* [PATCH v4] target: core: remove from tmr_list at lun unlink
@ 2021-10-18 13:57 Dmitry Bogdanov
2021-10-22 4:17 ` Mike Christie
2021-10-27 3:15 ` Martin K. Petersen
0 siblings, 2 replies; 3+ messages in thread
From: Dmitry Bogdanov @ 2021-10-18 13:57 UTC (permalink / raw)
To: Martin Petersen, target-devel
Cc: Mike Christie, linux-scsi, linux, Dmitry Bogdanov, Roman Bolshakov
Currently TMF commands are removed from de_device.dev_tmf_list at
the very end of se_cmd lifecycle. But se_lun unlinks from se_cmd
up on a command status (response) is queued in transport layer.
It means that LUN and backend device can be deleted meantime and at
the moment of repsonse completion a panic is occured:
target_tmr_work()
cmd->se_tfo->queue_tm_rsp(cmd); // send abort_rsp to a wire
transport_lun_remove_cmd(cmd) // unlink se_cmd from se_lun
- // - // - // -
<<<--- lun remove
<<<--- core backend device remove
- // - // - // -
qlt_handle_abts_completion()
tfo->free_mcmd()
transport_generic_free_cmd()
target_put_sess_cmd()
core_tmr_release_req() {
if (dev) { // backend device, can not be null
spin_lock_irqsave(&dev->se_tmr_lock, flags); //<<<--- CRASH
Call Trace:
NIP [c000000000e1683c] _raw_spin_lock_irqsave+0x2c/0xc0
LR [c00800000e433338] core_tmr_release_req+0x40/0xa0 [target_core_mod]
Call Trace:
(unreliable)
0x0
target_put_sess_cmd+0x2a0/0x370 [target_core_mod]
transport_generic_free_cmd+0x6c/0x1b0 [target_core_mod]
tcm_qla2xxx_complete_mcmd+0x28/0x50 [tcm_qla2xxx]
process_one_work+0x2c4/0x5c0
worker_thread+0x88/0x690
For FC protocol it is a race condition, but for iSCSI protocol it is
easyly reproduced by manual sending iSCSI commands:
- Send some SCSI sommand
- Send Abort of that command over iSCSI
- Remove LUN on target
- Send next iSCSI command to acknowledge the Abort_Response
- target panics
There is no sense to keep the command in tmr_list until response
completion, so move the removal from tmr_list from the response
completion to the response queueing when lun is unlinked.
Move the removal from state list too as it is a subject to the same
race condition.
Fixes: c66ac9db8d4a ("[SCSI] target: Add LIO target core v4.0.0-rc6")
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Dmitry Bogdanov <d.bogdanov@yadro.com>
---
v4:
remove superfluous transport_lookup_tmr_lun that could race with lun
reset
v3:
remove iscsi fix as not related to the issue
avoid double removal from tmr_list
v2:
fix stuck in tmr list in error case
The issue exists from the very begining.
I uploaded a scapy script that helps to reproduce the issue at
https://gist.github.com/logost/cb93df41dd2432454324449b390403c4
---
drivers/target/target_core_tmr.c | 17 +--------------
drivers/target/target_core_transport.c | 30 ++++++++++++++++++++------
2 files changed, 24 insertions(+), 23 deletions(-)
diff --git a/drivers/target/target_core_tmr.c b/drivers/target/target_core_tmr.c
index e7fcbc09f9db..bac111456fa1 100644
--- a/drivers/target/target_core_tmr.c
+++ b/drivers/target/target_core_tmr.c
@@ -50,15 +50,6 @@ EXPORT_SYMBOL(core_tmr_alloc_req);
void core_tmr_release_req(struct se_tmr_req *tmr)
{
- struct se_device *dev = tmr->tmr_dev;
- unsigned long flags;
-
- if (dev) {
- spin_lock_irqsave(&dev->se_tmr_lock, flags);
- list_del_init(&tmr->tmr_list);
- spin_unlock_irqrestore(&dev->se_tmr_lock, flags);
- }
-
kfree(tmr);
}
@@ -156,13 +147,6 @@ void core_tmr_abort_task(
se_cmd->state_active = false;
spin_unlock_irqrestore(&dev->queues[i].lock, flags);
- /*
- * Ensure that this ABORT request is visible to the LU
- * RESET code.
