Merge tag 'net-next-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next

Pull networking updates from Jakub Kicinski:
 "Core:

   - Enable memcg accounting for various networking objects.

  BPF:

   - Introduce bpf timers.

   - Add perf link and opaque bpf_cookie which the program can read out
     again, to be used in libbpf-based USDT library.

   - Add bpf_task_pt_regs() helper to access user space pt_regs in
     kprobes, to help user space stack unwinding.

   - Add support for UNIX sockets for BPF sockmap.

   - Extend BPF iterator support for UNIX domain sockets.

   - Allow BPF TCP congestion control progs and bpf iterators to call
     bpf_setsockopt(), e.g. to switch to another congestion control
     algorithm.

  Protocols:

   - Support IOAM Pre-allocated Trace with IPv6.

   - Support Management Component Transport Protocol.

   - bridge: multicast: add vlan support.

   - netfilter: add hooks for the SRv6 lightweight tunnel driver.

   - tcp:
       - enable mid-stream window clamping (by user space or BPF)
       - allow data-less, empty-cookie SYN with TFO_SERVER_COOKIE_NOT_REQD
       - more accurate DSACK processing for RACK-TLP

   - mptcp:
       - add full mesh path manager option
       - add partial support for MP_FAIL
       - improve use of backup subflows
       - optimize option processing

   - af_unix: add OOB notification support.

   - ipv6: add IFLA_INET6_RA_MTU to expose MTU value advertised by the
     router.

   - mac80211: Target Wake Time support in AP mode.

   - can: j1939: extend UAPI to notify about RX status.

  Driver APIs:

   - Add page frag support in page pool API.

   - Many improvements to the DSA (distributed switch) APIs.

   - ethtool: extend IRQ coalesce uAPI with timer reset modes.

   - devlink: control which auxiliary devices are created.

   - Support CAN PHYs via the generic PHY subsystem.

   - Proper cross-chip support for tag_8021q.

   - Allow TX forwarding for the software bridge data path to be
     offloaded to capable devices.

  Drivers:

   - veth: more flexible channels number configuration.

   - openvswitch: introduce per-cpu upcall dispatch.

   - Add internet mix (IMIX) mode to pktgen.

   - Transparently handle XDP operations in the bonding driver.

   - Add LiteETH network driver.

   - Renesas (ravb):
       - support Gigabit Ethernet IP

   - NXP Ethernet switch (sja1105):
       - fast aging support
       - support for "H" switch topologies
       - traffic termination for ports under VLAN-aware bridge

   - Intel 1G Ethernet
       - support getcrosststamp() with PCIe PTM (Precision Time
         Measurement) for better time sync
       - support Credit-Based Shaper (CBS) offload, enabling HW traffic
         prioritization and bandwidth reservation

   - Broadcom Ethernet (bnxt)
       - support pulse-per-second output
       - support larger Rx rings

   - Mellanox Ethernet (mlx5)
       - support ethtool RSS contexts and MQPRIO channel mode
       - support LAG offload with bridging
       - support devlink rate limit API
       - support packet sampling on tunnels

   - Huawei Ethernet (hns3):
       - basic devlink support
       - add extended IRQ coalescing support
       - report extended link state

   - Netronome Ethernet (nfp):
       - add conntrack offload support

   - Broadcom WiFi (brcmfmac):
       - add WPA3 Personal with FT to supported cipher suites
       - support 43752 SDIO device

   - Intel WiFi (iwlwifi):
       - support scanning hidden 6GHz networks
       - support for a new hardware family (Bz)

   - Xen pv driver:
       - harden netfront against malicious backends

   - Qualcomm mobile
       - ipa: refactor power management and enable automatic suspend
       - mhi: move MBIM to WWAN subsystem interfaces

  Refactor:

   - Ambient BPF run context and cgroup storage cleanup.

   - Compat rework for ndo_ioctl.

  Old code removal:

   - prism54 remove the obsoleted driver, deprecated by the p54 driver.

