From: Cunming Liang <cunming.liang@intel.com>
To: dev@dpdk.org
Subject: [dpdk-dev] [PATCH v1 06/11] ixgbe: fix rx intr compatible issue with PF mbox
Date: Thu, 24 Sep 2015 13:33:46 +0800 [thread overview]
Message-ID: <1443072831-19065-7-git-send-email-cunming.liang@intel.com> (raw)
In-Reply-To: <1443072831-19065-1-git-send-email-cunming.liang@intel.com>
When ixgbe runs as a PF, mbox interrupt is prerequisite to make VF start normally.
And PF sometimes won't 'dev_start', so the mbox interrupt register during 'dev_init' is required.
The patch rolls back the interrupt register for mbox,lsc to the 'dev_init'.
As UIO doesn't support multiple vector, mbox has to occupy the only one.
It adds condition check on 'dev_start', rxq interrupt is not allowed when PF running in IOV mode via UIO.
Signed-off-by: Cunming Liang <cunming.liang@intel.com>
---
drivers/net/ixgbe/ixgbe_ethdev.c | 45 +++++++++++++++++++++++++++-------------
1 file changed, 31 insertions(+), 14 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index dbc0cd4..f180d75 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -1028,6 +1028,13 @@ eth_ixgbe_dev_init(struct rte_eth_dev *eth_dev)
eth_dev->data->port_id, pci_dev->id.vendor_id,
pci_dev->id.device_id);
+ rte_intr_callback_register(&(pci_dev->intr_handle),
+ ixgbe_dev_interrupt_handler,
+ (void *)eth_dev);
+
+ /* enable uio/vfio intr/eventfd mapping */
+ rte_intr_enable(&(pci_dev->intr_handle));
+
/* enable support intr */
ixgbe_enable_intr(eth_dev);
@@ -1704,17 +1711,19 @@ ixgbe_dev_start(struct rte_eth_dev *dev)
ixgbe_pf_host_configure(dev);
/* check and configure queue intr-vector mapping */
- if (dev->data->dev_conf.intr_conf.rxq != 0)
+ if (((RTE_ETH_DEV_SRIOV(dev).active &&
+ rte_intr_cap_multiple(intr_handle)) ||
+ !RTE_ETH_DEV_SRIOV(dev).active) &&
+ dev->data->dev_conf.intr_conf.rxq != 0) {
intr_vector = dev->data->nb_rx_queues;
-
- if (rte_intr_efd_enable(intr_handle, intr_vector))
- return -1;
+ if (rte_intr_efd_enable(intr_handle, intr_vector))
+ return -1;
+ }
if (rte_intr_dp_is_en(intr_handle) && !intr_handle->intr_vec) {
intr_handle->intr_vec =
rte_zmalloc("intr_vec",
- dev->data->nb_rx_queues * sizeof(int),
- 0);
+ dev->data->nb_rx_queues * sizeof(int), 0);
if (intr_handle->intr_vec == NULL) {
PMD_INIT_LOG(ERR, "Failed to allocate %d rx_queues"
" intr_vec\n", dev->data->nb_rx_queues);
@@ -1801,20 +1810,22 @@ ixgbe_dev_start(struct rte_eth_dev *dev)
skip_link_setup:
- /* check if lsc interrupt is enabled */
- if (dev->data->dev_conf.intr_conf.lsc != 0) {
- if (rte_intr_allow_others(intr_handle)) {
- rte_intr_callback_register(intr_handle,
- ixgbe_dev_interrupt_handler,
- (void *)dev);
+ if (rte_intr_allow_others(intr_handle)) {
+ /* check if lsc interrupt is enabled */
+ if (dev->data->dev_conf.intr_conf.lsc != 0)
ixgbe_dev_lsc_interrupt_setup(dev);
- } else
+ } else {
+ rte_intr_callback_unregister(intr_handle,
+ ixgbe_dev_interrupt_handler,
+ (void *)dev);
+ if (dev->data->dev_conf.intr_conf.lsc != 0)
PMD_INIT_LOG(INFO, "lsc won't enable because of"
" no intr multiplex\n");
}
/* check if rxq interrupt is enabled */
- if (dev->data->dev_conf.intr_conf.rxq != 0)
+ if (dev->data->dev_conf.intr_conf.rxq != 0 &&
+ rte_intr_dp_is_en(intr_handle))
ixgbe_dev_rxq_interrupt_setup(dev);
/* enable uio/vfio intr/eventfd mapping */
@@ -1926,6 +1937,12 @@ ixgbe_dev_stop(struct rte_eth_dev *dev)
memset(filter_info->fivetuple_mask, 0,
sizeof(uint32_t) * IXGBE_5TUPLE_ARRAY_SIZE);
+ if (!rte_intr_allow_others(intr_handle))
+ /* resume to the default handler */
+ rte_intr_callback_register(intr_handle,
+ ixgbe_dev_interrupt_handler,
+ (void *)dev);
+
/* Clean datapath event and queue/vec mapping */
rte_intr_efd_disable(intr_handle);
if (intr_handle->intr_vec != NULL) {
--
2.4.3
next prev parent reply other threads:[~2015-09-24 5:34 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-24 5:33 [dpdk-dev] [PATCH v1 00/11] interrupt mode for i40e Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 01/11] eal/linux: vfio map misc intr to vector zero Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 02/11] ixgbe: reserve intr vector zero for misc cause Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 03/11] igb: " Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 04/11] eal/linux: not allow to enable zero intr efd Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 05/11] eal/linux: add intr api to report multi-vector capability Cunming Liang
2015-09-24 5:33 ` Cunming Liang [this message]
2015-11-02 16:03 ` [dpdk-dev] [PATCH v1 06/11] ixgbe: fix rx intr compatible issue with PF mbox David Marchand
2015-11-02 16:09 ` Ananyev, Konstantin
2015-11-02 16:22 ` David Marchand
[not found] ` <2601191342CEEE43887BDE71AB97725836AB862D@irsmsx105.ger.corp.intel.com>
2015-11-02 16:41 ` Ananyev, Konstantin
2015-11-02 17:06 ` David Marchand
2015-11-02 17:23 ` Ananyev, Konstantin
2015-11-02 17:45 ` David Marchand
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 07/11] ixgbevf: cleanup unnecessary interrupt handler Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 08/11] igb: fix rx intr compatible issue with PF mbox Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 09/11] i40e: add rx interrupt support Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 10/11] i40evf: " Cunming Liang
