From: Mark McLoughlin <[EMAIL PROTECTED]> The current virtio_net tx timer is 2ns, which doesn't make any sense. Set it to a more reasonable 250us instead.
However, even though we were requesting a 2ns tx timer, it was actually getting limited to MIN_TIMER_REARM_US which is currently 250us. So, even though the timer itself would only fire after 250us, expire_time was only set to +2ns, so we'd get the timeout callback next time qemu_run_timers() was called from the mainloop. This probably accounted for a lot of the jitter in the throughput numbers - the effective tx timer length was anywhere between 2ns and 250us depending on e.g. whether there was rx data available on the tap fd. Signed-off-by: Mark McLoughlin <[EMAIL PROTECTED]> Signed-off-by: Avi Kivity <[EMAIL PROTECTED]> diff --git a/qemu/hw/virtio-net.c b/qemu/hw/virtio-net.c index 2e57e5a..3a39c8f 100644 --- a/qemu/hw/virtio-net.c +++ b/qemu/hw/virtio-net.c @@ -26,7 +26,7 @@ #define VIRTIO_NET_F_MAC 5 #define VIRTIO_NET_F_GS0 6 -#define TX_TIMER_INTERVAL (1000 / 500) +#define TX_TIMER_INTERVAL 250000 /* 250 us */ /* The config defining mac address (6 bytes) */ struct virtio_net_config -- To unsubscribe from this list: send the line "unsubscribe kvm-commits" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html