[PATCH 0/4] debug enhancement and bug fix

[PATCH 0/4] debug enhancement and bug fix

[Chaoyong He]

This patch series includes two bug fix commits and two debug enhancement
commits.

Huaxing Zhu (2):
  net/nfp: add more logs to debug probe process
  net/nfp: clear errors status of aer after soft reset

Shihong Wang (1):
  net/nfp: do not set the IPv6 flag in transport mode

Zerun Fu (1):
  net/nfp: notify flower firmware about PF speed

 .mailmap                                      |  1 +
 .../net/nfp/flower/nfp_flower_representor.c   |  3 ++
 drivers/net/nfp/nfp_ethdev.c                  | 46 +++++++++++++++++--
 drivers/net/nfp/nfp_ipsec.c                   | 15 +-----
 drivers/net/nfp/nfp_net_common.c              |  2 +-
 drivers/net/nfp/nfp_net_common.h              |  2 +
 drivers/net/nfp/nfpcore/nfp_mutex.c           | 17 +++++--
 drivers/net/nfp/nfpcore/nfp_nsp.c             | 15 ++++--
 drivers/net/nfp/nfpcore/nfp_nsp_eth.c         |  5 +-
 drivers/net/nfp/nfpcore/nfp_resource.c        |  4 +-
 10 files changed, 83 insertions(+), 27 deletions(-)

-- 
2.39.1

[PATCH 1/4] net/nfp: notify flower firmware about PF speed

[Chaoyong He]

From: Zerun Fu <zerun.fu@corigine.com>

When using flower firmware, the VF speed is obtained from the
firmware and the firmware get the VF speed from the PF.

But the previous logic does not notify the firmware about PF speed,
and this cause VF speed to be unavailable.

Fix this by add the logic of notify firmware about PF speed.

Fixes: e1124c4f8a45 ("net/nfp: add flower representor framework")
Cc: chaoyong.he@corigine.com
Cc: stable@dpdk.org

Signed-off-by: Zerun Fu <zerun.fu@corigine.com>
Reviewed-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Long Wu <long.wu@corigine.com>
Reviewed-by: Peng Zhang <peng.zhang@corigine.com>
---
 drivers/net/nfp/flower/nfp_flower_representor.c | 3 +++
 drivers/net/nfp/nfp_net_common.c                | 2 +-
 drivers/net/nfp/nfp_net_common.h                | 2 ++
 3 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/net/nfp/flower/nfp_flower_representor.c b/drivers/net/nfp/flower/nfp_flower_representor.c
index eb0a02874b..eae6ba39e1 100644
--- a/drivers/net/nfp/flower/nfp_flower_representor.c
+++ b/drivers/net/nfp/flower/nfp_flower_representor.c
@@ -37,6 +37,9 @@ nfp_flower_repr_link_update(struct rte_eth_dev *dev,
 
 	ret = nfp_net_link_update_common(dev, link, link->link_status);
 
+	if (repr->repr_type == NFP_REPR_TYPE_PF)
+		nfp_net_notify_port_speed(repr->app_fw_flower->pf_hw, link);
+
 	return ret;
 }
 
diff --git a/drivers/net/nfp/nfp_net_common.c b/drivers/net/nfp/nfp_net_common.c
index b986ed4622..f76d5a6895 100644
--- a/drivers/net/nfp/nfp_net_common.c
+++ b/drivers/net/nfp/nfp_net_common.c
@@ -184,7 +184,7 @@ nfp_net_link_speed_nfp2rte_check(uint32_t speed)
 	return RTE_ETH_SPEED_NUM_NONE;
 }
 
-static void
+void
 nfp_net_notify_port_speed(struct nfp_net_hw *hw,
 		struct rte_eth_link *link)
 {
diff --git a/drivers/net/nfp/nfp_net_common.h b/drivers/net/nfp/nfp_net_common.h
index 8429db68f0..d4fe8338b9 100644
--- a/drivers/net/nfp/nfp_net_common.h
+++ b/drivers/net/nfp/nfp_net_common.h
@@ -383,6 +383,8 @@ int nfp_net_vf_config_app_init(struct nfp_net_hw *net_hw,
 bool nfp_net_version_check(struct nfp_hw *hw,
 		struct nfp_pf_dev *pf_dev);
 void nfp_net_ctrl_bar_size_set(struct nfp_pf_dev *pf_dev);
+void nfp_net_notify_port_speed(struct nfp_net_hw *hw,
+		struct rte_eth_link *link);
 
