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, 465 insertions, 0 deletions

diff --git a/drivers/char/agp/isoch.c b/drivers/char/agp/isoch.c
new file mode 100644
index 000000000..7ecf20a6d
--- /dev/null
+++ b/drivers/char/agp/isoch.c
@@ -0,0 +1,465 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Setup routines for AGP 3.5 compliant bridges.
+ */
+
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/agp_backend.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+
+#include "agp.h"
+
+/* Generic AGP 3.5 enabling routines */
+
+struct agp_3_5_dev {
+	struct list_head list;
+	u8 capndx;
+	u32 maxbw;
+	struct pci_dev *dev;
+};
+
+static void agp_3_5_dev_list_insert(struct list_head *head, struct list_head *new)
+{
+	struct agp_3_5_dev *cur, *n = list_entry(new, struct agp_3_5_dev, list);
+	struct list_head *pos;
+
+	list_for_each(pos, head) {
+		cur = list_entry(pos, struct agp_3_5_dev, list);
+		if (cur->maxbw > n->maxbw)
+			break;
+	}
+	list_add_tail(new, pos);
+}
+
+static void agp_3_5_dev_list_sort(struct agp_3_5_dev *list, unsigned int ndevs)
+{
+	struct agp_3_5_dev *cur;
+	struct pci_dev *dev;
+	struct list_head *pos, *tmp, *head = &list->list, *start = head->next;
+	u32 nistat;
+
+	INIT_LIST_HEAD(head);
+
+	for (pos=start; pos!=head; ) {
+		cur = list_entry(pos, struct agp_3_5_dev, list);
+		dev = cur->dev;
+
+		pci_read_config_dword(dev, cur->capndx+AGPNISTAT, &nistat);
+		cur->maxbw = (nistat >> 16) & 0xff;
+
+		tmp = pos;
+		pos = pos->next;
+		agp_3_5_dev_list_insert(head, tmp);
+	}
+}
+
+/*
+ * Initialize all isochronous transfer parameters for an AGP 3.0
+ * node (i.e. a host bridge in combination with the adapters
+ * lying behind it...)
+ */
+
+static int agp_3_5_isochronous_node_enable(struct agp_bridge_data *bridge,
+		struct agp_3_5_dev *dev_list, unsigned int ndevs)
+{
+	/*
+	 * Convenience structure to make the calculations clearer
+	 * here.  The field names come straight from the AGP 3.0 spec.
+	 */
+	struct isoch_data {
+		u32 maxbw;
+		u32 n;
+		u32 y;
+		u32 l;
+		u32 rq;
+		struct agp_3_5_dev *dev;
+	};
+
+	struct pci_dev *td = bridge->dev, *dev;
+	struct list_head *head = &dev_list->list, *pos;
+	struct agp_3_5_dev *cur;
+	struct isoch_data *master, target;
+	unsigned int cdev = 0;
+	u32 mnistat, tnistat, tstatus, mcmd;
+	u16 tnicmd, mnicmd;
+	u32 tot_bw = 0, tot_n = 0, tot_rq = 0, y_max, rq_isoch, rq_async;
+	u32 step, rem, rem_isoch, rem_async;
+	int ret = 0;
+
+	/*
+	 * We'll work with an array of isoch_data's (one for each
+	 * device in dev_list) throughout this function.
+	 */
+	master = kmalloc_array(ndevs, sizeof(*master), GFP_KERNEL);
+	if (master == NULL) {
+		ret = -ENOMEM;
+		goto get_out;
+	}
+
+	/*
+	 * Sort the device list by maxbw.  We need to do this because the
+	 * spec suggests that the devices with the smallest requirements
+	 * have their resources allocated first, with all remaining resources
+	 * falling to the device with the largest requirement.
+	 *
+	 * We don't exactly do this, we divide target resources by ndevs
+	 * and split them amongst the AGP 3.0 devices.  The remainder of such
+	 * division operations are dropped on the last device, sort of like
+	 * the spec mentions it should be done.
+	 *
+	 * We can't do this sort when we initially construct the dev_list
+	 * because we don't know until this function whether isochronous
+	 * transfers are enabled and consequently whether maxbw will mean
+	 * anything.
