Skip to contentSkip to conversation

hid: intel-ish-hid: ishtp: Fix ishtp client sending disordered message

PATCH Applied Even Xu 2 participants 3 messages Aug 4, 2022 – Aug 26, 2022

Subsystems (2)
lore
Share & tools
lore

Even Xu

author
#0 lore raw

PATCH 1/1 hid: intel-ish-hid: ishtp: Fix ishtp client sending disordered message

There is a timing issue captured during ishtp client sending stress tests. It was observed during stress tests that ISH firmware is getting out of ordered messages. This is a rare scenario as the current set of ISH client drivers don't send much data to firmware. But this may not be the case going forward.

When message size is bigger than IPC MTU, ishtp splits the message into fragments and uses serialized async method to send message fragments. The call stack:
ishtp_cl_send_msg_ipc->ipc_tx_callback(first fregment)-> ishtp_send_msg(with callback)->write_ipc_to_queue-> write_ipc_from_queue->callback->ipc_tx_callback(next fregment)......

When an ipc write complete interrupt is received, driver also calls write_ipc_from_queue->ipc_tx_callback in ISR to start sending of next fragment.

Through ipc_tx_callback uses spin_lock to protect message splitting, as the serialized sending method will call back to ipc_tx_callback again, so it doesn't put sending under spin_lock, it causes driver cannot guarantee all fragments be sent in order.

Considering this scenario:
ipc_tx_callback just finished a fragment splitting, and not call ishtp_send_msg yet, there is a write complete interrupt happens, then ISR->write_ipc_from_queue ->ipc_tx_callback->ishtp_send_msg->write_ipc_to_queue......

Because ISR has higher exec priority than normal thread, this causes the new fragment be sent out before previous fragment. This disordered message causes invalid message to firmware.

The solution is, to send fragments synchronously:
Use ishtp_write_message writing fragments into tx queue directly one by one, instead of ishtp_send_msg only writing one fragment with completion callback. As no completion callback be used, so change ipc_tx_callback to ipc_tx_send.

---
 drivers/hid/intel-ish-hid/ishtp/client.c | 68 ++++++++++++++++++--------------
 1 file changed, 39 insertions(+), 29 deletions(-)
drivers/hid/intel-ish-hid/ishtp/client.c +39 -29
index 405e0d5..df0a825 100644
@@ -626,13 +626,14 @@ static void ishtp_cl_read_complete(struct ishtp_cl_rb *rb)
 }
 
 /**
- * ipc_tx_callback() - IPC tx callback function
+ * ipc_tx_send() - IPC tx send function
  * @prm: Pointer to client device instance
  *
- * Send message over IPC either first time or on callback on previous message
- * completion
+ * Send message over IPC. Message will be split into fragments
+ * if message size is bigger than IPC FIFO size, and all
+ * fragments will be sent one by one.
  */
-static void ipc_tx_callback(void *prm)
+static void ipc_tx_send(void *prm)
 {
 	struct ishtp_cl	*cl = prm;
 	struct ishtp_cl_tx_ring	*cl_msg;
@@ -677,32 +678,41 @@ static void ipc_tx_callback(void *prm)
 			    list);
 	rem = cl_msg->send_buf.size - cl->tx_offs;
 
-	ishtp_hdr.host_addr = cl->host_client_id;
-	ishtp_hdr.fw_addr = cl->fw_client_id;
-	ishtp_hdr.reserved = 0;
-	pmsg = cl_msg->send_buf.data + cl->tx_offs;
+	while (rem > 0) {
+		ishtp_hdr.host_addr = cl->host_client_id;
+		ishtp_hdr.fw_addr = cl->fw_client_id;
+		ishtp_hdr.reserved = 0;
+		pmsg = cl_msg->send_buf.data + cl->tx_offs;
+
+		if (rem <= dev->mtu) {
+			/* Last fragment or only one packet */
+			ishtp_hdr.length = rem;
+			ishtp_hdr.msg_complete = 1;
+			/* Submit to IPC queue with no callback */
+			ishtp_write_message(dev, &ishtp_hdr, pmsg);
+			cl->tx_offs = 0;
+			cl->sending = 0;
 
-	if (rem <= dev->mtu) {
-		ishtp_hdr.length = rem;
-		ishtp_hdr.msg_complete = 1;
-		cl->sending = 0;
-		list_del_init(&cl_msg->list);	/* Must be before write */
-		spin_unlock_irqrestore(&cl->tx_list_spinlock, tx_flags);
-		/* Submit to IPC queue with no callback */
-		ishtp_write_message(dev, &ishtp_hdr, pmsg);
-		spin_lock_irqsave(&cl->tx_free_list_spinlock, tx_free_flags);
-		list_add_tail(&cl_msg->list, &cl->tx_free_list.list);
-		++cl->tx_ring_free_size;
-		spin_unlock_irqrestore(&cl->tx_free_list_spinlock,
-			tx_free_flags);
-	} else {
-		/* Send IPC fragment */
-		spin_unlock_irqrestore(&cl->tx_list_spinlock, tx_flags);
-		cl->tx_offs += dev->mtu;
-		ishtp_hdr.length = dev->mtu;
-		ishtp_hdr.msg_complete = 0;
-		ishtp_send_msg(dev, &ishtp_hdr, pmsg, ipc_tx_callback, cl);
+			break;
+		} else {
+			/* Send ipc fragment */
+			ishtp_hdr.length = dev->mtu;
+			ishtp_hdr.msg_complete = 0;
+			/* All fregments submitted to IPC queue with no callback */
+			ishtp_write_message(dev, &ishtp_hdr, pmsg);
+			cl->tx_offs += dev->mtu;
+			rem = cl_msg->send_buf.size - cl->tx_offs;
+		}
 	}
+
+	list_del_init(&cl_msg->list);
+	spin_unlock_irqrestore(&cl->tx_list_spinlock, tx_flags);
+
+	spin_lock_irqsave(&cl->tx_free_list_spinlock, tx_free_flags);
+	list_add_tail(&cl_msg->list, &cl->tx_free_list.list);
+	++cl->tx_ring_free_size;
+	spin_unlock_irqrestore(&cl->tx_free_list_spinlock,
+		tx_free_flags);
 }
 