 #define NFP_PRIV_TO_APP_FW_NIC(app_fw_priv)\
 	((struct nfp_app_fw_nic *)app_fw_priv)
-- 
2.39.1

[PATCH 2/4] net/nfp: do not set the IPv6 flag in transport mode

[Chaoyong He]

From: Shihong Wang <shihong.wang@corigine.com>

The transport only encapsulates the security protocol header,
does not pay attention to the IP protocol type, and need not
to set the IPv6 flag.

Fixes: 3d21da66c06b ("net/nfp: create security session")
Cc: stable@dpdk.org

Signed-off-by: Shihong Wang <shihong.wang@corigine.com>
Reviewed-by: Long Wu <long.wu@corigine.com>
Reviewed-by: Peng Zhang <peng.zhang@corigine.com>
Reviewed-by: Chaoyong He <chaoyong.he@corigine.com>
---
 drivers/net/nfp/nfp_ipsec.c | 15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

diff --git a/drivers/net/nfp/nfp_ipsec.c b/drivers/net/nfp/nfp_ipsec.c
index 647bc2bb6d..89116af1b2 100644
--- a/drivers/net/nfp/nfp_ipsec.c
+++ b/drivers/net/nfp/nfp_ipsec.c
@@ -1056,20 +1056,9 @@ nfp_ipsec_msg_build(struct rte_eth_dev *eth_dev,
 
 		break;
 	case RTE_SECURITY_IPSEC_SA_MODE_TRANSPORT:
-		type = conf->ipsec.tunnel.type;
 		cfg->ctrl_word.mode = NFP_IPSEC_MODE_TRANSPORT;
-		if (type == RTE_SECURITY_IPSEC_TUNNEL_IPV4) {
-			memset(&cfg->src_ip, 0, sizeof(cfg->src_ip));
-			memset(&cfg->dst_ip, 0, sizeof(cfg->dst_ip));
-			cfg->ipv6 = 0;
-		} else if (type == RTE_SECURITY_IPSEC_TUNNEL_IPV6) {
-			memset(&cfg->src_ip, 0, sizeof(cfg->src_ip));
-			memset(&cfg->dst_ip, 0, sizeof(cfg->dst_ip));
-			cfg->ipv6 = 1;
-		} else {
-			PMD_DRV_LOG(ERR, "Unsupported address family!");
-			return -EINVAL;
-		}
+		memset(&cfg->src_ip, 0, sizeof(cfg->src_ip));
+		memset(&cfg->dst_ip, 0, sizeof(cfg->dst_ip));
 
 		break;
 	default:
-- 
2.39.1

[PATCH 3/4] net/nfp: add more logs to debug probe process

[Chaoyong He]

From: Huaxing Zhu <huaxing.zhu@corigine.com>

Add more logs to debug probe process, and modify some log level.
Also add an new entry to the mailmap file.

Signed-off-by: Huaxing Zhu <huaxing.zhu@corigine.com>
Reviewed-by: Long Wu <long.wu@corigine.com>
Reviewed-by: Peng Zhang <peng.zhang@corigine.com>
Reviewed-by: Chaoyong He <chaoyong.he@corigine.com>
---
 .mailmap                               |  1 +
 drivers/net/nfp/nfp_ethdev.c           |  6 +++++-
 drivers/net/nfp/nfpcore/nfp_mutex.c    | 17 +++++++++++++----
 drivers/net/nfp/nfpcore/nfp_nsp.c      | 15 +++++++++++----
 drivers/net/nfp/nfpcore/nfp_nsp_eth.c  |  5 ++++-
 drivers/net/nfp/nfpcore/nfp_resource.c |  4 +++-
 6 files changed, 37 insertions(+), 11 deletions(-)

