On 2024-03-02 12:14, Mattias Rönnblom wrote:
On 2024-03-01 18:14, Stephen Hemminger wrote:
The DPDK has a lot of "cargo cult" usage of rte_memcpy.
This patch set replaces cases where rte_memcpy is used with a fixed
size constant size.
Typical example is:
rte_memcpy(mac_addrs, mac.addr_bytes, RTE_ETHER_ADDR_LEN);
which can be replaced with:
memcpy(mac_addrs, mac.addr_bytes, RTE_ETHER_ADDR_LEN);
This has two benefits. Gcc (and clang) are smart enough that for
all small fixed size values, they just generate the necessary
instructions
to do it inline. It also means that fortify, Coverity, and ASAN
analyzers can check these memcpy's.
Instead of smearing out the knowledge of when to use rte_memcpy(), and
when to use memcpy() across the code base, wouldn't it be better to
*always* call rte_memcpy() in the fast path, and leave the policy
decision to the rte_memcpy() implementation?
In rte_memcpy(), add:
if (__builtin_constant_p(n) && n < RTE_LIBC_MEMCPY_SIZE_THRESHOLD)
memcpy(/../);
..or something to that effect.
Could you have a #ifdef for dumb static analysis tools? To make it look
like you are always using memcpy()?
So faster, better, safer.
What is "faster" based on?
I ran some DSW benchmarks, and if you add
diff --git a/lib/eal/x86/include/rte_memcpy.h
b/lib/eal/x86/include/rte_memcpy.h
index 72a92290e0..64cd82d78d 100644
--- a/lib/eal/x86/include/rte_memcpy.h
+++ b/lib/eal/x86/include/rte_memcpy.h
@@ -862,6 +862,11 @@ rte_memcpy_aligned(void *dst, const void *src,
size_t n)
static __rte_always_inline void *
rte_memcpy(void *dst, const void *src, size_t n)
{
+ if (__builtin_constant_p(n) && n <= 32) {
+ memcpy(dst, src, n);
+ return dst;
+ }
+
if (!(((uintptr_t)dst | (uintptr_t)src) & ALIGNMENT_MASK))
return rte_memcpy_aligned(dst, src, n);
else
...the overhead increases from roughly 48 core clock cycles/event to 59
cc/event. The same for "n < 128". (I'm not sure what counts as "small"
here.)
So something rte_memcpy() does for small and constant memory copies does
make things go *significantly* faster, at least in certain cases.
(Linux, GCC 11.4, Intel Gracemont.)
My experience with replacing rte_memcpy() with memcpy() (or vice versa)
is mixed.
I've also tried just dropping the DPDK-custom memcpy() implementation
altogether, and that caused a performance drop (in a particular app, on
a particular compiler and CPU).
The first patch is a simple coccinelle script to do the replacement
and the rest are the results broken out by module.
The coccinelle script can be used again to make sure more bad
usage doesn't creep in with new drivers.
