Skip to content

Commit 0bcfd9a

Browse files
committed
tracefs: Clear EVENT_INODE flag in tracefs_drop_inode()
When the inode is being dropped from the dentry, the TRACEFS_EVENT_INODE flag needs to be cleared to prevent a remount from calling eventfs_remount() on the tracefs_inode private data. There's a race between the inode is dropped (and the dentry freed) to where the inode is actually freed. If a remount happens between the two, the eventfs_inode could be accessed after it is freed (only the dentry keeps a ref count on it). Currently the TRACEFS_EVENT_INODE flag is cleared from the dentry iput() function. But this is incorrect, as it is possible that the inode has another reference to it. The flag should only be cleared when the inode is really being dropped and has no more references. That happens in the drop_inode callback of the inode, as that gets called when the last reference of the inode is released. Remove the tracefs_d_iput() function and move its logic to the more appropriate tracefs_drop_inode() callback function. Link: https://lore.kernel.org/linux-trace-kernel/[email protected] Cc: [email protected] Cc: Masami Hiramatsu <[email protected]> Cc: Mark Rutland <[email protected]> Cc: Mathieu Desnoyers <[email protected]> Cc: Andrew Morton <[email protected]> Cc: Masahiro Yamada <[email protected]> Fixes: baa23a8 ("tracefs: Reset permissions on remount if permissions are options") Signed-off-by: Steven Rostedt (Google) <[email protected]>
1 parent 340f0c7 commit 0bcfd9a

File tree

1 file changed

+17
-16
lines changed

1 file changed

+17
-16
lines changed

fs/tracefs/inode.c

Lines changed: 17 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -426,10 +426,26 @@ static int tracefs_show_options(struct seq_file *m, struct dentry *root)
426426
return 0;
427427
}
428428

429+
static int tracefs_drop_inode(struct inode *inode)
430+
{
431+
struct tracefs_inode *ti = get_tracefs(inode);
432+
433+
/*
434+
* This inode is being freed and cannot be used for
435+
* eventfs. Clear the flag so that it doesn't call into
436+
* eventfs during the remount flag updates. The eventfs_inode
437+
* gets freed after an RCU cycle, so the content will still
438+
* be safe if the iteration is going on now.
439+
*/
440+
ti->flags &= ~TRACEFS_EVENT_INODE;
441+
442+
return 1;
443+
}
444+
429445
static const struct super_operations tracefs_super_operations = {
430446
.alloc_inode = tracefs_alloc_inode,
431447
.free_inode = tracefs_free_inode,
432-
.drop_inode = generic_delete_inode,
448+
.drop_inode = tracefs_drop_inode,
433449
.statfs = simple_statfs,
434450
.show_options = tracefs_show_options,
435451
};
@@ -455,22 +471,7 @@ static int tracefs_d_revalidate(struct dentry *dentry, unsigned int flags)
455471
return !(ei && ei->is_freed);
456472
}
457473

458-
static void tracefs_d_iput(struct dentry *dentry, struct inode *inode)
459-
{
460-
struct tracefs_inode *ti = get_tracefs(inode);
461-
462-
/*
463-
* This inode is being freed and cannot be used for
464-
* eventfs. Clear the flag so that it doesn't call into
465-
* eventfs during the remount flag updates. The eventfs_inode
466-
* gets freed after an RCU cycle, so the content will still
467-
* be safe if the iteration is going on now.
468-
*/
469-
ti->flags &= ~TRACEFS_EVENT_INODE;
470-
}
471-
472474
static const struct dentry_operations tracefs_dentry_operations = {
473-
.d_iput = tracefs_d_iput,
474475
.d_revalidate = tracefs_d_revalidate,
475476
.d_release = tracefs_d_release,
476477
};

0 commit comments

Comments
 (0)