diff --git a/.mailmap b/.mailmap
index ec12c42dd4..a4439c9020 100644
--- a/.mailmap
+++ b/.mailmap
@@ -561,6 +561,7 @@ Hrvoje Habjanic <hrvoje.habjanic@zg.ht.hr>
 Huaibin Wang <huaibin.wang@6wind.com>
 Huanle Han <hanxueluo@gmail.com>
 Huawei Xie <huawei.xhw@alibaba-inc.com> <huawei.xie@intel.com>
+Huaxing Zhu <huaxing.zhu@corigine.com>
 Huichao Cai <chcchc88@163.com>
 Huilong Xu <huilongx.xu@intel.com>
 Huisong Li <lihuisong@huawei.com>
diff --git a/drivers/net/nfp/nfp_ethdev.c b/drivers/net/nfp/nfp_ethdev.c
index a199988999..8cbbadb4de 100644
--- a/drivers/net/nfp/nfp_ethdev.c
+++ b/drivers/net/nfp/nfp_ethdev.c
@@ -1297,13 +1297,17 @@ nfp_fw_upload(struct nfp_nsp *nsp,
 static void
 nfp_fw_unload(struct nfp_cpp *cpp)
 {
+	int err;
 	struct nfp_nsp *nsp;
 
 	nsp = nfp_nsp_open(cpp);
 	if (nsp == NULL)
 		return;
 
-	nfp_nsp_device_soft_reset(nsp);
+	err = nfp_nsp_device_soft_reset(nsp);
+	if (err != 0)
+		PMD_DRV_LOG(WARNING, "Failed to do soft reset when nfp fw unload.");
+
 	nfp_nsp_close(nsp);
 }
 
diff --git a/drivers/net/nfp/nfpcore/nfp_mutex.c b/drivers/net/nfp/nfpcore/nfp_mutex.c
index edb78dfdc9..9a4dc1e14e 100644
--- a/drivers/net/nfp/nfpcore/nfp_mutex.c
+++ b/drivers/net/nfp/nfpcore/nfp_mutex.c
@@ -308,11 +308,16 @@ nfp_cpp_mutex_trylock(struct nfp_cpp_mutex *mutex)
 
 	/* Verify that the lock marker is not damaged */
 	err = nfp_cpp_readl(cpp, mur, mutex->address + 4, &key);
-	if (err < 0)
+	if (err < 0) {
+		PMD_DRV_LOG(ERR, "Failed to read key.");
 		return err;
+	}
 
-	if (key != mutex->key)
+	if (key != mutex->key) {
+		PMD_DRV_LOG(ERR, "Key: %x is not same with the mutex: %x.",
+				key, mutex->key);
 		return -EPERM;
+	}
 
 	/*
 	 * Compare against the unlocked state, and if true,
@@ -335,8 +340,10 @@ nfp_cpp_mutex_trylock(struct nfp_cpp_mutex *mutex)
 	 * atomic, which returns the original value.
 	 */
 	err = nfp_cpp_readl(cpp, mus, mutex->address, &tmp);
-	if (err < 0)
+	if (err < 0) {
+		PMD_DRV_LOG(ERR, "Failed to read tmp.");
 		return err;
+	}
 
 	/* Was it unlocked? */
 	if (nfp_mutex_is_unlocked(tmp)) {
@@ -350,8 +357,10 @@ nfp_cpp_mutex_trylock(struct nfp_cpp_mutex *mutex)
 		 * debug and bookkeeping.
 		 */
 		err = nfp_cpp_writel(cpp, muw, mutex->address, value);
-		if (err < 0)
+		if (err < 0) {
+			PMD_DRV_LOG(ERR, "Failed to write value.");
 			return err;
+		}
 
 		mutex->depth = 1;
 		return 0;
diff --git a/drivers/net/nfp/nfpcore/nfp_nsp.c b/drivers/net/nfp/nfpcore/nfp_nsp.c
index 2ac39b10b5..32f092eda1 100644
--- a/drivers/net/nfp/nfpcore/nfp_nsp.c
+++ b/drivers/net/nfp/nfpcore/nfp_nsp.c
@@ -252,6 +252,7 @@ nfp_nsp_open(struct nfp_cpp *cpp)
 
 	state = malloc(sizeof(*state));
 	if (state == NULL) {
+		PMD_DRV_LOG(ERR, "NSP - failed to malloc name %s", NFP_RESOURCE_NSP);
 		nfp_resource_release(res);
 		return NULL;
 	}
@@ -369,8 +370,10 @@ nfp_nsp_command_real(struct nfp_nsp *state,
 	}
 
