Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-nextgrafted

Pull networking updates from Jakub Kicinski:
 "Core:

   - Add dedicated kmem_cache for typical/small skb->head, avoid having
     to access struct page at kfree time, and improve memory use.

   - Introduce sysctl to set default RPS configuration for new netdevs.

   - Define Netlink protocol specification format which can be used to
     describe messages used by each family and auto-generate parsers.
     Add tools for generating kernel data structures and uAPI headers.

   - Expose all net/core sysctls inside netns.

   - Remove 4s sleep in netpoll if carrier is instantly detected on
     boot.

   - Add configurable limit of MDB entries per port, and port-vlan.

   - Continue populating drop reasons throughout the stack.

   - Retire a handful of legacy Qdiscs and classifiers.

  Protocols:

   - Support IPv4 big TCP (TSO frames larger than 64kB).

   - Add IP_LOCAL_PORT_RANGE socket option, to control local port range
     on socket by socket basis.

   - Track and report in procfs number of MPTCP sockets used.

   - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path
     manager.

   - IPv6: don't check net.ipv6.route.max_size and rely on garbage
     collection to free memory (similarly to IPv4).

   - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986).

   - ICMP: add per-rate limit counters.

   - Add support for user scanning requests in ieee802154.

   - Remove static WEP support.

   - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate
     reporting.

   - WiFi 7 EHT channel puncturing support (client & AP).

  BPF:

   - Add a rbtree data structure following the "next-gen data structure"
     precedent set by recently added linked list, that is, by using
     kfunc + kptr instead of adding a new BPF map type.

   - Expose XDP hints via kfuncs with initial support for RX hash and
     timestamp metadata.

   - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to
     better support decap on GRE tunnel devices not operating in collect
     metadata.

   - Improve x86 JIT's codegen for PROBE_MEM runtime error checks.

   - Remove the need for trace_printk_lock for bpf_trace_printk and
     bpf_trace_vprintk helpers.

   - Extend libbpf's bpf_tracing.h support for tracing arguments of
     kprobes/uprobes and syscall as a special case.

   - Significantly reduce the search time for module symbols by
     livepatch and BPF.

   - Enable cpumasks to be used as kptrs, which is useful for tracing
     programs tracking which tasks end up running on which CPUs in
     different time intervals.

   - Add support for BPF trampoline on s390x and riscv64.

   - Add capability to export the XDP features supported by the NIC.

   - Add __bpf_kfunc tag for marking kernel functions as kfuncs.

   - Add cgroup.memory=nobpf kernel parameter option to disable BPF
     memory accounting for container environments.

  Netfilter:

   - Remove the CLUSTERIP target. It has been marked as obsolete for
     years, and we still have WARN splats wrt races of the out-of-band
     /proc interface installed by this target.

   - Add 'destroy' commands to nf_tables. They are identical to the
     existing 'delete' commands, but do not return an error if the
     referenced object (set, chain, rule...) did not exist.

  Driver API:

   - Improve cpumask_local_spread() locality to help NICs set the right
     IRQ affinity on AMD platforms.

   - Separate C22 and C45 MDIO bus transactions more clearly.

   - Introduce new DCB table to control DSCP rewrite on egress.

   - Support configuration of Physical Layer Collision Avoidance (PLCA)
     Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of
     shared medium Ethernet.

   - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing
     preemption of low priority frames by high priority frames.

   - Add support for controlling MACSec offload using netlink SET.

   - Rework devlink instance refcounts to allow registration and
     de-registration under the instance lock. Split the code into
     multiple files, drop some of the unnecessarily granular locks and
     factor out common parts of netlink operation handling.

   - Add TX frame aggregation parameters (for USB drivers).

   - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning
     messages with notifications for debug.

   - Allow offloading of UDP NEW connections via act_ct.

   - Add support for per action HW stats in TC.

   - Support hardware miss to TC action (continue processing in SW from
     a specific point in the action chain).

   - Warn if old Wireless Extension user space interface is used with
     modern cfg80211/mac80211 drivers. Do not support Wireless
     Extensions for Wi-Fi 7 devices at all. Everyone should switch to
     using nl80211 interface instead.

   - Improve the CAN bit timing configuration. Use extack to return
     error messages directly to user space, update the SJW handling,
     including the definition of a new default value that will benefit
     CAN-FD controllers, by increasing their oscillator tolerance.

  New hardware / drivers:

   - Ethernet:
      - nVidia BlueField-3 support (control traffic driver)
      - Ethernet support for imx93 SoCs
      - Motorcomm yt8531 gigabit Ethernet PHY
      - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA)
      - Microchip LAN8841 PHY (incl. cable diagnostics and PTP)
      - Amlogic gxl MDIO mux

   - WiFi:
      - RealTek RTL8188EU (rtl8xxxu)
      - Qualcomm Wi-Fi 7 devices (ath12k)

   - CAN:
      - Renesas R-Car V4H

  Drivers:

   - Bluetooth:
      - Set Per Platform Antenna Gain (PPAG) for Intel controllers.