v2 - fix CI failure on some OS by adding string.h
remove rte_memcpy.h if no longer used
Stephen Hemminger (71):
cocci/rte_memcpy: add script to eliminate fixed size rte_memcpy
eal: replace use of fixed size rte_memcpy
ethdev: replace use of fixed size rte_memcpy
eventdev: replace use of fixed size rte_memcpy
cryptodev: replace use of fixed size rte_memcpy
ip_frag: replace use of fixed size rte_memcpy
net: replace use of fixed size rte_memcpy
lpm: replace use of fixed size rte_memcpy
node: replace use of fixed size rte_memcpy
pdcp: replace use of fixed size rte_memcpy
pipeline: replace use of fixed size rte_memcpy
rib: replace use of fixed size rte_memcpy
security: replace use of fixed size rte_memcpy
net/mlx5: replace use of fixed size rte_memcpy
net/nfp: replace use of fixed size rte_memcpy
net/ngbe: replace use of fixed size rte_memcpy
net/null: replace use of fixed size rte_memcpy
net/pcap: replace use of fixed size rte_memcpy
net/sfc: replace use of fixed size rte_memcpy
net/tap: replace use of fixed size rte_memcpy
net/txgbe: replace use of fixed size rte_memcpy
raw/ifpga: replace use of fixed size rte_memcpy
raw/skeleton: replace use of fixed size rte_memcpy
net/hns3: replace use of fixed size rte_memcpy
net/i40e: replace use of fixed size rte_memcpy
net/iavf: replace use of fixed size rte_memcpy
net/ice: replace use of fixed size rte_memcpy
net/idpf: replace use of fixed size rte_memcpy
net/ipn3ke: replace use of fixed size rte_memcpy
net/ixgbe: replace use of fixed size rte_memcpy
net/memif: replace use of fixed size rte_memcpy
net/qede: replace use of fixed size rte_memcpy
baseband/acc: replace use of fixed size rte_memcpy
baseband/la12xx: replace use of fixed size rte_memcpy
common/idpf: replace use of fixed size rte_memcpy
common/qat: replace use of fixed size rte_memcpy
compress/qat: replace use of fixed size rte_memcpy
crypto/ccp: replace use of fixed size rte_memcpy
crypto/cnxk: replace use of fixed size rte_memcpy
crypto/dpaa_sec: replace use of fixed size rte_memcpy
crypto/ipsec_mb: replace use of fixed size rte_memcpy
crypto/qat: replace use of fixed size rte_memcpy
crypto/scheduler: replace use of fixed size rte_memcpy
event/cnxk: replace use of fixed size rte_memcpy
event/dlb2: replace use of fixed size rte_memcpy
event/dpaa2: replace use of fixed size rte_memcpy
event/octeontx: replace use of fixed size rte_memcpy
mempool/dpaa: replace use of fixed size rte_memcpy
mempool/dpaa2: replace use of fixed size rte_memcpy
ml/cnxk: replace use of fixed size rte_memcpy
net/af_xdp: replace use of fixed size rte_memcpy
net/avp: replace use of fixed size rte_memcpy
net/axgbe: replace use of fixed size rte_memcpy
net/bnx2x: replace use of fixed size rte_memcpy
