date:20220904

[PATCH v2] app/proc-info: add rte_eal_cleanup() to avoid memory leak

2022-09-04 Thread Fidaullah Noonari

when app is launched with -m proc-info exit without
rte_eal_cleanup() causing memory leakage. This commit resolves the
memory leakage issue and closes app properly.

Bugzilla id: 898

Signed-off-by: Fidaullah Noonari 
---
 app/proc-info/main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/app/proc-info/main.c b/app/proc-info/main.c
index 1bfba5f60d..9cfcce31ab 100644
--- a/app/proc-info/main.c
+++ b/app/proc-info/main.c
@@ -1523,7 +1523,7 @@ main(int argc, char **argv)
 
if (mem_info) {
meminfo_display();
-   return 0;
+   goto cleanup;
}
 
nb_ports = rte_eth_dev_count_avail();
@@ -1593,6 +1593,7 @@ main(int argc, char **argv)
RTE_ETH_FOREACH_DEV(i)
rte_eth_dev_close(i);
 
+cleanup:
ret = rte_eal_cleanup();
if (ret)
printf("Error from rte_eal_cleanup(), %d\n", ret);
-- 
2.25.1

[PATCH] common/iavf: do not copy desc/buff in async mode

2022-09-04 Thread Steven Zou

If cmd_details are defined or async flag is set,
don't copy the desc/buff back to temp.

Signed-off-by: Steven Zou 
---
 drivers/common/iavf/iavf_adminq.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/common/iavf/iavf_adminq.c 
b/drivers/common/iavf/iavf_adminq.c
index 0bf5af0cbe..2e59820b3b 100644
--- a/drivers/common/iavf/iavf_adminq.c
+++ b/drivers/common/iavf/iavf_adminq.c
@@ -784,7 +784,8 @@ enum iavf_status iavf_asq_send_command(struct iavf_hw *hw,
}
 
/* if ready, copy the desc back to temp */
-   if (iavf_asq_done(hw)) {
+   if (iavf_asq_done(hw) &&
+   !details->async && !details->postpone) {
iavf_memcpy(desc, desc_on_ring, sizeof(struct iavf_aq_desc),
IAVF_DMA_TO_NONDMA);
if (buff != NULL)
-- 
2.34.1

RE: [PATCH] common/iavf: do not copy desc/buff in async mode

2022-09-04 Thread Zhang, Qi Z




> -Original Message-
> From: Zou, Steven 
> Sent: Monday, September 5, 2022 9:42 AM
> To: dev@dpdk.org
> Cc: Zhang, Qi Z ; Yang, Guang3
> ; Zou, Steven 
> Subject: [PATCH] common/iavf: do not copy desc/buff in async mode
> 
> If cmd_details are defined or async flag is set, don't copy the desc/buff back
> to temp.
> 

Added 
Fixes: e5b2a9e957e7 ("net/avf/base: add base code for avf PMD")
Cc: sta...@dpdk.org

> Signed-off-by: Steven Zou 

Acked-by: Qi Zhang 

Applied to dpdk-next-net-intel.