   - Ethernet NICs:
      - Intel (1G, igc):
         - support TSN / Qbv / packet scheduling features of i226 model
      - Intel (100G, ice):
         - use GNSS subsystem instead of TTY
         - multi-buffer XDP support
         - extend support for GPIO pins to E823 devices
      - nVidia/Mellanox:
         - update the shared buffer configuration on PFC commands
         - implement PTP adjphase function for HW offset control
         - TC support for Geneve and GRE with VF tunnel offload
         - more efficient crypto key management method
         - multi-port eswitch support
      - Netronome/Corigine:
         - add DCB IEEE support
         - support IPsec offloading for NFP3800
      - Freescale/NXP (enetc):
         - support XDP_REDIRECT for XDP non-linear buffers
         - improve reconfig, avoid link flap and waiting for idle
         - support MAC Merge layer
      - Other NICs:
         - sfc/ef100: add basic devlink support for ef100
         - ionic: rx_push mode operation (writing descriptors via MMIO)
         - bnxt: use the auxiliary bus abstraction for RDMA
         - r8169: disable ASPM and reset bus in case of tx timeout
         - cpsw: support QSGMII mode for J721e CPSW9G
         - cpts: support pulse-per-second output
         - ngbe: add an mdio bus driver
         - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing
         - r8152: handle devices with FW with NCM support
         - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation
         - virtio-net: support multi buffer XDP
         - virtio/vsock: replace virtio_vsock_pkt with sk_buff
         - tsnep: XDP support

   - Ethernet high-speed switches:
      - nVidia/Mellanox (mlxsw):
         - add support for latency TLV (in FW control messages)
      - Microchip (sparx5):
         - separate explicit and implicit traffic forwarding rules, make
           the implicit rules always active
         - add support for egress DSCP rewrite
         - IS0 VCAP support (Ingress Classification)
         - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS
           etc.)
         - ES2 VCAP support (Egress Access Control)
         - support for Per-Stream Filtering and Policing (802.1Q,
           8.6.5.1)

   - Ethernet embedded switches:
      - Marvell (mv88e6xxx):
         - add MAB (port auth) offload support
         - enable PTP receive for mv88e6390
      - NXP (ocelot):
         - support MAC Merge layer
         - support for the the vsc7512 internal copper phys
      - Microchip:
         - lan9303: convert to PHYLINK
         - lan966x: support TC flower filter statistics
         - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x
         - lan937x: support Credit Based Shaper configuration
         - ksz9477: support Energy Efficient Ethernet
      - other:
         - qca8k: convert to regmap read/write API, use bulk operations
         - rswitch: Improve TX timestamp accuracy

   - Intel WiFi (iwlwifi):
      - EHT (Wi-Fi 7) rate reporting
      - STEP equalizer support: transfer some STEP (connection to radio
        on platforms with integrated wifi) related parameters from the
        BIOS to the firmware.

   - Qualcomm 802.11ax WiFi (ath11k):
      - IPQ5018 support
      - Fine Timing Measurement (FTM) responder role support
      - channel 177 support

   - MediaTek WiFi (mt76):
      - per-PHY LED support
      - mt7996: EHT (Wi-Fi 7) support
      - Wireless Ethernet Dispatch (WED) reset support
      - switch to using page pool allocator

   - RealTek WiFi (rtw89):
      - support new version of Bluetooth co-existance

   - Mobile:
      - rmnet: support TX aggregation"

* tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits)
  page_pool: add a comment explaining the fragment counter usage
  net: ethtool: fix __ethtool_dev_mm_supported() implementation
  ethtool: pse-pd: Fix double word in comments
  xsk: add linux/vmalloc.h to xsk.c
  sefltests: netdevsim: wait for devlink instance after netns removal
  selftest: fib_tests: Always cleanup before exit
  net/mlx5e: Align IPsec ASO result memory to be as required by hardware
  net/mlx5e: TC, Set CT miss to the specific ct action instance
  net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG
  net/mlx5: Refactor tc miss handling to a single function
  net/mlx5: Kconfig: Make tc offload depend on tc skb extension
  net/sched: flower: Support hardware miss to tc action
  net/sched: flower: Move filter handle initialization earlier
  net/sched: cls_api: Support hardware miss to tc action
  net/sched: Rename user cookie and act cookie
  sfc: fix builds without CONFIG_RTC_LIB
  sfc: clean up some inconsistent indentings
  net/mlx4_en: Introduce flexible array to silence overflow warning
  net: lan966x: Fix possible deadlock inside PTP
  net/ulp: Remove redundant ->clone() test in inet_clone_ulp().
  ...

