Re: [PATCH] HV: properly delay KVP packets when negotiation is in progress

[Vitaly Kuznetsov <vkuznets@redhat.com>]

Long Li <longli@microsoft.com> writes:

> The host may send multiple KVP packets before the negotiation with daemon
> is finished. We need to keep those packets in ring buffer until the daemon
> is negotiated and connected.

The patch looks OK but previously we always presumed that this can't
happen for util drivers and host will never send a new request before we
answer to the previous one. If this is not true we may have more issues
which need fixing as all three drivers we have are written in a
'transaction' fashion.

So my question would be: can the host send multiple (KVP) packets
_after_ the negotiation with daemon is finished?


>
> This patch is based on the work of Nick Meier <Nick.Meier@microsoft.com>
>
> Signed-off-by: Long Li <longli@microsoft.com>
> ---
>  drivers/hv/hv_kvp.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/hv/hv_kvp.c b/drivers/hv/hv_kvp.c
> index de26371..b9f928d 100644
> --- a/drivers/hv/hv_kvp.c
> +++ b/drivers/hv/hv_kvp.c
> @@ -628,16 +628,17 @@ void hv_kvp_onchannelcallback(void *context)
>  		     NEGO_IN_PROGRESS,
>  		     NEGO_FINISHED} host_negotiatied = NEGO_NOT_STARTED;
>
> -	if (host_negotiatied == NEGO_NOT_STARTED &&
> -	    kvp_transaction.state < HVUTIL_READY) {
> +	if (kvp_transaction.state < HVUTIL_READY) {
>  		/*
>  		 * If userspace daemon is not connected and host is asking
>  		 * us to negotiate we need to delay to not lose messages.
>  		 * This is important for Failover IP setting.
>  		 */
> -		host_negotiatied = NEGO_IN_PROGRESS;
> -		schedule_delayed_work(&kvp_host_handshake_work,
> +		if (host_negotiatied == NEGO_NOT_STARTED) {
> +			host_negotiatied = NEGO_IN_PROGRESS;
> +			schedule_delayed_work(&kvp_host_handshake_work,
>  				      HV_UTIL_NEGO_TIMEOUT * HZ);
> +		}
>  		return;
>  	}
>  	if (kvp_transaction.state > HVUTIL_READY)

-- 
  Vitaly