Linux-mm Archive on lore.kernel.org
 help / color / Atom feed
* [PATCH v2] hugetlbfs: add O_TMPFILE support
@ 2019-11-09  7:50 Piotr Sarna
  0 siblings, 0 replies; only message in thread
From: Piotr Sarna @ 2019-11-09  7:50 UTC (permalink / raw)
  To: mike.kravetz, linux-kernel
  Cc: Piotr Sarna, linux-mm, viro, linux-fsdevel, mhocko,
	syzbot+136d2439a4e6561ea00c

With hugetlbfs, a common pattern for mapping anonymous huge pages
is to create a temporary file first. Currently libraries like
libhugetlbfs and seastar create these with a standard mkstemp+unlink
trick, but it would be more robust to be able to simply pass
the O_TMPFILE flag to open(). O_TMPFILE is already supported by several
file systems like ext4 and xfs. The implementation simply uses the existing
d_tmpfile utility function to instantiate the dcache entry for the file.

Tested manually by successfully creating a temporary file by opening
it with (O_TMPFILE|O_RDWR) on mounted hugetlbfs and successfully
mapping 2M huge pages with it. Without the patch, trying to open
a file with O_TMPFILE results in -ENOSUP.

v2 changes:
 * syzkaller thankfully discovered a bug during unmount - tmpfile
erroneously called dget() on a dentry when creating a tmpfile,
and it was never countered by a dput(), because tmpfile is never
explicitly unlinked. In v2, dget() is simply not called for tmpfile.
Verified manually, and also with the reproducer provided by syzkaller.
Reported-by: syzbot+136d2439a4e6561ea00c@syzkaller.appspotmail.com

Signed-off-by: Piotr Sarna <p.sarna@tlen.pl>
---
 fs/hugetlbfs/inode.c | 28 ++++++++++++++++++++++++----
 1 file changed, 24 insertions(+), 4 deletions(-)

diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index a478df035651..a39d7a0a158e 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -815,8 +815,11 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb,
 /*
  * File creation. Allocate an inode, and we're done..
  */
-static int hugetlbfs_mknod(struct inode *dir,
-			struct dentry *dentry, umode_t mode, dev_t dev)
+static int do_hugetlbfs_mknod(struct inode *dir,
+			struct dentry *dentry,
+			umode_t mode,
+			dev_t dev,
+			bool tmpfile)
 {
 	struct inode *inode;
 	int error = -ENOSPC;
@@ -824,13 +827,23 @@ static int hugetlbfs_mknod(struct inode *dir,
 	inode = hugetlbfs_get_inode(dir->i_sb, dir, mode, dev);
 	if (inode) {
 		dir->i_ctime = dir->i_mtime = current_time(dir);
-		d_instantiate(dentry, inode);
-		dget(dentry);	/* Extra count - pin the dentry in core */
+		if (tmpfile) {
+			d_tmpfile(dentry, inode);
+		} else {
+			d_instantiate(dentry, inode);
+			dget(dentry);/* Extra count - pin the dentry in core */
+		}
 		error = 0;
 	}
 	return error;
 }
 
+static int hugetlbfs_mknod(struct inode *dir,
+			struct dentry *dentry, umode_t mode, dev_t dev)
+{
+	return do_hugetlbfs_mknod(dir, dentry, mode, dev, false);
+}
+
 static int hugetlbfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
 {
 	int retval = hugetlbfs_mknod(dir, dentry, mode | S_IFDIR, 0);
@@ -844,6 +857,12 @@ static int hugetlbfs_create(struct inode *dir, struct dentry *dentry, umode_t mo
 	return hugetlbfs_mknod(dir, dentry, mode | S_IFREG, 0);
 }
 
+static int hugetlbfs_tmpfile(struct inode *dir,
+			struct dentry *dentry, umode_t mode)
+{
+	return do_hugetlbfs_mknod(dir, dentry, mode | S_IFREG, 0, true);
+}
+
 static int hugetlbfs_symlink(struct inode *dir,
 			struct dentry *dentry, const char *symname)
 {
@@ -1102,6 +1121,7 @@ static const struct inode_operations hugetlbfs_dir_inode_operations = {
 	.mknod		= hugetlbfs_mknod,
 	.rename		= simple_rename,
 	.setattr	= hugetlbfs_setattr,
+	.tmpfile	= hugetlbfs_tmpfile,
 };
 
 static const struct inode_operations hugetlbfs_inode_operations = {
-- 
2.23.0



^ permalink raw reply	[flat|nested] only message in thread

only message in thread, back to index

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-09  7:50 [PATCH v2] hugetlbfs: add O_TMPFILE support Piotr Sarna

Linux-mm Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-mm/0 linux-mm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-mm linux-mm/ https://lore.kernel.org/linux-mm \
		linux-mm@kvack.org
	public-inbox-index linux-mm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kvack.linux-mm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git