net/bnxt: replace use of fixed size rte_memcpy
net/bonding: replace use of fixed size rte_memcpy
net/cnxk: replace use of fixed size rte_memcpy
net/cpfl: replace use of fixed size rte_memcpy
net/cxgbe: replace use of fixed size rte_memcpy
net/dpaa2: replace use of fixed size rte_memcpy
net/e1000: replace use of fixed size rte_memcpy
net/enic: replace use of fixed size rte_memcpy
net/failsafe: replace use of fixed size rte_memcpy
net/gve/base: replace use of fixed size rte_memcpy
net/hinic: replace use of fixed size rte_memcpy
net/mvpp2: replace use of fixed size rte_memcpy
app/test-pmd: replace use of fixed size rte_memcpy
app/graph: replace use of fixed size rte_memcpy
app/test-eventdev: replace use of fixed size rte_memcpy
app/test: replace use of fixed size rte_memcpy
examples: replace use of fixed size rte_memcpy
app/graph/neigh.c | 8 +-
app/test-eventdev/test_pipeline_common.c | 19 ++-
app/test-pmd/cmdline.c | 48 ++++----
app/test-pmd/cmdline_flow.c | 24 ++--
app/test-pmd/config.c | 8 +-
app/test-pmd/csumonly.c | 1 -
app/test-pmd/flowgen.c | 1 -
app/test-pmd/iofwd.c | 1 -
app/test-pmd/macfwd.c | 1 -
app/test-pmd/macswap.c | 1 -
app/test-pmd/noisy_vnf.c | 1 -
app/test-pmd/rxonly.c | 1 -
app/test-pmd/testpmd.c | 1 -
app/test/commands.c | 1 -
app/test/packet_burst_generator.c | 4 +-
app/test/test_crc.c | 5 +-
app/test/test_cryptodev.c | 18 ++-
app/test/test_cryptodev_asym.c | 1 -
app/test/test_cryptodev_security_pdcp.c | 1 -
app/test/test_efd.c | 1 -
app/test/test_efd_perf.c | 1 -
app/test/test_event_crypto_adapter.c | 12 +-
app/test/test_event_dma_adapter.c | 4 +-
app/test/test_eventdev.c | 1 -
app/test/test_ipsec.c | 6 +-
app/test/test_link_bonding_mode4.c | 8 +-
app/test/test_mbuf.c | 1 -
app/test/test_member.c | 1 -
app/test/test_member_perf.c | 1 -
app/test/test_rawdev.c | 1 -
app/test/test_security_inline_proto.c | 36 +++---
app/test/test_service_cores.c | 1 -
app/test/virtual_pmd.c | 3 +-
devtools/cocci/rte_memcpy.cocci | 11 ++
drivers/baseband/acc/rte_acc100_pmd.c | 17 ++-
drivers/baseband/acc/rte_vrb_pmd.c | 21 ++--
drivers/baseband/la12xx/bbdev_la12xx.c | 4 +-
drivers/common/idpf/idpf_common_device.c | 4 +-
drivers/common/idpf/idpf_common_virtchnl.c | 8 +-
drivers/common/qat/qat_qp.c | 10 +-
drivers/compress/qat/qat_comp.c | 8 +-
drivers/crypto/ccp/ccp_crypto.c | 14 +--
drivers/crypto/cnxk/cnxk_cryptodev_ops.c | 2 +-
drivers/crypto/cnxk/cnxk_se.h | 2 +-
drivers/crypto/dpaa_sec/dpaa_sec.c | 2 +-
drivers/crypto/ipsec_mb/pmd_snow3g.c | 4 +-
drivers/crypto/qat/qat_sym_session.c | 52 ++++-----
.../scheduler/rte_cryptodev_scheduler.c | 6 +-
drivers/crypto/scheduler/scheduler_failover.c | 12 +-
drivers/event/cnxk/cnxk_tim_evdev.c | 4 +-
drivers/event/dlb2/dlb2.c | 6 +-
