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, 1235 insertions, 0 deletions

diff --git a/drivers/usb/host/ohci-q.c b/drivers/usb/host/ohci-q.c
new file mode 100644
index 000000000..3b445312b
--- /dev/null
+++ b/drivers/usb/host/ohci-q.c
@@ -0,0 +1,1235 @@
+// SPDX-License-Identifier: GPL-1.0+
+/*
+ * OHCI HCD (Host Controller Driver) for USB.
+ *
+ * (C) Copyright 1999 Roman Weissgaerber <weissg@vienna.at>
+ * (C) Copyright 2000-2002 David Brownell <dbrownell@users.sourceforge.net>
+ *
+ * This file is licenced under the GPL.
+ */
+
+#include <linux/irq.h>
+#include <linux/slab.h>
+
+static void urb_free_priv (struct ohci_hcd *hc, urb_priv_t *urb_priv)
+{
+	int		last = urb_priv->length - 1;
+
+	if (last >= 0) {
+		int		i;
+		struct td	*td;
+
+		for (i = 0; i <= last; i++) {
+			td = urb_priv->td [i];
+			if (td)
+				td_free (hc, td);
+		}
+	}
+
+	list_del (&urb_priv->pending);
+	kfree (urb_priv);
+}
+
+/*-------------------------------------------------------------------------*/
+
+/*
+ * URB goes back to driver, and isn't reissued.
+ * It's completely gone from HC data structures.
+ * PRECONDITION:  ohci lock held, irqs blocked.
+ */
+static void
+finish_urb(struct ohci_hcd *ohci, struct urb *urb, int status)
+__releases(ohci->lock)
+__acquires(ohci->lock)
+{
+	struct device *dev = ohci_to_hcd(ohci)->self.controller;
+	struct usb_host_endpoint *ep = urb->ep;
+	struct urb_priv *urb_priv;
+
+	// ASSERT (urb->hcpriv != 0);
+
+ restart:
+	urb_free_priv (ohci, urb->hcpriv);
+	urb->hcpriv = NULL;
+	if (likely(status == -EINPROGRESS))
+		status = 0;
+
+	switch (usb_pipetype (urb->pipe)) {
+	case PIPE_ISOCHRONOUS:
+		ohci_to_hcd(ohci)->self.bandwidth_isoc_reqs--;
+		if (ohci_to_hcd(ohci)->self.bandwidth_isoc_reqs == 0) {
+			if (quirk_amdiso(ohci))
+				usb_amd_quirk_pll_enable();
+			if (quirk_amdprefetch(ohci))
+				sb800_prefetch(dev, 0);
+		}
+		break;
+	case PIPE_INTERRUPT:
+		ohci_to_hcd(ohci)->self.bandwidth_int_reqs--;
+		break;
+	}
+
+	/* urb->complete() can reenter this HCD */
+	usb_hcd_unlink_urb_from_ep(ohci_to_hcd(ohci), urb);
+	spin_unlock (&ohci->lock);
+	usb_hcd_giveback_urb(ohci_to_hcd(ohci), urb, status);
+	spin_lock (&ohci->lock);
+
+	/* stop periodic dma if it's not needed */
+	if (ohci_to_hcd(ohci)->self.bandwidth_isoc_reqs == 0
+			&& ohci_to_hcd(ohci)->self.bandwidth_int_reqs == 0) {
+		ohci->hc_control &= ~(OHCI_CTRL_PLE|OHCI_CTRL_IE);
+		ohci_writel (ohci, ohci->hc_control, &ohci->regs->control);
+	}
+
+	/*
+	 * An isochronous URB that is sumitted too late won't have any TDs
+	 * (marked by the fact that the td_cnt value is larger than the
+	 * actual number of TDs).  If the next URB on this endpoint is like
+	 * that, give it back now.
+	 */
+	if (!list_empty(&ep->urb_list)) {
+		urb = list_first_entry(&ep->urb_list, struct urb, urb_list);
+		urb_priv = urb->hcpriv;
+		if (urb_priv->td_cnt > urb_priv->length) {
+			status = 0;
+			goto restart;
+		}
+	}
+}
+
+
+/*-------------------------------------------------------------------------*
+ * ED handling functions
+ *-------------------------------------------------------------------------*/
+
+/* search for the right schedule branch to use for a periodic ed.
+ * does some load balancing; returns the branch, or negative errno.
+ */
+static int balance (struct ohci_hcd *ohci, int interval, int load)
+{
+	int	i, branch = -ENOSPC;
+
+	/* iso periods can be huge; iso tds specify frame numbers */
+	if (interval > NUM_INTS)
+		interval = NUM_INTS;
+
+	/* search for the least loaded schedule branch of that period
+	 * that has enough bandwidth left unreserved.
+	 */
+	for (i = 0; i < interval ; i++) {
+		if (branch < 0 || ohci->load [branch] > ohci->load [i]) {
+			int	j;
+
+			/* usb 1.1 says 90% of one frame */
+			for (j = i; j < NUM_INTS; j += interval) {
+				if ((ohci->load [j] + load) > 900)
+					break;
+			}
+			if (j < NUM_INTS)
+				continue;
+			branch = i;
+		}
+	}
+	return branch;
+}
+
+/*-------------------------------------------------------------------------*/
+
+/* both iso and interrupt requests have periods; this routine puts them
+ * into the schedule tree in the apppropriate place.  most iso devices use
+ * 1msec periods, but that's not required.
+ */
+static void periodic_link (struct ohci_hcd *ohci, struct ed *ed)
+{
+	unsigned	i;
+
+	ohci_dbg(ohci, "link %sed %p branch %d [%dus.], interval %d\n",
+		(ed->hwINFO & cpu_to_hc32 (ohci, ED_ISO)) ? "iso " : "",
+		ed, ed->branch, ed->load, ed->interval);
+
+	for (i = ed->branch; i < NUM_INTS; i += ed->interval) {
+		struct ed	**prev = &ohci->periodic [i];
+		__hc32		*prev_p = &ohci->hcca->int_table [i];
+		struct ed	*here = *prev;
+
+		/* sorting each branch by period (slow before fast)
+		 * lets us share the faster parts of the tree.