1 files changed, 755 insertions, 0 deletions

diff --git a/tools/virtio/vringh_test.c b/tools/virtio/vringh_test.c
new file mode 100644
index 000000000..98ff808d6
--- /dev/null
+++ b/tools/virtio/vringh_test.c
@@ -0,0 +1,755 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Simple test of virtio code, entirely in userpsace. */
+#define _GNU_SOURCE
+#include <sched.h>
+#include <err.h>
+#include <linux/kernel.h>
+#include <linux/err.h>
+#include <linux/virtio.h>
+#include <linux/vringh.h>
+#include <linux/virtio_ring.h>
+#include <linux/virtio_config.h>
+#include <linux/uaccess.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/mman.h>
+#include <sys/wait.h>
+#include <fcntl.h>
+
+#define USER_MEM (1024*1024)
+void *__user_addr_min, *__user_addr_max;
+void *__kmalloc_fake, *__kfree_ignore_start, *__kfree_ignore_end;
+static u64 user_addr_offset;
+
+#define RINGSIZE 256
+#define ALIGN 4096
+
+static bool never_notify_host(struct virtqueue *vq)
+{
+	abort();
+}
+
+static void never_callback_guest(struct virtqueue *vq)
+{
+	abort();
+}
+
+static bool getrange_iov(struct vringh *vrh, u64 addr, struct vringh_range *r)
+{
+	if (addr < (u64)(unsigned long)__user_addr_min - user_addr_offset)
+		return false;
+	if (addr >= (u64)(unsigned long)__user_addr_max - user_addr_offset)
+		return false;
+
+	r->start = (u64)(unsigned long)__user_addr_min - user_addr_offset;
+	r->end_incl = (u64)(unsigned long)__user_addr_max - 1 - user_addr_offset;
+	r->offset = user_addr_offset;
+	return true;
+}
+
+/* We return single byte ranges. */
+static bool getrange_slow(struct vringh *vrh, u64 addr, struct vringh_range *r)
+{
+	if (addr < (u64)(unsigned long)__user_addr_min - user_addr_offset)
+		return false;
+	if (addr >= (u64)(unsigned long)__user_addr_max - user_addr_offset)
+		return false;
+
+	r->start = addr;
+	r->end_incl = r->start;
+	r->offset = user_addr_offset;
+	return true;
+}
+
+struct guest_virtio_device {
+	struct virtio_device vdev;
+	int to_host_fd;
+	unsigned long notifies;
+};
+
+static bool parallel_notify_host(struct virtqueue *vq)
+{
+	int rc;
+	struct guest_virtio_device *gvdev;
+
+	gvdev = container_of(vq->vdev, struct guest_virtio_device, vdev);
+	rc = write(gvdev->to_host_fd, "", 1);
+	if (rc < 0)
+		return false;
+	gvdev->notifies++;
+	return true;
+}
+
+static bool no_notify_host(struct virtqueue *vq)
+{
+	return true;
+}
+
+#define NUM_XFERS (10000000)
+
+/* We aim for two "distant" cpus. */
+static void find_cpus(unsigned int *first, unsigned int *last)
+{
+	unsigned int i;
+
+	*first = -1U;
+	*last = 0;
+	for (i = 0; i < 4096; i++) {
+		cpu_set_t set;
+		CPU_ZERO(&set);
+		CPU_SET(i, &set);
+		if (sched_setaffinity(getpid(), sizeof(set), &set) == 0) {
+			if (i < *first)
+				*first = i;
+			if (i > *last)
+				*last = i;
+		}
+	}
+}
+
+/* Opencoded version for fast mode */
+static inline int vringh_get_head(struct vringh *vrh, u16 *head)
+{
+	u16 avail_idx, i;
+	int err;
+
+	err = get_user(avail_idx, &vrh->vring.avail->idx);
+	if (err)
+		return err;
+
+	if (vrh->last_avail_idx == avail_idx)
+		return 0;
+
+	/* Only get avail ring entries after they have been exposed by guest. */
+	virtio_rmb(vrh->weak_barriers);
+
+	i = vrh->last_avail_idx & (vrh->vring.num - 1);
+
+	err = get_user(*head, &vrh->vring.avail->ring[i]);
+	if (err)
+		return err;
+
+	vrh->last_avail_idx++;
+	return 1;
+}
+
+static int parallel_test(u64 features,
+			 bool (*getrange)(struct vringh *vrh,
+					  u64 addr, struct vringh_range *r),
+			 bool fast_vringh)
+{
+	void *host_map, *guest_map;
+	int fd, mapsize, to_guest[2], to_host[2];