drivers/event/dpaa2/dpaa2_eventdev.c | 6 +-
drivers/event/octeontx/timvf_evdev.c | 4 +-
drivers/mempool/dpaa/dpaa_mempool.c | 4 +-
drivers/mempool/dpaa2/dpaa2_hw_mempool.c | 4 +-
drivers/ml/cnxk/cn10k_ml_model.c | 8 +-
drivers/ml/cnxk/cn10k_ml_ops.c | 11 +-
drivers/ml/cnxk/cnxk_ml_ops.c | 2 +-
drivers/ml/cnxk/mvtvm_ml_model.c | 8 +-
drivers/ml/cnxk/mvtvm_ml_ops.c | 8 +-
drivers/net/af_xdp/rte_eth_af_xdp.c | 2 +-
drivers/net/avp/avp_ethdev.c | 4 +-
drivers/net/axgbe/axgbe_ethdev.c | 4 +-
drivers/net/bnx2x/bnx2x.c | 32 +++--
drivers/net/bnx2x/bnx2x_stats.c | 10 +-
drivers/net/bnx2x/bnx2x_vfpf.c | 19 +--
drivers/net/bnxt/bnxt_flow.c | 34 +++---
drivers/net/bonding/rte_eth_bond_8023ad.c | 4 +-
drivers/net/bonding/rte_eth_bond_flow.c | 2 +-
drivers/net/cnxk/cnxk_ethdev_ops.c | 2 +-
drivers/net/cnxk/cnxk_tm.c | 5 +-
drivers/net/cpfl/cpfl_ethdev.c | 3 +-
drivers/net/cpfl/cpfl_vchnl.c | 4 +-
drivers/net/cxgbe/clip_tbl.c | 2 +-
drivers/net/cxgbe/cxgbe_filter.c | 8 +-
drivers/net/cxgbe/l2t.c | 4 +-
drivers/net/cxgbe/smt.c | 20 ++--
drivers/net/dpaa2/base/dpaa2_hw_dpni.c | 1 -
drivers/net/dpaa2/dpaa2_ethdev.c | 1 -
drivers/net/dpaa2/dpaa2_recycle.c | 1 -
drivers/net/dpaa2/dpaa2_rxtx.c | 1 -
drivers/net/dpaa2/dpaa2_sparser.c | 1 -
drivers/net/dpaa2/dpaa2_tm.c | 2 +-
drivers/net/e1000/em_rxtx.c | 1 -
drivers/net/e1000/igb_flow.c | 22 ++--
drivers/net/e1000/igb_pf.c | 7 +-
drivers/net/e1000/igb_rxtx.c | 1 -
drivers/net/enic/enic_main.c | 8 +-
drivers/net/failsafe/failsafe_ops.c | 6 +-
drivers/net/gve/base/gve_adminq.c | 2 +-
drivers/net/hinic/hinic_pmd_flow.c | 40 +++----
drivers/net/hns3/hns3_fdir.c | 2 +-
drivers/net/hns3/hns3_flow.c | 4 +-
drivers/net/i40e/i40e_ethdev.c | 109 ++++++++----------
drivers/net/i40e/i40e_fdir.c | 28 +++--
drivers/net/i40e/i40e_flow.c | 56 +++++----
drivers/net/i40e/i40e_pf.c | 3 +-
drivers/net/i40e/i40e_tm.c | 11 +-
drivers/net/i40e/rte_pmd_i40e.c | 34 +++---
drivers/net/iavf/iavf_fdir.c | 93 +++++++--------
drivers/net/iavf/iavf_fsub.c | 50 ++++----
drivers/net/iavf/iavf_generic_flow.c | 2 +-
drivers/net/iavf/iavf_tm.c | 11 +-
drivers/net/iavf/iavf_vchnl.c | 9 +-
drivers/net/ice/ice_dcf.c | 5 +-
drivers/net/ice/ice_dcf_parent.c | 2 +-
drivers/net/ice/ice_dcf_sched.c | 11 +-
drivers/net/ice/ice_diagnose.c | 4 +-
drivers/net/ice/ice_ethdev.c | 14 +--
drivers/net/ice/ice_fdir_filter.c | 37 +++---
drivers/net/ice/ice_generic_flow.c | 2 +-
drivers/net/ice/ice_hash.c | 2 +-
drivers/net/ice/ice_tm.c | 11 +-
drivers/net/idpf/idpf_ethdev.c | 7 +-
drivers/net/idpf/idpf_rxtx.c | 10 +-
drivers/net/ipn3ke/ipn3ke_flow.c | 32 +++--
drivers/net/ipn3ke/ipn3ke_representor.c | 16 +--
drivers/net/ipn3ke/ipn3ke_tm.c | 6 +-
drivers/net/ixgbe/ixgbe_ethdev.c | 9 +-