Thanks
Qi

RE: [PATCH v2] net/iavf: enable inner and outer Tx checksum offload

2022-09-04 Thread Yang, Qiming

Test fail.
Nacked-by: Qiming Yang 

> -Original Message-
> From: Zhang, Peng1X 
> Sent: Thursday, September 1, 2022 5:33 PM
> To: dev@dpdk.org
> Cc: Yang, Qiming ; Zhang, Qi Z
> ; Zhang, Peng1X 
> Subject: [PATCH v2] net/iavf: enable inner and outer Tx checksum offload
> 
> Enable inner and outer Tx checksum offload for tunnel packet by configure
> ol_flags.
> 
> Signed-off-by: Peng Zhang 
> 
> ---
> v2: add outer udp cksum flag and remove unrelated code
> ---
>  drivers/net/iavf/iavf_ethdev.c |  1 +
>  drivers/net/iavf/iavf_rxtx.c   | 48 --
>  drivers/net/iavf/iavf_rxtx.h   |  9 ++-
>  3 files changed, 55 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/iavf/iavf_ethdev.c b/drivers/net/iavf/iavf_ethdev.c
> index 506fcff6e3..fa040766e5 100644
> --- a/drivers/net/iavf/iavf_ethdev.c
> +++ b/drivers/net/iavf/iavf_ethdev.c
> @@ -1134,6 +1134,7 @@ iavf_dev_info_get(struct rte_eth_dev *dev, struct
> rte_eth_dev_info *dev_info)
>   RTE_ETH_TX_OFFLOAD_TCP_CKSUM |
>   RTE_ETH_TX_OFFLOAD_SCTP_CKSUM |
>   RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM |
> + RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM |
>   RTE_ETH_TX_OFFLOAD_TCP_TSO |
>   RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO |
>   RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO |
> diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c index
> 3deabe1d7e..b784c5cc18 100644
> --- a/drivers/net/iavf/iavf_rxtx.c
> +++ b/drivers/net/iavf/iavf_rxtx.c
> @@ -2334,7 +2334,8 @@ static inline uint16_t
> iavf_calc_context_desc(uint64_t flags, uint8_t vlan_flag)  {
>   if (flags & (RTE_MBUF_F_TX_TCP_SEG | RTE_MBUF_F_TX_UDP_SEG |
> - RTE_MBUF_F_TX_TUNNEL_MASK))
> + RTE_MBUF_F_TX_TUNNEL_MASK |
> RTE_MBUF_F_TX_OUTER_IP_CKSUM |
> + RTE_MBUF_F_TX_OUTER_UDP_CKSUM))
>   return 1;
>   if (flags & RTE_MBUF_F_TX_VLAN &&
>   vlan_flag & IAVF_TX_FLAGS_VLAN_TAG_LOC_L2TAG2)
> @@ -2399,6 +2400,44 @@ iavf_fill_ctx_desc_tunnelling_field(volatile
> uint64_t *qw0,
>   break;
>   }
> 
> + /* L4TUNT: L4 Tunneling Type */
> + switch (m->ol_flags & RTE_MBUF_F_TX_TUNNEL_MASK) {
> + case RTE_MBUF_F_TX_TUNNEL_IPIP:
> + /* for non UDP / GRE tunneling, set to 00b */
> + break;
> + case RTE_MBUF_F_TX_TUNNEL_VXLAN:
> + case RTE_MBUF_F_TX_TUNNEL_GTP:
> + case RTE_MBUF_F_TX_TUNNEL_GENEVE:
> + eip_typ |= IAVF_TXD_CTX_UDP_TUNNELING;
> + break;
> + case RTE_MBUF_F_TX_TUNNEL_GRE:
> + eip_typ |= IAVF_TXD_CTX_GRE_TUNNELING;
> + break;
> + default:
> + PMD_TX_LOG(ERR, "Tunnel type not supported");
> + return;
> + }
> +
> + /* L4TUNLEN: L4 Tunneling Length, in Words
> +  *
> +  * We depend on app to set rte_mbuf.l2_len correctly.
> +  * For IP in GRE it should be set to the length of the GRE
> +  * header;
> +  * For MAC in GRE or MAC in UDP it should be set to the length
> +  * of the GRE or UDP headers plus the inner MAC up to including
> +  * its last Ethertype.
> +  * If MPLS labels exists, it should include them as well.
> +  */
> + eip_typ |= (m->l2_len >> 1) << IAVF_TXD_CTX_QW0_NATLEN_SHIFT;
> +
> + /**
> +  * Calculate the tunneling UDP checksum.
> +  * Shall be set only if L4TUNT = 01b and EIPT is not zero
> +  */
> + if (!(eip_typ & IAVF_TX_CTX_EXT_IP_NONE) &&
> + (eip_typ & IAVF_TXD_CTX_UDP_TUNNELING))
> + eip_typ |= IAVF_TXD_CTX_QW0_L4T_CS_MASK;
> +
>   *qw0 = eip_typ << IAVF_TXD_CTX_QW0_TUN_PARAMS_EIPT_SHIFT |
>   eip_len <<
> IAVF_TXD_CTX_QW0_TUN_PARAMS_EIPLEN_SHIFT |
>   eip_noinc <<
> IAVF_TXD_CTX_QW0_TUN_PARAMS_EIP_NOINC_SHIFT;
> @@ -2535,7 +2574,12 @@ iavf_build_data_desc_cmd_offset_fields(volatile
> uint64_t *qw1,
>   }
> 
>   /* Set MACLEN */
> - offset |= (m->l2_len >> 1) << IAVF_TX_DESC_LENGTH_MACLEN_SHIFT;
> + if (m->ol_flags & RTE_MBUF_F_TX_TUNNEL_MASK)
> + offset |= (m->outer_l2_len >> 1)
> + << IAVF_TX_DESC_LENGTH_MACLEN_SHIFT;
> + else
> + offset |= (m->l2_len >> 1)
> + << IAVF_TX_DESC_LENGTH_MACLEN_SHIFT;
> 
>   /* Enable L3 checksum offloading inner */
>   if (m->ol_flags & RTE_MBUF_F_TX_IP_CKSUM) { diff --git
> a/drivers/net/iavf/iavf_rxtx.h b/drivers/net/iavf/iavf_rxtx.h index
> 1695e43cd5..66e832713c 100644
> --- a/drivers/net/iavf/iavf_rxtx.h
> +++ b/drivers/net/iavf/iavf_rxtx.h
> @@ -26,6 +26,8 @@
>  #define IAVF_TX_NO_VECTOR_FLAGS ( \
>   RTE_ETH_TX_OFFLOAD_MULTI_SEGS |  \
>   RTE_ETH_TX_OFFLOAD_TCP_TSO | \
> + RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM |\
> + RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM |\
>   RTE_ETH_TX_OFFLOAD_SECURITY)
> 
>  #define IAVF_TX_VECTOR_OFFL

