Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 7 May 2020 17:02:42 -0700
From:   Andrew Morton <akpm@linux-foundation.org>
To:     Vasily Averin <vvs@virtuozzo.com>
Cc:     linux-kernel@vger.kernel.org, Waiman Long <longman@redhat.com>,
        Andreas Schwab <schwab@suse.de>,
        Matthew Wilcox <willy@infradead.org>
Subject: Re: [PATCH] ipc/util.c: sysvipc_find_ipc() incorrectly updates
 position index
Message-Id: <20200507170242.6cbb88ae672deed67152e221@linux-foundation.org>
In-Reply-To: <4921fe9b-9385-a2b4-1dc4-1099be6d2e39@virtuozzo.com>
References: <ad24ee52-21a8-3ff0-e968-63d315472b91@virtuozzo.com>
        <4921fe9b-9385-a2b4-1dc4-1099be6d2e39@virtuozzo.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, 6 May 2020 09:25:54 +0300 Vasily Averin <vvs@virtuozzo.com> wrote:

> new_pos should jump through hole of unused ids,
> pos can be updated inside "for" cycle.
> 
> Cc: stable@vger.kernel.org
> Fixes: 89163f93c6f9 ("ipc/util.c: sysvipc_find_ipc() should increase position index")

This:

> --- a/ipc/util.c
> +++ b/ipc/util.c
> @@ -764,21 +764,21 @@ static struct kern_ipc_perm *sysvipc_find_ipc(struct ipc_ids *ids, loff_t pos,
>  			total++;
>  	}
>  
> -	*new_pos = pos + 1;
> +	ipc = NULL;
>  	if (total >= ids->in_use)
> -		return NULL;
> +		goto out;
>  
>  	for (; pos < ipc_mni; pos++) {
>  		ipc = idr_find(&ids->ipcs_idr, pos);
>  		if (ipc != NULL) {
>  			rcu_read_lock();
>  			ipc_lock_object(ipc);
> -			return ipc;
> +			break;
>  		}
>  	}
> -
> -	/* Out of range - return NULL to terminate iteration */
> -	return NULL;
> +out:
> +	*new_pos = pos + 1;
> +	return ipc;
>  }
>  
>  static void *sysvipc_proc_next(struct seq_file *s, void *it, loff_t *pos)

Messes up Matthew's "ipc: convert ipcs_idr to XArray"
(http://lkml.kernel.org/r/20200326151418.27545-1-willy@infradead.org).

Here's how I resolved things.  Please check?

static struct kern_ipc_perm *sysvipc_find_ipc(struct ipc_ids *ids, loff_t pos,
					      loff_t *new_pos)
{
	unsigned long index = pos;
	struct kern_ipc_perm *ipc;

	rcu_read_lock();
	ipc = xa_find(&ids->ipcs, &index, ULONG_MAX, XA_PRESENT);
	if (ipc)
		ipc_lock_object(ipc);
	else
		rcu_read_unlock();
	*new_pos = pos + 1;
	return ipc;
}