+		 * (plus maybe: put interrupt eds before iso)
+		 */
+		while (here && ed != here) {
+			if (ed->interval > here->interval)
+				break;
+			prev = &here->ed_next;
+			prev_p = &here->hwNextED;
+			here = *prev;
+		}
+		if (ed != here) {
+			ed->ed_next = here;
+			if (here)
+				ed->hwNextED = *prev_p;
+			wmb ();
+			*prev = ed;
+			*prev_p = cpu_to_hc32(ohci, ed->dma);
+			wmb();
+		}
+		ohci->load [i] += ed->load;
+	}
+	ohci_to_hcd(ohci)->self.bandwidth_allocated += ed->load / ed->interval;
+}
+
+/* link an ed into one of the HC chains */
+
+static int ed_schedule (struct ohci_hcd *ohci, struct ed *ed)
+{
+	int	branch;
+
+	ed->ed_prev = NULL;
+	ed->ed_next = NULL;
+	ed->hwNextED = 0;
+	wmb ();
+
+	/* we care about rm_list when setting CLE/BLE in case the HC was at
+	 * work on some TD when CLE/BLE was turned off, and isn't quiesced
+	 * yet.  finish_unlinks() restarts as needed, some upcoming INTR_SF.
+	 *
+	 * control and bulk EDs are doubly linked (ed_next, ed_prev), but
+	 * periodic ones are singly linked (ed_next). that's because the
+	 * periodic schedule encodes a tree like figure 3-5 in the ohci
+	 * spec:  each qh can have several "previous" nodes, and the tree
+	 * doesn't have unused/idle descriptors.
+	 */
+	switch (ed->type) {
+	case PIPE_CONTROL:
+		if (ohci->ed_controltail == NULL) {
+			WARN_ON (ohci->hc_control & OHCI_CTRL_CLE);
+			ohci_writel (ohci, ed->dma,
+					&ohci->regs->ed_controlhead);
+		} else {
+			ohci->ed_controltail->ed_next = ed;
+			ohci->ed_controltail->hwNextED = cpu_to_hc32 (ohci,
+								ed->dma);
+		}
+		ed->ed_prev = ohci->ed_controltail;
+		if (!ohci->ed_controltail && !ohci->ed_rm_list) {
+			wmb();
+			ohci->hc_control |= OHCI_CTRL_CLE;
+			ohci_writel (ohci, 0, &ohci->regs->ed_controlcurrent);
+			ohci_writel (ohci, ohci->hc_control,
+					&ohci->regs->control);
+		}
+		ohci->ed_controltail = ed;
+		break;
+
+	case PIPE_BULK:
+		if (ohci->ed_bulktail == NULL) {
+			WARN_ON (ohci->hc_control & OHCI_CTRL_BLE);
+			ohci_writel (ohci, ed->dma, &ohci->regs->ed_bulkhead);
+		} else {
+			ohci->ed_bulktail->ed_next = ed;
+			ohci->ed_bulktail->hwNextED = cpu_to_hc32 (ohci,
+								ed->dma);
+		}
+		ed->ed_prev = ohci->ed_bulktail;
+		if (!ohci->ed_bulktail && !ohci->ed_rm_list) {
+			wmb();
+			ohci->hc_control |= OHCI_CTRL_BLE;
+			ohci_writel (ohci, 0, &ohci->regs->ed_bulkcurrent);
+			ohci_writel (ohci, ohci->hc_control,
+					&ohci->regs->control);
+		}
+		ohci->ed_bulktail = ed;
+		break;
+
+	// case PIPE_INTERRUPT:
+	// case PIPE_ISOCHRONOUS:
+	default:
+		branch = balance (ohci, ed->interval, ed->load);
+		if (branch < 0) {
+			ohci_dbg (ohci,
+				"ERR %d, interval %d msecs, load %d\n",
+				branch, ed->interval, ed->load);
+			// FIXME if there are TDs queued, fail them!
+			return branch;
+		}
+		ed->branch = branch;
+		periodic_link (ohci, ed);
+	}
+
+	/* the HC may not see the schedule updates yet, but if it does
+	 * then they'll be properly ordered.
+	 */
+
+	ed->state = ED_OPER;
+	return 0;
+}
+
+/*-------------------------------------------------------------------------*/
+
+/* scan the periodic table to find and unlink this ED */
+static void periodic_unlink (struct ohci_hcd *ohci, struct ed *ed)
+{
+	int	i;
+
+	for (i = ed->branch; i < NUM_INTS; i += ed->interval) {
+		struct ed	*temp;
+		struct ed	**prev = &ohci->periodic [i];
+		__hc32		*prev_p = &ohci->hcca->int_table [i];
+
+		while (*prev && (temp = *prev) != ed) {
+			prev_p = &temp->hwNextED;
+			prev = &temp->ed_next;
+		}
+		if (*prev) {
+			*prev_p = ed->hwNextED;
+			*prev = ed->ed_next;
+		}
+		ohci->load [i] -= ed->load;
+	}
+	ohci_to_hcd(ohci)->self.bandwidth_allocated -= ed->load / ed->interval;
+
+	ohci_dbg(ohci, "unlink %sed %p branch %d [%dus.], interval %d\n",
+		(ed->hwINFO & cpu_to_hc32 (ohci, ED_ISO)) ? "iso " : "",
+		ed, ed->branch, ed->load, ed->interval);
+}
+
+/* unlink an ed from one of the HC chains.
+ * just the link to the ed is unlinked.
+ * the link from the ed still points to another operational ed or 0
+ * so the HC can eventually finish the processing of the unlinked ed
+ * (assuming it already started that, which needn't be true).
+ *
+ * ED_UNLINK is a transient state: the HC may still see this ED, but soon
+ * it won't.  ED_SKIP means the HC will finish its current transaction,
+ * but won't start anything new.  The TD queue may still grow; device
+ * drivers don't know about this HCD-internal state.
