OVS_FORCE does not affect Clang, only sparse.
On Fri, Apr 13, 2018 at 11:43:24AM +, Gabor Halász wrote:
> Hi Ben,
> Indeed there was some problem with that line.
> Instead of using the eth_addr_to_uint64() func, I have added OVS_FORCE
> compiler flag to keep the logic more visible.
> +
Hi Ben,
Indeed there was some problem with that line.
Instead of using the eth_addr_to_uint64() func, I have added OVS_FORCE compiler
flag to keep the logic more visible.
+hash = hash_add64(hash, *((OVS_FORCE uint64_t *) &flow->dl_dst));
+hash = hash_add(hash, (OVS_FORCE uint32_t)
On Thu, Apr 05, 2018 at 09:02:22AM +, Gabor Halász wrote:
> In real-world vSwitch deployments, handling a few thousand flows,
> EMC is quickly saturated, so it's optimal usage is critical to
> reach the highest packet forwarding speed of the vSwitch.
>
> EMC lookup is initiated based on the ha
Hi,
I was focusing on keeping the algorithm of miniflow_hash_5tuple() and
flow_hash_5tuple () functions the same (as noted in the original comment).
Looking into the side-effects of the proposed modification:
1. miniflow_hash_5tuple is being called from dpif_netdev_packet_get_rss_hash
during e