patch 'net/hns3: fix mailbox sync' has been queued to stable release 21.11.6

Kevin Traynor ktraynor at redhat.com
Thu Nov 16 14:23:37 CET 2023


Hi,

FYI, your patch has been queued to stable release 21.11.6

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 11/21/23. So please
shout if anyone has objections.

Also note that after the patch there's a diff of the upstream commit vs the
patch applied to the branch. This will indicate if there was any rebasing
needed to apply to the stable branch. If there were code changes for rebasing
(ie: not only metadata diffs), please double check that the rebase was
correctly done.

Queued patches are on a temporary branch at:
https://github.com/kevintraynor/dpdk-stable

This queued commit can be viewed at:
https://github.com/kevintraynor/dpdk-stable/commit/1d9d4b5b633d2349f711750dfdf1ef65e9809db1

Thanks.

Kevin

---
>From 1d9d4b5b633d2349f711750dfdf1ef65e9809db1 Mon Sep 17 00:00:00 2001
From: Dengdui Huang <huangdengdui at huawei.com>
Date: Sat, 11 Nov 2023 09:59:14 +0800
Subject: [PATCH] net/hns3: fix mailbox sync

[ upstream commit be3590f54d0e415c23d4ed6ea55d967139c3ad10 ]

Currently, hns3 VF driver uses the following points to match
the response and request message for the mailbox synchronous
message between VF and PF.
1. req_msg_data which is consist of message code and subcode,
   is used to match request and response.
2. head means the number of send success for VF.
3. tail means the number of receive success for VF.
4. lost means the number of send timeout for VF.
And 'head', 'tail' and 'lost' are dynamically updated during
the communication.

Now there is a issue that all sync mailbox message will
send failure forever at the flollowing case:
1. VF sends the message A
    then head=UINT32_MAX-1, tail=UINT32_MAX-3, lost=2.
2. VF sends the message B
    then head=UINT32_MAX, tail=UINT32_MAX-2, lost=2.
3. VF sends the message C, the message will be timeout because
   it can't get the response within 500ms.
   then head=0, tail=0, lost=2
   note: tail is assigned to head if tail > head according to
   current code logic. From now on, all subsequent sync milbox
   messages fail to be sent.

It's very complicated to use the fields 'lost','tail','head'.
The code and subcode of the request sync mailbox are used as the
matching code of the message, which is used to match the response
message for receiving the synchronization response.

This patch drops these fields and uses the following solution
to solve this issue:
In the handling response message process, using the req_msg_data
of the request and response message to judge whether the sync
mailbox message has been received.

Fixes: 463e748964f5 ("net/hns3: support mailbox")

Signed-off-by: Dengdui Huang <huangdengdui at huawei.com>
Signed-off-by: Jie Hai <haijie1 at huawei.com>
Acked-by: Huisong Li <lihuisong at huawei.com>
---
 drivers/net/hns3/hns3_cmd.c |  3 --
 drivers/net/hns3/hns3_mbx.c | 81 ++++++-------------------------------
 drivers/net/hns3/hns3_mbx.h | 10 -----
 3 files changed, 13 insertions(+), 81 deletions(-)

diff --git a/drivers/net/hns3/hns3_cmd.c b/drivers/net/hns3/hns3_cmd.c
index 3495e2acc1..12f95ba4bc 100644
--- a/drivers/net/hns3/hns3_cmd.c
+++ b/drivers/net/hns3/hns3_cmd.c
@@ -664,7 +664,4 @@ hns3_cmd_init(struct hns3_hw *hw)
 	hw->cmq.crq.next_to_clean = 0;
 	hw->cmq.crq.next_to_use = 0;
-	hw->mbx_resp.head = 0;
-	hw->mbx_resp.tail = 0;
-	hw->mbx_resp.lost = 0;
 	hns3_cmd_init_regs(hw);
 
diff --git a/drivers/net/hns3/hns3_mbx.c b/drivers/net/hns3/hns3_mbx.c
index 8e0a58aa02..f1743c195e 100644
--- a/drivers/net/hns3/hns3_mbx.c
+++ b/drivers/net/hns3/hns3_mbx.c
@@ -41,21 +41,4 @@ hns3_resp_to_errno(uint16_t resp_code)
 }
 
