xen-netback: use default TX queue size for vifs

Commit Message

Roger Pau Monné Oct. 5, 2023, 2:08 p.m. UTC

Do not set netback interfaces (vifs) default TX queue size to the ring size.
The TX queue size is not related to the ring size, and using the ring size (32)
as the queue size can lead to packet drops.  Note the TX side of the vif
interface in the netback domain is the one receiving packets to be injected
to the guest.

Do not explicitly set the TX queue length to any value when creating the
interface, and instead use the system default.  Note that the queue length can
also be adjusted at runtime.

Fixes: f942dc2552b8 ('xen network backend driver')
Signed-off-by: Roger Pau Monné <[e-mail]>
---
 drivers/net/xen-netback/interface.c | 4 ----
 1 file changed, 4 deletions(-)

Comments

Ross Lagerwall Oct. 6, 2023, 1:15 p.m. UTC | #1

On Thu, Oct 5, 2023 at 3:08 PM Roger Pau Monne <[e-mail]> wrote:
>
> Do not set netback interfaces (vifs) default TX queue size to the ring size.
> The TX queue size is not related to the ring size, and using the ring size (32)
> as the queue size can lead to packet drops.  Note the TX side of the vif
> interface in the netback domain is the one receiving packets to be injected
> to the guest.
>
> Do not explicitly set the TX queue length to any value when creating the
> interface, and instead use the system default.  Note that the queue length can
> also be adjusted at runtime.
>
> Fixes: f942dc2552b8 ('xen network backend driver')
> Signed-off-by: Roger Pau Monné <[e-mail]>
> ---

Reviewed-by: Ross Lagerwall <[e-mail]>

Wei Liu Oct. 6, 2023, 9:44 p.m. UTC | #2

On Thu, Oct 05, 2023 at 04:08:31PM +0200, Roger Pau Monne wrote:
> Do not set netback interfaces (vifs) default TX queue size to the ring size.
> The TX queue size is not related to the ring size, and using the ring size (32)
> as the queue size can lead to packet drops.  Note the TX side of the vif
> interface in the netback domain is the one receiving packets to be injected
> to the guest.
> 
> Do not explicitly set the TX queue length to any value when creating the
> interface, and instead use the system default.  Note that the queue length can
> also be adjusted at runtime.
> 
> Fixes: f942dc2552b8 ('xen network backend driver')
> Signed-off-by: Roger Pau Monné <[e-mail]>

Acked-by: Wei Liu <[e-mail]>

[e-mail] Oct. 8, 2023, 1:01 p.m. UTC | #3

Hello:

This patch was applied to netdev/net.git (main)
by David S. Miller <[e-mail]>:

On Thu,  5 Oct 2023 16:08:31 +0200 you wrote:
> Do not set netback interfaces (vifs) default TX queue size to the ring size.
> The TX queue size is not related to the ring size, and using the ring size (32)
> as the queue size can lead to packet drops.  Note the TX side of the vif
> interface in the netback domain is the one receiving packets to be injected
> to the guest.
> 
> Do not explicitly set the TX queue length to any value when creating the
> interface, and instead use the system default.  Note that the queue length can
> also be adjusted at runtime.
> 
> [...]

Here is the summary with links:
  - xen-netback: use default TX queue size for vifs
    https://git.kernel.org/netdev/net/c/66cf7435a269

You are awesome, thank you!

Patch

diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c
index f3f2c07423a6..fc3bb63b9ac3 100644
--- a/drivers/net/xen-netback/interface.c
+++ b/drivers/net/xen-netback/interface.c
@@ -41,8 +41,6 @@ 
 #include <asm/xen/hypercall.h>
 #include <xen/balloon.h>
 
-#define XENVIF_QUEUE_LENGTH 32
-
 /* Number of bytes allowed on the internal guest Rx queue. */
 #define XENVIF_RX_QUEUE_BYTES (XEN_NETIF_RX_RING_SIZE/2 * PAGE_SIZE)
 
@@ -530,8 +528,6 @@  struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
 	dev->features = dev->hw_features | NETIF_F_RXCSUM;
 	dev->ethtool_ops = &xenvif_ethtool_ops;
 
-	dev->tx_queue_len = XENVIF_QUEUE_LENGTH;
-
 	dev->min_mtu = ETH_MIN_MTU;
 	dev->max_mtu = ETH_MAX_MTU - VLAN_ETH_HLEN;