- */
- if (!tmr->tmr_dev)
- WARN_ON_ONCE(transport_lookup_tmr_lun(tmr->task_cmd) < 0);
-
if (dev->transport->tmr_notify)
dev->transport->tmr_notify(dev, TMR_ABORT_TASK,
&aborted_list);
@@ -234,6 +218,7 @@ static void core_tmr_drain_tmr_list(
}
list_move_tail(&tmr_p->tmr_list, &drain_tmr_list);
+ tmr_p->tmr_dev = NULL;
}
spin_unlock_irqrestore(&dev->se_tmr_lock, flags);
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index 14c6f2bb1b01..e60abd230e90 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -676,6 +676,21 @@ static void target_remove_from_state_list(struct se_cmd *cmd)
spin_unlock_irqrestore(&dev->queues[cmd->cpuid].lock, flags);
}
+static void target_remove_from_tmr_list(struct se_cmd *cmd)
+{
+ struct se_device *dev = NULL;
+ unsigned long flags;
+
+ if (cmd->se_cmd_flags & SCF_SCSI_TMR_CDB)
+ dev = cmd->se_tmr_req->tmr_dev;
+
+ if (dev) {
+ spin_lock_irqsave(&dev->se_tmr_lock, flags);
+ if (cmd->se_tmr_req->tmr_dev)
+ list_del_init(&cmd->se_tmr_req->tmr_list);
+ spin_unlock_irqrestore(&dev->se_tmr_lock, flags);
+ }
+}
/*
* This function is called by the target core after the target core has
* finished processing a SCSI command or SCSI TMF. Both the regular command
@@ -687,13 +702,6 @@ static int transport_cmd_check_stop_to_fabric(struct se_cmd *cmd)
{
unsigned long flags;
- target_remove_from_state_list(cmd);
-
- /*
- * Clear struct se_cmd->se_lun before the handoff to FE.
- */
- cmd->se_lun = NULL;
-
spin_lock_irqsave(&cmd->t_state_lock, flags);
/*
* Determine if frontend context caller is requesting the stopping of
@@ -728,8 +736,16 @@ static void transport_lun_remove_cmd(struct se_cmd *cmd)
if (!lun)
return;
+ target_remove_from_state_list(cmd);
+ target_remove_from_tmr_list(cmd);
+
if (cmpxchg(&cmd->lun_ref_active, true, false))
percpu_ref_put(&lun->lun_ref);
+
+ /*
+ * Clear struct se_cmd->se_lun before the handoff to FE.
+ */
+ cmd->se_lun = NULL;
}
static void target_complete_failure_work(struct work_struct *work)
--
2.25.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v4] target: core: remove from tmr_list at lun unlink
2021-10-18 13:57 [PATCH v4] target: core: remove from tmr_list at lun unlink Dmitry Bogdanov
@ 2021-10-22 4:17 ` Mike Christie
2021-10-27 3:15 ` Martin K. Petersen
1 sibling, 0 replies; 3+ messages in thread
From: Mike Christie @ 2021-10-22 4:17 UTC (permalink / raw)
To: Dmitry Bogdanov, Martin Petersen, target-devel
Cc: linux-scsi, linux, Roman Bolshakov
On 10/18/21 8:57 AM, Dmitry Bogdanov wrote:
> Currently TMF commands are removed from de_device.dev_tmf_list at
> the very end of se_cmd lifecycle. But se_lun unlinks from se_cmd
> up on a command status (response) is queued in transport layer.