+ *
+ * When the HC can't see the ED, something changes ED_UNLINK to one of:
+ *
+ *  - ED_OPER: when there's any request queued, the ED gets rescheduled
+ *    immediately.  HC should be working on them.
+ *
+ *  - ED_IDLE: when there's no TD queue or the HC isn't running.
+ *
+ * When finish_unlinks() runs later, after SOF interrupt, it will often
+ * complete one or more URB unlinks before making that state change.
+ */
+static void ed_deschedule (struct ohci_hcd *ohci, struct ed *ed)
+{
+	ed->hwINFO |= cpu_to_hc32 (ohci, ED_SKIP);
+	wmb ();
+	ed->state = ED_UNLINK;
+
+	/* To deschedule something from the control or bulk list, just
+	 * clear CLE/BLE and wait.  There's no safe way to scrub out list
+	 * head/current registers until later, and "later" isn't very
+	 * tightly specified.  Figure 6-5 and Section 6.4.2.2 show how
+	 * the HC is reading the ED queues (while we modify them).
+	 *
+	 * For now, ed_schedule() is "later".  It might be good paranoia
+	 * to scrub those registers in finish_unlinks(), in case of bugs
+	 * that make the HC try to use them.
+	 */
+	switch (ed->type) {
+	case PIPE_CONTROL:
+		/* remove ED from the HC's list: */
+		if (ed->ed_prev == NULL) {
+			if (!ed->hwNextED) {
+				ohci->hc_control &= ~OHCI_CTRL_CLE;
+				ohci_writel (ohci, ohci->hc_control,
+						&ohci->regs->control);
+				// a ohci_readl() later syncs CLE with the HC
+			} else
+				ohci_writel (ohci,
+					hc32_to_cpup (ohci, &ed->hwNextED),
+					&ohci->regs->ed_controlhead);
+		} else {
+			ed->ed_prev->ed_next = ed->ed_next;
+			ed->ed_prev->hwNextED = ed->hwNextED;
+		}
+		/* remove ED from the HCD's list: */
+		if (ohci->ed_controltail == ed) {
+			ohci->ed_controltail = ed->ed_prev;
+			if (ohci->ed_controltail)
+				ohci->ed_controltail->ed_next = NULL;
+		} else if (ed->ed_next) {
+			ed->ed_next->ed_prev = ed->ed_prev;
+		}
+		break;
+
+	case PIPE_BULK:
+		/* remove ED from the HC's list: */
+		if (ed->ed_prev == NULL) {
+			if (!ed->hwNextED) {
+				ohci->hc_control &= ~OHCI_CTRL_BLE;
+				ohci_writel (ohci, ohci->hc_control,
+						&ohci->regs->control);
+				// a ohci_readl() later syncs BLE with the HC
+			} else
+				ohci_writel (ohci,
+					hc32_to_cpup (ohci, &ed->hwNextED),
+					&ohci->regs->ed_bulkhead);
+		} else {
+			ed->ed_prev->ed_next = ed->ed_next;
+			ed->ed_prev->hwNextED = ed->hwNextED;
+		}
+		/* remove ED from the HCD's list: */
+		if (ohci->ed_bulktail == ed) {
+			ohci->ed_bulktail = ed->ed_prev;
+			if (ohci->ed_bulktail)
+				ohci->ed_bulktail->ed_next = NULL;
+		} else if (ed->ed_next) {
+			ed->ed_next->ed_prev = ed->ed_prev;
+		}
+		break;
+
+	// case PIPE_INTERRUPT:
+	// case PIPE_ISOCHRONOUS:
+	default:
+		periodic_unlink (ohci, ed);
+		break;
+	}
+}
+
+
+/*-------------------------------------------------------------------------*/
+
+/* get and maybe (re)init an endpoint. init _should_ be done only as part
+ * of enumeration, usb_set_configuration() or usb_set_interface().
+ */
+static struct ed *ed_get (
+	struct ohci_hcd		*ohci,
+	struct usb_host_endpoint *ep,
+	struct usb_device	*udev,
+	unsigned int		pipe,
+	int			interval
+) {
+	struct ed		*ed;
+	unsigned long		flags;
+
+	spin_lock_irqsave (&ohci->lock, flags);
+
+	ed = ep->hcpriv;
+	if (!ed) {
+		struct td	*td;
+		int		is_out;
+		u32		info;
+
+		ed = ed_alloc (ohci, GFP_ATOMIC);
+		if (!ed) {
+			/* out of memory */
+			goto done;
+		}
+
+		/* dummy td; end of td list for ed */
+		td = td_alloc (ohci, GFP_ATOMIC);
+		if (!td) {
+			/* out of memory */
+			ed_free (ohci, ed);
+			ed = NULL;
+			goto done;
+		}
+		ed->dummy = td;
+		ed->hwTailP = cpu_to_hc32 (ohci, td->td_dma);
+		ed->hwHeadP = ed->hwTailP;	/* ED_C, ED_H zeroed */
+		ed->state = ED_IDLE;
+
+		is_out = !(ep->desc.bEndpointAddress & USB_DIR_IN);
+
+		/* FIXME usbcore changes dev->devnum before SET_ADDRESS
+		 * succeeds ... otherwise we wouldn't need "pipe".
+		 */
+		info = usb_pipedevice (pipe);
+		ed->type = usb_pipetype(pipe);
+
+		info |= (ep->desc.bEndpointAddress & ~USB_DIR_IN) << 7;
+		info |= usb_endpoint_maxp(&ep->desc) << 16;
+		if (udev->speed == USB_SPEED_LOW)
+			info |= ED_LOWSPEED;
+		/* only control transfers store pids in tds */
+		if (ed->type != PIPE_CONTROL) {
+			info |= is_out ? ED_OUT : ED_IN;
+			if (ed->type != PIPE_BULK) {
+				/* periodic transfers... */
+				if (ed->type == PIPE_ISOCHRONOUS)
+					info |= ED_ISO;
+				else if (interval > 32)	/* iso can be bigger */
+					interval = 32;
+				ed->interval = interval;
+				ed->load = usb_calc_bus_time (
+					udev->speed, !is_out,
+					ed->type == PIPE_ISOCHRONOUS,
+					usb_endpoint_maxp(&ep->desc))
+						/ 1000;
+			}
+		}
+		ed->hwINFO = cpu_to_hc32(ohci, info);
+
+		ep->hcpriv = ed;
+	}
+
+done:
+	spin_unlock_irqrestore (&ohci->lock, flags);
+	return ed;
+}
+
+/*-------------------------------------------------------------------------*/
+
+/* request unlinking of an endpoint from an operational HC.
