From: Chaoyong He <chaoyong.he@corigine.com>
To: dev@dpdk.org
Cc: oss-drivers@corigine.com, Peng Zhang <peng.zhang@corigine.com>,
stable@dpdk.org, Chaoyong He <chaoyong.he@corigine.com>,
Long Wu <long.wu@corigine.com>
Subject: [PATCH 04/23] net/nfp: fix repeat disable the port
Date: Wed, 19 Jun 2024 17:58:11 +0800 [thread overview]
Message-ID: <20240619095830.3479757-5-chaoyong.he@corigine.com> (raw)
In-Reply-To: <20240619095830.3479757-1-chaoyong.he@corigine.com>
From: Peng Zhang <peng.zhang@corigine.com>
For firmware with multiple PFs, all the PFs share
the same 'nfp_eth_table' data structure. So the original
logic loop the ports in 'nfp_eth_table' will make other
PFs suddently down, which will cause problem.
Fix this by adding the special logic for firmware with
multiple PFs.
Fixes: 3b00109d2b65 ("net/nfp: add PF ID used to format symbols")
Cc: stable@dpdk.org
Signed-off-by: Peng Zhang <peng.zhang@corigine.com>
Reviewed-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Long Wu <long.wu@corigine.com>
---
drivers/net/nfp/nfp_ethdev.c | 35 ++++++++++++++++++++++++++------
drivers/net/nfp/nfp_net_common.c | 10 +++++++++
drivers/net/nfp/nfp_net_common.h | 2 ++
3 files changed, 41 insertions(+), 6 deletions(-)
diff --git a/drivers/net/nfp/nfp_ethdev.c b/drivers/net/nfp/nfp_ethdev.c
index 181798e8e3..7479802a52 100644
--- a/drivers/net/nfp/nfp_ethdev.c
+++ b/drivers/net/nfp/nfp_ethdev.c
@@ -1773,6 +1773,30 @@ nfp_net_speed_capa_get(struct nfp_pf_dev *pf_dev,
return 0;
}
+/* Force the physical port down to clear the possible DMA error */
+static int
+nfp_net_force_port_down(struct nfp_pf_dev *pf_dev,
+ struct nfp_eth_table *nfp_eth_table,
+ struct nfp_cpp *cpp)
+{
+ int ret;
+ uint32_t i;
+ uint32_t id;
+ uint32_t index;
+ uint32_t count;
+
+ count = nfp_net_get_port_num(pf_dev, nfp_eth_table);
+ for (i = 0; i < count; i++) {
+ id = nfp_function_id_get(pf_dev, i);
+ index = nfp_eth_table->ports[id].index;
+ ret = nfp_eth_set_configured(cpp, index, 0);
+ if (ret < 0)
+ return ret;
+ }
+
+ return 0;
+}
+
static int
nfp_pf_init(struct rte_pci_device *pci_dev)
{
@@ -1781,7 +1805,6 @@ nfp_pf_init(struct rte_pci_device *pci_dev)
uint32_t id;
int ret = 0;
uint64_t addr;
- uint32_t index;
uint32_t cpp_id;
uint8_t function_id;
struct nfp_cpp *cpp;
@@ -1875,11 +1898,11 @@ nfp_pf_init(struct rte_pci_device *pci_dev)
pf_dev->multi_pf.enabled = nfp_check_multi_pf_from_nsp(pci_dev, cpp);
pf_dev->multi_pf.function_id = function_id;
- /* Force the physical port down to clear the possible DMA error */
- for (i = 0; i < nfp_eth_table->count; i++) {
- id = nfp_function_id_get(pf_dev, i);
- index = nfp_eth_table->ports[id].index;
- nfp_eth_set_configured(cpp, index, 0);
+ ret = nfp_net_force_port_down(pf_dev, nfp_eth_table, cpp);
+ if (ret != 0) {
+ PMD_INIT_LOG(ERR, "Failed to force port down");
+ ret = -EIO;
+ goto eth_table_cleanup;
}
ret = nfp_devargs_parse(&pf_dev->devargs, pci_dev->device.devargs);
diff --git a/drivers/net/nfp/nfp_net_common.c b/drivers/net/nfp/nfp_net_common.c
index 107306a4e3..d8b7045c3a 100644
--- a/drivers/net/nfp/nfp_net_common.c
+++ b/drivers/net/nfp/nfp_net_common.c
@@ -2554,3 +2554,13 @@ nfp_net_fec_set(struct rte_eth_dev *dev,
return nfp_eth_set_fec(hw_priv->pf_dev->cpp, eth_port->index, fec);
}
+
+uint32_t
+nfp_net_get_port_num(struct nfp_pf_dev *pf_dev,
+ struct nfp_eth_table *nfp_eth_table)
+{
+ if (pf_dev->multi_pf.enabled)
+ return 1;
+ else
+ return nfp_eth_table->count;
+}
diff --git a/drivers/net/nfp/nfp_net_common.h b/drivers/net/nfp/nfp_net_common.h
index 2feeb6f5bd..4016652cf9 100644
--- a/drivers/net/nfp/nfp_net_common.h
+++ b/drivers/net/nfp/nfp_net_common.h
@@ -326,6 +326,8 @@ void nfp_net_get_fw_version(struct nfp_cpp *cpp,
uint32_t *fw_version);
int nfp_net_txrwb_alloc(struct rte_eth_dev *eth_dev);
void nfp_net_txrwb_free(struct rte_eth_dev *eth_dev);
+uint32_t nfp_net_get_port_num(struct nfp_pf_dev *pf_dev,
+ struct nfp_eth_table *nfp_eth_table);
#define NFP_PRIV_TO_APP_FW_NIC(app_fw_priv)\
((struct nfp_app_fw_nic *)app_fw_priv)
--
2.39.1
next prev parent reply other threads:[~2024-06-19 9:59 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20240619095830.3479757-1-chaoyong.he@corigine.com>
2024-06-19 9:58 ` [PATCH 01/23] net/nfp: fix dereference of null pointer Chaoyong He
2024-07-07 18:56 ` Ferruh Yigit
2024-07-08 1:35 ` Chaoyong He
2024-06-19 9:58 ` [PATCH 02/23] net/nfp: disable ctrl VNIC queues Chaoyong He
2024-06-19 9:58 ` [PATCH 03/23] net/nfp: fix dereference of null pointer Chaoyong He
2024-06-19 9:58 ` Chaoyong He [this message]
2024-06-19 9:58 ` [PATCH 05/23] net/nfp: fix repeat set the speed configure Chaoyong He
2024-06-19 9:58 ` [PATCH 18/23] net/nfp: add check for numbers of VF representor port Chaoyong He
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240619095830.3479757-5-chaoyong.he@corigine.com \
--to=chaoyong.he@corigine.com \
--cc=dev@dpdk.org \
--cc=long.wu@corigine.com \
--cc=oss-drivers@corigine.com \
--cc=peng.zhang@corigine.com \
--cc=stable@dpdk.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).