[PATCH v2 0/2] vhost: support VIRTIO_F_RING_RESET for vhost-user

2022-09-04 Thread Kangjie Xu

Add VIRTIO_F_RING_RESET, which indicates that the driver can reset a
virtqueue individually.

VIRTIO_F_RING_RESET feature is added to virtio-spec 1.2. The relevant
information is in
https://github.com/oasis-tcs/virtio-spec/issues/124
https://github.com/oasis-tcs/virtio-spec/issues/139

The implementation only adds the feature bit in supported features.
It also adds a new vhost-user message VHOST_USER_RESET_VRING. The
related definition is defined in the related QEMU patch set:
  
https://lore.kernel.org/qemu-devel/cover.1661510725.git.kangjie...@linux.alibaba.com/T/#t

The virtqueue reset process can be concluded as two parts:
1. The driver can reset a virtqueue. It will send VHOST_USER_RESET_VRING
to DPDK. After received the message, DPDK will reset the virtqueue. The
new message VHOST_USER_RESET_VRING has been acked by the QEMU virtio
maintainer:

https://lore.kernel.org/qemu-devel/57362274-b2fb-a47d-fea7-d3ebcfad9...@redhat.com/

2. After the virtqueue is disabled, the driver may optionally re-enable
it. To avoid confusion with VHOST_USER_SET_VRING_ENABLE, we call this
part as "restart". The virtqueue's information may be changed when
restarting it. Thus, the information of the reset virtqueue should be
updated. This part is basically similar to when the virtqueue is started
for the first time, except that the restart process does not need to set
features and set mem table since they do not change. QEMU will send
messages containing size, base, addr, kickfd and callfd of the virtqueue
in order. Specifically, the DPDK will receive these messages in order:
a. VHOST_USER_SET_VRING_NUM
b. VHOST_USER_SET_VRING_BASE
c. VHOST_USER_SET_VRING_ADDR
d. VHOST_USER_SET_VRING_KICK
e. VHOST_USER_SET_VRING_CALL
f. VHOST_USER_SET_VRING_ENABLE
The last VHOST_USER_SET_VRING_ENABLE message with "payload.state.num" set
to 1, will be sent to enable the virtqueue and the restart process is
finished.

