Re: [PATCH 2/2] target: iscsi: free cmds before session free

From: Forza <forza@tnonline.net>
To: Dmitry Bogdanov <d.bogdanov@yadro.com>,
	Martin Petersen <martin.petersen@oracle.com>,
	target-devel@vger.kernel.org
Cc: linux-scsi@vger.kernel.org, linux@yadro.com
Subject: Re: [PATCH 2/2] target: iscsi: free cmds before session free
Date: Sat, 11 Feb 2023 09:42:43 +0100	[thread overview]
Message-ID: <b2e0cb60-c8d5-2bd2-2954-d3b32252572a@tnonline.net> (raw)
In-Reply-To: <20230125083309.24678-3-d.bogdanov@yadro.com>


On 2023-01-25 09:33, Dmitry Bogdanov wrote:
> Commands from recovery entries are freed after its session has been
> closed. That leads to use-after-free at command free or NPE with such
> call trace:
> 
> Time2Retain timer expired for SID: 1, cleaning up iSCSI session.
> BUG: kernel NULL pointer dereference, address: 0000000000000140
> RIP: 0010:sbitmap_queue_clear+0x3a/0xa0
> Call Trace:
>   target_release_cmd_kref+0xd1/0x1f0 [target_core_mod]
>   transport_generic_free_cmd+0xd1/0x180 [target_core_mod]
>   iscsit_free_cmd+0x53/0xd0 [iscsi_target_mod]
>   iscsit_free_connection_recovery_entries+0x29d/0x320 [iscsi_target_mod]
>   iscsit_close_session+0x13a/0x140 [iscsi_target_mod]
>   iscsit_check_post_dataout+0x440/0x440 [iscsi_target_mod]
>   call_timer_fn+0x24/0x140
> 
> Move cleanup of recovery enrties to before session freeing.
> 
> Reported-by: Forza <forza@tnonline.net>
> Signed-off-by: Dmitry Bogdanov <d.bogdanov@yadro.com>
> ---
>   drivers/target/iscsi/iscsi_target.c | 6 +++---
>   1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c
> index f6008675dd3f..0748cbfb9631 100644
> --- a/drivers/target/iscsi/iscsi_target.c
> +++ b/drivers/target/iscsi/iscsi_target.c
> @@ -4521,6 +4521,9 @@ int iscsit_close_session(struct iscsit_session *sess, bool can_sleep)
>   	iscsit_stop_time2retain_timer(sess);
>   	spin_unlock_bh(&se_tpg->session_lock);
>   
> +	if (sess->sess_ops->ErrorRecoveryLevel == 2)
> +		iscsit_free_connection_recovery_entries(sess);
> +
>   	/*
>   	 * transport_deregister_session_configfs() will clear the
>   	 * struct se_node_acl->nacl_sess pointer now as a iscsi_np process context
> @@ -4544,9 +4547,6 @@ int iscsit_close_session(struct iscsit_session *sess, bool can_sleep)
>   
>   	transport_deregister_session(sess->se_sess);
>   
> -	if (sess->sess_ops->ErrorRecoveryLevel == 2)
> -		iscsit_free_connection_recovery_entries(sess);
> -
>   	iscsit_free_all_ooo_cmdsns(sess);
>   
>   	spin_lock_bh(&se_tpg->session_lock);

I have not had any more issues since applying the patch, which is good 
news :)

Will this be backported to earlier stable kernels like 5.15.x and 6.1.x

Regards,
Forza