This patchset change ndo_xdp_xmit API to take a bulk of xdp frames.
When kernel is compiled with CONFIG_RETPOLINE, every indirect function
pointer (branch) call hurts performance. For XDP this have a huge
negative performance impact.
This patchset reduce the needed (indirect) calls to ndo_xdp_xmit, but
also prepares for further optimizations. The DMA APIs use of indirect
function pointer calls is the primary source the regression. It is
left for a followup patchset, to use bulking calls towards the DMA API
(via the scatter-gatter calls).
The other advantage of this API change is that drivers can easier
amortize the cost of any sync/locking scheme, over the bulk of
packets. The assumption of the current API is that the driver
implemementing the NDO will also allocate a dedicated XDP TX queue for
every CPU in the system. Which is not always possible or practical to
configure. E.g. ixgbe cannot load an XDP program on a machine with
more than 96 CPUs, due to limited hardware TX queues. E.g. virtio_net
is hard to configure as it requires manually increasing the
queues. E.g. tun driver chooses to use a per XDP frame producer lock
modulo smp_processor_id over avail queues.
---
Jesper Dangaard Brouer (4):
bpf: devmap introduce dev_map_enqueue
bpf: devmap prepare xdp frames for bulking
xdp: add tracepoint for devmap like cpumap have
xdp: change ndo_xdp_xmit API to support bulking
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 26 ++++-
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 2
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 21 +++-
drivers/net/tun.c | 37 ++++---
drivers/net/virtio_net.c | 66 +++++++++---
include/linux/bpf.h | 15 ++-
include/linux/netdevice.h | 14 ++-
include/net/xdp.h | 1
include/trace/events/xdp.h | 50 +++++++++
kernel/bpf/devmap.c | 134 ++++++++++++++++++++++++-
net/core/filter.c | 19 +---
net/core/xdp.c | 14 ++-
samples/bpf/xdp_monitor_kern.c | 49 +++++++++
samples/bpf/xdp_monitor_user.c | 69 +++++++++++++
14 files changed, 436 insertions(+), 81 deletions(-)
--