Test environment:
Host: 5.4.189
Qemu: QEMU emulator version 7.0.50 (With vq reset support)
Guest: 5.19.0-rc3 (With vq reset support)
DPDK: 22.07-rc1
Test Cmd: ethtool -g eth1; ethtool -G eth1 rx $1 tx $2;
ethtool -g eth1;

The driver can resize the virtio queue, then virtio queue reset
function should be triggered.

Guest Kernel Patch:

https://lore.kernel.org/bpf/20220801063902.129329-1-xuanz...@linux.alibaba.com/

QEMU Patch:

https://lore.kernel.org/qemu-devel/cover.1661510725.git.kangjie...@linux.alibaba.com/T/

https://lore.kernel.org/qemu-devel/cover.1661414345.git.kangjie...@linux.alibaba.com/T/

Looking forward to your review and comments.

Kangjie Xu (2):
  vhost: support VIRTIO_F_RING_RESET for vhost-user
  vhost: introduce VHOST_USER_RESET_VRING

 lib/vhost/vhost.c  |  2 +-
 lib/vhost/vhost.h  |  9 -
 lib/vhost/vhost_user.c | 27 ++-
 lib/vhost/vhost_user.h |  1 +
 4 files changed, 36 insertions(+), 3 deletions(-)

-- 
2.32.0

[PATCH v2 2/2] vhost: introduce VHOST_USER_RESET_VRING

2022-09-04 Thread Kangjie Xu

To support the reset operation for an individual virtqueue, we
introduce a new message VHOST_USER_RESET_VRING. When the feature
VIRTIO_F_RING_RESET feature has been successfully negotiated, This
message is submitted by the front-end to reset an individual
virtqueue to initial states in the back-end. The reply is needed
to ensure that the reset operation is complete.

Signed-off-by: Kangjie Xu 
Signed-off-by: Xuan Zhuo 
---
 lib/vhost/vhost.c  |  2 +-
 lib/vhost/vhost.h  |  1 +
 lib/vhost/vhost_user.c | 27 ++-
 lib/vhost/vhost_user.h |  1 +
 4 files changed, 29 insertions(+), 2 deletions(-)

diff --git a/lib/vhost/vhost.c b/lib/vhost/vhost.c
index 60cb05a0ff..215a1ca355 100644
--- a/lib/vhost/vhost.c
+++ b/lib/vhost/vhost.c
@@ -610,7 +610,7 @@ init_vring_queue(struct virtio_net *dev, uint32_t vring_idx)
vhost_user_iotlb_init(dev, vring_idx);
 }
 
-static void
+void
 reset_vring_queue(struct virtio_net *dev, uint32_t vring_idx)
 {
struct vhost_virtqueue *vq;
diff --git a/lib/vhost/vhost.h b/lib/vhost/vhost.h
index 76461a3406..eccb52842d 100644
--- a/lib/vhost/vhost.h
+++ b/lib/vhost/vhost.h
@@ -791,6 +791,7 @@ get_device(int vid)
 
 int vhost_new_device(void);
 void cleanup_device(struct virtio_net *dev, int destroy);
+void reset_vring_queue(struct virtio_net *dev, uint32_t vring_idx);
 void reset_device(struct virtio_net *dev);
 void vhost_destroy_device(int);
 void vhost_destroy_device_notify(struct virtio_net *dev);
diff --git a/lib/vhost/vhost_user.c b/lib/vhost/vhost_user.c
index 4ad28bac45..5f7743d9d9 100644
--- a/lib/vhost/vhost_user.c
+++ b/lib/vhost/vhost_user.c
@@ -2771,6 +2771,30 @@ vhost_user_set_status(struct virtio_net **pdev,
return RTE_VHOST_MSG_RESULT_OK;
 }
 