 /**
@@ -720,7 +730,7 @@ static void ishtp_cl_send_msg_ipc(struct ishtp_device *dev,
 		return;
 
 	cl->tx_offs = 0;
-	ipc_tx_callback(cl);
+	ipc_tx_send(cl);
 	++cl->send_msg_cnt_ipc;
 }
 
Signature · 1 line
2.7.4
maintainer Applied
#1 lore raw
Quoted 21 lines · Even Xu

There is a timing issue captured during ishtp client sending stress tests. It was observed during stress tests that ISH firmware is getting out of ordered messages. This is a rare scenario as the current set of ISH client drivers don't send much data to firmware. But this may not be the case going forward.

When message size is bigger than IPC MTU, ishtp splits the message into fragments and uses serialized async method to send message fragments. The call stack:
ishtp_cl_send_msg_ipc->ipc_tx_callback(first fregment)-> ishtp_send_msg(with callback)->write_ipc_to_queue-> write_ipc_from_queue->callback->ipc_tx_callback(next fregment)......

When an ipc write complete interrupt is received, driver also calls write_ipc_from_queue->ipc_tx_callback in ISR to start sending of next fragment.

Through ipc_tx_callback uses spin_lock to protect message splitting, as the serialized sending method will call back to ipc_tx_callback again, so it doesn't put sending under spin_lock, it causes driver cannot guarantee all fragments be sent in order.

Quoted 8 lines

Considering this scenario:
ipc_tx_callback just finished a fragment splitting, and not call ishtp_send_msg yet, there is a write complete interrupt happens, then ISR->write_ipc_from_queue ->ipc_tx_callback->ishtp_send_msg->write_ipc_to_queue......

Because ISR has higher exec priority than normal thread, this causes the new fragment be sent out before previous fragment. This disordered message causes invalid message to firmware.

I can imagine that this must have been nightmare to debug :)

Quoted 5 lines

The solution is, to send fragments synchronously: Use ishtp_write_message writing fragments into tx queue directly one by one, instead of ishtp_send_msg only writing one fragment with completion callback. As no completion callback be used, so change ipc_tx_callback to ipc_tx_send.

Applied, thank you.

Signature · 2 lines
Jiri Kosina
SUSE Labs
author Applied
#2 lore raw

Thanks Jiri!

Best Regards,
Even Xu

-----Original Message-----
From: Jiri Kosina <[email protected]>
Sent: Thursday, August 25, 2022 5:36 PM
To: Xu, Even <[email protected]>
Cc: [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH] hid: intel-ish-hid: ishtp: Fix ishtp client sending disordered message

Quoted 27 lines · Even Xu

There is a timing issue captured during ishtp client sending stress tests. It was observed during stress tests that ISH firmware is getting out of ordered messages. This is a rare scenario as the current set of ISH client drivers don't send much data to firmware. But this may not be the case going forward.

When message size is bigger than IPC MTU, ishtp splits the message into fragments and uses serialized async method to send message fragments. The call stack:
ishtp_cl_send_msg_ipc->ipc_tx_callback(first fregment)-> ishtp_send_msg(with callback)->write_ipc_to_queue-> write_ipc_from_queue->callback->ipc_tx_callback(next fregment)......

When an ipc write complete interrupt is received, driver also calls write_ipc_from_queue->ipc_tx_callback in ISR to start sending of next fragment.

Through ipc_tx_callback uses spin_lock to protect message splitting, as the serialized sending method will call back to ipc_tx_callback again, so it doesn't put sending under spin_lock, it causes driver cannot guarantee all fragments be sent in order.

Considering this scenario:
ipc_tx_callback just finished a fragment splitting, and not call ishtp_send_msg yet, there is a write complete interrupt happens, then ISR->write_ipc_from_queue ->ipc_tx_callback->ishtp_send_msg->write_ipc_to_queue......

Quoted 3 lines

Because ISR has higher exec priority than normal thread, this causes the new fragment be sent out before previous fragment. This disordered message causes invalid message to firmware.

I can imagine that this must have been nightmare to debug :)

Quoted 5 lines

The solution is, to send fragments synchronously: Use ishtp_write_message writing fragments into tx queue directly one by one, instead of ishtp_send_msg only writing one fragment with completion callback. As no completion callback be used, so change ipc_tx_callback to ipc_tx_send.

Applied, thank you.

--
Jiri Kosina
SUSE Labs