On Mon, Oct 15, 2018 at 10:47 AM Florian Fainelli <f.fainelli@xxxxxxxxx> wrote: > > > > On 10/10/2018 12:29 PM, Florian Fainelli wrote: > > This is based on Stephen's v4.14 patches, with the necessary merge > > conflicts, and the lack of timer_setup() on the 4.9 baseline. > > > > Perf results on a gigabit capable system, before and after are below. > > > > Series can also be found here: > > > > https://github.com/ffainelli/linux/commits/fragment-stack-v4.9-v2 > > > > Changes in v2: > > > > - drop "net: sk_buff rbnode reorg" > > - added original "ip: use rb trees for IP frag queue." commit > > Eric, does this look reasonable to you? Yes, thanks a lot Florian. > > > > > Before patches: > > > > PerfTop: 180 irqs/sec kernel:78.9% exact: 0.0% [4000Hz cycles:ppp], (all, 4 CPUs) > > ------------------------------------------------------------------------------- > > > > 34.81% [kernel] [k] ip_defrag > > 4.57% [kernel] [k] arch_cpu_idle > > 2.09% [kernel] [k] fib_table_lookup > > 1.74% [kernel] [k] finish_task_switch > > 1.57% [kernel] [k] v7_dma_inv_range > > 1.47% [kernel] [k] __netif_receive_skb_core > > 1.06% [kernel] [k] __slab_free > > 1.04% [kernel] [k] __netdev_alloc_skb > > 0.99% [kernel] [k] ip_route_input_noref > > 0.96% [kernel] [k] dev_gro_receive > > 0.96% [kernel] [k] tick_nohz_idle_enter > > 0.93% [kernel] [k] bcm_sysport_poll > > 0.92% [kernel] [k] skb_release_data > > 0.91% [kernel] [k] __memzero > > 0.90% [kernel] [k] __free_page_frag > > 0.87% [kernel] [k] ip_rcv > > 0.77% [kernel] [k] eth_type_trans > > 0.71% [kernel] [k] _raw_spin_unlock_irqrestore > > 0.68% [kernel] [k] tick_nohz_idle_exit > > 0.65% [kernel] [k] bcm_sysport_rx_refill > > > > After patches: > > > > PerfTop: 214 irqs/sec kernel:80.4% exact: 0.0% [4000Hz cycles:ppp], (all, 4 CPUs) > > ------------------------------------------------------------------------------- > > > > 6.61% [kernel] [k] arch_cpu_idle > > 3.77% [kernel] [k] ip_defrag > > 3.65% [kernel] [k] v7_dma_inv_range > > 3.18% [kernel] [k] fib_table_lookup > > 3.04% [kernel] [k] __netif_receive_skb_core > > 2.31% [kernel] [k] finish_task_switch > > 2.31% [kernel] [k] _raw_spin_unlock_irqrestore > > 1.65% [kernel] [k] bcm_sysport_poll > > 1.63% [kernel] [k] ip_route_input_noref > > 1.63% [kernel] [k] __memzero > > 1.58% [kernel] [k] __netdev_alloc_skb > > 1.47% [kernel] [k] tick_nohz_idle_enter > > 1.40% [kernel] [k] __slab_free > > 1.32% [kernel] [k] ip_rcv > > 1.32% [kernel] [k] __softirqentry_text_start > > 1.30% [kernel] [k] dev_gro_receive > > 1.23% [kernel] [k] bcm_sysport_rx_refill > > 1.11% [kernel] [k] tick_nohz_idle_exit > > 1.06% [kernel] [k] memcmp > > 1.02% [kernel] [k] dma_cache_maint_page > > > > > > Dan Carpenter (1): > > ipv4: frags: precedence bug in ip_expire() > > > > Eric Dumazet (21): > > inet: frags: change inet_frags_init_net() return value > > inet: frags: add a pointer to struct netns_frags > > inet: frags: refactor ipfrag_init() > > inet: frags: refactor ipv6_frag_init() > > inet: frags: refactor lowpan_net_frag_init() > > ipv6: export ip6 fragments sysctl to unprivileged users > > rhashtable: add schedule points > > inet: frags: use rhashtables for reassembly units > > inet: frags: remove some helpers > > inet: frags: get rif of inet_frag_evicting() > > inet: frags: remove inet_frag_maybe_warn_overflow() > > inet: frags: break the 2GB limit for frags storage > > inet: frags: do not clone skb in ip_expire() > > ipv6: frags: rewrite ip6_expire_frag_queue() > > rhashtable: reorganize struct rhashtable layout > > inet: frags: reorganize struct netns_frags > > inet: frags: get rid of ipfrag_skb_cb/FRAG_CB > > inet: frags: fix ip6frag_low_thresh boundary > > net: speed up skb_rbtree_purge() > > net: pskb_trim_rcsum() and CHECKSUM_COMPLETE are friends > > net: add rb_to_skb() and other rb tree helpers > > > > Florian Westphal (1): > > ipv6: defrag: drop non-last frags smaller than min mtu > > > > Peter Oskolkov (5): > > ip: discard IPv4 datagrams with overlapping segments. > > net: modify skb_rbtree_purge to return the truesize of all purged > > skbs. > > ip: use rb trees for IP frag queue. > > ip: add helpers to process in-order fragments faster. > > ip: process in-order fragments efficiently > > > > Taehee Yoo (1): > > ip: frags: fix crash in ip_do_fragment() > > > > Documentation/networking/ip-sysctl.txt | 13 +- > > include/linux/rhashtable.h | 4 +- > > include/linux/skbuff.h | 34 +- > > include/net/inet_frag.h | 133 +++--- > > include/net/ip.h | 1 - > > include/net/ipv6.h | 26 +- > > include/uapi/linux/snmp.h | 1 + > > lib/rhashtable.c | 5 +- > > net/core/skbuff.c | 31 +- > > net/ieee802154/6lowpan/6lowpan_i.h | 26 +- > > net/ieee802154/6lowpan/reassembly.c | 148 +++--- > > net/ipv4/inet_fragment.c | 379 ++++------------ > > net/ipv4/ip_fragment.c | 573 +++++++++++++----------- > > net/ipv4/proc.c | 7 +- > > net/ipv4/tcp_input.c | 33 +- > > net/ipv6/netfilter/nf_conntrack_reasm.c | 100 ++--- > > net/ipv6/proc.c | 5 +- > > net/ipv6/reassembly.c | 212 ++++----- > > 18 files changed, 774 insertions(+), 957 deletions(-) > > > > -- > Florian