+static int
+vhost_user_reset_vring(struct virtio_net **pdev,
+   struct vhu_msg_context *ctx __rte_unused,
+   int main_fd __rte_unused)
+{
+   struct virtio_net *dev = *pdev;
+   int index = (int)ctx->msg.payload.state.index;
+
+   VHOST_LOG_CONFIG(dev->ifname, INFO, "reset queue: queue idx: %d\n", 
index);
+
+   if (!(dev->features & (1ULL << VIRTIO_F_RING_RESET))) {
+   return RTE_VHOST_MSG_RESULT_ERR;
+   }
+
+   dev->virtqueue[index]->enabled = false;
+   reset_vring_queue(dev, index);
+
+   ctx->msg.payload.state.num = 0;
+   ctx->msg.size = sizeof(ctx->msg.payload.u64);
+   ctx->fd_num = 0;
+
+   return RTE_VHOST_MSG_RESULT_REPLY;
+}
+
 #define VHOST_MESSAGE_HANDLERS \
 VHOST_MESSAGE_HANDLER(VHOST_USER_NONE, NULL, false) \
 VHOST_MESSAGE_HANDLER(VHOST_USER_GET_FEATURES, vhost_user_get_features, false) 
\
@@ -2803,7 +2827,8 @@ VHOST_MESSAGE_HANDLER(VHOST_USER_POSTCOPY_END, 
vhost_user_postcopy_end, false) \
 VHOST_MESSAGE_HANDLER(VHOST_USER_GET_INFLIGHT_FD, vhost_user_get_inflight_fd, 
false) \
 VHOST_MESSAGE_HANDLER(VHOST_USER_SET_INFLIGHT_FD, vhost_user_set_inflight_fd, 
true) \
 VHOST_MESSAGE_HANDLER(VHOST_USER_SET_STATUS, vhost_user_set_status, false) \
-VHOST_MESSAGE_HANDLER(VHOST_USER_GET_STATUS, vhost_user_get_status, false)
+VHOST_MESSAGE_HANDLER(VHOST_USER_GET_STATUS, vhost_user_get_status, false) \
+VHOST_MESSAGE_HANDLER(VHOST_USER_RESET_VRING, vhost_user_reset_vring, false)
 
 #define VHOST_MESSAGE_HANDLER(id, handler, accepts_fd) \
[id] = { #id, handler, accepts_fd },
diff --git a/lib/vhost/vhost_user.h b/lib/vhost/vhost_user.h
index 8ecca68597..51cb2fc74a 100644
--- a/lib/vhost/vhost_user.h
+++ b/lib/vhost/vhost_user.h
@@ -60,6 +60,7 @@ typedef enum VhostUserRequest {
VHOST_USER_SET_INFLIGHT_FD = 32,
VHOST_USER_SET_STATUS = 39,
VHOST_USER_GET_STATUS = 40,
+   VHOST_USER_RESET_VRING = 41
 } VhostUserRequest;
 
 typedef enum VhostUserSlaveRequest {
-- 
2.32.0

[PATCH v2 1/2] vhost: support VIRTIO_F_RING_RESET for vhost-user

2022-09-04 Thread Kangjie Xu

Add VIRTIO_F_RING_RESET, which indicates that the driver can reset a
queue individually.

The feature is added to virtio-spec 1.2. The relevant information is
in https://github.com/oasis-tcs/virtio-spec/issues/124

Signed-off-by: Kangjie Xu 
Signed-off-by: Xuan Zhuo 
---
 lib/vhost/vhost.h | 8 +++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/lib/vhost/vhost.h b/lib/vhost/vhost.h
index 40fac3b7c6..76461a3406 100644
--- a/lib/vhost/vhost.h
+++ b/lib/vhost/vhost.h
@@ -376,6 +376,11 @@ struct vhost_msg {
  #define VIRTIO_F_VERSION_1 32
 #endif
 
+/* This feature indicates that the driver can reset a queue individually. */
+#ifndef VIRTIO_F_RING_RESET
+#define VIRTIO_F_RING_RESET 40
+#endif
+
 /* Declare packed ring related bits for older kernels */
 #ifndef VIRTIO_F_RING_PACKED
 
@@ -438,7 +443,8 @@ struct vring_packed_desc_event {
(1ULL << VIRTIO_NET_F_MTU)  | \
(1ULL << VIRTIO_F_IN_ORDER) | \
(1ULL << VIRTIO_F_IOMMU_PLATFORM) | \
-   (1ULL << VIRTIO_F_RING_PACKED))
+   (1ULL << VIRTIO_F_RING_PACKED)  | \
+   (1ULL << VIRTIO_F_RING_RESET))
 
 
 struct guest_page {
-- 
2.32.0

RE: [PATCH] gro: fix gro with tcp push flag

2022-09-04 Thread Hu, Jiayu

Hi Qiu,

I cannot find the original mail, and please see replies inline.