2015-09-24 5:33 ` [dpdk-dev] [PATCH v1 11/11] doc: release note update for intr mode Cunming Liang
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 00/11] interrupt mode for i40e Cunming Liang
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 01/11] eal/linux: vfio map misc intr to vector zero Cunming Liang
2015-10-30 7:11 ` He, Shaopeng
2015-10-30 7:33 ` Zhang, Helin
2015-10-30 14:22 ` Liang, Cunming
2015-11-02 15:53 ` David Marchand
2015-11-04 1:17 ` Liang, Cunming
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 00/13] interrupt mode for i40e Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 01/13] eal: vfio map misc intr to vector zero Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 02/13] ixgbe: reserve intr vector zero for misc cause Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 03/13] igb: " Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 04/13] eal/linux: not allow to enable zero intr efd Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 05/13] ixgbe: fix efd_enable with zero number Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 06/13] igb: " Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 07/13] eal: add intr api to report multi-vector capability Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 08/13] ixgbe: fix rx intr compatible issue with PF mbox Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 09/13] ixgbe: fix unnecessary intr_vec free in dev_close Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 10/13] ixgbevf: cleanup unnecessary interrupt handler Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 11/13] igb: fix rx intr compatible issue with PF mbox Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 12/13] i40e: add rx interrupt support Cunming Liang
2015-11-04 6:07 ` [dpdk-dev] [PATCH v3 13/13] i40evf: " Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 00/13] interrupt mode for i40e Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 01/13] eal: vfio map misc intr to vector zero Cunming Liang
2015-11-04 13:06 ` David Marchand
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 02/13] ixgbe: reserve intr vector zero for misc cause Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 03/13] igb: " Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 04/13] eal/linux: not allow to enable zero intr efd Cunming Liang
2015-11-04 13:06 ` David Marchand
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 05/13] ixgbe: fix efd_enable with zero number Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 06/13] igb: " Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 07/13] eal: add intr api to report multi-vector capability Cunming Liang
2015-11-04 13:07 ` David Marchand
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 08/13] ixgbe: fix rx intr compatible issue with PF mbox Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 09/13] ixgbe: fix unnecessary intr_vec free in dev_close Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 10/13] ixgbevf: cleanup unnecessary interrupt handler Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 11/13] igb: fix rx intr compatible issue with PF mbox Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 12/13] i40e: add rx interrupt support Cunming Liang
2015-11-04 8:45 ` [dpdk-dev] [PATCH v4 13/13] i40evf: " Cunming Liang
2015-11-04 14:29 ` [dpdk-dev] [PATCH v4 00/13] interrupt mode for i40e Thomas Monjalon
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 02/11] ixgbe: reserve intr vector zero for misc cause Cunming Liang
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 03/11] igb: " Cunming Liang
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 04/11] eal/linux: not allow to enable zero intr efd Cunming Liang
2015-10-30 7:11 ` He, Shaopeng
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 05/11] eal/linux: add intr api to report multi-vector capability Cunming Liang
2015-10-30 7:13 ` He, Shaopeng
2015-11-02 15:59 ` David Marchand
2015-11-04 1:21 ` Liang, Cunming
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 06/11] ixgbe: fix rx intr compatible issue with PF mbox Cunming Liang
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 07/11] ixgbevf: cleanup unnecessary interrupt handler Cunming Liang
2015-11-02 16:06 ` David Marchand
2015-11-04 1:37 ` Liang, Cunming
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 08/11] igb: fix rx intr compatible issue with PF mbox Cunming Liang
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 09/11] i40e: add rx interrupt support Cunming Liang
2015-10-30 7:33 ` Zhang, Helin
2015-10-30 7:35 ` Wu, Jingjing
2015-10-30 5:27 ` [dpdk-dev] [PATCH v2 10/11] i40evf: " Cunming Liang
2015-10-30 7:36 ` Wu, Jingjing
2015-10-30 7:38 ` Zhang, Helin
2015-10-30 5:27 ` [dpdk-dev] [PATCH 11/11] doc: release note update for intr mode Cunming Liang
2015-10-30 8:21 ` [dpdk-dev] [PATCH v2 00/11] interrupt mode for i40e Zhang, Helin
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=1443072831-19065-7-git-send-email-cunming.liang@intel.com \
--to=cunming.liang@intel.com \
--cc=dev@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).