patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Xiao Wang <xiao.w.wang@intel.com>
To: stable@dpdk.org
Cc: yuanhan.liu@linux.intel.com, Xiao Wang <xiao.w.wang@intel.com>
Subject: [dpdk-stable] [PATCH] net/fm10k: fix secondary process crash
Date: Mon, 22 May 2017 08:22:30 -0700	[thread overview]
Message-ID: <1495466550-64354-1-git-send-email-xiao.w.wang@intel.com> (raw)

[ backported from upstream commit
88e4ed70b3a0b279383ddac43ef231ec518d6b4d ]

If the primary process has initialized all the queues to vector
pmd mode, the secondary process should not use scalar code path,
because the per queue data structures haven't been prepared for
that, e.g. txq->ops is for vector Tx rather than scalar Tx.

Fixes: a6ce64a97520 ("fm10k: introduce vector driver")

Signed-off-by: Xiao Wang <xiao.w.wang@intel.com>
Acked-by: Jing Chen <jing.d.chen@intel.com>
---
 drivers/net/fm10k/fm10k_ethdev.c | 26 ++++++++++++++++++++++++--
 1 file changed, 24 insertions(+), 2 deletions(-)

diff --git a/drivers/net/fm10k/fm10k_ethdev.c b/drivers/net/fm10k/fm10k_ethdev.c
index cd051ae..a4eefe3 100644
--- a/drivers/net/fm10k/fm10k_ethdev.c
+++ b/drivers/net/fm10k/fm10k_ethdev.c
@@ -2738,6 +2738,19 @@ static void __attribute__((cold))
 	int use_sse = 1;
 	uint16_t tx_ftag_en = 0;
 
+	if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
+		/* primary process has set the ftag flag and txq_flags */
+		txq = dev->data->tx_queues[0];
+		if (fm10k_tx_vec_condition_check(txq)) {
+			dev->tx_pkt_burst = fm10k_xmit_pkts;
+			PMD_INIT_LOG(DEBUG, "Use regular Tx func");
+		} else {
+			PMD_INIT_LOG(DEBUG, "Use vector Tx func");
+			dev->tx_pkt_burst = fm10k_xmit_pkts_vec;
+		}
+		return;
+	}
+
 	if (fm10k_check_ftag(dev->pci_dev->device.devargs))
 		tx_ftag_en = 1;
 
@@ -2796,6 +2809,9 @@ static void __attribute__((cold))
 	else
 		PMD_INIT_LOG(DEBUG, "Use regular Rx func");
 
+	if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+		return;
+
 	for (i = 0; i < dev->data->nb_rx_queues; i++) {
 		struct fm10k_rx_queue *rxq = dev->data->rx_queues[i];
 
@@ -2839,9 +2855,15 @@ static void __attribute__((cold))
 	dev->rx_pkt_burst = &fm10k_recv_pkts;
 	dev->tx_pkt_burst = &fm10k_xmit_pkts;
 
-	/* only initialize in the primary process */
-	if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+	/*
+	 * Primary process does the whole initialization, for secondary
+	 * processes, we just select the same Rx and Tx function as primary.
+	 */
+	if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
+		fm10k_set_rx_function(dev);
+		fm10k_set_tx_function(dev);
 		return 0;
+	}
 
 	rte_eth_copy_pci_info(dev, dev->pci_dev);
 
-- 
1.8.3.1

             reply	other threads:[~2017-05-22  6:35 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-22 15:22 Xiao Wang [this message]
  -- strict thread matches above, loose matches on Subject: below --
2017-03-28  3:58 Xiao Wang
2017-03-31  1:30 ` Chen, Jing D
2017-04-01  0:20   ` Wang, Xiao W
2017-04-01  3:22 ` Chen, Jing D
2017-04-03  9:51   ` 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=1495466550-64354-1-git-send-email-xiao.w.wang@intel.com \
    --to=xiao.w.wang@intel.com \
    --cc=stable@dpdk.org \
    --cc=yuanhan.liu@linux.intel.com \
    /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).