+	 */
+	agp_3_5_dev_list_sort(dev_list, ndevs);
+
+	pci_read_config_dword(td, bridge->capndx+AGPNISTAT, &tnistat);
+	pci_read_config_dword(td, bridge->capndx+AGPSTAT, &tstatus);
+
+	/* Extract power-on defaults from the target */
+	target.maxbw = (tnistat >> 16) & 0xff;
+	target.n     = (tnistat >> 8)  & 0xff;
+	target.y     = (tnistat >> 6)  & 0x3;
+	target.l     = (tnistat >> 3)  & 0x7;
+	target.rq    = (tstatus >> 24) & 0xff;
+
+	y_max = target.y;
+
+	/*
+	 * Extract power-on defaults for each device in dev_list.  Along
+	 * the way, calculate the total isochronous bandwidth required
+	 * by these devices and the largest requested payload size.
+	 */
+	list_for_each(pos, head) {
+		cur = list_entry(pos, struct agp_3_5_dev, list);
+		dev = cur->dev;
+
+		pci_read_config_dword(dev, cur->capndx+AGPNISTAT, &mnistat);
+
+		master[cdev].maxbw = (mnistat >> 16) & 0xff;
+		master[cdev].n     = (mnistat >> 8)  & 0xff;
+		master[cdev].y     = (mnistat >> 6)  & 0x3;
+		master[cdev].dev   = cur;
+
+		tot_bw += master[cdev].maxbw;
+		y_max = max(y_max, master[cdev].y);
+
+		cdev++;
+	}
+
+	/* Check if this configuration has any chance of working */
+	if (tot_bw > target.maxbw) {
+		dev_err(&td->dev, "isochronous bandwidth required "
+			"by AGP 3.0 devices exceeds that which is supported by "
+			"the AGP 3.0 bridge!\n");
+		ret = -ENODEV;
+		goto free_and_exit;
+	}
+
+	target.y = y_max;
+
+	/*
+	 * Write the calculated payload size into the target's NICMD
+	 * register.  Doing this directly effects the ISOCH_N value
+	 * in the target's NISTAT register, so we need to do this now
+	 * to get an accurate value for ISOCH_N later.
+	 */
+	pci_read_config_word(td, bridge->capndx+AGPNICMD, &tnicmd);
+	tnicmd &= ~(0x3 << 6);
+	tnicmd |= target.y << 6;
+	pci_write_config_word(td, bridge->capndx+AGPNICMD, tnicmd);
+
+	/* Reread the target's ISOCH_N */
+	pci_read_config_dword(td, bridge->capndx+AGPNISTAT, &tnistat);
+	target.n = (tnistat >> 8) & 0xff;
+
+	/* Calculate the minimum ISOCH_N needed by each master */
+	for (cdev=0; cdev<ndevs; cdev++) {
+		master[cdev].y = target.y;
+		master[cdev].n = master[cdev].maxbw / (master[cdev].y + 1);
+
+		tot_n += master[cdev].n;
+	}
+
+	/* Exit if the minimal ISOCH_N allocation among the masters is more
+	 * than the target can handle. */
+	if (tot_n > target.n) {
+		dev_err(&td->dev, "number of isochronous "
+			"transactions per period required by AGP 3.0 devices "
+			"exceeds that which is supported by the AGP 3.0 "
+			"bridge!\n");
+		ret = -ENODEV;
+		goto free_and_exit;
+	}
+
+	/* Calculate left over ISOCH_N capability in the target.  We'll give
+	 * this to the hungriest device (as per the spec) */
+	rem  = target.n - tot_n;
+
+	/*
+	 * Calculate the minimum isochronous RQ depth needed by each master.
+	 * Along the way, distribute the extra ISOCH_N capability calculated
+	 * above.
+	 */
+	for (cdev=0; cdev<ndevs; cdev++) {
+		/*
+		 * This is a little subtle.  If ISOCH_Y > 64B, then ISOCH_Y
+		 * byte isochronous writes will be broken into 64B pieces.
+		 * This means we need to budget more RQ depth to account for
+		 * these kind of writes (each isochronous write is actually
+		 * many writes on the AGP bus).
