Re: [dpdk-dev] [PATCH] bus/pci: fix mmap PCI resource
Tested-by: Xiao Qimai
Regards,
Xiao Qimai
> -Original Message-
> From: dev On Behalf Of alvinx.zh...@intel.com
> Sent: Wednesday, July 8, 2020 5:25 PM
> To: dev@dpdk.org
> Cc: Xing, Beilei ; Guo, Jia
> Subject: [dpdk-dev] [PATCH] bus/pci: fix mmap PCI resource
>
> From: Alvin Zhang
>
> When mapping a PCI BAR containing an MSI-X table, some devices do not
> need to actually map this BAR or only need to map part of them, which may
> cause the mapping to fail. Now some checks are added and a non-NULL initial
> value is set to the variable to avoid this situation.
>
> Fixes: 2fd3567e5425 ("pci: use OS generic memory mapping functions")
> Cc: tal...@mellanox.com
>
> Signed-off-by: Alvin Zhang
> ---
> drivers/bus/pci/linux/pci_vfio.c | 12 +++-
> 1 file changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/bus/pci/linux/pci_vfio.c
> b/drivers/bus/pci/linux/pci_vfio.c
> index fdeb9a8..9143bfc 100644
> --- a/drivers/bus/pci/linux/pci_vfio.c
> +++ b/drivers/bus/pci/linux/pci_vfio.c
> @@ -547,6 +547,14 @@
> bar_index,
> memreg[0].offset, memreg[0].size,
> memreg[1].offset, memreg[1].size);
> +
> + if (memreg[0].size == 0 && memreg[1].size == 0) {
> + /* No need to map this BAR */
> + RTE_LOG(DEBUG, EAL, "Skipping BAR%d\n",
> bar_index);
> + bar->size = 0;
> + bar->addr = 0;
> + return 0;
> + }
> } else {
> memreg[0].offset = bar->offset;
> memreg[0].size = bar->size;
> @@ -556,7 +564,9 @@
> bar_addr = mmap(bar->addr, bar->size, 0, MAP_PRIVATE |
> MAP_ANONYMOUS | additional_flags, -1, 0);
> if (bar_addr != MAP_FAILED) {
> - void *map_addr = NULL;
> + /* Set non NULL initial value for in case of no PCI mapping */
> + void *map_addr = bar_addr;
> +
> if (memreg[0].size) {
> /* actual map of first part */
> map_addr = pci_map_resource(bar_addr,
> vfio_dev_fd,
> --
> 1.8.3.1
Re: [dpdk-dev] [PATCH] net/ice: fix core dumped issue in switch filter
Tested-by: Xiao, QimaiX
Regards,
Xiao Qimai
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Junyu Jiang
> Sent: Wednesday, April 29, 2020 3:25 PM
> To: dev@dpdk.org
> Cc: Yang, Qiming ; Jiang, JunyuX
> ; sta...@dpdk.org
> Subject: [dpdk-dev] [PATCH] net/ice: fix core dumped issue in switch filter
>
> The number of queues in queue group should be checked before using it to
> avoid NULL pointer. This patch fixed the issue.
>
> Fixes: 47d460d63233 ("net/ice: rework switch filter")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Junyu Jiang
> ---
> drivers/net/ice/ice_switch_filter.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/net/ice/ice_switch_filter.c
> b/drivers/net/ice/ice_switch_filter.c
> index 179430136..c2762e331 100644
> --- a/drivers/net/ice/ice_switch_filter.c
> +++ b/drivers/net/ice/ice_switch_filter.c
> @@ -1296,6 +1296,8 @@ ice_switch_parse_action(struct ice_pf *pf,
> switch (action_type) {
> case RTE_FLOW_ACTION_TYPE_RSS:
> act_qgrop = action->conf;
> + if (act_qgrop->queue_num <= 1)
> + goto error;
> rule_info->sw_act.fltr_act =
> ICE_FWD_TO_QGRP;
> rule_info->sw_act.fwd_id.q_id =
> --
> 2.17.1
Re: [dpdk-dev] [PATCH] net/ice: fix switch action number check
Tested-by: Xiao, QimaiX
Regards,
Xiao Qimai
> -Original Message-
> From: dev On Behalf Of Wei Zhao
> Sent: Thursday, May 21, 2020 3:34 PM
> To: dev@dpdk.org
> Cc: sta...@dpdk.org; Zhang, Qi Z ; Ye, Xiaolong
> ; Zhao1, Wei
> Subject: [dpdk-dev] [PATCH] net/ice: fix switch action number check
>
> The action number for switch filter should be 1, any other such as 0 or more
> than 1 is invalid.