-static void
-hns3_mbx_proc_timeout(struct hns3_hw *hw, uint16_t code, uint16_t subcode)
-{
-	if (hw->mbx_resp.matching_scheme ==
-	    HNS3_MBX_RESP_MATCHING_SCHEME_OF_ORIGINAL) {
-		hw->mbx_resp.lost++;
-		hns3_err(hw,
-			 "VF could not get mbx(%u,%u) head(%u) tail(%u) "
-			 "lost(%u) from PF",
-			 code, subcode, hw->mbx_resp.head, hw->mbx_resp.tail,
-			 hw->mbx_resp.lost);
-		return;
-	}
-
-	hns3_err(hw, "VF could not get mbx(%u,%u) from PF", code, subcode);
-}
-
 static int
 hns3_get_mbx_resp(struct hns3_hw *hw, uint16_t code, uint16_t subcode,
@@ -68,5 +51,4 @@ hns3_get_mbx_resp(struct hns3_hw *hw, uint16_t code, uint16_t subcode,
 	struct hns3_mbx_resp_status *mbx_resp;
 	uint32_t wait_time = 0;
-	bool received;
 
 	if (resp_len > HNS3_MBX_MAX_RESP_DATA_SIZE) {
@@ -94,11 +76,5 @@ hns3_get_mbx_resp(struct hns3_hw *hw, uint16_t code, uint16_t subcode,
 		rte_delay_us(HNS3_WAIT_RESP_US);
 
-		if (hw->mbx_resp.matching_scheme ==
-		    HNS3_MBX_RESP_MATCHING_SCHEME_OF_ORIGINAL)
-			received = (hw->mbx_resp.head ==
-				    hw->mbx_resp.tail + hw->mbx_resp.lost);
-		else
-			received = hw->mbx_resp.received_match_resp;
-		if (received)
+		if (hw->mbx_resp.received_match_resp)
 			break;
 
@@ -107,5 +83,5 @@ hns3_get_mbx_resp(struct hns3_hw *hw, uint16_t code, uint16_t subcode,
 	hw->mbx_resp.req_msg_data = 0;
 	if (wait_time >= mbx_time_limit) {
-		hns3_mbx_proc_timeout(hw, code, subcode);
+		hns3_err(hw, "VF could not get mbx(%u,%u) from PF", code, subcode);
 		return -ETIME;
 	}
@@ -133,5 +109,4 @@ hns3_mbx_prepare_resp(struct hns3_hw *hw, uint16_t code, uint16_t subcode)
 	 */
 	hw->mbx_resp.req_msg_data = (uint32_t)code << 16 | subcode;
-	hw->mbx_resp.head++;
 
 	/* Update match_id and ensure the value of match_id is not zero */
@@ -186,5 +161,4 @@ hns3_send_mbx_msg(struct hns3_hw *hw, uint16_t code, uint16_t subcode,
 		ret = hns3_cmd_send(hw, &desc, 1);
 		if (ret) {
-			hw->mbx_resp.head--;
 			rte_spinlock_unlock(&hw->mbx_resp.lock);
 			hns3_err(hw, "VF failed(=%d) to send mbx message to PF",
@@ -255,39 +229,8 @@ hns3_handle_asserting_reset(struct hns3_hw *hw,
 }
 
