From: Ajit Khaparde <ajit.khaparde@broadcom.com>
To: dev@dpdk.org
Cc: Somnath Kotur <somnath.kotur@broadcom.com>,
stable@dpdk.org,
Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>,
Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Subject: [dpdk-dev] [PATCH v2 25/25] net/bnxt: add separate mutex for FW health check
Date: Tue, 15 Sep 2020 21:28:51 -0700 [thread overview]
Message-ID: <20200916042851.32914-26-ajit.khaparde@broadcom.com> (raw)
In-Reply-To: <20200916042851.32914-1-ajit.khaparde@broadcom.com>
From: Somnath Kotur <somnath.kotur@broadcom.com>
def_cp_lock was added to sync race between dev_configure and
int_handler. It should not be used to synchronize scheduling of FW
health check between dev_start and async event handler as well,
use a separate mutex for the same.
Fixes: a73b8e939f10 ("net/bnxt: fix race between start and interrupt handler")
Cc: stable@dpdk.org
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
---
drivers/net/bnxt/bnxt.h | 1 +
drivers/net/bnxt/bnxt_ethdev.c | 16 ++++++++++++----
2 files changed, 13 insertions(+), 4 deletions(-)
diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index f0b080015..bb265999d 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -712,6 +712,7 @@ struct bnxt {
rte_iova_t hwrm_short_cmd_req_dma_addr;
rte_spinlock_t hwrm_lock;
pthread_mutex_t def_cp_lock;
+ pthread_mutex_t health_check_lock;
uint16_t max_req_len;
uint16_t max_resp_len;
uint16_t hwrm_max_ext_req_len;
diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c
index 7c27e2435..05e9a6abb 100644
--- a/drivers/net/bnxt/bnxt_ethdev.c
+++ b/drivers/net/bnxt/bnxt_ethdev.c
@@ -1252,9 +1252,7 @@ static int bnxt_dev_start_op(struct rte_eth_dev *eth_dev)
eth_dev->rx_pkt_burst = bnxt_receive_function(eth_dev);
eth_dev->tx_pkt_burst = bnxt_transmit_function(eth_dev);
- pthread_mutex_lock(&bp->def_cp_lock);
bnxt_schedule_fw_health_check(bp);
- pthread_mutex_unlock(&bp->def_cp_lock);
return 0;
@@ -4675,17 +4673,22 @@ void bnxt_schedule_fw_health_check(struct bnxt *bp)
{
uint32_t polling_freq;
+ pthread_mutex_lock(&bp->health_check_lock);
+
if (!bnxt_is_recovery_enabled(bp))
- return;
+ goto done;
if (bp->flags & BNXT_FLAG_FW_HEALTH_CHECK_SCHEDULED)
- return;
+ goto done;
polling_freq = bp->recovery_info->driver_polling_freq;
rte_eal_alarm_set(US_PER_MS * polling_freq,
bnxt_check_fw_health, (void *)bp);
bp->flags |= BNXT_FLAG_FW_HEALTH_CHECK_SCHEDULED;
+
+done:
+ pthread_mutex_unlock(&bp->health_check_lock);
}
static void bnxt_cancel_fw_health_check(struct bnxt *bp)
@@ -5473,6 +5476,10 @@ bnxt_init_locks(struct bnxt *bp)
err = pthread_mutex_init(&bp->def_cp_lock, NULL);
if (err)
PMD_DRV_LOG(ERR, "Unable to initialize def_cp_lock\n");
+
+ err = pthread_mutex_init(&bp->health_check_lock, NULL);
+ if (err)
+ PMD_DRV_LOG(ERR, "Unable to initialize health_check_lock\n");
return err;
}
@@ -5884,6 +5891,7 @@ bnxt_uninit_locks(struct bnxt *bp)
{
pthread_mutex_destroy(&bp->flow_lock);
pthread_mutex_destroy(&bp->def_cp_lock);
+ pthread_mutex_destroy(&bp->health_check_lock);
if (bp->rep_info) {
pthread_mutex_destroy(&bp->rep_info->vfr_lock);
pthread_mutex_destroy(&bp->rep_info->vfr_start_lock);
--
2.21.1 (Apple Git-122.3)
next prev parent reply other threads:[~2020-09-16 4:33 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-11 1:55 [dpdk-dev] [PATCH 00/25] patchset for bnxt Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 01/25] net/bnxt: fix port stop process and cleanup resources Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 02/25] net/bnxt: fix the drop action flow to support count action Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 03/25] net/bnxt: reject offload flows with invalid MAC address Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 04/25] net/bnxt: reduce debug log messages Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 05/25] net/bnxt: fix to break the ipv4 and ipv6 ingress rule Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 06/25] net/bnxt: free the em index on failure Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 07/25] net/bnxt: add a null ptr check for the resource manager Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 08/25] net/bnxt: change default flow rule to use 8B encap Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 09/25] net/bnxt: fix the function id used in the flow flush Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 10/25] net/bnxt: vfr port clean up during port stop Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 11/25] net/bnxt: fix crash in VF rep queue selection Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 12/25] net/bnxt: fix to conditionally rollback added