>
> Fixes: 3428c6b6ec1f ("net/ice: add action number check for switch")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Wei Zhao
> ---
> drivers/net/ice/ice_switch_filter.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/ice/ice_switch_filter.c
> b/drivers/net/ice/ice_switch_filter.c
> index dd3f4847a..20e8187d3 100644
> --- a/drivers/net/ice/ice_switch_filter.c
> +++ b/drivers/net/ice/ice_switch_filter.c
> @@ -1388,7 +1388,7 @@ ice_switch_check_action(const struct
> rte_flow_action *actions,
> }
> }
>
> - if (actions_num > 1) {
> + if (actions_num != 1) {
> rte_flow_error_set(error,
> EINVAL, RTE_FLOW_ERROR_TYPE_ACTION,
> actions,
> --
> 2.19.1
Re: [dpdk-dev] [PATCH] vhost: fix packed ring zero-copy
Tested-by: Xiao, QimaiX
Regards,
Xiao Qimai
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Marvin Liu
> Sent: Monday, March 16, 2020 11:38 PM
> To: maxime.coque...@redhat.com; Wang, Zhihong
> ; Ye, Xiaolong
> Cc: dev@dpdk.org; sta...@dpdk.org; Liu, Yong
> Subject: [dpdk-dev] [PATCH] vhost: fix packed ring zero-copy
>
> Available buffer ID should be stored in the zmbuf in the packed-ring
> dequeue path. There's no guarantee that local queue avail index is equal to
> buffer ID.
>
> Fixes: d1eafb532268 ("vhost: add packed ring zcopy batch and single
> dequeue")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Marvin Liu
> Reported-by: Yinan Wang
> Reviewed-by: Maxime Coquelin
>
> diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c
> index
> 37c47c7dc..210415904 100644
> --- a/lib/librte_vhost/virtio_net.c
> +++ b/lib/librte_vhost/virtio_net.c
> @@ -2004,7 +2004,7 @@ virtio_dev_tx_batch_packed_zmbuf(struct
> virtio_net *dev,
>
> vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) {
> zmbufs[i]->mbuf = pkts[i];
> - zmbufs[i]->desc_idx = avail_idx + i;
> + zmbufs[i]->desc_idx = ids[i];
> zmbufs[i]->desc_count = 1;
> }
>
> @@ -2045,7 +2045,7 @@ virtio_dev_tx_single_packed_zmbuf(struct
> virtio_net *dev,
> return -1;
> }
> zmbuf->mbuf = *pkts;
> - zmbuf->desc_idx = vq->last_avail_idx;
> + zmbuf->desc_idx = buf_id;
> zmbuf->desc_count = desc_count;
>
> rte_mbuf_refcnt_update(*pkts, 1);
> --
> 2.17.1
Re: [dpdk-dev] [PATCH] net/i40e: remove useless check for queue number
Tested-by: Xiao, QimaiX
-Original Message-
From: Zhao1, Wei
Sent: Sunday, April 28, 2019 2:13 PM
To: dev@dpdk.org
Cc: sta...@dpdk.org; Zhang, Qi Z ; Xiao, QimaiX
; Peng, Yuan ; Zhao1, Wei
Subject: [PATCH] net/i40e: remove useless check for queue number
There is no need to do such a check of queue number, so remove it.
Fixes: ac8d22de2394 ("ethdev: flatten RSS configuration in flow API")
Signed-off-by: Wei Zhao
---
drivers/net/i40e/i40e_ethdev.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index dca61f0..5059bad 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -12699,9 +12699,6 @@ i40e_config_rss_filter(struct i40e_pf *pf,
return -EINVAL;
}
- if (rss_info->conf.queue_num)
- return -EINVAL;
-
/* If both VMDQ and RSS enabled, not all of PF queues are configured.
* It's necessary to calculate the actual PF queues that are configured.
