Before 2011 there was no meaningful synchronization between
read/readdir/write/seek. Only in commit
ef3d0fd27e90 ("vfs: do (nearly) lockless generic_file_llseek")
synchronization was added for SEEK_CUR by taking f_lock around
vfs_setpos().
Then in 2014 full synchronization between read/readdir/write/seek was
added in commit 9c225f2655e3 ("vfs: atomic f_pos accesses as per POSIX")
by introducing f_pos_lock for regular files with FMODE_ATOMIC_POS and
for directories. At that point taking f_lock became unnecessary for such
files.
So only acquire f_lock for SEEK_CUR if this isn't a file that would have
acquired f_pos_lock if necessary.
Signed-off-by: Christian Brauner <brauner@xxxxxxxxxx>
---
fs/read_write.c | 20 +++++++++++++++-----
1 file changed, 15 insertions(+), 5 deletions(-)
diff --git a/fs/read_write.c b/fs/read_write.c
index a6133241dfb8..816189f9c56d 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -168,13 +168,23 @@ generic_file_llseek_size(struct file *file, loff_t offset, int whence,
return offset;
if (whence == SEEK_CUR) {
+ bool locked;
+
/*
- * f_lock protects against read/modify/write race with
- * other SEEK_CURs. Note that parallel writes and reads
- * behave like SEEK_SET.
+ * If the file requires locking via f_pos_lock we know
+ * that mutual exclusion for SEEK_CUR on the same file
+ * is guaranteed. If the file isn't locked, we take
+ * f_lock to protect against f_pos races with other
+ * SEEK_CURs.
*/
- guard(spinlock)(&file->f_lock);
- return vfs_setpos(file, file->f_pos + offset, maxsize);
+ locked = (file->f_mode & FMODE_ATOMIC_POS) ||
+ file->f_op->iterate_shared;
+ if (!locked)
+ spin_lock(&file->f_lock);
+ offset = vfs_setpos(file, file->f_pos + offset, maxsize);
+ if (!locked)
+ spin_unlock(&file->f_lock);
+ return offset;
}
return vfs_setpos(file, offset, maxsize);
--
2.47.2
