|
| 64 | 2.5 | 4.0 | 60% |
| 512 | 2.1 | 3.6 | 71% |
| 1024 | 1.9 | 3.3 | 73% |
+--+-+-+-+
Signed-off-by: Yunjian Wang
---
drivers/net/tun.c | 177 +++--
drivers/vhost/net.c| 4 +
inc
If TUN supports AF_XDP TX zero-copy, the XDP program will enqueue
packets to the XDP ring and wake up the vhost worker. This requires
the vhost worker to call peek_len(), which can be used to consume
XDP descriptors.
Signed-off-by: Yunjian Wang
---
drivers/vhost/net.c | 17 -
1
Now dma mappings are used by the physical NICs. However the vNIC
maybe do not need them. So remove non-zero 'dma_page' check in
xp_assign_dev.
Signed-off-by: Yunjian Wang
---
net/xdp/xsk_buff_pool.c | 7 ---
1 file changed, 7 deletions(-)
diff --git a/net/xdp/xsk_buff_pool.c b/net/xdp
).
+--+-+-+-+
| | copy |zero-copy| speedup |
+--+-+-+-+
| UDP | Mpps | Mpps |%|
| 64 | 2.5 | 4.0 | 60% |
| 512 | 2.1 | 3.6 | 71% |
| 1024 | 1.9 | 3.3 | 73% |
+--+-+-+-+
Yunjian Wang
length.
- When the tun support AF_XDP Rx zero-copy, the vq's array maybe empty.
So add a check for empty vq's array in vhost_net_buf_produce().
- add XDP_SETUP_XSK_POOL and ndo_xsk_wakeup callback support
- add tun_put_user_desc function to copy the Rx data to VM
Signed-off-by: Yunjian Wang
---
drive
Now dma mappings are used by the physical NICs. However the vNIC
maybe do not need them. So remove non-zero 'dma_page' check in
xp_assign_dev.
Signed-off-by: Yunjian Wang
---
net/xdp/xsk_buff_pool.c | 7 ---
1 file changed, 7 deletions(-)
diff --git a/net/xdp/xsk_buff_pool.c b/net/xdp
age for OVS' PMD has
decreased from 59.8% to 8.8% while forwarding 70pps.
Yunjian Wang (2):
xsk: Remove non-zero 'dma_page' check in xp_assign_dev
tun: AF_XDP Rx zero-copy support
drivers/net/tun.c | 165 +++-
drivers/vhost/net.c | 18 +++--
When we meet an error(err=-EBADFD) recvmsg, the error handling in vhost
handle_rx() will continue. This will cause a soft CPU lockup in vhost thread.
Signed-off-by: Yunjian Wang <wangyunj...@huawei.com>
---
drivers/vhost/net.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers
When we meet an error(err=-EBADFD) recvmsg, the error handling in vhost
handle_rx() will continue. This will cause a soft CPU lockup in vhost thread.
Signed-off-by: Yunjian Wang
---
drivers/vhost/net.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/vhost/net.c b/drivers/vhost
9 matches
Mail list logo