> It means that LUN and backend device can be deleted meantime and at
> the moment of repsonse completion a panic is occured:
>
> target_tmr_work()
> cmd->se_tfo->queue_tm_rsp(cmd); // send abort_rsp to a wire
> transport_lun_remove_cmd(cmd) // unlink se_cmd from se_lun
> - // - // - // -
> <<<--- lun remove
> <<<--- core backend device remove
> - // - // - // -
> qlt_handle_abts_completion()
> tfo->free_mcmd()
> transport_generic_free_cmd()
> target_put_sess_cmd()
> core_tmr_release_req() {
> if (dev) { // backend device, can not be null
> spin_lock_irqsave(&dev->se_tmr_lock, flags); //<<<--- CRASH
>
> Call Trace:
> NIP [c000000000e1683c] _raw_spin_lock_irqsave+0x2c/0xc0
> LR [c00800000e433338] core_tmr_release_req+0x40/0xa0 [target_core_mod]
> Call Trace:
> (unreliable)
> 0x0
> target_put_sess_cmd+0x2a0/0x370 [target_core_mod]
> transport_generic_free_cmd+0x6c/0x1b0 [target_core_mod]
> tcm_qla2xxx_complete_mcmd+0x28/0x50 [tcm_qla2xxx]
> process_one_work+0x2c4/0x5c0
> worker_thread+0x88/0x690
>
> For FC protocol it is a race condition, but for iSCSI protocol it is
> easyly reproduced by manual sending iSCSI commands:
> - Send some SCSI sommand
> - Send Abort of that command over iSCSI
> - Remove LUN on target
> - Send next iSCSI command to acknowledge the Abort_Response
> - target panics
>
> There is no sense to keep the command in tmr_list until response
> completion, so move the removal from tmr_list from the response
> completion to the response queueing when lun is unlinked.
> Move the removal from state list too as it is a subject to the same
> race condition.
>
> Fixes: c66ac9db8d4a ("[SCSI] target: Add LIO target core v4.0.0-rc6")
> Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
> Signed-off-by: Dmitry Bogdanov <d.bogdanov@yadro.com>
>
> ---
> v4:
> remove superfluous transport_lookup_tmr_lun that could race with lun
> reset
> v3:
> remove iscsi fix as not related to the issue
> avoid double removal from tmr_list
> v2:
> fix stuck in tmr list in error case
>
> The issue exists from the very begining.
> I uploaded a scapy script that helps to reproduce the issue at
> https://gist.github.com/logost/cb93df41dd2432454324449b390403c4
Reviewed-by: Mike Christie <michael.christie@oracle.com>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v4] target: core: remove from tmr_list at lun unlink
2021-10-18 13:57 [PATCH v4] target: core: remove from tmr_list at lun unlink Dmitry Bogdanov
2021-10-22 4:17 ` Mike Christie
@ 2021-10-27 3:15 ` Martin K. Petersen
1 sibling, 0 replies; 3+ messages in thread
From: Martin K. Petersen @ 2021-10-27 3:15 UTC (permalink / raw)
To: Dmitry Bogdanov
Cc: Martin Petersen, target-devel, Mike Christie, linux-scsi, linux,
Roman Bolshakov
Dmitry,
> Currently TMF commands are removed from de_device.dev_tmf_list at the
> very end of se_cmd lifecycle. But se_lun unlinks from se_cmd up on a
> command status (response) is queued in transport layer. It means that
> LUN and backend device can be deleted meantime and at the moment of
> repsonse completion a panic is occured:
Applied to 5.16/scsi-staging, thanks!
--
Martin K. Petersen Oracle Linux Engineering
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-10-27 3:16 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-18 13:57 [PATCH v4] target: core: remove from tmr_list at lun unlink Dmitry Bogdanov
2021-10-22 4:17 ` Mike Christie
2021-10-27 3:15 ` Martin K. Petersen
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.