+	unsigned long xfers = 0, notifies = 0, receives = 0;
+	unsigned int first_cpu, last_cpu;
+	cpu_set_t cpu_set;
+	char buf[128];
+
+	/* Create real file to mmap. */
+	fd = open("/tmp/vringh_test-file", O_RDWR|O_CREAT|O_TRUNC, 0600);
+	if (fd < 0)
+		err(1, "Opening /tmp/vringh_test-file");
+
+	/* Extra room at the end for some data, and indirects */
+	mapsize = vring_size(RINGSIZE, ALIGN)
+		+ RINGSIZE * 2 * sizeof(int)
+		+ RINGSIZE * 6 * sizeof(struct vring_desc);
+	mapsize = (mapsize + getpagesize() - 1) & ~(getpagesize() - 1);
+	ftruncate(fd, mapsize);
+
+	/* Parent and child use separate addresses, to check our mapping logic! */
+	host_map = mmap(NULL, mapsize, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);
+	guest_map = mmap(NULL, mapsize, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);
+
+	pipe(to_guest);
+	pipe(to_host);
+
+	CPU_ZERO(&cpu_set);
+	find_cpus(&first_cpu, &last_cpu);
+	printf("Using CPUS %u and %u\n", first_cpu, last_cpu);
+	fflush(stdout);
+
+	if (fork() != 0) {
+		struct vringh vrh;
+		int status, err, rlen = 0;
+		char rbuf[5];
+
+		/* We are the host: never access guest addresses! */
+		munmap(guest_map, mapsize);
+
+		__user_addr_min = host_map;
+		__user_addr_max = __user_addr_min + mapsize;
+		user_addr_offset = host_map - guest_map;
+		assert(user_addr_offset);
+
+		close(to_guest[0]);
+		close(to_host[1]);
+
+		vring_init(&vrh.vring, RINGSIZE, host_map, ALIGN);
+		vringh_init_user(&vrh, features, RINGSIZE, true,
+				 vrh.vring.desc, vrh.vring.avail, vrh.vring.used);
+		CPU_SET(first_cpu, &cpu_set);
+		if (sched_setaffinity(getpid(), sizeof(cpu_set), &cpu_set))
+			errx(1, "Could not set affinity to cpu %u", first_cpu);
+
+		while (xfers < NUM_XFERS) {
+			struct iovec host_riov[2], host_wiov[2];
+			struct vringh_iov riov, wiov;
+			u16 head, written;
+
+			if (fast_vringh) {
+				for (;;) {
+					err = vringh_get_head(&vrh, &head);
+					if (err != 0)
+						break;
+					err = vringh_need_notify_user(&vrh);
+					if (err < 0)
+						errx(1, "vringh_need_notify_user: %i",
+						     err);
+					if (err) {
+						write(to_guest[1], "", 1);
+						notifies++;
+					}
+				}
+				if (err != 1)
+					errx(1, "vringh_get_head");
+				written = 0;
+				goto complete;
+			} else {
+				vringh_iov_init(&riov,
+						host_riov,
+						ARRAY_SIZE(host_riov));
+				vringh_iov_init(&wiov,
+						host_wiov,
+						ARRAY_SIZE(host_wiov));
+
+				err = vringh_getdesc_user(&vrh, &riov, &wiov,
+							  getrange, &head);
+			}
+			if (err == 0) {
+				err = vringh_need_notify_user(&vrh);
+				if (err < 0)
+					errx(1, "vringh_need_notify_user: %i",
+					     err);
+				if (err) {
+					write(to_guest[1], "", 1);
+					notifies++;
+				}
+
+				if (!vringh_notify_enable_user(&vrh))
+					continue;
+
+				/* Swallow all notifies at once. */
+				if (read(to_host[0], buf, sizeof(buf)) < 1)
+					break;
+
+				vringh_notify_disable_user(&vrh);
+				receives++;
+				continue;
+			}
+			if (err != 1)
+				errx(1, "vringh_getdesc_user: %i", err);
+
+			/* We simply copy bytes. */
+			if (riov.used) {
+				rlen = vringh_iov_pull_user(&riov, rbuf,
+							    sizeof(rbuf));
+				if (rlen != 4)
+					errx(1, "vringh_iov_pull_user: %i",
+					     rlen);
+				assert(riov.i == riov.used);
+				written = 0;
+			} else {
+				err = vringh_iov_push_user(&wiov, rbuf, rlen);
+				if (err != rlen)
+					errx(1, "vringh_iov_push_user: %i",
+					     err);
+				assert(wiov.i == wiov.used);
+				written = err;
+			}
+		complete:
+			xfers++;
+
+			err = vringh_complete_user(&vrh, head, written);
