kernel/git/stable/linux.git/net/dsa/dsa.c, branch linux-4.19.y

net: dsa: add GRO support via gro_cells

2021-03-11T13:05:00+00:00

commit e131a5634830047923c694b4ce0c3b31745ff01b upstream. gro_cells lib is used by different encapsulating netdevices, such as geneve, macsec, vxlan etc. to speed up decapsulated traffic processing. CPU tag is a sort of "encapsulation", and we can use the same mechs to greatly improve overall DSA performance. skbs are passed to the GRO layer after removing CPU tags, so we don't need any new packet offload types as it was firstly proposed by me in the first GRO-over-DSA variant [1]. The size of struct gro_cells is sizeof(void *), so hot struct dsa_slave_priv becomes only 4/8 bytes bigger, and all critical fields remain in one 32-byte cacheline. The other positive side effect is that drivers for network devices that can be shipped as CPU ports of DSA-driven switches can now use napi_gro_frags() to pass skbs to kernel. Packets built that way are completely non-linear and are likely being dropped without GRO. This was tested on to-be-mainlined-soon Ethernet driver that uses napi_gro_frags(), and the overall performance was on par with the variant from [1], sometimes even better due to minimal overhead. net.core.gro_normal_batch tuning may help to push it to the limit on particular setups and platforms. iperf3 IPoE VLAN NAT TCP forwarding (port1.218 -> port0) setup on 1.2 GHz MIPS board: 5.7-rc2 baseline: [ID] Interval Transfer Bitrate Retr [ 5] 0.00-120.01 sec 9.00 GBytes 644 Mbits/sec 413 sender [ 5] 0.00-120.00 sec 8.99 GBytes 644 Mbits/sec receiver Iface RX packets TX packets eth0 7097731 7097702 port0 426050 6671829 port1 6671681 425862 port1.218 6671677 425851 With this patch: [ID] Interval Transfer Bitrate Retr [ 5] 0.00-120.01 sec 12.2 GBytes 870 Mbits/sec 122 sender [ 5] 0.00-120.00 sec 12.2 GBytes 870 Mbits/sec receiver Iface RX packets TX packets eth0 9474792 9474777 port0 455200 353288 port1 9019592 455035 port1.218 353144 455024 v2: - Add some performance examples in the commit message; - No functional changes. [1] https://lore.kernel.org/netdev/20191230143028.27313-1-alobakin@dlink.ru/ Signed-off-by: Alexander Lobakin Signed-off-by: David S. Miller Cc: Pali Rohár Signed-off-by: Greg Kroah-Hartman

net: dsa: Fix error cleanup path in dsa_init_module

2019-05-16T17:41:29+00:00

[ Upstream commit 68be930249d051fd54d3d99156b3dcadcb2a1f9b ] BUG: unable to handle kernel paging request at ffffffffa01c5430 PGD 3270067 P4D 3270067 PUD 3271063 PMD 230bc5067 PTE 0 Oops: 0000 [#1 CPU: 0 PID: 6159 Comm: modprobe Not tainted 5.1.0+ #33 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.9.3-0-ge2fc41e-prebuilt.qemu-project.org 04/01/2014 RIP: 0010:raw_notifier_chain_register+0x16/0x40 Code: 63 f8 66 90 e9 5d ff ff ff 90 90 90 90 90 90 90 90 90 90 90 55 48 8b 07 48 89 e5 48 85 c0 74 1c 8b 56 10 3b 50 10 7e 07 eb 12 <39> 50 10 7c 0d 48 8d 78 08 48 8b 40 08 48 85 c0 75 ee 48 89 46 08 RSP: 0018:ffffc90001c33c08 EFLAGS: 00010282 RAX: ffffffffa01c5420 RBX: ffffffffa01db420 RCX: 4fcef45928070a8b RDX: 0000000000000000 RSI: ffffffffa01db420 RDI: ffffffffa01b0068 RBP: ffffc90001c33c08 R08: 000000003e0a33d0 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000094443661 R12: ffff88822c320700 R13: ffff88823109be80 R14: 0000000000000000 R15: ffffc90001c33e78 FS: 00007fab8bd08540(0000) GS:ffff888237a00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffffa01c5430 CR3: 00000002297ea000 CR4: 00000000000006f0 Call Trace: register_netdevice_notifier+0x43/0x250 ? 0xffffffffa01e0000 dsa_slave_register_notifier+0x13/0x70 [dsa_core ? 0xffffffffa01e0000 dsa_init_module+0x2e/0x1000 [dsa_core do_one_initcall+0x6c/0x3cc ? do_init_module+0x22/0x1f1 ? rcu_read_lock_sched_held+0x97/0xb0 ? kmem_cache_alloc_trace+0x325/0x3b0 do_init_module+0x5b/0x1f1 load_module+0x1db1/0x2690 ? m_show+0x1d0/0x1d0 __do_sys_finit_module+0xc5/0xd0 __x64_sys_finit_module+0x15/0x20 do_syscall_64+0x6b/0x1d0 entry_SYSCALL_64_after_hwframe+0x49/0xbe Cleanup allocated resourses if there are errors, otherwise it will trgger memleak. Fixes: c9eb3e0f8701 ("net: dsa: Add support for learning FDB through notification") Signed-off-by: YueHaibing Reviewed-by: Vivien Didelot Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