+		 */
+		master[cdev].rq = master[cdev].n;
+		if (master[cdev].y > 0x1)
+			master[cdev].rq *= (1 << (master[cdev].y - 1));
+
+		tot_rq += master[cdev].rq;
+	}
+	master[ndevs-1].n += rem;
+
+	/* Figure the number of isochronous and asynchronous RQ slots the
+	 * target is providing. */
+	rq_isoch = (target.y > 0x1) ? target.n * (1 << (target.y - 1)) : target.n;
+	rq_async = target.rq - rq_isoch;
+
+	/* Exit if the minimal RQ needs of the masters exceeds what the target
+	 * can provide. */
+	if (tot_rq > rq_isoch) {
+		dev_err(&td->dev, "number of request queue slots "
+			"required by the isochronous bandwidth requested by "
+			"AGP 3.0 devices exceeds the number provided by the "
+			"AGP 3.0 bridge!\n");
+		ret = -ENODEV;
+		goto free_and_exit;
+	}
+
+	/* Calculate asynchronous RQ capability in the target (per master) as
+	 * well as the total number of leftover isochronous RQ slots. */
+	step      = rq_async / ndevs;
+	rem_async = step + (rq_async % ndevs);
+	rem_isoch = rq_isoch - tot_rq;
+
+	/* Distribute the extra RQ slots calculated above and write our
+	 * isochronous settings out to the actual devices. */
+	for (cdev=0; cdev<ndevs; cdev++) {
+		cur = master[cdev].dev;
+		dev = cur->dev;
+
+		master[cdev].rq += (cdev == ndevs - 1)
+		              ? (rem_async + rem_isoch) : step;
+
+		pci_read_config_word(dev, cur->capndx+AGPNICMD, &mnicmd);
+		pci_read_config_dword(dev, cur->capndx+AGPCMD, &mcmd);
+
+		mnicmd &= ~(0xff << 8);
+		mnicmd &= ~(0x3  << 6);
+		mcmd   &= ~(0xff << 24);
+
+		mnicmd |= master[cdev].n  << 8;
+		mnicmd |= master[cdev].y  << 6;
+		mcmd   |= master[cdev].rq << 24;
+
+		pci_write_config_dword(dev, cur->capndx+AGPCMD, mcmd);
+		pci_write_config_word(dev, cur->capndx+AGPNICMD, mnicmd);
+	}
+
+free_and_exit:
+	kfree(master);
+
+get_out:
+	return ret;
+}
+
+/*
+ * This function basically allocates request queue slots among the
+ * AGP 3.0 systems in nonisochronous nodes.  The algorithm is
+ * pretty stupid, divide the total number of RQ slots provided by the
+ * target by ndevs.  Distribute this many slots to each AGP 3.0 device,
+ * giving any left over slots to the last device in dev_list.
+ */
+static void agp_3_5_nonisochronous_node_enable(struct agp_bridge_data *bridge,
+		struct agp_3_5_dev *dev_list, unsigned int ndevs)
+{
+	struct agp_3_5_dev *cur;
+	struct list_head *head = &dev_list->list, *pos;
+	u32 tstatus, mcmd;
+	u32 trq, mrq, rem;
+	unsigned int cdev = 0;
+
+	pci_read_config_dword(bridge->dev, bridge->capndx+AGPSTAT, &tstatus);
+
+	trq = (tstatus >> 24) & 0xff;
+	mrq = trq / ndevs;
+
+	rem = mrq + (trq % ndevs);
+
+	for (pos=head->next; cdev<ndevs; cdev++, pos=pos->next) {
+		cur = list_entry(pos, struct agp_3_5_dev, list);
+
+		pci_read_config_dword(cur->dev, cur->capndx+AGPCMD, &mcmd);
+		mcmd &= ~(0xff << 24);
+		mcmd |= ((cdev == ndevs - 1) ? rem : mrq) << 24;
+		pci_write_config_dword(cur->dev, cur->capndx+AGPCMD, mcmd);
+	}
+}
+
+/*
+ * Fully configure and enable an AGP 3.0 host bridge and all the devices
+ * lying behind it.
+ */
+int agp_3_5_enable(struct agp_bridge_data *bridge)
+{
+	struct pci_dev *td = bridge->dev, *dev = NULL;
+	u8 mcapndx;
+	u32 isoch;
+	u32 tstatus, mstatus, ncapid;
+	u32 mmajor;
+	u16 mpstat;
+	struct agp_3_5_dev *dev_list, *cur;
+	struct list_head *head, *pos;
+	unsigned int ndevs = 0;
+	int ret = 0;
+
+	/* Extract some power-on defaults from the target */
+	pci_read_config_dword(td, bridge->capndx+AGPSTAT, &tstatus);
+	isoch     = (tstatus >> 17) & 0x1;
+	if (isoch == 0)	/* isoch xfers not available, bail out. */
+		return -ENODEV;
+
+	/*
+	 * Allocate a head for our AGP 3.5 device list
+	 * (multiple AGP v3 devices are allowed behind a single bridge).