   - wan: remove sbni/granch driver"

* tag 'net-next-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1715 commits)
  net: Add depends on OF_NET for LiteX's LiteETH
  ipv6: seg6: remove duplicated include
  net: hns3: remove unnecessary spaces
  net: hns3: add some required spaces
  net: hns3: clean up a type mismatch warning
  net: hns3: refine function hns3_set_default_feature()
  ipv6: remove duplicated 'net/lwtunnel.h' include
  net: w5100: check return value after calling platform_get_resource()
  net/mlxbf_gige: Make use of devm_platform_ioremap_resourcexxx()
  net: mdio: mscc-miim: Make use of the helper function devm_platform_ioremap_resource()
  net: mdio-ipq4019: Make use of devm_platform_ioremap_resource()
  fou: remove sparse errors
  ipv4: fix endianness issue in inet_rtm_getroute_build_skb()
  octeontx2-af: Set proper errorcode for IPv4 checksum errors
  octeontx2-af: Fix static code analyzer reported issues
  octeontx2-af: Fix mailbox errors in nix_rss_flowkey_cfg
  octeontx2-af: Fix loop in free and unmap counter
  af_unix: fix potential NULL deref in unix_dgram_connect()
  dpaa2-eth: Replace strlcpy with strscpy
  octeontx2-af: Use NDC TX for transmit packet data
  ...