+			if (err != 0)
+				errx(1, "vringh_complete_user: %i", err);
+		}
+
+		err = vringh_need_notify_user(&vrh);
+		if (err < 0)
+			errx(1, "vringh_need_notify_user: %i", err);
+		if (err) {
+			write(to_guest[1], "", 1);
+			notifies++;
+		}
+		wait(&status);
+		if (!WIFEXITED(status))
+			errx(1, "Child died with signal %i?", WTERMSIG(status));
+		if (WEXITSTATUS(status) != 0)
+			errx(1, "Child exited %i?", WEXITSTATUS(status));
+		printf("Host: notified %lu, pinged %lu\n", notifies, receives);
+		return 0;
+	} else {
+		struct guest_virtio_device gvdev;
+		struct virtqueue *vq;
+		unsigned int *data;
+		struct vring_desc *indirects;
+		unsigned int finished = 0;
+
+		/* We pass sg[]s pointing into here, but we need RINGSIZE+1 */
+		data = guest_map + vring_size(RINGSIZE, ALIGN);
+		indirects = (void *)data + (RINGSIZE + 1) * 2 * sizeof(int);
+
+		/* We are the guest. */
+		munmap(host_map, mapsize);
+
+		close(to_guest[1]);
+		close(to_host[0]);
+
+		gvdev.vdev.features = features;
+		INIT_LIST_HEAD(&gvdev.vdev.vqs);
+		spin_lock_init(&gvdev.vdev.vqs_list_lock);
+		gvdev.to_host_fd = to_host[1];
+		gvdev.notifies = 0;
+
+		CPU_SET(first_cpu, &cpu_set);
+		if (sched_setaffinity(getpid(), sizeof(cpu_set), &cpu_set))
+			err(1, "Could not set affinity to cpu %u", first_cpu);
+
+		vq = vring_new_virtqueue(0, RINGSIZE, ALIGN, &gvdev.vdev, true,
+					 false, guest_map,
+					 fast_vringh ? no_notify_host
+					 : parallel_notify_host,
+					 never_callback_guest, "guest vq");
+
+		/* Don't kfree indirects. */
+		__kfree_ignore_start = indirects;
+		__kfree_ignore_end = indirects + RINGSIZE * 6;
+
+		while (xfers < NUM_XFERS) {
+			struct scatterlist sg[4];
+			unsigned int num_sg, len;
+			int *dbuf, err;
+			bool output = !(xfers % 2);
+
+			/* Consume bufs. */
+			while ((dbuf = virtqueue_get_buf(vq, &len)) != NULL) {
+				if (len == 4)
+					assert(*dbuf == finished - 1);
+				else if (!fast_vringh)
+					assert(*dbuf == finished);
+				finished++;
+			}
+
+			/* Produce a buffer. */
+			dbuf = data + (xfers % (RINGSIZE + 1));
+
+			if (output)
+				*dbuf = xfers;
+			else
+				*dbuf = -1;
+
+			switch ((xfers / sizeof(*dbuf)) % 4) {
+			case 0:
+				/* Nasty three-element sg list. */
+				sg_init_table(sg, num_sg = 3);
+				sg_set_buf(&sg[0], (void *)dbuf, 1);
+				sg_set_buf(&sg[1], (void *)dbuf + 1, 2);
+				sg_set_buf(&sg[2], (void *)dbuf + 3, 1);
+				break;
+			case 1:
+				sg_init_table(sg, num_sg = 2);
+				sg_set_buf(&sg[0], (void *)dbuf, 1);
+				sg_set_buf(&sg[1], (void *)dbuf + 1, 3);
+				break;
+			case 2:
+				sg_init_table(sg, num_sg = 1);
+				sg_set_buf(&sg[0], (void *)dbuf, 4);
+				break;
+			case 3:
+				sg_init_table(sg, num_sg = 4);
+				sg_set_buf(&sg[0], (void *)dbuf, 1);
+				sg_set_buf(&sg[1], (void *)dbuf + 1, 1);
+				sg_set_buf(&sg[2], (void *)dbuf + 2, 1);
+				sg_set_buf(&sg[3], (void *)dbuf + 3, 1);
+				break;
+			}
+
+			/* May allocate an indirect, so force it to allocate
+			 * user addr */
+			__kmalloc_fake = indirects + (xfers % RINGSIZE) * 4;
+			if (output)
+				err = virtqueue_add_outbuf(vq, sg, num_sg, dbuf,
+							   GFP_KERNEL);
+			else
+				err = virtqueue_add_inbuf(vq, sg, num_sg,
+							  dbuf, GFP_KERNEL);
+
+			if (err == -ENOSPC) {
+				if (!virtqueue_enable_cb_delayed(vq))
+					continue;
+				/* Swallow all notifies at once. */
+				if (read(to_guest[0], buf, sizeof(buf)) < 1)
+					break;
+				
+				receives++;
+				virtqueue_disable_cb(vq);
+				continue;
+			}
+
+			if (err)
+				errx(1, "virtqueue_add_in/outbuf: %i", err);
+
+			xfers++;
+			virtqueue_kick(vq);
