[PATCH] r/o bind mounts: debugging for missed calls

There have been a few oopses caused by 'struct file's with NULL f_vfsmnts.
There was also a set of potentially missed mnt_want_write()s from
dentry_open() calls.

This patch provides a very simple debugging framework to catch these kinds of
bugs.  It will WARN_ON() them, but should stop us from having any oopses or
mnt_writer count imbalances.

I'm quite convinced that this is a good thing because it found bugs in the
stuff I was working on as soon as I wrote it.

[hch: made it conditional on a debug option.
      But it's still a little bit too ugly]

[hch: merged forced remount r/o fix from Dave and akpm's fix for the fix]

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

1 files changed, 49 insertions, 0 deletions

diff --git a/include/linux/fs.h b/include/linux/fs.h
index 013b9c2b88e..d1eeea669d2 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -776,6 +776,9 @@ static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
                 index <  ra->start + ra->size);
 }
 
+#define FILE_MNT_WRITE_TAKEN    1
+#define FILE_MNT_WRITE_RELEASED 2
+
 struct file {
         /*
          * fu_list becomes invalid after file_free is called and queued via
@@ -810,6 +813,9 @@ struct file {
         spinlock_t              f_ep_lock;
 #endif /* #ifdef CONFIG_EPOLL */
         struct address_space    *f_mapping;
+#ifdef CONFIG_DEBUG_WRITECOUNT
+        unsigned long f_mnt_write_state;
+#endif
 };
 extern spinlock_t files_lock;
 #define file_list_lock() spin_lock(&files_lock);
@@ -818,6 +824,49 @@ extern spinlock_t files_lock;
 #define get_file(x)     atomic_inc(&(x)->f_count)
 #define file_count(x)   atomic_read(&(x)->f_count)
 
+#ifdef CONFIG_DEBUG_WRITECOUNT
+static inline void file_take_write(struct file *f)
+{
+        WARN_ON(f->f_mnt_write_state != 0);
+        f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
+}
+static inline void file_release_write(struct file *f)
+{
+        f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
+}
+static inline void file_reset_write(struct file *f)
+{
+        f->f_mnt_write_state = 0;
+}
+static inline void file_check_state(struct file *f)
+{
+        /*
+         * At this point, either both or neither of these bits
+         * should be set.
+         */
+        WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
+        WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
+}
+static inline int file_check_writeable(struct file *f)
+{
+        if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
+                return 0;
+        printk(KERN_WARNING "writeable file with no "
+                            "mnt_want_write()\n");
+        WARN_ON(1);
+        return -EINVAL;
+}
+#else /* !CONFIG_DEBUG_WRITECOUNT */
+static inline void file_take_write(struct file *filp) {}
+static inline void file_release_write(struct file *filp) {}
+static inline void file_reset_write(struct file *filp) {}
+static inline void file_check_state(struct file *filp) {}
+static inline int file_check_writeable(struct file *filp)
+{
+        return 0;
+}
+#endif /* CONFIG_DEBUG_WRITECOUNT */
+
 #define MAX_NON_LFS     ((1UL<<31) - 1)
 
 /* Page cache limit. The filesystems should put that into their s_maxbytes