+	 */
+	if ((dev_list = kmalloc(sizeof(*dev_list), GFP_KERNEL)) == NULL) {
+		ret = -ENOMEM;
+		goto get_out;
+	}
+	head = &dev_list->list;
+	INIT_LIST_HEAD(head);
+
+	/* Find all AGP devices, and add them to dev_list. */
+	for_each_pci_dev(dev) {
+		mcapndx = pci_find_capability(dev, PCI_CAP_ID_AGP);
+		if (mcapndx == 0)
+			continue;
+
+		switch ((dev->class >>8) & 0xff00) {
+			case 0x0600:    /* Bridge */
+				/* Skip bridges. We should call this function for each one. */
+				continue;
+
+			case 0x0001:    /* Unclassified device */
+				/* Don't know what this is, but log it for investigation. */
+				if (mcapndx != 0) {
+					dev_info(&td->dev, "wacky, found unclassified AGP device %s [%04x/%04x]\n",
+						 pci_name(dev),
+						 dev->vendor, dev->device);
+				}
+				continue;
+
+			case 0x0300:    /* Display controller */
+			case 0x0400:    /* Multimedia controller */
+				if ((cur = kmalloc(sizeof(*cur), GFP_KERNEL)) == NULL) {
+					ret = -ENOMEM;
+					goto free_and_exit;
+				}
+				cur->dev = dev;
+
+				pos = &cur->list;
+				list_add(pos, head);
+				ndevs++;
+				continue;
+
+			default:
+				continue;
+		}
+	}
+
+	/*
+	 * Take an initial pass through the devices lying behind our host
+	 * bridge.  Make sure each one is actually an AGP 3.0 device, otherwise
+	 * exit with an error message.  Along the way store the AGP 3.0
+	 * cap_ptr for each device
+	 */
+	list_for_each(pos, head) {
+		cur = list_entry(pos, struct agp_3_5_dev, list);
+		dev = cur->dev;
+
+		pci_read_config_word(dev, PCI_STATUS, &mpstat);
+		if ((mpstat & PCI_STATUS_CAP_LIST) == 0)
+			continue;
+
+		pci_read_config_byte(dev, PCI_CAPABILITY_LIST, &mcapndx);
+		if (mcapndx != 0) {
+			do {
+				pci_read_config_dword(dev, mcapndx, &ncapid);
+				if ((ncapid & 0xff) != 2)
+					mcapndx = (ncapid >> 8) & 0xff;
+			}
+			while (((ncapid & 0xff) != 2) && (mcapndx != 0));
+		}
+
+		if (mcapndx == 0) {
+			dev_err(&td->dev, "woah!  Non-AGP device %s on "
+				"secondary bus of AGP 3.5 bridge!\n",
+				pci_name(dev));
+			ret = -ENODEV;
+			goto free_and_exit;
+		}
+
+		mmajor = (ncapid >> AGP_MAJOR_VERSION_SHIFT) & 0xf;
+		if (mmajor < 3) {
+			dev_err(&td->dev, "woah!  AGP 2.0 device %s on "
+				"secondary bus of AGP 3.5 bridge operating "
+				"with AGP 3.0 electricals!\n", pci_name(dev));
+			ret = -ENODEV;
+			goto free_and_exit;
+		}
+
+		cur->capndx = mcapndx;
+
+		pci_read_config_dword(dev, cur->capndx+AGPSTAT, &mstatus);
+
+		if (((mstatus >> 3) & 0x1) == 0) {
+			dev_err(&td->dev, "woah!  AGP 3.x device %s not "
+				"operating in AGP 3.x mode on secondary bus "
+				"of AGP 3.5 bridge operating with AGP 3.0 "
+				"electricals!\n", pci_name(dev));
+			ret = -ENODEV;
+			goto free_and_exit;
+		}
+	}		
+
+	/*
+	 * Call functions to divide target resources amongst the AGP 3.0
+	 * masters.  This process is dramatically different depending on
+	 * whether isochronous transfers are supported.
+	 */
+	if (isoch) {
+		ret = agp_3_5_isochronous_node_enable(bridge, dev_list, ndevs);
+		if (ret) {
+			dev_info(&td->dev, "something bad happened setting "
+				 "up isochronous xfers; falling back to "
+				 "non-isochronous xfer mode\n");
+		} else {
+			goto free_and_exit;
+		}
+	}
+	agp_3_5_nonisochronous_node_enable(bridge, dev_list, ndevs);
+
+free_and_exit:
+	/* Be sure to free the dev_list */
+	for (pos=head->next; pos!=head; ) {
+		cur = list_entry(pos, struct agp_3_5_dev, list);
+
+		pos = pos->next;
+		kfree(cur);
+	}
+	kfree(dev_list);
+
+get_out:
+	return ret;
+}