drivers/net/ixgbe/ixgbe_fdir.c | 7 +-
drivers/net/ixgbe/ixgbe_flow.c | 65 +++++------
drivers/net/ixgbe/ixgbe_ipsec.c | 8 +-
drivers/net/ixgbe/ixgbe_pf.c | 5 +-
drivers/net/ixgbe/ixgbe_tm.c | 11 +-
drivers/net/ixgbe/rte_pmd_ixgbe.c | 4 +-
drivers/net/memif/memif_socket.c | 4 +-
drivers/net/mlx5/mlx5_devx.c | 4 +-
drivers/net/mlx5/mlx5_flow.c | 38 +++---
drivers/net/mlx5/mlx5_flow_aso.c | 6 +-
drivers/net/mlx5/mlx5_flow_hw.c | 16 +--
drivers/net/mlx5/mlx5_rx.c | 6 +-
drivers/net/mlx5/mlx5_rxtx_vec.c | 8 +-
drivers/net/mvpp2/mrvl_tm.c | 2 +-
drivers/net/nfp/flower/nfp_conntrack.c | 2 +-
drivers/net/nfp/flower/nfp_flower_flow.c | 16 +--
.../net/nfp/flower/nfp_flower_representor.c | 2 +-
drivers/net/nfp/nfp_mtr.c | 10 +-
drivers/net/ngbe/ngbe_pf.c | 4 +-
drivers/net/null/rte_eth_null.c | 6 +-
drivers/net/pcap/pcap_ethdev.c | 2 +-
drivers/net/pcap/pcap_osdep_freebsd.c | 2 +-
drivers/net/pcap/pcap_osdep_linux.c | 2 +-
drivers/net/qede/qede_main.c | 2 +-
drivers/net/sfc/sfc.c | 2 +-
drivers/net/sfc/sfc_ef10_tx.c | 2 +-
drivers/net/sfc/sfc_ethdev.c | 11 +-
drivers/net/sfc/sfc_flow.c | 20 ++--
drivers/net/sfc/sfc_flow_rss.c | 2 +-
drivers/net/sfc/sfc_mae.c | 2 +-
drivers/net/sfc/sfc_rx.c | 2 +-
drivers/net/sfc/sfc_tso.c | 2 +-
drivers/net/sfc/sfc_tso.h | 9 +-
drivers/net/tap/rte_eth_tap.c | 14 +--
drivers/net/txgbe/txgbe_ethdev.c | 9 +-
drivers/net/txgbe/txgbe_fdir.c | 6 +-
drivers/net/txgbe/txgbe_flow.c | 65 +++++------
drivers/net/txgbe/txgbe_ipsec.c | 8 +-
drivers/net/txgbe/txgbe_pf.c | 4 +-
drivers/net/txgbe/txgbe_tm.c | 11 +-
drivers/raw/ifpga/afu_pmd_he_hssi.c | 3 +-
drivers/raw/ifpga/afu_pmd_he_lpbk.c | 3 +-
drivers/raw/ifpga/afu_pmd_he_mem.c | 3 +-
drivers/raw/ifpga/afu_pmd_n3000.c | 8 +-
drivers/raw/ifpga/ifpga_rawdev.c | 11 +-
drivers/raw/skeleton/skeleton_rawdev.c | 7 +-
examples/bbdev_app/main.c | 2 +-
examples/l2fwd-cat/cat.c | 4 +-
examples/ptpclient/ptpclient.c | 11 +-
examples/vhost/main.c | 5 +-
examples/vmdq/main.c | 6 +-
examples/vmdq_dcb/main.c | 15 +--
lib/cryptodev/rte_cryptodev.c | 2 +-
lib/eal/common/eal_common_options.c | 7 +-
lib/ethdev/rte_ethdev.c | 3 +-
lib/ethdev/rte_flow.c | 5 +-
lib/eventdev/rte_event_crypto_adapter.c | 2 +-
lib/eventdev/rte_event_dma_adapter.c | 4 +-
lib/eventdev/rte_event_timer_adapter.c | 2 +-
lib/fib/trie.c | 2 +-
lib/ip_frag/rte_ipv6_fragmentation.c | 4 +-
lib/ip_frag/rte_ipv6_reassembly.c | 6 +-
lib/lpm/rte_lpm6.c | 3 +-
lib/net/rte_ether.c | 2 +-
lib/node/ip6_lookup.c | 8 +-
lib/pdcp/pdcp_process.c | 36 +++---
lib/pipeline/rte_table_action.c | 8 +-
lib/rib/rte_rib6.h | 5 +-
lib/security/rte_security.c | 4 +-
188 files changed, 881 insertions(+), 998 deletions(-)
create mode 100644 devtools/cocci/rte_memcpy.cocci