2023-09-12 17:45:44

by NeilBrown

Subject: [PATCH] NFSD: fix possible oops when nfsd/pool_stats is closed.

If /proc/fs/nfsd/pool_stats is open when the last nfsd thread exits, then
when the file is closed a NULL pointer is dereferenced.
This is because nfsd_pool_stats_release() assumes that the
pointer to the svc_serv cannot become NULL while a reference is held.

This used to be the case but a recent patch split nfsd_last_thread() out
from nfsd_put(), and clearing the pointer is done in nfsd_last_thread().

This is easily reproduced by running
rpc.nfsd 8 ; ( rpc.nfsd 0;true) < /proc/fs/nfsd/pool_stats

Fortunately nfsd_pool_stats_release() has easy access to the svc_serv
pointer, and so can call svc_put() on it directly.

Fixes: 9f28a971ee9f ("nfsd: separate nfsd_last_thread() from nfsd_put()")
Signed-off-by: NeilBrown <[email protected]>
---
fs/nfsd/nfssvc.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index 1582af33e204..551c16a72012 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -1082,11 +1082,12 @@ int nfsd_pool_stats_open(struct inode *inode, struct file *file)

int nfsd_pool_stats_release(struct inode *inode, struct file *file)
{
+ struct svc_serv *serv =
+ ((struct seq_file *) file->private_data)->private;
int ret = seq_release(inode, file);
- struct net *net = inode->i_sb->s_fs_info;

mutex_lock(&nfsd_mutex);
- nfsd_put(net);
+ svc_put(serv);
mutex_unlock(&nfsd_mutex);
return ret;
}
--
2.42.0

2023-09-13 12:41:28

by Chuck Lever III

Subject: Re: [PATCH] NFSD: fix possible oops when nfsd/pool_stats is closed.

On Tue, Sep 12, 2023 at 11:25:00AM +1000, NeilBrown wrote:
>
> If /proc/fs/nfsd/pool_stats is open when the last nfsd thread exits, then
> when the file is closed a NULL pointer is dereferenced.
> This is because nfsd_pool_stats_release() assumes that the
> pointer to the svc_serv cannot become NULL while a reference is held.
>
> This used to be the case but a recent patch split nfsd_last_thread() out
> from nfsd_put(), and clearing the pointer is done in nfsd_last_thread().
>
> This is easily reproduced by running
> rpc.nfsd 8 ; ( rpc.nfsd 0;true) < /proc/fs/nfsd/pool_stats
>
> Fortunately nfsd_pool_stats_release() has easy access to the svc_serv
> pointer, and so can call svc_put() on it directly.
>
> Fixes: 9f28a971ee9f ("nfsd: separate nfsd_last_thread() from nfsd_put()")
> Signed-off-by: NeilBrown <[email protected]>
> ---
> fs/nfsd/nfssvc.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)

Thanks, applied to nfsd-fixes (for v6.6-rc).


> diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
> index 1582af33e204..551c16a72012 100644
> --- a/fs/nfsd/nfssvc.c
> +++ b/fs/nfsd/nfssvc.c
> @@ -1082,11 +1082,12 @@ int nfsd_pool_stats_open(struct inode *inode, struct file *file)
>
> int nfsd_pool_stats_release(struct inode *inode, struct file *file)
> {
> + struct svc_serv *serv =
> + ((struct seq_file *) file->private_data)->private;
> int ret = seq_release(inode, file);
> - struct net *net = inode->i_sb->s_fs_info;
>
> mutex_lock(&nfsd_mutex);
> - nfsd_put(net);
> + svc_put(serv);
> mutex_unlock(&nfsd_mutex);
> return ret;
> }
> --
> 2.42.0
>

--
Chuck Lever