Re: [PATCH] proc: add locking checks in proc_inode_is_dead

[Oleg Nesterov]

On 11/29, Wen Yang wrote:
>
> The proc_inode_is_dead function might race with __unhash_process.
> This will result in a whole bunch of stale proc entries being cached.
> To prevent that, add the required locking.

I leave this to Eric but I don't understand how can this patch help,
__unhash_process() can be called right after proc_inode_is_dead().

And in any case, we certainly do not want to take tasklist_lock in
proc_inode_is_dead().

> 
> Signed-off-by: Wen Yang <wenyang@xxxxxxxxxxxxxxxxx>
> Cc: Oleg Nesterov <oleg@xxxxxxxxxx>
> Cc: "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx>
> Cc: Alexey Dobriyan <adobriyan@xxxxxxxxx>
> Cc: Christian Brauner <christian@xxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Cc: linux-fsdevel@xxxxxxxxxxxxxxx
> ---
>  fs/proc/base.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index 1bc9bcd..59720bc 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -1994,7 +1994,13 @@ static int pid_revalidate(struct dentry *dentry, unsigned int flags)
>  
>  static inline bool proc_inode_is_dead(struct inode *inode)
>  {
> -	return !proc_pid(inode)->tasks[PIDTYPE_PID].first;
> +	bool has_task;
> +
> +	read_lock(&tasklist_lock);
> +	has_task = pid_has_task(proc_pid(inode), PIDTYPE_PID);
> +	read_unlock(&tasklist_lock);
> +
> +	return !has_task;
>  }
>  
>  int pid_delete_dentry(const struct dentry *dentry)
> -- 
> 1.8.3.1
>