1 files changed, 254 insertions, 0 deletions

diff --git a/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c b/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
new file mode 100644
index 000000000000..5eea3c3a40fe
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
@@ -0,0 +1,254 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2021 Facebook */
+#define _GNU_SOURCE
+#include <pthread.h>
+#include <sched.h>
+#include <sys/syscall.h>
+#include <unistd.h>
+#include <test_progs.h>
+#include "test_bpf_cookie.skel.h"
+
+static void kprobe_subtest(struct test_bpf_cookie *skel)
+{
+	DECLARE_LIBBPF_OPTS(bpf_kprobe_opts, opts);
+	struct bpf_link *link1 = NULL, *link2 = NULL;
+	struct bpf_link *retlink1 = NULL, *retlink2 = NULL;
+
+	/* attach two kprobes */
+	opts.bpf_cookie = 0x1;
+	opts.retprobe = false;
+	link1 = bpf_program__attach_kprobe_opts(skel->progs.handle_kprobe,
+						 SYS_NANOSLEEP_KPROBE_NAME, &opts);
+	if (!ASSERT_OK_PTR(link1, "link1"))
+		goto cleanup;
+
+	opts.bpf_cookie = 0x2;
+	opts.retprobe = false;
+	link2 = bpf_program__attach_kprobe_opts(skel->progs.handle_kprobe,
+						 SYS_NANOSLEEP_KPROBE_NAME, &opts);
+	if (!ASSERT_OK_PTR(link2, "link2"))
+		goto cleanup;
+
+	/* attach two kretprobes */
+	opts.bpf_cookie = 0x10;
+	opts.retprobe = true;
+	retlink1 = bpf_program__attach_kprobe_opts(skel->progs.handle_kretprobe,
+						    SYS_NANOSLEEP_KPROBE_NAME, &opts);
+	if (!ASSERT_OK_PTR(retlink1, "retlink1"))
+		goto cleanup;
+
+	opts.bpf_cookie = 0x20;
+	opts.retprobe = true;
+	retlink2 = bpf_program__attach_kprobe_opts(skel->progs.handle_kretprobe,
+						    SYS_NANOSLEEP_KPROBE_NAME, &opts);
+	if (!ASSERT_OK_PTR(retlink2, "retlink2"))
+		goto cleanup;
+
+	/* trigger kprobe && kretprobe */
+	usleep(1);
+
+	ASSERT_EQ(skel->bss->kprobe_res, 0x1 | 0x2, "kprobe_res");
+	ASSERT_EQ(skel->bss->kretprobe_res, 0x10 | 0x20, "kretprobe_res");
+
+cleanup:
+	bpf_link__destroy(link1);
+	bpf_link__destroy(link2);
+	bpf_link__destroy(retlink1);
+	bpf_link__destroy(retlink2);
+}
+
+static void uprobe_subtest(struct test_bpf_cookie *skel)
+{
+	DECLARE_LIBBPF_OPTS(bpf_uprobe_opts, opts);
+	struct bpf_link *link1 = NULL, *link2 = NULL;
+	struct bpf_link *retlink1 = NULL, *retlink2 = NULL;
+	size_t uprobe_offset;
+	ssize_t base_addr;
+
+	base_addr = get_base_addr();
+	uprobe_offset = get_uprobe_offset(&get_base_addr, base_addr);
+
+	/* attach two uprobes */
+	opts.bpf_cookie = 0x100;
+	opts.retprobe = false;
+	link1 = bpf_program__attach_uprobe_opts(skel->progs.handle_uprobe, 0 /* self pid */,
+						"/proc/self/exe", uprobe_offset, &opts);
+	if (!ASSERT_OK_PTR(link1, "link1"))
+		goto cleanup;
+
+	opts.bpf_cookie = 0x200;
+	opts.retprobe = false;
+	link2 = bpf_program__attach_uprobe_opts(skel->progs.handle_uprobe, -1 /* any pid */,
+						"/proc/self/exe", uprobe_offset, &opts);
+	if (!ASSERT_OK_PTR(link2, "link2"))
+		goto cleanup;
+
+	/* attach two uretprobes */
+	opts.bpf_cookie = 0x1000;
+	opts.retprobe = true;
+	retlink1 = bpf_program__attach_uprobe_opts(skel->progs.handle_uretprobe, -1 /* any pid */,
+						   "/proc/self/exe", uprobe_offset, &opts);
+	if (!ASSERT_OK_PTR(retlink1, "retlink1"))
+		goto cleanup;
+
+	opts.bpf_cookie = 0x2000;
+	opts.retprobe = true;
+	retlink2 = bpf_program__attach_uprobe_opts(skel->progs.handle_uretprobe, 0 /* self pid */,
+						   "/proc/self/exe", uprobe_offset, &opts);
+	if (!ASSERT_OK_PTR(retlink2, "retlink2"))
+		goto cleanup;
+
+	/* trigger uprobe && uretprobe */
+	get_base_addr();
+
+	ASSERT_EQ(skel->bss->uprobe_res, 0x100 | 0x200, "uprobe_res");
+	ASSERT_EQ(skel->bss->uretprobe_res, 0x1000 | 0x2000, "uretprobe_res");
+
+cleanup:
+	bpf_link__destroy(link1);
+	bpf_link__destroy(link2);
+	bpf_link__destroy(retlink1);
+	bpf_link__destroy(retlink2);
+}
+