+ * put the ep on the rm_list
+ * real work is done at the next start frame (SF) hardware interrupt
+ * caller guarantees HCD is running, so hardware access is safe,
+ * and that ed->state is ED_OPER
+ */
+static void start_ed_unlink (struct ohci_hcd *ohci, struct ed *ed)
+{
+	ed->hwINFO |= cpu_to_hc32 (ohci, ED_DEQUEUE);
+	ed_deschedule (ohci, ed);
+
+	/* rm_list is just singly linked, for simplicity */
+	ed->ed_next = ohci->ed_rm_list;
+	ed->ed_prev = NULL;
+	ohci->ed_rm_list = ed;
+
+	/* enable SOF interrupt */
+	ohci_writel (ohci, OHCI_INTR_SF, &ohci->regs->intrstatus);
+	ohci_writel (ohci, OHCI_INTR_SF, &ohci->regs->intrenable);
+	// flush those writes, and get latest HCCA contents
+	(void) ohci_readl (ohci, &ohci->regs->control);
+
+	/* SF interrupt might get delayed; record the frame counter value that
+	 * indicates when the HC isn't looking at it, so concurrent unlinks
+	 * behave.  frame_no wraps every 2^16 msec, and changes right before
+	 * SF is triggered.
+	 */
+	ed->tick = ohci_frame_no(ohci) + 1;
+
+}
+
+/*-------------------------------------------------------------------------*
+ * TD handling functions
+ *-------------------------------------------------------------------------*/
+
+/* enqueue next TD for this URB (OHCI spec 5.2.8.2) */
+
+static void
+td_fill (struct ohci_hcd *ohci, u32 info,
+	dma_addr_t data, int len,
+	struct urb *urb, int index)
+{
+	struct td		*td, *td_pt;
+	struct urb_priv		*urb_priv = urb->hcpriv;
+	int			is_iso = info & TD_ISO;
+	int			hash;
+
+	// ASSERT (index < urb_priv->length);
+
+	/* aim for only one interrupt per urb.  mostly applies to control
+	 * and iso; other urbs rarely need more than one TD per urb.
+	 * this way, only final tds (or ones with an error) cause IRQs.
+	 * at least immediately; use DI=6 in case any control request is
+	 * tempted to die part way through.  (and to force the hc to flush
+	 * its donelist soonish, even on unlink paths.)
+	 *
+	 * NOTE: could delay interrupts even for the last TD, and get fewer
+	 * interrupts ... increasing per-urb latency by sharing interrupts.
+	 * Drivers that queue bulk urbs may request that behavior.
+	 */
+	if (index != (urb_priv->length - 1)
+			|| (urb->transfer_flags & URB_NO_INTERRUPT))
+		info |= TD_DI_SET (6);
+
+	/* use this td as the next dummy */
+	td_pt = urb_priv->td [index];
+
+	/* fill the old dummy TD */
+	td = urb_priv->td [index] = urb_priv->ed->dummy;
+	urb_priv->ed->dummy = td_pt;
+
+	td->ed = urb_priv->ed;
+	td->next_dl_td = NULL;
+	td->index = index;
+	td->urb = urb;
+	td->data_dma = data;
+	if (!len)
+		data = 0;
+
+	td->hwINFO = cpu_to_hc32 (ohci, info);
+	if (is_iso) {
+		td->hwCBP = cpu_to_hc32 (ohci, data & 0xFFFFF000);
+		*ohci_hwPSWp(ohci, td, 0) = cpu_to_hc16 (ohci,
+						(data & 0x0FFF) | 0xE000);
+	} else {
+		td->hwCBP = cpu_to_hc32 (ohci, data);
+	}
+	if (data)
+		td->hwBE = cpu_to_hc32 (ohci, data + len - 1);
+	else
+		td->hwBE = 0;
+	td->hwNextTD = cpu_to_hc32 (ohci, td_pt->td_dma);
+
+	/* append to queue */
+	list_add_tail (&td->td_list, &td->ed->td_list);
+
+	/* hash it for later reverse mapping */
+	hash = TD_HASH_FUNC (td->td_dma);
+	td->td_hash = ohci->td_hash [hash];
+	ohci->td_hash [hash] = td;
+
+	/* HC might read the TD (or cachelines) right away ... */
+	wmb ();
+	td->ed->hwTailP = td->hwNextTD;
+}
+
+/*-------------------------------------------------------------------------*/
+
+/* Prepare all TDs of a transfer, and queue them onto the ED.
+ * Caller guarantees HC is active.
+ * Usually the ED is already on the schedule, so TDs might be
+ * processed as soon as they're queued.
+ */
+static void td_submit_urb (
+	struct ohci_hcd	*ohci,
+	struct urb	*urb
+) {
+	struct urb_priv	*urb_priv = urb->hcpriv;
+	struct device *dev = ohci_to_hcd(ohci)->self.controller;
+	dma_addr_t	data;
+	int		data_len = urb->transfer_buffer_length;
+	int		cnt = 0;
+	u32		info = 0;
+	int		is_out = usb_pipeout (urb->pipe);
+	int		periodic = 0;
+	int		i, this_sg_len, n;
+	struct scatterlist	*sg;
+
+	/* OHCI handles the bulk/interrupt data toggles itself.  We just
+	 * use the device toggle bits for resetting, and rely on the fact
+	 * that resetting toggle is meaningless if the endpoint is active.