*/
--
2.7.5
Re: [dpdk-dev] [PATCH v1] vhost: set zmbufs to NULL when freed
Tested-by: Xiao Qimai
Regards,
Xiao Qimai
> -Original Message-
> From: dev On Behalf Of patrick...@intel.com
> Sent: Wednesday, July 15, 2020 11:59 PM
> To: dev@dpdk.org; maxime.coque...@redhat.com; Xia, Chenbo
>
> Cc: Wang, Yinan ; Fu, Patrick
> Subject: [dpdk-dev] [PATCH v1] vhost: set zmbufs to NULL when freed
>
> From: Patrick Fu
>
> zmbufs should be set to NULL when getting freed to avoid double free on
> the same buffer pointer
>
> Fixes: b0a985d1f340 ("vhost: add dequeue zero copy")
>
> Signed-off-by: Patrick Fu
> ---
> lib/librte_vhost/vhost_user.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
> index df0db8a07..9ddeae362 100644
> --- a/lib/librte_vhost/vhost_user.c
> +++ b/lib/librte_vhost/vhost_user.c
> @@ -1934,6 +1934,7 @@ free_zmbufs(struct vhost_virtqueue *vq)
> drain_zmbuf_list(vq);
>
> rte_free(vq->zmbufs);
> + vq->zmbufs = NULL;
> }
>
> /*
> --
> 2.17.1
Re: [dpdk-dev] [PATCH v4 3/3] net/vhost: fix interrupt mode
Tested-by: Xiao Qimai
Regards,
Xiao Qimai
> -Original Message-
> From: dev On Behalf Of Xia, Chenbo
> Sent: Wednesday, July 29, 2020 10:19 PM
> To: Maxime Coquelin ; dev@dpdk.org;
> ma...@mellanox.com; Liu, Yong ; Wang, Yinan
>
> Cc: tho...@monjalon.net; Yigit, Ferruh ;
> david.march...@redhat.com
> Subject: Re: [dpdk-dev] [PATCH v4 3/3] net/vhost: fix interrupt mode
>
>
> > -Original Message-
> > From: Maxime Coquelin
> > Sent: Wednesday, July 29, 2020 9:36 PM
> > To: dev@dpdk.org; ma...@mellanox.com; Xia, Chenbo
> > ; Liu, Yong ; Wang, Yinan
> >
> > Cc: tho...@monjalon.net; Yigit, Ferruh ;
> > david.march...@redhat.com; Maxime Coquelin
> >
> > Subject: [PATCH v4 3/3] net/vhost: fix interrupt mode
> >
> > At .new_device() time, only the first vring pair is now ready, other
> > vrings are configured later.
> >
> > Problem is that when application will setup and enable interrupts,
> > only the first queue pair Rx interrupt will be enabled.
> >
> > This patches fixes the issue by setting the number of max interrupts
> > to the number of Rx queues that will be later initialized. Then, as
> > soon as a Rx vring is ready and interrupt enabled by the application,
> > it removes the corresponding uninitialized epoll event, and installs a new
> one with the valid FD.
> >
> > Fixes: 604052ae5395 ("net/vhost: support queue update")
> >
> > Signed-off-by: Maxime Coquelin
> > ---
> > drivers/net/vhost/rte_eth_vhost.c | 91
> > +++
> > 1 file changed, 81 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/net/vhost/rte_eth_vhost.c
> > b/drivers/net/vhost/rte_eth_vhost.c
> > index 951929c663..e55278af69 100644
> > --- a/drivers/net/vhost/rte_eth_vhost.c
> > +++ b/drivers/net/vhost/rte_eth_vhost.c
> > @@ -5,6 +5,7 @@
> > #include
> > #include
> > #include
> > +#include
> >
> > #include
> > #include
> > @@ -95,6 +96,8 @@ struct vhost_queue {
> > uint16_t port;
> > uint16_t virtqueue_id;
> > struct vhost_stats stats;
> > + int intr_enable;
> > + rte_spinlock_t intr_lock;
> > };
> >
> > struct pmd_internal {
> > @@ -524,12 +527,58 @@ find_internal_resource(char *ifname)
> > return list;
> > }
> >
> > +static int
> > +eth_vhost_update_intr(struct rte_eth_dev *eth_dev, uint16_t rxq_idx) {
> > + struct rte_intr_handle *handle = eth_dev->intr_handle;
> > + struct rte_epoll_event rev;
> > + int epfd, ret;
> > +
> > + if (!handle)
> > + return 0;
> > +
> > + if (handle->efds[rxq_idx] == handle->elist[rxq_idx].fd)
> > + return 0;
> > +
> > + VHOST_LOG(INFO, "kickfd for rxq-%d was changed, updating
> > handler.\n",
> > + rxq_idx);
> > +
> > + if (handle->elist[rxq_idx].fd != -1)
> > + VHOST_LOG(ERR, "Unexpected previous kickfd value
> (Got %d,
> > expected -1).\n",
> > + handle->elist[rxq_idx].fd);
> > +
> > + /*
> > +* First remove invalid epoll event, and then install
> > +* the new one. May be solved with a proper API in the
> > +* future.
