[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <691d5adf.a70a0220.d98e3.0008.GAE@google.com>
Date: Tue, 18 Nov 2025 21:51:27 -0800
From: syzbot <syzbot+a72c325b042aae6403c7@...kaller.appspotmail.com>
To: linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com
Subject: Forwarded: [PATCH] tracing: Fix WARN_ON in tracing_buffers_mmap_close
for split VMAs
For archival purposes, forwarding an incoming command email to
linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com.
***
Subject: [PATCH] tracing: Fix WARN_ON in tracing_buffers_mmap_close for split VMAs
Author: kartikey406@...il.com
#syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
When a VMA is split (e.g., by partial munmap or MAP_FIXED), the kernel
calls vm_ops->close on each portion. For trace buffer mappings, this
results in ring_buffer_unmap() being called multiple times while
ring_buffer_map() was only called once.
This causes ring_buffer_unmap() to return -ENODEV on subsequent calls
because user_mapped is already 0, triggering a WARN_ON.
Fix this by handling -ENODEV gracefully in tracing_buffers_mmap_close().
When ring_buffer_unmap() returns -ENODEV, it means this VMA was a split
portion that doesn't hold a reference, so simply return without calling
put_snapshot_map().
Closes: https://syzkaller.appspot.com/bug?extid=a72c325b042aae6403c7
Reported-by: syzbot+a72c325b042aae6403c7@...kaller.appspotmail.com
Signed-off-by: Deepanshu Kartikey <kartikey406@...il.com>
---
kernel/trace/trace.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index d1e527cf2aae..fe593dd2c387 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -8776,8 +8776,17 @@ static void tracing_buffers_mmap_close(struct vm_area_struct *vma)
{
struct ftrace_buffer_info *info = vma->vm_file->private_data;
struct trace_iterator *iter = &info->iter;
+ int ret;
- WARN_ON(ring_buffer_unmap(iter->array_buffer->buffer, iter->cpu_file));
+ ret = ring_buffer_unmap(iter->array_buffer->buffer, iter->cpu_file);
+ if (ret == -ENODEV) {
+ /*
+ * This VMA was split from the original mapping. Since
+ * ring buffer mappings do not support partial mappings,
+ * the split VMA does not hold a reference.
+ */
+ return;
+ }
put_snapshot_map(iter->tr);
}
--
2.43.0
Powered by blists - more mailing lists