+	 */
+	if (!usb_gettoggle (urb->dev, usb_pipeendpoint (urb->pipe), is_out)) {
+		usb_settoggle (urb->dev, usb_pipeendpoint (urb->pipe),
+			is_out, 1);
+		urb_priv->ed->hwHeadP &= ~cpu_to_hc32 (ohci, ED_C);
+	}
+
+	list_add (&urb_priv->pending, &ohci->pending);
+
+	i = urb->num_mapped_sgs;
+	if (data_len > 0 && i > 0) {
+		sg = urb->sg;
+		data = sg_dma_address(sg);
+
+		/*
+		 * urb->transfer_buffer_length may be smaller than the
+		 * size of the scatterlist (or vice versa)
+		 */
+		this_sg_len = min_t(int, sg_dma_len(sg), data_len);
+	} else {
+		sg = NULL;
+		if (data_len)
+			data = urb->transfer_dma;
+		else
+			data = 0;
+		this_sg_len = data_len;
+	}
+
+	/* NOTE:  TD_CC is set so we can tell which TDs the HC processed by
+	 * using TD_CC_GET, as well as by seeing them on the done list.
+	 * (CC = NotAccessed ... 0x0F, or 0x0E in PSWs for ISO.)
+	 */
+	switch (urb_priv->ed->type) {
+
+	/* Bulk and interrupt are identical except for where in the schedule
+	 * their EDs live.
+	 */
+	case PIPE_INTERRUPT:
+		/* ... and periodic urbs have extra accounting */
+		periodic = ohci_to_hcd(ohci)->self.bandwidth_int_reqs++ == 0
+			&& ohci_to_hcd(ohci)->self.bandwidth_isoc_reqs == 0;
+		fallthrough;
+	case PIPE_BULK:
+		info = is_out
+			? TD_T_TOGGLE | TD_CC | TD_DP_OUT
+			: TD_T_TOGGLE | TD_CC | TD_DP_IN;
+		/* TDs _could_ transfer up to 8K each */
+		for (;;) {
+			n = min(this_sg_len, 4096);
+
+			/* maybe avoid ED halt on final TD short read */
+			if (n >= data_len || (i == 1 && n >= this_sg_len)) {
+				if (!(urb->transfer_flags & URB_SHORT_NOT_OK))
+					info |= TD_R;
+			}
+			td_fill(ohci, info, data, n, urb, cnt);
+			this_sg_len -= n;
+			data_len -= n;
+			data += n;
+			cnt++;
+
+			if (this_sg_len <= 0) {
+				if (--i <= 0 || data_len <= 0)
+					break;
+				sg = sg_next(sg);
+				data = sg_dma_address(sg);
+				this_sg_len = min_t(int, sg_dma_len(sg),
+						data_len);
+			}
+		}
+		if ((urb->transfer_flags & URB_ZERO_PACKET)
+				&& cnt < urb_priv->length) {
+			td_fill (ohci, info, 0, 0, urb, cnt);
+			cnt++;
+		}
+		/* maybe kickstart bulk list */
+		if (urb_priv->ed->type == PIPE_BULK) {
+			wmb ();
+			ohci_writel (ohci, OHCI_BLF, &ohci->regs->cmdstatus);
+		}
+		break;
+
+	/* control manages DATA0/DATA1 toggle per-request; SETUP resets it,
+	 * any DATA phase works normally, and the STATUS ack is special.
+	 */
+	case PIPE_CONTROL:
+		info = TD_CC | TD_DP_SETUP | TD_T_DATA0;
+		td_fill (ohci, info, urb->setup_dma, 8, urb, cnt++);
+		if (data_len > 0) {
+			info = TD_CC | TD_R | TD_T_DATA1;
+			info |= is_out ? TD_DP_OUT : TD_DP_IN;
+			/* NOTE:  mishandles transfers >8K, some >4K */
+			td_fill (ohci, info, data, data_len, urb, cnt++);
+		}
+		info = (is_out || data_len == 0)
+			? TD_CC | TD_DP_IN | TD_T_DATA1
+			: TD_CC | TD_DP_OUT | TD_T_DATA1;
+		td_fill (ohci, info, data, 0, urb, cnt++);
+		/* maybe kickstart control list */
+		wmb ();
+		ohci_writel (ohci, OHCI_CLF, &ohci->regs->cmdstatus);
+		break;
+
+	/* ISO has no retransmit, so no toggle; and it uses special TDs.
+	 * Each TD could handle multiple consecutive frames (interval 1);
+	 * we could often reduce the number of TDs here.
+	 */
+	case PIPE_ISOCHRONOUS:
+		for (cnt = urb_priv->td_cnt; cnt < urb->number_of_packets;
+				cnt++) {
+			int	frame = urb->start_frame;
+
+			// FIXME scheduling should handle frame counter
+			// roll-around ... exotic case (and OHCI has
+			// a 2^16 iso range, vs other HCs max of 2^10)
+			frame += cnt * urb->interval;
+			frame &= 0xffff;
+			td_fill (ohci, TD_CC | TD_ISO | frame,
+				data + urb->iso_frame_desc [cnt].offset,
+				urb->iso_frame_desc [cnt].length, urb, cnt);
+		}
+		if (ohci_to_hcd(ohci)->self.bandwidth_isoc_reqs == 0) {
+			if (quirk_amdiso(ohci))
+				usb_amd_quirk_pll_disable();
+			if (quirk_amdprefetch(ohci))
+				sb800_prefetch(dev, 1);
+		}
+		periodic = ohci_to_hcd(ohci)->self.bandwidth_isoc_reqs++ == 0
+			&& ohci_to_hcd(ohci)->self.bandwidth_int_reqs == 0;
+		break;
+	}
+
+	/* start periodic dma if needed */
+	if (periodic) {
+		wmb ();
+		ohci->hc_control |= OHCI_CTRL_PLE|OHCI_CTRL_IE;
+		ohci_writel (ohci, ohci->hc_control, &ohci->regs->control);
+	}
+
+	// ASSERT (urb_priv->length == cnt);
+}
+
+/*-------------------------------------------------------------------------*
+ * Done List handling functions
+ *-------------------------------------------------------------------------*/
+
+/* calculate transfer length/status and update the urb */
+static int td_done(struct ohci_hcd *ohci, struct urb *urb, struct td *td)
+{
+	u32	tdINFO = hc32_to_cpup (ohci, &td->hwINFO);
+	int	cc = 0;
+	int	status = -EINPROGRESS;
+
+	list_del (&td->td_list);
+
+	/* ISO ... drivers see per-TD length/status */
+	if (tdINFO & TD_ISO) {
+		u16	tdPSW = ohci_hwPSW(ohci, td, 0);
+		int	dlen = 0;
+
+		/* NOTE:  assumes FC in tdINFO == 0, and that
+		 * only the first of 0..MAXPSW psws is used.
