s390/qeth: fixes 2018-03-20

Hi Dave,

Please apply one final set of qeth patches for 4.16.
All of these fix long-standing bugs, so please queue them up for -stable as well.

Thank you,
Julian

Julian Wiedmann (4):
s390/qeth: free netdevice when removing a card s390/qeth: when thread completes, wake up all waiters s390/qeth: lock read device while queueing next buffer s390/qeth: on channel error, reject further cmd requests

 drivers/s390/net/qeth_core_main.c | 21 +++++++++++++++------
 drivers/s390/net/qeth_l2_main.c   |  2 +-
 drivers/s390/net/qeth_l3_main.c   |  2 +-
 3 files changed, 17 insertions(+), 8 deletions(-)

2.13.5

On removal, a qeth card's netdevice is currently not properly freed because the call chain looks as follows:

qeth_core_remove_device(card)

	lx_remove_device(card)
		unregister_netdev(card->dev)
		card->dev = NULL			!!!
	qeth_core_free_card(card)
		if (card->dev)				!!!
			free_netdev(card->dev)

Fix it by free'ing the netdev straight after unregistering. This also fixes the sysfs-driven layer switch case (qeth_dev_layer2_store()), where the need to free the current netdevice was not considered at all.

Note that free_netdev() takes care of the netif_napi_del() for us too.

---
 drivers/s390/net/qeth_core_main.c | 2 --
 drivers/s390/net/qeth_l2_main.c   | 2 +-
 drivers/s390/net/qeth_l3_main.c   | 2 +-
 3 files changed, 2 insertions(+), 4 deletions(-)

index c8b308cfabf1..4a820afececd 100644
@@ -5087,8 +5087,6 @@ static void qeth_core_free_card(struct qeth_card *card)
 	QETH_DBF_HEX(SETUP, 2, &card, sizeof(void *));
 	qeth_clean_channel(&card->read);
 	qeth_clean_channel(&card->write);
-	if (card->dev)
-		free_netdev(card->dev);
 	qeth_free_qdio_buffers(card);
 	unregister_service_level(&card->qeth_service_level);
 	kfree(card);

index 7f236440483f..5ef4c978ad19 100644
@@ -915,8 +915,8 @@ static void qeth_l2_remove_device(struct ccwgroup_device *cgdev)
 		qeth_l2_set_offline(cgdev);
 
 	if (card->dev) {
-		netif_napi_del(&card->napi);
 		unregister_netdev(card->dev);
+		free_netdev(card->dev);
 		card->dev = NULL;
 	}
 	return;

index 962a04b68dd2..b6b12220da71 100644
@@ -2865,8 +2865,8 @@ static void qeth_l3_remove_device(struct ccwgroup_device *cgdev)
 		qeth_l3_set_offline(cgdev);
 
 	if (card->dev) {
-		netif_napi_del(&card->napi);
 		unregister_netdev(card->dev);
+		free_netdev(card->dev);
 		card->dev = NULL;
 	}
 

2.13.5

qeth_wait_for_threads() is potentially called by multiple users, make sure to notify all of them after qeth_clear_thread_running_bit() adjusted the thread_running_mask. With no timeout, callers would otherwise stall.

---
 drivers/s390/net/qeth_core_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 4a820afececd..42ee8806fa53 100644
@@ -960,7 +960,7 @@ void qeth_clear_thread_running_bit(struct qeth_card *card, unsigned long thread)
 	spin_lock_irqsave(&card->thread_mask_lock, flags);
 	card->thread_running_mask &= ~thread;
 	spin_unlock_irqrestore(&card->thread_mask_lock, flags);
-	wake_up(&card->wait_q);
+	wake_up_all(&card->wait_q);
 }
 EXPORT_SYMBOL_GPL(qeth_clear_thread_running_bit);
 

2.13.5

For calling ccw_device_start(), issue_next_read() needs to hold the device's ccwlock.
This is satisfied for the IRQ handler path (where qeth_irq() gets called under the ccwlock), but we need explicit locking for the initial call by the MPC initialization.

---
 drivers/s390/net/qeth_core_main.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

index 42ee8806fa53..2a9afaf8f264 100644
@@ -527,8 +527,7 @@ static inline int qeth_is_cq(struct qeth_card *card, unsigned int queue)
 	    queue == card->qdio.no_in_queues - 1;
 }
 
-
-static int qeth_issue_next_read(struct qeth_card *card)
+static int __qeth_issue_next_read(struct qeth_card *card)
 {
 	int rc;
 	struct qeth_cmd_buffer *iob;

@@ -559,6 +558,17 @@ static int qeth_issue_next_read(struct qeth_card *card)
 	return rc;
 }
 
+static int qeth_issue_next_read(struct qeth_card *card)
+{
+	int ret;
+
+	spin_lock_irq(get_ccwdev_lock(CARD_RDEV(card)));
+	ret = __qeth_issue_next_read(card);
+	spin_unlock_irq(get_ccwdev_lock(CARD_RDEV(card)));
+
+	return ret;
+}
+
 static struct qeth_reply *qeth_alloc_reply(struct qeth_card *card)
 {
 	struct qeth_reply *reply;

@@ -1182,7 +1192,7 @@ static void qeth_irq(struct ccw_device *cdev, unsigned long intparm,
 		return;
 	if (channel == &card->read &&
 	    channel->state == CH_STATE_UP)
-		qeth_issue_next_read(card);
+		__qeth_issue_next_read(card);
 
 	iob = channel->iob;
 	index = channel->buf_no;

2.13.5

When the IRQ handler determines that one of the cmd IO channels has failed and schedules recovery, block any further cmd requests from being submitted. The request would inevitably stall, and prevent the recovery from making progress until the request times out.

This sort of error was observed after Live Guest Relocation, where the pending IO on the READ channel intentionally gets terminated to kick-start recovery. Simultaneously the guest executed SIOCETHTOOL, triggering qeth to issue a QUERY CARD INFO command. The command then stalled in the inoperabel WRITE channel.

---
 drivers/s390/net/qeth_core_main.c | 1 +
 1 file changed, 1 insertion(+)

index 2a9afaf8f264..3653bea38470 100644
@@ -1174,6 +1174,7 @@ static void qeth_irq(struct ccw_device *cdev, unsigned long intparm,
 		}
 		rc = qeth_get_problem(cdev, irb);
 		if (rc) {
+			card->read_or_write_problem = 1;
 			qeth_clear_ipacmd_list(card);
 			qeth_schedule_recovery(card);
 			goto out;

2.13.5

From: Julian Wiedmann <[email protected]>
Date: Tue, 20 Mar 2018 07:59:11 +0100

Series applied and queued up for -stable, thank you.