 ]

ntfs_read_inode_mount invokes ntfs_malloc_nofs with zero allocation size. It triggers one BUG in the __ntfs_malloc function.

Fix this by adding sanity check on ni->attr_list_size.

---
 fs/ntfs/inode.c | 4 ++++
 1 file changed, 4 insertions(+)
fs/ntfs/inode.c +4 -0
index e844b43f2eac..0acd1f02b146 100644
@@ -1906,6 +1906,10 @@ int ntfs_read_inode_mount(struct inode *vi)
 		}
 		/* Now allocate memory for the attribute list. */
 		ni->attr_list_size = (u32)ntfs_attr_size(a);
+		if (!ni->attr_list_size) {
+			ntfs_error(sb, "Attr_list_size is zero");
+			goto put_err_out;
+		}
 		ni->attr_list = ntfs_malloc_nofs(ni->attr_list_size);
 		if (!ni->attr_list) {
 			ntfs_error(sb, "Not enough memory to allocate buffer "
Signature · 1 line
2.34.1