> -Original Message-
> From: Hu, Jiayu
> Sent: Tuesday, August 30, 2022 4:44 PM
> To: 'Thomas Monjalon' 
> Cc: kumaraparameshwaran rathinavel ;
> dev@dpdk.org; Jun Qiu 
> Subject: RE: [PATCH] gro: fix gro with tcp push flag
> 
> Sure, I will review it.
> 
> > -Original Message-
> > From: Thomas Monjalon 
> > Sent: Monday, August 29, 2022 6:36 PM
> > To: Hu, Jiayu 
> > Cc: kumaraparameshwaran rathinavel ;
> > dev@dpdk.org; Jun Qiu 
> > Subject: Re: [PATCH] gro: fix gro with tcp push flag
> >
> > Jiayu, please could you help in this review?
> >
> >
> > 27/07/2022 10:44, Jun Qiu:
> > > I think this delay is tolerable.
> > > Many TCP stacks do not take special care of PUSH packets when
> > > receiving
> > them. All received packets with data will trigger Poll events.
> > >
> > > The patch is simple to implement and easy to understand, similar to
> > > how
> > the kernel stack is handled.
> > >
> > > From: kumaraparameshwaran rathinavel
> 
> > > Sent: Tuesday, July 26, 2022 3:08 PM
> > > To: Jun Qiu 
> > > Cc: dev@dpdk.org; jiayu...@intel.com; sta...@dpdk.org
> > > Subject: Re: [PATCH] gro: fix gro with tcp push flag
> > >
> > > We should do it for the rte_gro_reassemble as well, with timer mode
> > > it
> > could lead to more duplicate ACKs. I had a proposal for the
> > enhancement which would handle both  rte_gro_reassemble and
> > rte_gro_reassemble_burst but have not got any response yet.
> > >
> > > I have a custom patch which is working fine for timer mode where
> > > there is
> > no packet reordering, earlier without the patch there were DUP-ACKs
> > and this could potentially affect the window scaling.
> > >
> > > On Tue, Jul 26, 2022 at 12:27 PM Jun Qiu
> > mailto:jun@jaguarmicro.com>> wrote:
> > > May be in rte_gro_reassemble_burst, where no delay is introduced,
> > > PUSH packets can be merged
> > >
> > > 发件人: kumaraparameshwaran rathinavel
> > >
> mailto:kumaraparames...@gmail.com>>
> > > 发送时间: 2022年7月26日 14:41
> > > 收件人: Jun Qiu
> > mailto:jun@jaguarmicro.com>>
> > > 抄送: dev@dpdk.org;
> > > jiayu...@intel.com;
> > > sta...@dpdk.org
> > > 主题: Re: [PATCH] gro: fix gro with tcp push flag
> > >
> > >
> > >
> > > On Tue, Jul 26, 2022 at 11:48 AM Jun Qiu
> > mailto:jun@jaguarmicro.com>> wrote:
> > > TCP data packets sometimes carry a PUSH flag. Currently, only the
> > > packets that do not have PUSH flag can be GROed.
> > > The packets that have a PUSH flag cannot be GROed, the packets that
> > > cannot be processed by GRO are placed last.
> > > In this case, the received packets may be out of order.
> > > For example, there are two packets mbuf1 and mbuf2. mbuf1 contains
> > > PUSH flag, mbuf2 does not contain PUSH flag.
> > > After GRO processing, mbuf2 is sent for processing before mbuf1.
> > > This out-of-order will affect TCP processing performance and lead to
> > > unnecessary dup-ACK.
> > >
> > > Referring to the Linux kernel implementation, packets with PUSH flag
> > > can also perform GRO. And if one of the packets containing PUSH
> > > flag, the packets after GRO will carry PUSH flag.
> > >
> > > In case of smaller transfers in which the TCP segment size is not
> > > more than
> > one MTU, it is a single TCP packet with PSH flag set, so in those
> > cases  we are introducing unwanted delay.  I think the better approach
> > would be if there are previous packets in the flow and the current
> > packet received has PSH flag then coalesce with the previous packet,
> > if lookup is failure and the current packet has PSH flag set then deliver it
> immediately.