-/*
- * Case1: receive response after timeout, req_msg_data
- *        is 0, not equal resp_msg, do lost--
- * Case2: receive last response during new send_mbx_msg,
- *	  req_msg_data is different with resp_msg, let
- *	  lost--, continue to wait for response.
- */
-static void
-hns3_update_resp_position(struct hns3_hw *hw, uint32_t resp_msg)
-{
-	struct hns3_mbx_resp_status *resp = &hw->mbx_resp;
-	uint32_t tail = resp->tail + 1;
-
-	if (tail > resp->head)
-		tail = resp->head;
-	if (resp->req_msg_data != resp_msg) {
-		if (resp->lost)
-			resp->lost--;
-		hns3_warn(hw, "Received a mismatched response req_msg(%x) "
-			  "resp_msg(%x) head(%u) tail(%u) lost(%u)",
-			  resp->req_msg_data, resp_msg, resp->head, tail,
-			  resp->lost);
-	} else if (tail + resp->lost > resp->head) {
-		resp->lost--;
-		hns3_warn(hw, "Received a new response again resp_msg(%x) "
-			  "head(%u) tail(%u) lost(%u)", resp_msg,
-			  resp->head, tail, resp->lost);
-	}
-	rte_io_wmb();
-	resp->tail = tail;
-}
-
 static void
 hns3_handle_mbx_response(struct hns3_hw *hw, struct hns3_mbx_pf_to_vf_cmd *req)
 {
+#define HNS3_MBX_RESP_CODE_OFFSET 16
 	struct hns3_mbx_resp_status *resp = &hw->mbx_resp;
 	uint32_t msg_data;
@@ -299,10 +242,4 @@ hns3_handle_mbx_response(struct hns3_hw *hw, struct hns3_mbx_pf_to_vf_cmd *req)
 		 * to match the request.
 		 */
-		if (resp->matching_scheme !=
-		    HNS3_MBX_RESP_MATCHING_SCHEME_OF_MATCH_ID) {
-			resp->matching_scheme =
-				HNS3_MBX_RESP_MATCHING_SCHEME_OF_MATCH_ID;
-			hns3_info(hw, "detect mailbox support match id!");
-		}
 		if (req->match_id == resp->match_id) {
 			resp->resp_status = hns3_resp_to_errno(req->msg[3]);
@@ -320,9 +257,17 @@ hns3_handle_mbx_response(struct hns3_hw *hw, struct hns3_mbx_pf_to_vf_cmd *req)
 	 * original scheme to process.
 	 */
+	msg_data = (uint32_t)req->msg[1] << HNS3_MBX_RESP_CODE_OFFSET | req->msg[2];
+	if (resp->req_msg_data != msg_data) {
+		hns3_warn(hw,
+			"received response tag (%u) is mismatched with requested tag (%u)",
+			msg_data, resp->req_msg_data);
+		return;
+	}
+
 	resp->resp_status = hns3_resp_to_errno(req->msg[3]);
 	memcpy(resp->additional_info, &req->msg[4],
 	       HNS3_MBX_MAX_RESP_DATA_SIZE);
-	msg_data = (uint32_t)req->msg[1] << 16 | req->msg[2];
-	hns3_update_resp_position(hw, msg_data);
+	rte_io_wmb();
+	resp->received_match_resp = true;
 }
 
diff --git a/drivers/net/hns3/hns3_mbx.h b/drivers/net/hns3/hns3_mbx.h
index 0172a2e288..1d9a788df5 100644
--- a/drivers/net/hns3/hns3_mbx.h
+++ b/drivers/net/hns3/hns3_mbx.h
@@ -90,19 +90,9 @@ enum hns3_mbx_link_fail_subcode {
 #define HNS3_MBX_DEF_TIME_LIMIT_MS	500
 
-enum {
-	HNS3_MBX_RESP_MATCHING_SCHEME_OF_ORIGINAL = 0,
-	HNS3_MBX_RESP_MATCHING_SCHEME_OF_MATCH_ID
-};
-
 struct hns3_mbx_resp_status {
 	rte_spinlock_t lock; /* protects against contending sync cmd resp */
 
-	uint8_t matching_scheme;
-
 	/* The following fields used in the matching scheme for original */
 	uint32_t req_msg_data;
-	uint32_t head;
-	uint32_t tail;
-	uint32_t lost;
 
 	/* The following fields used in the matching scheme for match_id */
-- 
2.41.0

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2023-11-16 13:21:53.853050242 +0000
+++ 0055-net-hns3-fix-mailbox-sync.patch	2023-11-16 13:21:52.533946617 +0000
@@ -1 +1 @@
-From be3590f54d0e415c23d4ed6ea55d967139c3ad10 Mon Sep 17 00:00:00 2001
+From 1d9d4b5b633d2349f711750dfdf1ef65e9809db1 Mon Sep 17 00:00:00 2001
@@ -5,0 +6,2 @@
+[ upstream commit be3590f54d0e415c23d4ed6ea55d967139c3ad10 ]
+
@@ -42 +43,0 @@
-Cc: stable at dpdk.org
@@ -54 +55 @@
-index a5c4c11dc8..2c1664485b 100644
+index 3495e2acc1..12f95ba4bc 100644
@@ -57 +58 @@
-@@ -732,7 +732,4 @@ hns3_cmd_init(struct hns3_hw *hw)
+@@ -664,7 +664,4 @@ hns3_cmd_init(struct hns3_hw *hw)
@@ -202 +203 @@
-index c378783c6c..4a328802b9 100644
+index 0172a2e288..1d9a788df5 100644
@@ -205 +206 @@
-@@ -94,19 +94,9 @@ enum hns3_mbx_link_fail_subcode {
+@@ -90,19 +90,9 @@ enum hns3_mbx_link_fail_subcode {



More information about the stable mailing list