 	err = nfp_cpp_writeq(cpp, nsp_cpp, nsp_buffer, arg->buf);
-	if (err < 0)
+	if (err < 0) {
+		PMD_DRV_LOG(ERR, "CPP write buffer failed. err %d", err);
 		return err;
+	}
 
 	err = nfp_cpp_writeq(cpp, nsp_cpp, nsp_command,
 			FIELD_PREP(NSP_COMMAND_OPTION, arg->option) |
@@ -378,8 +381,10 @@ nfp_nsp_command_real(struct nfp_nsp *state,
 			FIELD_PREP(NSP_COMMAND_CODE, arg->code) |
 			FIELD_PREP(NSP_COMMAND_DMA_BUF, arg->dma) |
 			FIELD_PREP(NSP_COMMAND_START, 1));
-	if (err < 0)
+	if (err < 0) {
+		PMD_DRV_LOG(ERR, "CPP write command failed. err %d", err);
 		return err;
+	}
 
 	/* Wait for NSP_COMMAND_START to go to 0 */
 	err = nfp_nsp_wait_reg(cpp, &reg, nsp_cpp, nsp_command,
@@ -400,15 +405,17 @@ nfp_nsp_command_real(struct nfp_nsp *state,
 	}
 
 	err = nfp_cpp_readq(cpp, nsp_cpp, nsp_command, &ret_val);
-	if (err < 0)
+	if (err < 0) {
+		PMD_DRV_LOG(ERR, "CPP read return value failed. err %d", err);
 		return err;
+	}
 
 	ret_val = FIELD_GET(NSP_COMMAND_OPTION, ret_val);
 
 	err = FIELD_GET(NSP_STATUS_RESULT, reg);
 	if (err != 0) {
 		if (!arg->error_quiet)
-			PMD_DRV_LOG(WARNING, "Result (error) code set: %d (%d) command: %d",
+			PMD_DRV_LOG(ERR, "Result (error) code set: %d (%d) command: %d",
 					-err, (int)ret_val, arg->code);
 
 		if (arg->error_cb != 0)
diff --git a/drivers/net/nfp/nfpcore/nfp_nsp_eth.c b/drivers/net/nfp/nfpcore/nfp_nsp_eth.c
index 87db928c08..d1b43a75f0 100644
--- a/drivers/net/nfp/nfpcore/nfp_nsp_eth.c
+++ b/drivers/net/nfp/nfpcore/nfp_nsp_eth.c
@@ -347,12 +347,15 @@ nfp_eth_config_start(struct nfp_cpp *cpp,
 	union eth_table_entry *entries;
 
 	entries = malloc(NSP_ETH_TABLE_SIZE);
-	if (entries == NULL)
+	if (entries == NULL) {
+		PMD_DRV_LOG(ERR, "Malloc entries failed.");
 		return NULL;
+	}
 
 	memset(entries, 0, NSP_ETH_TABLE_SIZE);
 	nsp = nfp_nsp_open(cpp);
 	if (nsp == NULL) {
+		PMD_DRV_LOG(ERR, "NSP open failed.");
 		free(entries);
 		return nsp;
 	}
diff --git a/drivers/net/nfp/nfpcore/nfp_resource.c b/drivers/net/nfp/nfpcore/nfp_resource.c
index d59d2d6c1e..b05144036a 100644
--- a/drivers/net/nfp/nfpcore/nfp_resource.c
+++ b/drivers/net/nfp/nfpcore/nfp_resource.c
@@ -161,8 +161,10 @@ nfp_resource_acquire(struct nfp_cpp *cpp,
 	struct nfp_cpp_mutex *dev_mutex;
 
 	res = malloc(sizeof(*res));
-	if (res == NULL)
+	if (res == NULL) {
+		PMD_DRV_LOG(ERR, "RESOURCE - Malloc NSP memory failed.");
 		return NULL;
+	}
 
 	memset(res, 0, sizeof(*res));
 
-- 
2.39.1

[PATCH 4/4] net/nfp: clear errors status of aer after soft reset

[Chaoyong He]

From: Huaxing Zhu <huaxing.zhu@corigine.com>

Accessing device memory during soft reset may result in some
errors being recorded in PCIE's AER register, which is normal.
Therefore, after the soft reset is completed, these errors
should be cleared.