I agree with Kuma. The implementation in the patch will make PSH packet
wait in the gro table. For a PSH packet, we need to try to merge it with
previous packets in the table first. If merge successfully, the merged packet
needs to mark with PSH flag and deliver it to app then; If not, this
PSH packet also needs to be delivered to app immediately.

Thanks,
Jiayu

> > >
> > > Fixes: 0d2cbe59b719 ("lib/gro: support TCP/IPv4")
> > > Cc: sta...@dpdk.org
> > >
> > > Signed-off-by: Jun Qiu
> > > mailto:jun@jaguarmicro.com>>
> > > ---
> > >  lib/gro/gro_tcp4.c   |  4 ++--
> > >  lib/gro/gro_tcp4.h   | 16 +---
> > >  lib/gro/gro_vxlan_tcp4.c |  4 ++--
> > >  3 files changed, 17 insertions(+), 7 deletions(-)
> > >
> > > diff --git a/lib/gro/gro_tcp4.c b/lib/gro/gro_tcp4.c index
> > > 7498c66141..7849a2bd1d 100644
> > > --- a/lib/gro/gro_tcp4.c
> > > +++ b/lib/gro/gro_tcp4.c
> > > @@ -220,10 +220,10 @@ gro_tcp4_reassemble(struct rte_mbuf *pkt,
> > > hdr_len = pkt->l2_len + pkt->l3_len + pkt->l4_len;
> > >
> > > /*
> > > -* Don't process the packet which has FIN, SYN, RST, PSH, URG, ECE
> > > +* Don't process the packet which has FIN, SYN, RST, URG,
> > > + ECE
> > >  * or CWR set.
> > >

RE: [PATCH v3 1/4] net/axgbe: fix scattered Rx

2022-09-04 Thread Namburu, Chandu-babu

[Public]

Acked-by: Chandubabu Namburu 

-Original Message-
From: Modali, Bhagyada  
Sent: Friday, September 2, 2022 2:18 PM
To: Namburu, Chandu-babu ; Yigit, Ferruh 
Cc: dev@dpdk.org; sta...@dpdk.org; Modali, Bhagyada 
Subject: [PATCH v3 1/4] net/axgbe: fix scattered Rx

Error check needs to be done only for last segment of Jumbo packet.
Freed first_seg and reset eop to 0 in error case

Fixes: 965b3127d425 ("net/axgbe: support scattered Rx")
Cc: sta...@dpdk.org

Signed-off-by: Bhagyada Modali 

---
v2:
* removed the extra paranthesis
* initialised the variables
* updated the commit logs
---
 drivers/net/axgbe/axgbe_rxtx.c | 35 +++---
 1 file changed, 20 insertions(+), 15 deletions(-)

diff --git a/drivers/net/axgbe/axgbe_rxtx.c b/drivers/net/axgbe/axgbe_rxtx.c 
index f38bb64fab..89ed6fd74a 100644
--- a/drivers/net/axgbe/axgbe_rxtx.c
+++ b/drivers/net/axgbe/axgbe_rxtx.c
@@ -343,8 +343,8 @@ uint16_t eth_axgbe_recv_scattered_pkts(void *rx_queue,
uint64_t old_dirty = rxq->dirty;
struct rte_mbuf *first_seg = NULL;
struct rte_mbuf *mbuf, *tmbuf;
-   unsigned int err, etlt;
-   uint32_t error_status;
+   unsigned int err = 0, etlt;
+   uint32_t error_status = 0;
uint16_t idx, pidx, data_len = 0, pkt_len = 0;
uint64_t offloads;
 