+		}
+
+		/* Any extra? */
+		while (finished != xfers) {
+			int *dbuf;
+			unsigned int len;
+
+			/* Consume bufs. */
+			dbuf = virtqueue_get_buf(vq, &len);
+			if (dbuf) {
+				if (len == 4)
+					assert(*dbuf == finished - 1);
+				else
+					assert(len == 0);
+				finished++;
+				continue;
+			}
+
+			if (!virtqueue_enable_cb_delayed(vq))
+				continue;
+			if (read(to_guest[0], buf, sizeof(buf)) < 1)
+				break;
+				
+			receives++;
+			virtqueue_disable_cb(vq);
+		}
+
+		printf("Guest: notified %lu, pinged %lu\n",
+		       gvdev.notifies, receives);
+		vring_del_virtqueue(vq);
+		return 0;
+	}
+}
+
+int main(int argc, char *argv[])
+{
+	struct virtio_device vdev;
+	struct virtqueue *vq;
+	struct vringh vrh;
+	struct scatterlist guest_sg[RINGSIZE], *sgs[2];
+	struct iovec host_riov[2], host_wiov[2];
+	struct vringh_iov riov, wiov;
+	struct vring_used_elem used[RINGSIZE];
+	char buf[28];
+	u16 head;
+	int err;
+	unsigned i;
+	void *ret;
+	bool (*getrange)(struct vringh *vrh, u64 addr, struct vringh_range *r);
+	bool fast_vringh = false, parallel = false;
+
+	getrange = getrange_iov;
+	vdev.features = 0;
+	INIT_LIST_HEAD(&vdev.vqs);
+	spin_lock_init(&vdev.vqs_list_lock);
+
+	while (argv[1]) {
+		if (strcmp(argv[1], "--indirect") == 0)
+			__virtio_set_bit(&vdev, VIRTIO_RING_F_INDIRECT_DESC);
+		else if (strcmp(argv[1], "--eventidx") == 0)
+			__virtio_set_bit(&vdev, VIRTIO_RING_F_EVENT_IDX);
+		else if (strcmp(argv[1], "--virtio-1") == 0)
+			__virtio_set_bit(&vdev, VIRTIO_F_VERSION_1);
+		else if (strcmp(argv[1], "--slow-range") == 0)
+			getrange = getrange_slow;
+		else if (strcmp(argv[1], "--fast-vringh") == 0)
+			fast_vringh = true;
+		else if (strcmp(argv[1], "--parallel") == 0)
+			parallel = true;
+		else
+			errx(1, "Unknown arg %s", argv[1]);
+		argv++;
+	}
+
+	if (parallel)
+		return parallel_test(vdev.features, getrange, fast_vringh);
+
+	if (posix_memalign(&__user_addr_min, PAGE_SIZE, USER_MEM) != 0)
+		abort();
+	__user_addr_max = __user_addr_min + USER_MEM;
+	memset(__user_addr_min, 0, vring_size(RINGSIZE, ALIGN));
+
+	/* Set up guest side. */
+	vq = vring_new_virtqueue(0, RINGSIZE, ALIGN, &vdev, true, false,
+				 __user_addr_min,
+				 never_notify_host, never_callback_guest,
+				 "guest vq");
+
+	/* Set up host side. */
+	vring_init(&vrh.vring, RINGSIZE, __user_addr_min, ALIGN);
+	vringh_init_user(&vrh, vdev.features, RINGSIZE, true,
+			 vrh.vring.desc, vrh.vring.avail, vrh.vring.used);
+
+	/* No descriptor to get yet... */
+	err = vringh_getdesc_user(&vrh, &riov, &wiov, getrange, &head);
+	if (err != 0)
+		errx(1, "vringh_getdesc_user: %i", err);
+
+	/* Guest puts in a descriptor. */
+	memcpy(__user_addr_max - 1, "a", 1);
+	sg_init_table(guest_sg, 1);
+	sg_set_buf(&guest_sg[0], __user_addr_max - 1, 1);
+	sg_init_table(guest_sg+1, 1);
+	sg_set_buf(&guest_sg[1], __user_addr_max - 3, 2);
+	sgs[0] = &guest_sg[0];
+	sgs[1] = &guest_sg[1];
+
+	/* May allocate an indirect, so force it to allocate user addr */
+	__kmalloc_fake = __user_addr_min + vring_size(RINGSIZE, ALIGN);
+	err = virtqueue_add_sgs(vq, sgs, 1, 1, &err, GFP_KERNEL);
+	if (err)
+		errx(1, "virtqueue_add_sgs: %i", err);
+	__kmalloc_fake = NULL;
+
+	/* Host retreives it. */
+	vringh_iov_init(&riov, host_riov, ARRAY_SIZE(host_riov));
+	vringh_iov_init(&wiov, host_wiov, ARRAY_SIZE(host_wiov));
+
+	err = vringh_getdesc_user(&vrh, &riov, &wiov, getrange, &head);
+	if (err != 1)
+		errx(1, "vringh_getdesc_user: %i", err);
+
+	assert(riov.used == 1);
+	assert(riov.iov[0].iov_base == __user_addr_max - 1);
+	assert(riov.iov[0].iov_len == 1);