> > +*/
> > + epfd = handle->elist[rxq_idx].epfd;
> > + rev = handle->elist[rxq_idx];
> > + ret = rte_epoll_ctl(epfd, EPOLL_CTL_DEL, rev.fd,
> > + &handle->elist[rxq_idx]);
> > + if (ret) {
> > + VHOST_LOG(ERR, "Delete epoll event failed.\n");
> > + return ret;
> > + }
> > +
> > + rev.fd = handle->efds[rxq_idx];
> > + handle->elist[rxq_idx] = rev;
> > + ret = rte_epoll_ctl(epfd, EPOLL_CTL_ADD, rev.fd,
> > + &handle->elist[rxq_idx]);
> > + if (ret) {
> > + VHOST_LOG(ERR, "Add epoll event failed.\n");
> > + return ret;
> > + }
> > +
> > + return 0;
> > +}
> > +
> > static int
> > eth_rxq_intr_enable(struct rte_eth_dev *dev, uint16_t qid) {
> > struct rte_vhost_vring vring;
> > struct vhost_queue *vq;
> > - int ret = 0;
> > + int old_intr_enable, ret = 0;
> >
> > vq = dev->data->rx_queues[qid];
> > if (!vq) {
> > @@ -537,6 +586,18 @@ eth_rxq_intr_enable(struct rte_eth_dev *dev,
> > uint16_t
> > qid)
> > return -1;
> > }
> >
> > + rte_spinlock_lock(&vq->intr_lock);
> > + old_intr_enable = vq->intr_enable;
> > + vq->intr_enable = 1;
> > + ret = eth_vhost_update_intr(dev, qid);
> > + rte_spinlock_unlock(&vq->intr_lock);
> > +
> > + if (ret < 0) {
> > + VHOST_LOG(ERR, "Failed to update rxq%d's intr\n", qid);
> > + vq->intr_enable = old_intr_enable;
> > + return ret;
> > + }
> > +
> > ret = rte_vhost_get_vhost_vring(vq->vid, (qid << 1) + 1, &vring);
> > if (ret < 0) {
> > VHOST_LOG(ERR, "Failed to get rxq%d's vring\n", qid); @@ -
> > 571,6 +632,8 @@ eth_rxq_intr_disable(struct rte_eth_dev *dev, uint16_t
> qid)
> > rte_vhost_enable_guest_notification(vq->vid, (qid << 1) + 1, 0);
> > rte_wmb();
> >
> > + vq->intr_enable = 0;
> > +
> > return 0;
> > }
> >
> > @@ -593,7 +656,6 @@ eth_vhost_ins
Re: [dpdk-dev] [PATCH] net/bnxt: fix variable size of port id
Tested-by: Xiao, Qimai
Regards,
Xiao Qimai
> -Original Message-
> From: dev On Behalf Of Chenbo Xia
> Sent: Wednesday, August 5, 2020 1:42 AM
> To: dev@dpdk.org
> Cc: ajit.khapa...@broadcom.com; somnath.ko...@broadcom.com;
> kishore.padmana...@broadcom.com;
> venkatkumar.duvv...@broadcom.com; sta...@dpdk.org
> Subject: [dpdk-dev] [PATCH] net/bnxt: fix variable size of port id
>
> Currenly the variable size of ethdev port id is 8 bits. This patch
> standarizes it
> to 16 bits.
>
> Fixes: 769de16872ab ("net/bnxt: fix port default rule create/destroy")
> Cc: sta...@dpdk.org
>
> Reported-by: Yinan Wang
> Signed-off-by: Chenbo Xia
> ---
> drivers/net/bnxt/tf_ulp/ulp_def_rules.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/bnxt/tf_ulp/ulp_def_rules.c
> b/drivers/net/bnxt/tf_ulp/ulp_def_rules.c
> index 9fb1a028f..ff5aed3d7 100644
> --- a/drivers/net/bnxt/tf_ulp/ulp_def_rules.c
> +++ b/drivers/net/bnxt/tf_ulp/ulp_def_rules.c
> @@ -397,7 +397,7 @@ void
> bnxt_ulp_destroy_df_rules(struct bnxt *bp, bool global) {
> struct bnxt_ulp_df_rule_info *info;
> - uint8_t port_id;
> + uint16_t port_id;
>
> if (!BNXT_TRUFLOW_EN(bp) ||
> BNXT_ETH_DEV_IS_REPRESENTOR(bp->eth_dev))
> --
> 2.17.1
