Hi Olivier, Do you plan to integrate this patch set with latest mbuf changes into DPDK1.8?
If yes, I can develop TSO feature of tunneling packet based on this patch set and the Mirek's i40e TSO patch set in the next release. Thanks Jijiang Liu > -----Original Message----- > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Olivier Matz > Sent: Monday, May 19, 2014 9:56 PM > To: dev at dpdk.org > Subject: [dpdk-dev] [PATCH v2 00/17] ixgbe/mbuf: add TSO support > > This series add TSO support in ixgbe DPDK driver. This is the second version > of the > series. The first version (RFC) was posted 10 days ago and discussed on the > list > [1]. The list of changes is provided at the end of this cover letter. > > Also, as discussed previously on the list [2], one problem is that there is > not > enough room in rte_mbuf today to store the required information to implement > this feature: > - a new ol_flag > - the MSS > - the L4 header len > > A solution would be to increase the size of the mbuf to 2 cache lines but it > could > have a bad impact on performance. This series proposes some rework to > drastically reduce the size of the rte_mbuf structures before implementing the > TSO, avoiding to change the mbuf size to 128 bytes. > > After the rework of mbuf structures, the size of rte_mbuf structure is > reduced by > 9 bytes. The implementation of TSO requires to double the size of ol_flags > (16 to > 32 bits) and to double the size of offload information in order to add the > mss and > the l4 header length (32 to 64 bits). At the end of the whole series, > sizeof(rte_mbuf) is still 64 bytes and 4 bytes are available for future use. > > This rework causes a lot of modifications in the mbuf structure, implying some > changes in the applications that directly use the mbuf structure fields > instead of > using the API functions (sometimes there is no function). In my opinion, it's > the > proper moment for this evolution as the 1.7.0 window is open. > > About TSO, the new fields in mbuf try to be generic enough to apply to other > hardware in the future. To delegate the TCP segmentation to the hardware, the > user has to: > > - set the PKT_TX_TCP_SEG flag in mbuf->ol_flags (this flag implies > PKT_TX_IP_CKSUM and PKT_TX_TCP_CKSUM) > - fill the mbuf->hw_offload information: l2_len, l3_len, l4_len, mss > - calculate the pseudo header checksum and set it in the TCP header, > as required when doing hardware TCP checksum offload > - set the IP checksum to 0 > > Compilation of DPDK and examples is tested for the following > targets: x86_64-*-linuxapp-gcc, i686-*-linuxapp-gcc, x86_64-*-bsdapp-gcc > > The mbuf rework series is validated with autotests: > > cd dpdk.org/ > make install T=x86_64-default-linuxapp-gcc > cd x86_64-default-linuxapp-gcc/ > modprobe uio > insmod kmod/igb_uio.ko > python ../tools/igb_uio_bind.py -b igb_uio 0000:02:00.0 > echo 0 > /proc/sys/kernel/randomize_va_space > echo 1000 > /sys/devices/system/node/node0/hugepages/hugepages- > 2048kB/nr_hugepages > echo 1000 > /sys/devices/system/node/node1/hugepages/hugepages- > 2048kB/nr_hugepages > mount -t hugetlbfs none /mnt/huge > make test > > TSO is validated with IPv4 and IPv6 with testpmd (see the commit log of patch > "ixgbe: support TCP segmentation offload" for details). > > The performance non-regression has been tested with 6WINDGate fast path, and > with test-pmd [3]. > > [1] http://dpdk.org/ml/archives/dev/2014-May/002322.html > [2] http://dpdk.org/ml/archives/dev/2013-October/thread.html#572 > [3] http://dpdk.org/ml/archives/dev/2014-May/002516.html > > > Changes included in v2: > - rebase on head, resolve conflicts in 3 test-pmd engines > - add more comments in rte_mbuf.h about how to use TSO > - ixgbe: advertise the TSO feature > - ixgbe: small optimization, move some tso code in the if (tx_ol_req) > - split "mbuf: replace data pointer by an offset" (cosmetics vs functional) > - split "ixgbe/mbuf: add TSO support" (ixgbe vs generic changes) > - modifications of external PMDs (memnic, virtio, vmxnet3) > - I did not add the new checksum flags PKT_RX_L4_CKSUM_GOOD and > PKT_RX_IP_CKSUM_GOOD as proposed on the list. The patch is ready but > I don't think it should be included in this series. > > Olivier Matz (17): > dpdk igb/ixgbe: fix IP checksum calculation > dpdk mbuf: rename RTE_MBUF_SCATTER_GATHER into > dpdk mbuf: remove rte_ctrlmbuf > dpdk mbuf: remove the rte_pktmbuf structure > dpdk mbuf: merge physaddr and buf_len in a bitfield > dpdk mbuf: cosmetic changes in rte_mbuf structure > dpdk mbuf: replace data pointer by an offset > dpdk mbuf: add functions to get the name of an ol_flag > dpdk mbuf: change ol_flags to 32 bits > dpdk mbuf: rename vlan_macip_len in hw_offload and increase > dpdk testpmd: modify source address to validate checksum > dpdk mbuf: generic support of TCP segmentation offload > dpdk ixgbe: support TCP segmentation offload > virtio-net-pmd pmd: adapt to new rte_mbuf structure > vmxnet3-usermap pmd: remove support of old dpdk versions vmxnet3-usermap > pmd: adapt to new rte_mbuf structure > memnic pmd: adapt to new rte_mbuf structure > > Diffstat: > dpdk.org/app/test-pmd/cmdline.c | > 60 - > dpdk.org/app/test-pmd/config.c | 18 > dpdk.org/app/test-pmd/csumonly.c | 56 > dpdk.org/app/test-pmd/flowgen.c | 26 > dpdk.org/app/test-pmd/icmpecho.c | 6 > dpdk.org/app/test-pmd/ieee1588fwd.c | 6 > dpdk.org/app/test-pmd/macfwd-retry.c | 4 > dpdk.org/app/test-pmd/macfwd.c | 16 > dpdk.org/app/test-pmd/macswap.c | 16 > dpdk.org/app/test-pmd/rxonly.c | 53 > dpdk.org/app/test-pmd/testpmd.c | 14 > dpdk.org/app/test-pmd/testpmd.h | 15 > dpdk.org/app/test-pmd/txonly.c | > 59 - > dpdk.org/app/test/commands.c | 2 > dpdk.org/app/test/test_mbuf.c | > 106 - > dpdk.org/app/test/test_sched.c | 4 > dpdk.org/config/defconfig_i686-default-linuxapp-gcc | 2 > dpdk.org/config/defconfig_i686-default-linuxapp-icc | 2 > dpdk.org/config/defconfig_x86_64-default-bsdapp-gcc | 2 > dpdk.org/config/defconfig_x86_64-default-linuxapp-gcc | 2 > dpdk.org/config/defconfig_x86_64-default-linuxapp-icc | 2 > dpdk.org/doc/doxy-api.conf | 2 > dpdk.org/examples/dpdk_qat/crypto.c | 22 > dpdk.org/examples/dpdk_qat/main.c | 2 > dpdk.org/examples/exception_path/main.c | 13 > dpdk.org/examples/ip_reassembly/ipv4_rsmbl.h | 30 > dpdk.org/examples/ip_reassembly/main.c | 10 > dpdk.org/examples/ipv4_frag/Makefile | 4 > dpdk.org/examples/ipv4_frag/main.c | 4 > dpdk.org/examples/ipv4_frag/rte_ipv4_frag.h | 42 > dpdk.org/examples/ipv4_multicast/Makefile | 4 > dpdk.org/examples/ipv4_multicast/main.c | 16 > dpdk.org/examples/l3fwd-power/main.c | 2 > dpdk.org/examples/l3fwd-vf/main.c | 2 > dpdk.org/examples/l3fwd/main.c | 10 > dpdk.org/examples/load_balancer/runtime.c | 2 > dpdk.org/examples/multi_process/client_server_mp/mp_client/client.c | 2 > dpdk.org/examples/quota_watermark/qw/main.c | 4 > dpdk.org/examples/vhost/main.c | 43 > dpdk.org/examples/vhost_xen/main.c | 24 > dpdk.org/lib/librte_eal/bsdapp/eal/include/exec-env/rte_kni_common.h | 2 > dpdk.org/lib/librte_eal/linuxapp/eal/include/exec-env/rte_kni_common.h | 2 > dpdk.org/lib/librte_mbuf/rte_mbuf.c | > 105 - > dpdk.org/lib/librte_mbuf/rte_mbuf.h | > 580 ++++------ > dpdk.org/lib/librte_pmd_e1000/em_rxtx.c | > 163 +- > dpdk.org/lib/librte_pmd_e1000/igb_rxtx.c | > 212 +-- > dpdk.org/lib/librte_pmd_ixgbe/ixgbe_ethdev.c | 3 > dpdk.org/lib/librte_pmd_ixgbe/ixgbe_rxtx.c | > 419 ++++--- > dpdk.org/lib/librte_pmd_ixgbe/ixgbe_rxtx.h | 10 > dpdk.org/lib/librte_pmd_pcap/rte_eth_pcap.c | 14 > dpdk.org/lib/librte_pmd_virtio/virtio_rxtx.c | 20 > dpdk.org/lib/librte_pmd_virtio/virtqueue.h | 11 > dpdk.org/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c | 37 > dpdk.org/lib/librte_pmd_xenvirt/rte_eth_xenvirt.c | 14 > dpdk.org/lib/librte_pmd_xenvirt/virtqueue.h | 4 > dpdk.org/lib/librte_sched/rte_sched.c | 14 > dpdk.org/lib/librte_sched/rte_sched.h | 10 > memnic/pmd/pmd_memnic.c | 14 > vmxnet3-usermap/pmd/vmxnet3.c | > 119 -- > virtio-net-pmd/virtio_user.c | 44 > 60 files changed, 1217 insertions(+), 1289 deletions(-) > > > -- > 1.9.2