net: dsa: Drop GPIO includes

2018-08-27T22:24:33+00:00

Commit 52638f71fcff ("dsa: Move gpio reset into switch driver") moved the GPIO handling into the switch drivers but forgot to remove the GPIO header includes. Signed-off-by: Linus Walleij Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller

net: dsa: forward timestamping callbacks to switch drivers

2018-02-14T19:33:37+00:00

Forward the rx/tx timestamp machinery from the dsa infrastructure to the switch driver. On the rx side, defer delivery of skbs until we have an rx timestamp. This mimicks the behavior of skb_defer_rx_timestamp. On the tx side, identify PTP packets, clone them, and pass them to the underlying switch driver before we transmit. This mimicks the behavior of skb_tx_timestamp. Adjusted txstamp API to keep the allocation and freeing of the clone in the same central function by Richard Cochran Signed-off-by: Brandon Streiff Signed-off-by: Richard Cochran Signed-off-by: Andrew Lunn Signed-off-by: David S. Miller

net: dsa: Support prepended Broadcom tag

2017-11-13T01:34:54+00:00

Add a new type: DSA_TAG_PROTO_PREPEND which allows us to support for the 4-bytes Broadcom tag that we already support, but in a format where it is pre-pended to the packet instead of located between the MAC SA and the Ethertyper (DSA_TAG_PROTO_BRCM). Signed-off-by: Florian Fainelli Reviewed-by: Andrew Lunn Signed-off-by: David S. Miller

net: dsa: move fixed link registration helpers

2017-10-28T09:50:12+00:00

The new bindings (dsa2.c) and the old bindings (legacy.c) share two helpers dsa_cpu_dsa_setup and dsa_cpu_dsa_destroy, used to register or deregister a fixed PHY if a given port has a corresponding device node. Unclutter the code by moving them into two new port.c helpers, dsa_port_fixed_link_register_of and dsa_port_fixed_link_(un)register_of. Signed-off-by: Vivien Didelot Signed-off-by: David S. Miller

net: dsa: use dsa_is_user_port everywhere

2017-10-27T15:00:09+00:00

Most of the DSA code still check ds->enabled_port_mask directly to inspect a given port type instead of using the provided dsa_is_user_port helper. Change this. Signed-off-by: Vivien Didelot Reviewed-by: Florian Fainelli Signed-off-by: David S. Miller

net: dsa: split dsa_port's netdev member

2017-10-18T11:24:33+00:00

The dsa_port structure has a "netdev" member, which can be used for either the master device, or the slave device, depending on its type. It is true that today, CPU port are not exposed to userspace, thus the port's netdev member can be used to point to its master interface. But it is still slightly confusing, so split it into more explicit "master" and "slave" members inside an anonymous union. Signed-off-by: Vivien Didelot Signed-off-by: David S. Miller

net: dsa: Add support for DSA specific notifiers

2017-10-12T19:10:01+00:00

In preparation for communicating a given DSA network device's port number and switch index, create a specialized DSA notifier and two events: DSA_PORT_REGISTER and DSA_PORT_UNREGISTER that communicate: the slave network device (slave_dev), port number and switch number in the tree. This will be later used for network device drivers like bcmsysport which needs to cooperate with its DSA network devices to set-up queue mapping and scheduling. Signed-off-by: Florian Fainelli Signed-off-by: David S. Miller

net: dsa: change dsa_ptr for a dsa_port

2017-10-01T03:15:07+00:00

With DSA, a master net device (CPU facing interface) has a dsa_ptr pointer to which hangs a dsa_switch_tree. This is not correct because a master interface is wired to a dedicated switch port, and because we can theoretically have several master interfaces pointing to several CPU ports of the same switch fabric. Change the master interface's dsa_ptr for the CPU dsa_port pointer. This is a step towards supporting multiple CPU ports. Signed-off-by: Vivien Didelot Signed-off-by: David S. Miller