+	if (getrange != getrange_slow) {
+		assert(wiov.used == 1);
+		assert(wiov.iov[0].iov_base == __user_addr_max - 3);
+		assert(wiov.iov[0].iov_len == 2);
+	} else {
+		assert(wiov.used == 2);
+		assert(wiov.iov[0].iov_base == __user_addr_max - 3);
+		assert(wiov.iov[0].iov_len == 1);
+		assert(wiov.iov[1].iov_base == __user_addr_max - 2);
+		assert(wiov.iov[1].iov_len == 1);
+	}
+
+	err = vringh_iov_pull_user(&riov, buf, 5);
+	if (err != 1)
+		errx(1, "vringh_iov_pull_user: %i", err);
+	assert(buf[0] == 'a');
+	assert(riov.i == 1);
+	assert(vringh_iov_pull_user(&riov, buf, 5) == 0);
+
+	memcpy(buf, "bcdef", 5);
+	err = vringh_iov_push_user(&wiov, buf, 5);
+	if (err != 2)
+		errx(1, "vringh_iov_push_user: %i", err);
+	assert(memcmp(__user_addr_max - 3, "bc", 2) == 0);
+	assert(wiov.i == wiov.used);
+	assert(vringh_iov_push_user(&wiov, buf, 5) == 0);
+
+	/* Host is done. */
+	err = vringh_complete_user(&vrh, head, err);
+	if (err != 0)
+		errx(1, "vringh_complete_user: %i", err);
+
+	/* Guest should see used token now. */
+	__kfree_ignore_start = __user_addr_min + vring_size(RINGSIZE, ALIGN);
+	__kfree_ignore_end = __kfree_ignore_start + 1;
+	ret = virtqueue_get_buf(vq, &i);
+	if (ret != &err)
+		errx(1, "virtqueue_get_buf: %p", ret);
+	assert(i == 2);
+
+	/* Guest puts in a huge descriptor. */
+	sg_init_table(guest_sg, RINGSIZE);
+	for (i = 0; i < RINGSIZE; i++) {
+		sg_set_buf(&guest_sg[i],
+			   __user_addr_max - USER_MEM/4, USER_MEM/4);
+	}
+
+	/* Fill contents with recognisable garbage. */
+	for (i = 0; i < USER_MEM/4; i++)
+		((char *)__user_addr_max - USER_MEM/4)[i] = i;
+
+	/* This will allocate an indirect, so force it to allocate user addr */
+	__kmalloc_fake = __user_addr_min + vring_size(RINGSIZE, ALIGN);
+	err = virtqueue_add_outbuf(vq, guest_sg, RINGSIZE, &err, GFP_KERNEL);
+	if (err)
+		errx(1, "virtqueue_add_outbuf (large): %i", err);
+	__kmalloc_fake = NULL;
+
+	/* Host picks it up (allocates new iov). */
+	vringh_iov_init(&riov, host_riov, ARRAY_SIZE(host_riov));
+	vringh_iov_init(&wiov, host_wiov, ARRAY_SIZE(host_wiov));
+
+	err = vringh_getdesc_user(&vrh, &riov, &wiov, getrange, &head);
+	if (err != 1)
+		errx(1, "vringh_getdesc_user: %i", err);
+
+	assert(riov.max_num & VRINGH_IOV_ALLOCATED);
+	assert(riov.iov != host_riov);
+	if (getrange != getrange_slow)
+		assert(riov.used == RINGSIZE);
+	else
+		assert(riov.used == RINGSIZE * USER_MEM/4);
+
+	assert(!(wiov.max_num & VRINGH_IOV_ALLOCATED));
+	assert(wiov.used == 0);
+
+	/* Pull data back out (in odd chunks), should be as expected. */
+	for (i = 0; i < RINGSIZE * USER_MEM/4; i += 3) {
+		err = vringh_iov_pull_user(&riov, buf, 3);
+		if (err != 3 && i + err != RINGSIZE * USER_MEM/4)
+			errx(1, "vringh_iov_pull_user large: %i", err);
+		assert(buf[0] == (char)i);
+		assert(err < 2 || buf[1] == (char)(i + 1));
+		assert(err < 3 || buf[2] == (char)(i + 2));
+	}
+	assert(riov.i == riov.used);
+	vringh_iov_cleanup(&riov);
+	vringh_iov_cleanup(&wiov);
+
+	/* Complete using multi interface, just because we can. */
+	used[0].id = head;
+	used[0].len = 0;
+	err = vringh_complete_multi_user(&vrh, used, 1);
+	if (err)
+		errx(1, "vringh_complete_multi_user(1): %i", err);
+
+	/* Free up those descriptors. */
+	ret = virtqueue_get_buf(vq, &i);
+	if (ret != &err)
+		errx(1, "virtqueue_get_buf: %p", ret);
+
+	/* Add lots of descriptors. */
+	sg_init_table(guest_sg, 1);
+	sg_set_buf(&guest_sg[0], __user_addr_max - 1, 1);
+	for (i = 0; i < RINGSIZE; i++) {
+		err = virtqueue_add_outbuf(vq, guest_sg, 1, &err, GFP_KERNEL);
+		if (err)