+static void tp_subtest(struct test_bpf_cookie *skel)
+{
+	DECLARE_LIBBPF_OPTS(bpf_tracepoint_opts, opts);
+	struct bpf_link *link1 = NULL, *link2 = NULL, *link3 = NULL;
+
+	/* attach first tp prog */
+	opts.bpf_cookie = 0x10000;
+	link1 = bpf_program__attach_tracepoint_opts(skel->progs.handle_tp1,
+						    "syscalls", "sys_enter_nanosleep", &opts);
+	if (!ASSERT_OK_PTR(link1, "link1"))
+		goto cleanup;
+
+	/* attach second tp prog */
+	opts.bpf_cookie = 0x20000;
+	link2 = bpf_program__attach_tracepoint_opts(skel->progs.handle_tp2,
+						    "syscalls", "sys_enter_nanosleep", &opts);
+	if (!ASSERT_OK_PTR(link2, "link2"))
+		goto cleanup;
+
+	/* trigger tracepoints */
+	usleep(1);
+
+	ASSERT_EQ(skel->bss->tp_res, 0x10000 | 0x20000, "tp_res1");
+
+	/* now we detach first prog and will attach third one, which causes
+	 * two internal calls to bpf_prog_array_copy(), shuffling
+	 * bpf_prog_array_items around. We test here that we don't lose track
+	 * of associated bpf_cookies.
+	 */
+	bpf_link__destroy(link1);
+	link1 = NULL;
+	kern_sync_rcu();
+	skel->bss->tp_res = 0;
+
+	/* attach third tp prog */
+	opts.bpf_cookie = 0x40000;
+	link3 = bpf_program__attach_tracepoint_opts(skel->progs.handle_tp3,
+						    "syscalls", "sys_enter_nanosleep", &opts);
+	if (!ASSERT_OK_PTR(link3, "link3"))
+		goto cleanup;
+
+	/* trigger tracepoints */
+	usleep(1);
+
+	ASSERT_EQ(skel->bss->tp_res, 0x20000 | 0x40000, "tp_res2");
+
+cleanup:
+	bpf_link__destroy(link1);
+	bpf_link__destroy(link2);
+	bpf_link__destroy(link3);
+}
+
+static void burn_cpu(void)
+{
+	volatile int j = 0;
+	cpu_set_t cpu_set;
+	int i, err;
+
+	/* generate some branches on cpu 0 */
+	CPU_ZERO(&cpu_set);
+	CPU_SET(0, &cpu_set);
+	err = pthread_setaffinity_np(pthread_self(), sizeof(cpu_set), &cpu_set);
+	ASSERT_OK(err, "set_thread_affinity");
+
+	/* spin the loop for a while (random high number) */
+	for (i = 0; i < 1000000; ++i)
+		++j;
+}
+
+static void pe_subtest(struct test_bpf_cookie *skel)
+{
+	DECLARE_LIBBPF_OPTS(bpf_perf_event_opts, opts);
+	struct bpf_link *link = NULL;
+	struct perf_event_attr attr;
+	int pfd = -1;
+
+	/* create perf event */
+	memset(&attr, 0, sizeof(attr));
+	attr.size = sizeof(attr);
+	attr.type = PERF_TYPE_SOFTWARE;
+	attr.config = PERF_COUNT_SW_CPU_CLOCK;
+	attr.freq = 1;
+	attr.sample_freq = 4000;
+	pfd = syscall(__NR_perf_event_open, &attr, -1, 0, -1, PERF_FLAG_FD_CLOEXEC);
+	if (!ASSERT_GE(pfd, 0, "perf_fd"))
+		goto cleanup;
+
+	opts.bpf_cookie = 0x100000;
+	link = bpf_program__attach_perf_event_opts(skel->progs.handle_pe, pfd, &opts);
+	if (!ASSERT_OK_PTR(link, "link1"))
+		goto cleanup;
+
+	burn_cpu(); /* trigger BPF prog */
+
+	ASSERT_EQ(skel->bss->pe_res, 0x100000, "pe_res1");
+
+	/* prevent bpf_link__destroy() closing pfd itself */
+	bpf_link__disconnect(link);
+	/* close BPF link's FD explicitly */
+	close(bpf_link__fd(link));
+	/* free up memory used by struct bpf_link */
+	bpf_link__destroy(link);
+	link = NULL;
+	kern_sync_rcu();
+	skel->bss->pe_res = 0;
+
+	opts.bpf_cookie = 0x200000;
+	link = bpf_program__attach_perf_event_opts(skel->progs.handle_pe, pfd, &opts);
+	if (!ASSERT_OK_PTR(link, "link2"))
+		goto cleanup;
+
+	burn_cpu(); /* trigger BPF prog */
+
+	ASSERT_EQ(skel->bss->pe_res, 0x200000, "pe_res2");
+
+cleanup:
+	close(pfd);
+	bpf_link__destroy(link);
+}
+
+void test_bpf_cookie(void)
+{
+	struct test_bpf_cookie *skel;
+
+	skel = test_bpf_cookie__open_and_load();
+	if (!ASSERT_OK_PTR(skel, "skel_open"))
+		return;
+
+	skel->bss->my_tid = syscall(SYS_gettid);
+
+	if (test__start_subtest("kprobe"))
+		kprobe_subtest(skel);
+	if (test__start_subtest("uprobe"))
+		uprobe_subtest(skel);
+	if (test__start_subtest("tracepoint"))
+		tp_subtest(skel);
+	if (test__start_subtest("perf_event"))
+		pe_subtest(skel);
+
+	test_bpf_cookie__destroy(skel);
+}