VF-rep ports Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 13/25] net/bnxt: update resource allocation settings Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 14/25] net/bnxt: move IF tbl from tunneled to direct HWRM msg Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 15/25] net/bnxt: remove VLAN pop action for egress flows Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 16/25] net/bnxt: increase counter support from 8K to 16K Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 17/25] net/bnxt: fix to explicitly check and set for start cntr ID Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 18/25] net/bnxt: enable support for VXLAN ipv6 encapsulation Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 19/25] net/bnxt: enable support for nat action with tagged traffic Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 20/25] net/bnxt: fix out of bound access in action bit handling Ajit Khaparde
2020-09-11 1:55 ` [dpdk-dev] [PATCH 21/25] net/bnxt: provide switch info while VF-Reps are configured Ajit Khaparde
2020-09-11 1:56 ` [dpdk-dev] [PATCH 22/25] net/bnxt: fix bugs in representor data path Ajit Khaparde
2020-09-11 1:56 ` [dpdk-dev] [PATCH 23/25] net/bnxt: add support for locks in flow database Ajit Khaparde
2020-09-11 1:56 ` [dpdk-dev] [PATCH 24/25] net/bnxt: fix to check for vnic ptr in bnxt shutdown path Ajit Khaparde
2020-09-11 1:56 ` [dpdk-dev] [PATCH 25/25] net/bnxt: fix to have a separate mutex for FW health check Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 00/25] patchset for bnxt Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 01/25] net/bnxt: fix resource cleanup in port stop Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 02/25] net/bnxt: fix the drop action flow to support count Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 03/25] net/bnxt: reject flow offload with invalid MAC Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 04/25] net/bnxt: reduce debug log messages Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 05/25] net/bnxt: fix coexistence of ipv4 and ipv6 ingress rules Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 06/25] net/bnxt: free the EM index on failure Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 07/25] net/bnxt: add null pointer check for resource manager Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 08/25] net/bnxt: modify default flow rule creation Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 09/25] net/bnxt: fix the function id used in flow flush Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 10/25] net/bnxt: refactor VFR port clean up Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 11/25] net/bnxt: fix crash in VFR queue select Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 12/25] net/bnxt: fix VFR cleanup during init failure Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 13/25] net/bnxt: update resource settings Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 14/25] net/bnxt: use direct HWRM message for interface table Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 15/25] net/bnxt: remove VLAN pop action for egress flows Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 16/25] net/bnxt: increase counter support from 8K to 16K Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 17/25] net/bnxt: check and set initial counter ID Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 18/25] net/bnxt: enable VXLAN ipv6 encapsulation Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 19/25] net/bnxt: enable NAT action with tagged traffic Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 20/25] net/bnxt: fix out of bound access in bit handling Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 21/25] net/bnxt: provide switch info if VFR are configured Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 22/25] net/bnxt: fix bugs in representor data path Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 23/25] net/bnxt: add locks in flow database Ajit Khaparde
2020-09-16 4:28 ` [dpdk-dev] [PATCH v2 24/25] net/bnxt: fix to check VNIC in shutdown path Ajit Khaparde
2020-09-16 4:28 ` Ajit Khaparde [this message]
2020-09-16 16:21 ` [dpdk-dev] [PATCH v2 00/25] patchset for bnxt Ajit Khaparde
2020-09-16 23:57 ` Ferruh Yigit
2020-09-17 0:13 ` Ajit Khaparde
2020-09-17 7:39 ` Ferruh Yigit
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=20200916042851.32914-26-ajit.khaparde@broadcom.com \
--to=ajit.khaparde@broadcom.com \
--cc=dev@dpdk.org \
--cc=kalesh-anakkur.purayil@broadcom.com \
--cc=somnath.kotur@broadcom.com \
--cc=sriharsha.basavapatna@broadcom.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).