+			errx(1, "virtqueue_add_outbuf (multiple): %i", err);
+	}
+
+	/* Now get many, and consume them all at once. */
+	vringh_iov_init(&riov, host_riov, ARRAY_SIZE(host_riov));
+	vringh_iov_init(&wiov, host_wiov, ARRAY_SIZE(host_wiov));
+
+	for (i = 0; i < RINGSIZE; i++) {
+		err = vringh_getdesc_user(&vrh, &riov, &wiov, getrange, &head);
+		if (err != 1)
+			errx(1, "vringh_getdesc_user: %i", err);
+		used[i].id = head;
+		used[i].len = 0;
+	}
+	/* Make sure it wraps around ring, to test! */
+	assert(vrh.vring.used->idx % RINGSIZE != 0);
+	err = vringh_complete_multi_user(&vrh, used, RINGSIZE);
+	if (err)
+		errx(1, "vringh_complete_multi_user: %i", err);
+
+	/* Free those buffers. */
+	for (i = 0; i < RINGSIZE; i++) {
+		unsigned len;
+		assert(virtqueue_get_buf(vq, &len) != NULL);
+	}
+
+	/* Test weird (but legal!) indirect. */
+	if (__virtio_test_bit(&vdev, VIRTIO_RING_F_INDIRECT_DESC)) {
+		char *data = __user_addr_max - USER_MEM/4;
+		struct vring_desc *d = __user_addr_max - USER_MEM/2;
+		struct vring vring;
+
+		/* Force creation of direct, which we modify. */
+		__virtio_clear_bit(&vdev, VIRTIO_RING_F_INDIRECT_DESC);
+		vq = vring_new_virtqueue(0, RINGSIZE, ALIGN, &vdev, true,
+					 false, __user_addr_min,
+					 never_notify_host,
+					 never_callback_guest,
+					 "guest vq");
+
+		sg_init_table(guest_sg, 4);
+		sg_set_buf(&guest_sg[0], d, sizeof(*d)*2);
+		sg_set_buf(&guest_sg[1], d + 2, sizeof(*d)*1);
+		sg_set_buf(&guest_sg[2], data + 6, 4);
+		sg_set_buf(&guest_sg[3], d + 3, sizeof(*d)*3);
+
+		err = virtqueue_add_outbuf(vq, guest_sg, 4, &err, GFP_KERNEL);
+		if (err)
+			errx(1, "virtqueue_add_outbuf (indirect): %i", err);
+
+		vring_init(&vring, RINGSIZE, __user_addr_min, ALIGN);
+
+		/* They're used in order, but double-check... */
+		assert(vring.desc[0].addr == (unsigned long)d);
+		assert(vring.desc[1].addr == (unsigned long)(d+2));
+		assert(vring.desc[2].addr == (unsigned long)data + 6);
+		assert(vring.desc[3].addr == (unsigned long)(d+3));
+		vring.desc[0].flags |= VRING_DESC_F_INDIRECT;
+		vring.desc[1].flags |= VRING_DESC_F_INDIRECT;
+		vring.desc[3].flags |= VRING_DESC_F_INDIRECT;
+
+		/* First indirect */
+		d[0].addr = (unsigned long)data;
+		d[0].len = 1;
+		d[0].flags = VRING_DESC_F_NEXT;
+		d[0].next = 1;
+		d[1].addr = (unsigned long)data + 1;
+		d[1].len = 2;
+		d[1].flags = 0;
+
+		/* Second indirect */
+		d[2].addr = (unsigned long)data + 3;
+		d[2].len = 3;
+		d[2].flags = 0;
+
+		/* Third indirect */
+		d[3].addr = (unsigned long)data + 10;
+		d[3].len = 5;
+		d[3].flags = VRING_DESC_F_NEXT;
+		d[3].next = 1;
+		d[4].addr = (unsigned long)data + 15;
+		d[4].len = 6;
+		d[4].flags = VRING_DESC_F_NEXT;
+		d[4].next = 2;
+		d[5].addr = (unsigned long)data + 21;
+		d[5].len = 7;
+		d[5].flags = 0;
+
+		/* Host picks it up (allocates new iov). */
+		vringh_iov_init(&riov, host_riov, ARRAY_SIZE(host_riov));
+		vringh_iov_init(&wiov, host_wiov, ARRAY_SIZE(host_wiov));
+
+		err = vringh_getdesc_user(&vrh, &riov, &wiov, getrange, &head);
+		if (err != 1)
+			errx(1, "vringh_getdesc_user: %i", err);
+
+		if (head != 0)
+			errx(1, "vringh_getdesc_user: head %i not 0", head);
+
+		assert(riov.max_num & VRINGH_IOV_ALLOCATED);
+		if (getrange != getrange_slow)
+			assert(riov.used == 7);
+		else
+			assert(riov.used == 28);
+		err = vringh_iov_pull_user(&riov, buf, 29);
+		assert(err == 28);
+
+		/* Data should be linear. */
+		for (i = 0; i < err; i++)
+			assert(buf[i] == i);
+		vringh_iov_cleanup(&riov);
+	}
+
+	/* Don't leak memory... */
+	vring_del_virtqueue(vq);
+	free(__user_addr_min);
+
+	return 0;
+}