+		 */
+
+		cc = (tdPSW >> 12) & 0xF;
+		if (tdINFO & TD_CC)	/* hc didn't touch? */
+			return status;
+
+		if (usb_pipeout (urb->pipe))
+			dlen = urb->iso_frame_desc [td->index].length;
+		else {
+			/* short reads are always OK for ISO */
+			if (cc == TD_DATAUNDERRUN)
+				cc = TD_CC_NOERROR;
+			dlen = tdPSW & 0x3ff;
+		}
+		urb->actual_length += dlen;
+		urb->iso_frame_desc [td->index].actual_length = dlen;
+		urb->iso_frame_desc [td->index].status = cc_to_error [cc];
+
+		if (cc != TD_CC_NOERROR)
+			ohci_dbg(ohci,
+				"urb %p iso td %p (%d) len %d cc %d\n",
+				urb, td, 1 + td->index, dlen, cc);
+
+	/* BULK, INT, CONTROL ... drivers see aggregate length/status,
+	 * except that "setup" bytes aren't counted and "short" transfers
+	 * might not be reported as errors.
+	 */
+	} else {
+		int	type = usb_pipetype (urb->pipe);
+		u32	tdBE = hc32_to_cpup (ohci, &td->hwBE);
+
+		cc = TD_CC_GET (tdINFO);
+
+		/* update packet status if needed (short is normally ok) */
+		if (cc == TD_DATAUNDERRUN
+				&& !(urb->transfer_flags & URB_SHORT_NOT_OK))
+			cc = TD_CC_NOERROR;
+		if (cc != TD_CC_NOERROR && cc < 0x0E)
+			status = cc_to_error[cc];
+
+		/* count all non-empty packets except control SETUP packet */
+		if ((type != PIPE_CONTROL || td->index != 0) && tdBE != 0) {
+			if (td->hwCBP == 0)
+				urb->actual_length += tdBE - td->data_dma + 1;
+			else
+				urb->actual_length +=
+					  hc32_to_cpup (ohci, &td->hwCBP)
+					- td->data_dma;
+		}
+
+		if (cc != TD_CC_NOERROR && cc < 0x0E)
+			ohci_dbg(ohci,
+				"urb %p td %p (%d) cc %d, len=%d/%d\n",
+				urb, td, 1 + td->index, cc,
+				urb->actual_length,
+				urb->transfer_buffer_length);
+	}
+	return status;
+}
+
+/*-------------------------------------------------------------------------*/
+
+static void ed_halted(struct ohci_hcd *ohci, struct td *td, int cc)
+{
+	struct urb		*urb = td->urb;
+	urb_priv_t		*urb_priv = urb->hcpriv;
+	struct ed		*ed = td->ed;
+	struct list_head	*tmp = td->td_list.next;
+	__hc32			toggle = ed->hwHeadP & cpu_to_hc32 (ohci, ED_C);
+
+	/* clear ed halt; this is the td that caused it, but keep it inactive
+	 * until its urb->complete() has a chance to clean up.
+	 */
+	ed->hwINFO |= cpu_to_hc32 (ohci, ED_SKIP);
+	wmb ();
+	ed->hwHeadP &= ~cpu_to_hc32 (ohci, ED_H);
+
+	/* Get rid of all later tds from this urb.  We don't have
+	 * to be careful: no errors and nothing was transferred.
+	 * Also patch the ed so it looks as if those tds completed normally.
+	 */
+	while (tmp != &ed->td_list) {
+		struct td	*next;
+
+		next = list_entry (tmp, struct td, td_list);
+		tmp = next->td_list.next;
+
+		if (next->urb != urb)
+			break;
+
+		/* NOTE: if multi-td control DATA segments get supported,
+		 * this urb had one of them, this td wasn't the last td
+		 * in that segment (TD_R clear), this ed halted because
+		 * of a short read, _and_ URB_SHORT_NOT_OK is clear ...
+		 * then we need to leave the control STATUS packet queued
+		 * and clear ED_SKIP.
+		 */
+
+		list_del(&next->td_list);
+		urb_priv->td_cnt++;
+		ed->hwHeadP = next->hwNextTD | toggle;
+	}
+
+	/* help for troubleshooting:  report anything that
+	 * looks odd ... that doesn't include protocol stalls
+	 * (or maybe some other things)
+	 */
+	switch (cc) {
+	case TD_DATAUNDERRUN:
+		if ((urb->transfer_flags & URB_SHORT_NOT_OK) == 0)
+			break;
+		fallthrough;
+	case TD_CC_STALL:
+		if (usb_pipecontrol (urb->pipe))
+			break;
+		fallthrough;
+	default:
+		ohci_dbg (ohci,
+			"urb %p path %s ep%d%s %08x cc %d --> status %d\n",
+			urb, urb->dev->devpath,
+			usb_pipeendpoint (urb->pipe),
+			usb_pipein (urb->pipe) ? "in" : "out",
+			hc32_to_cpu (ohci, td->hwINFO),
+			cc, cc_to_error [cc]);
+	}
+}
+
+/* Add a TD to the done list */
+static void add_to_done_list(struct ohci_hcd *ohci, struct td *td)
+{
+	struct td	*td2, *td_prev;
+	struct ed	*ed;
+
+	if (td->next_dl_td)
+		return;		/* Already on the list */
+
+	/* Add all the TDs going back until we reach one that's on the list */
+	ed = td->ed;
+	td2 = td_prev = td;
+	list_for_each_entry_continue_reverse(td2, &ed->td_list, td_list) {
+		if (td2->next_dl_td)
+			break;
+		td2->next_dl_td = td_prev;
+		td_prev = td2;
+	}
+
+	if (ohci->dl_end)
+		ohci->dl_end->next_dl_td = td_prev;
+	else
+		ohci->dl_start = td_prev;
+
+	/*
+	 * Make td->next_dl_td point to td itself, to mark the fact
+	 * that td is on the done list.
