Re: [PATCH 2/5 -tip] cifs: umount_begin BKL pushdown

[Al Viro <viro@ZenIV.linux.org.uk>]

On Thu, Apr 23, 2009 at 09:12:02PM +0200, Alessio Igor Bogani wrote:
>  	struct cifs_sb_info *cifs_sb = CIFS_SB(sb);
>  	struct cifsTconInfo *tcon;
>  
> -	if (cifs_sb == NULL)
> +	lock_kernel();
> +
> +	if (cifs_sb == NULL) {
> +		unlock_kernel();
>  		return;
> +	}

Huh?  Why would a bleeding *local* *variable* care about BKL at all?

>  	tcon = cifs_sb->tcon;
> -	if (tcon == NULL)
> +	if (tcon == NULL) {
> +		unlock_kernel();
>  		return;
> +	}
>  
>  	read_lock(&cifs_tcp_ses_lock);
>  	if (tcon->tc_count == 1)
>  		tcon->tidStatus = CifsExiting;
>  	read_unlock(&cifs_tcp_ses_lock);
>  
> +	unlock_kernel();
> +
>  	/* cancel_brl_requests(tcon); */ /* BB mark all brl mids as exiting */
>  	/* cancel_notify_requests(tcon); */

... and why would the rest of it *not* care?

>  	if (tcon->ses && tcon->ses->server) {
> -- 
> 1.6.0.4
>