Signed-off-by: Huaxing Zhu <huaxing.zhu@corigine.com>
Reviewed-by: Long Wu <long.wu@corigine.com>
Reviewed-by: Peng Zhang <peng.zhang@corigine.com>
Reviewed-by: Chaoyong He <chaoyong.he@corigine.com>
---
 drivers/net/nfp/nfp_ethdev.c | 40 ++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_ethdev.c b/drivers/net/nfp/nfp_ethdev.c
index 8cbbadb4de..b16fbe7db7 100644
--- a/drivers/net/nfp/nfp_ethdev.c
+++ b/drivers/net/nfp/nfp_ethdev.c
@@ -1338,9 +1338,37 @@ nfp_fw_check_change(struct nfp_cpp *cpp,
 	return 0;
 }
 
+static void
+nfp_pcie_reg32_write_clear(struct rte_pci_device *pci_dev,
+		int position)
+{
+	int ret;
+	uint32_t capability;
+
+	ret = rte_pci_read_config(pci_dev, &capability, 4, position);
+	if (ret < 0)
+		capability = 0xffffffff;
+
+	(void)rte_pci_write_config(pci_dev, &capability, 4, position);
+}
+
+static void
+nfp_pcie_aer_clear(struct rte_pci_device *pci_dev)
+{
+	int pos;
+
+	pos = rte_pci_find_ext_capability(pci_dev, RTE_PCI_EXT_CAP_ID_ERR);
+	if (pos <= 0)
+		return;
+
+	nfp_pcie_reg32_write_clear(pci_dev, pos + RTE_PCI_ERR_UNCOR_STATUS);
+	nfp_pcie_reg32_write_clear(pci_dev, pos + RTE_PCI_ERR_COR_STATUS);
+}
+
 static int
 nfp_fw_reload(struct nfp_nsp *nsp,
 		char *fw_name,
+		struct rte_pci_device *pci_dev,
 		int reset)
 {
 	int err;
@@ -1357,6 +1385,14 @@ nfp_fw_reload(struct nfp_nsp *nsp,
 		}
 	}
 
+	/*
+	 * Accessing device memory during soft reset may result in some
+	 * errors being recorded in PCIE's AER register, which is normal.
+	 * Therefore, after the soft reset is completed, these errors
+	 * should be cleared.
+	 */
+	nfp_pcie_aer_clear(pci_dev);
+
 	err = nfp_fw_upload(nsp, fw_name);
 	if (err != 0) {
 		PMD_DRV_LOG(ERR, "NFP firmware load failed");
@@ -1463,7 +1499,7 @@ nfp_fw_reload_for_single_pf_from_disk(struct nfp_nsp *nsp,
 	if (!fw_changed)
 		return 0;
 
-	ret = nfp_fw_reload(nsp, fw_name, reset);
+	ret = nfp_fw_reload(nsp, fw_name, pf_dev->pci_dev, reset);
 	if (ret != 0)
 		return ret;
 
@@ -1523,7 +1559,7 @@ nfp_fw_reload_for_multi_pf_from_disk(struct nfp_nsp *nsp,
 	if (skip_load_fw && !reload_fw)
 		return 0;
 
-	err = nfp_fw_reload(nsp, fw_name, reset);
+	err = nfp_fw_reload(nsp, fw_name, pf_dev->pci_dev, reset);
 	if (err != 0)
 		return err;
 
-- 
2.39.1

Re: [PATCH 0/4] debug enhancement and bug fix

[Ferruh Yigit]

On 10/14/2024 3:43 AM, Chaoyong He wrote:
> This patch series includes two bug fix commits and two debug enhancement
> commits.
> 
> Huaxing Zhu (2):
>   net/nfp: add more logs to debug probe process
>   net/nfp: clear errors status of aer after soft reset
> 
> Shihong Wang (1):
>   net/nfp: do not set the IPv6 flag in transport mode
> 
> Zerun Fu (1):
>   net/nfp: notify flower firmware about PF speed
>

Series applied to dpdk-next-net/main, thanks.