+	 */
+	ohci->dl_end = td->next_dl_td = td;
+
+	/* Did we just add the latest pending TD? */
+	td2 = ed->pending_td;
+	if (td2 && td2->next_dl_td)
+		ed->pending_td = NULL;
+}
+
+/* Get the entries on the hardware done queue and put them on our list */
+static void update_done_list(struct ohci_hcd *ohci)
+{
+	u32		td_dma;
+	struct td	*td = NULL;
+
+	td_dma = hc32_to_cpup (ohci, &ohci->hcca->done_head);
+	ohci->hcca->done_head = 0;
+	wmb();
+
+	/* get TD from hc's singly linked list, and
+	 * add to ours.  ed->td_list changes later.
+	 */
+	while (td_dma) {
+		int		cc;
+
+		td = dma_to_td (ohci, td_dma);
+		if (!td) {
+			ohci_err (ohci, "bad entry %8x\n", td_dma);
+			break;
+		}
+
+		td->hwINFO |= cpu_to_hc32 (ohci, TD_DONE);
+		cc = TD_CC_GET (hc32_to_cpup (ohci, &td->hwINFO));
+
+		/* Non-iso endpoints can halt on error; un-halt,
+		 * and dequeue any other TDs from this urb.
+		 * No other TD could have caused the halt.
+		 */
+		if (cc != TD_CC_NOERROR
+				&& (td->ed->hwHeadP & cpu_to_hc32 (ohci, ED_H)))
+			ed_halted(ohci, td, cc);
+
+		td_dma = hc32_to_cpup (ohci, &td->hwNextTD);
+		add_to_done_list(ohci, td);
+	}
+}
+
+/*-------------------------------------------------------------------------*/
+
+/* there are some urbs/eds to unlink; called in_irq(), with HCD locked */
+static void finish_unlinks(struct ohci_hcd *ohci)
+{
+	unsigned	tick = ohci_frame_no(ohci);
+	struct ed	*ed, **last;
+
+rescan_all:
+	for (last = &ohci->ed_rm_list, ed = *last; ed != NULL; ed = *last) {
+		struct list_head	*entry, *tmp;
+		int			completed, modified;
+		__hc32			*prev;
+
+		/* only take off EDs that the HC isn't using, accounting for
+		 * frame counter wraps and EDs with partially retired TDs
+		 */
+		if (likely(ohci->rh_state == OHCI_RH_RUNNING) &&
+				tick_before(tick, ed->tick)) {
+skip_ed:
+			last = &ed->ed_next;
+			continue;
+		}
+		if (!list_empty(&ed->td_list)) {
+			struct td	*td;
+			u32		head;
+
+			td = list_first_entry(&ed->td_list, struct td, td_list);
+
+			/* INTR_WDH may need to clean up first */
+			head = hc32_to_cpu(ohci, ed->hwHeadP) & TD_MASK;
+			if (td->td_dma != head &&
+					ohci->rh_state == OHCI_RH_RUNNING)
+				goto skip_ed;
+
+			/* Don't mess up anything already on the done list */
+			if (td->next_dl_td)
+				goto skip_ed;
+		}
+
+		/* ED's now officially unlinked, hc doesn't see */
+		ed->hwHeadP &= ~cpu_to_hc32(ohci, ED_H);
+		ed->hwNextED = 0;
+		wmb();
+		ed->hwINFO &= ~cpu_to_hc32(ohci, ED_SKIP | ED_DEQUEUE);
+
+		/* reentrancy:  if we drop the schedule lock, someone might
+		 * have modified this list.  normally it's just prepending
+		 * entries (which we'd ignore), but paranoia won't hurt.
+		 */
+		*last = ed->ed_next;
+		ed->ed_next = NULL;
+		modified = 0;
+
+		/* unlink urbs as requested, but rescan the list after
+		 * we call a completion since it might have unlinked
+		 * another (earlier) urb
+		 *
+		 * When we get here, the HC doesn't see this ed.  But it
+		 * must not be rescheduled until all completed URBs have
+		 * been given back to the driver.
+		 */
+rescan_this:
+		completed = 0;
+		prev = &ed->hwHeadP;
+		list_for_each_safe (entry, tmp, &ed->td_list) {
+			struct td	*td;
+			struct urb	*urb;
+			urb_priv_t	*urb_priv;
+			__hc32		savebits;
+			u32		tdINFO;
+
+			td = list_entry (entry, struct td, td_list);
+			urb = td->urb;
+			urb_priv = td->urb->hcpriv;
+
+			if (!urb->unlinked) {
+				prev = &td->hwNextTD;
+				continue;
+			}
+
+			/* patch pointer hc uses */
+			savebits = *prev & ~cpu_to_hc32 (ohci, TD_MASK);
+			*prev = td->hwNextTD | savebits;
+
+			/* If this was unlinked, the TD may not have been
+			 * retired ... so manually save the data toggle.
+			 * The controller ignores the value we save for
+			 * control and ISO endpoints.