@@ -381,19 +381,6 @@ uint16_t eth_axgbe_recv_scattered_pkts(void *rx_queue,
}
 
mbuf = rxq->sw_ring[idx];
-   /* Check for any errors and free mbuf*/
-   err = AXGMAC_GET_BITS_LE(desc->write.desc3,
-RX_NORMAL_DESC3, ES);
-   error_status = 0;
-   if (unlikely(err)) {
-   error_status = desc->write.desc3 & AXGBE_ERR_STATUS;
-   if ((error_status != AXGBE_L3_CSUM_ERR)
-   && (error_status != AXGBE_L4_CSUM_ERR)) 
{
-   rxq->errors++;
-   rte_pktmbuf_free(mbuf);
-   goto err_set;
-   }
-   }
rte_prefetch1(rte_pktmbuf_mtod(mbuf, void *));
 
if (!AXGMAC_GET_BITS_LE(desc->write.desc3,
@@ -406,6 +393,24 @@ uint16_t eth_axgbe_recv_scattered_pkts(void *rx_queue,
pkt_len = AXGMAC_GET_BITS_LE(desc->write.desc3,
RX_NORMAL_DESC3, PL);
data_len = pkt_len - rxq->crc_len;
+   /* Check for any errors and free mbuf*/
+   err = AXGMAC_GET_BITS_LE(desc->write.desc3,
+   RX_NORMAL_DESC3, ES);
+   error_status = 0;
+   if (unlikely(err)) {
+   error_status = desc->write.desc3 &
+   AXGBE_ERR_STATUS;
+   if (error_status != AXGBE_L3_CSUM_ERR &&
+   error_status != 
AXGBE_L4_CSUM_ERR) {
+   rxq->errors++;
+   rte_pktmbuf_free(mbuf);
+   rte_pktmbuf_free(first_seg);
+   first_seg = NULL;
+   eop = 0;
+   goto err_set;
+   }
+   }
+
}
 
if (first_seg != NULL) {
--
2.25.1

Re: [PATCH v2] vhost: compilation fix for GCC-12

2022-09-04 Thread Bagas Sanjaya

On Fri, Sep 02, 2022 at 06:11:52PM +0530, Amit Prakash Shukla wrote:
> ../lib/vhost/virtio_net.c:941:35: error:
>   'buf_vec[0].buf_len' may be used uninitialized
>   [-Werror=maybe-uninitialized]
>   941 | buf_len = buf_vec[vec_idx].buf_len;
>   |   ^~~~
> ../lib/vhost/virtio_net.c: In function 'virtio_dev_rx_packed':
> ../lib/vhost/virtio_net.c:1285:27: note: 'buf_vec' declared here
>  1285 | struct buf_vector buf_vec[BUF_VECTOR_MAX];
>   |   ^~~
> cc1: all warnings being treated as errors
> 
> Fixes: 93520085efda ("vhost: add packed ring single enqueue")
> Cc: sta...@dpdk.org

Please describe what this patch is doing (the current state of code,
why it errored, and how it is fixing the error). Write the description
in imperative mood. I don't see the description other than error message
above.

Also, for stable patches submission, Cc sta...@vger.kernel.org.

Thanks.

-- 
An old man doll... just what I always wanted! - Clara


signature.asc
Description: PGP signature

[PATCH v2] app/proc-info: add rte_eal_cleanup() to avoid memory leak

[PATCH] common/iavf: do not copy desc/buff in async mode

RE: [PATCH] common/iavf: do not copy desc/buff in async mode

RE: [PATCH v2] net/iavf: enable inner and outer Tx checksum offload

[PATCH v2 0/2] vhost: support VIRTIO_F_RING_RESET for vhost-user

[PATCH v2 2/2] vhost: introduce VHOST_USER_RESET_VRING

[PATCH v2 1/2] vhost: support VIRTIO_F_RING_RESET for vhost-user

RE: [PATCH] gro: fix gro with tcp push flag

RE: [PATCH v3 1/4] net/axgbe: fix scattered Rx

Re: [PATCH v2] vhost: compilation fix for GCC-12

10 matches

Site Navigation

Mail list logo

Footer information