+			 */
+			tdINFO = hc32_to_cpup(ohci, &td->hwINFO);
+			if ((tdINFO & TD_T) == TD_T_DATA0)
+				ed->hwHeadP &= ~cpu_to_hc32(ohci, ED_C);
+			else if ((tdINFO & TD_T) == TD_T_DATA1)
+				ed->hwHeadP |= cpu_to_hc32(ohci, ED_C);
+
+			/* HC may have partly processed this TD */
+			td_done (ohci, urb, td);
+			urb_priv->td_cnt++;
+
+			/* if URB is done, clean up */
+			if (urb_priv->td_cnt >= urb_priv->length) {
+				modified = completed = 1;
+				finish_urb(ohci, urb, 0);
+			}
+		}
+		if (completed && !list_empty (&ed->td_list))
+			goto rescan_this;
+
+		/*
+		 * If no TDs are queued, ED is now idle.
+		 * Otherwise, if the HC is running, reschedule.
+		 * If the HC isn't running, add ED back to the
+		 * start of the list for later processing.
+		 */
+		if (list_empty(&ed->td_list)) {
+			ed->state = ED_IDLE;
+			list_del(&ed->in_use_list);
+		} else if (ohci->rh_state == OHCI_RH_RUNNING) {
+			ed_schedule(ohci, ed);
+		} else {
+			ed->ed_next = ohci->ed_rm_list;
+			ohci->ed_rm_list = ed;
+			/* Don't loop on the same ED */
+			if (last == &ohci->ed_rm_list)
+				last = &ed->ed_next;
+		}
+
+		if (modified)
+			goto rescan_all;
+	}
+
+	/* maybe reenable control and bulk lists */
+	if (ohci->rh_state == OHCI_RH_RUNNING && !ohci->ed_rm_list) {
+		u32	command = 0, control = 0;
+
+		if (ohci->ed_controltail) {
+			command |= OHCI_CLF;
+			if (quirk_zfmicro(ohci))
+				mdelay(1);
+			if (!(ohci->hc_control & OHCI_CTRL_CLE)) {
+				control |= OHCI_CTRL_CLE;
+				ohci_writel (ohci, 0,
+					&ohci->regs->ed_controlcurrent);
+			}
+		}
+		if (ohci->ed_bulktail) {
+			command |= OHCI_BLF;
+			if (quirk_zfmicro(ohci))
+				mdelay(1);
+			if (!(ohci->hc_control & OHCI_CTRL_BLE)) {
+				control |= OHCI_CTRL_BLE;
+				ohci_writel (ohci, 0,
+					&ohci->regs->ed_bulkcurrent);
+			}
+		}
+
+		/* CLE/BLE to enable, CLF/BLF to (maybe) kickstart */
+		if (control) {
+			ohci->hc_control |= control;
+			if (quirk_zfmicro(ohci))
+				mdelay(1);
+			ohci_writel (ohci, ohci->hc_control,
+					&ohci->regs->control);
+		}
+		if (command) {
+			if (quirk_zfmicro(ohci))
+				mdelay(1);
+			ohci_writel (ohci, command, &ohci->regs->cmdstatus);
+		}
+	}
+}
+
+
+
+/*-------------------------------------------------------------------------*/
+
+/* Take back a TD from the host controller */
+static void takeback_td(struct ohci_hcd *ohci, struct td *td)
+{
+	struct urb	*urb = td->urb;
+	urb_priv_t	*urb_priv = urb->hcpriv;
+	struct ed	*ed = td->ed;
+	int		status;
+
+	/* update URB's length and status from TD */
+	status = td_done(ohci, urb, td);
+	urb_priv->td_cnt++;
+
+	/* If all this urb's TDs are done, call complete() */
+	if (urb_priv->td_cnt >= urb_priv->length)
+		finish_urb(ohci, urb, status);
+
+	/* clean schedule:  unlink EDs that are no longer busy */
+	if (list_empty(&ed->td_list)) {
+		if (ed->state == ED_OPER)
+			start_ed_unlink(ohci, ed);
+
+	/* ... reenabling halted EDs only after fault cleanup */
+	} else if ((ed->hwINFO & cpu_to_hc32(ohci, ED_SKIP | ED_DEQUEUE))
+			== cpu_to_hc32(ohci, ED_SKIP)) {
+		td = list_entry(ed->td_list.next, struct td, td_list);
+		if (!(td->hwINFO & cpu_to_hc32(ohci, TD_DONE))) {
+			ed->hwINFO &= ~cpu_to_hc32(ohci, ED_SKIP);
+			/* ... hc may need waking-up */
+			switch (ed->type) {
+			case PIPE_CONTROL:
+				ohci_writel(ohci, OHCI_CLF,
+						&ohci->regs->cmdstatus);
+				break;
+			case PIPE_BULK:
+				ohci_writel(ohci, OHCI_BLF,
+						&ohci->regs->cmdstatus);
+				break;
+			}
+		}
+	}
+}
+
+/*
+ * Process normal completions (error or success) and clean the schedules.
+ *
+ * This is the main path for handing urbs back to drivers.  The only other
+ * normal path is finish_unlinks(), which unlinks URBs using ed_rm_list,
+ * instead of scanning the (re-reversed) donelist as this does.
+ */
+static void process_done_list(struct ohci_hcd *ohci)
+{
+	struct td	*td;
+
+	while (ohci->dl_start) {
+		td = ohci->dl_start;
+		if (td == ohci->dl_end)
+			ohci->dl_start = ohci->dl_end = NULL;
+		else
+			ohci->dl_start = td->next_dl_td;
+
+		takeback_td(ohci, td);
+	}
+}
+
+/*
+ * TD takeback and URB giveback must be single-threaded.
+ * This routine takes care of it all.
+ */
+static void ohci_work(struct ohci_hcd *ohci)
+{
+	if (ohci->working) {
+		ohci->restart_work = 1;
+		return;
+	}
+	ohci->working = 1;
+
+ restart:
+	process_done_list(ohci);
+	if (ohci->ed_rm_list)
+		finish_unlinks(ohci);
+
+	if (ohci->restart_work) {
+		ohci->restart_work = 0;
+		goto restart;
+	}
+	ohci->working = 0;
+}