aboutsummaryrefslogtreecommitdiff
path: root/sys
diff options
context:
space:
mode:
authorAndrew Turner <andrew@FreeBSD.org>2020-07-06 08:51:55 +0000
committerAndrew Turner <andrew@FreeBSD.org>2020-07-06 08:51:55 +0000
commiteed8b80f64f0f419c6cff47a00e2811d0171bcea (patch)
tree00f96372cf259882be4b2c04a0c7e873634cacf2 /sys
parent1866c98e64d654649d6e134c20e3e6db7823915d (diff)
downloadsrc-eed8b80f64f0f419c6cff47a00e2811d0171bcea.tar.gz
src-eed8b80f64f0f419c6cff47a00e2811d0171bcea.zip
Add a driver for bcm2838 PCI express controller
This adds support for the Broadcom bcm2711 PCI express controller, found on the Raspberry Pi 4 (aka the bcm2838 SoC). The driver has only been developed against the soldered-on VIA XHCI controller and not tested with other end points. Submitted by: Robert Crowston <crowston_protonmail.com> Differential Revision: https://reviews.freebsd.org/D25068
Notes
Notes: svn path=/head/; revision=362954
Diffstat (limited to 'sys')
-rw-r--r--sys/arm/broadcom/bcm2835/bcm2838_pci.c743
-rw-r--r--sys/arm/broadcom/bcm2835/files.bcm283x1
-rw-r--r--sys/conf/files.arm641
-rw-r--r--sys/dev/pci/pci_host_generic_fdt.c18
-rw-r--r--sys/dev/pci/pci_host_generic_fdt.h1
5 files changed, 763 insertions, 1 deletions
diff --git a/sys/arm/broadcom/bcm2835/bcm2838_pci.c b/sys/arm/broadcom/bcm2835/bcm2838_pci.c
new file mode 100644
index 000000000000..59f2d264de22
--- /dev/null
+++ b/sys/arm/broadcom/bcm2835/bcm2838_pci.c
@@ -0,0 +1,743 @@
+/*-
+ * SPDX-License-Identifier: ISC
+ *
+ * Copyright (c) 2020 Dr Robert Harvey Crowston <crowston@protonmail.com>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ *
+ *
+ * $FreeBSD$
+ *
+ */
+
+/*
+ * BCM2838-compatible PCI-express controller.
+ *
+ * Broadcom likes to give the same chip lots of different names. The name of
+ * this driver is taken from the Raspberry Pi 4 Broadcom 2838 chip.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/systm.h>
+#include <sys/endian.h>
+#include <sys/kernel.h>
+#include <sys/module.h>
+#include <sys/bus.h>
+#include <sys/proc.h>
+#include <sys/rman.h>
+#include <sys/intr.h>
+#include <sys/mutex.h>
+
+#include <dev/ofw/openfirm.h>
+#include <dev/ofw/ofw_bus.h>
+#include <dev/ofw/ofw_bus_subr.h>
+
+#include <dev/pci/pci_host_generic.h>
+#include <dev/pci/pci_host_generic_fdt.h>
+#include <dev/pci/pcivar.h>
+#include <dev/pci/pcireg.h>
+#include <dev/pci/pcib_private.h>
+
+#include <machine/bus.h>
+#include <machine/intr.h>
+
+#include "pcib_if.h"
+#include "msi_if.h"
+
+extern struct bus_space memmap_bus;
+
+#define BUS_SPACE_3G_MAXADDR 0xc0000000
+#define PCI_ID_VAL3 0x43c
+#define CLASS_SHIFT 0x10
+#define SUBCLASS_SHIFT 0x8
+
+#define REG_CONTROLLER_HW_REV 0x406c
+#define REG_BRIDGE_CTRL 0x9210
+#define BRIDGE_DISABLE_FLAG 0x1
+#define BRIDGE_RESET_FLAG 0x2
+#define REG_BRIDGE_SERDES_MODE 0x4204
+#define REG_BRIDGE_CONFIG 0x4008
+#define REG_BRIDGE_MEM_WINDOW_LOW 0x4034
+#define REG_BRIDGE_MEM_WINDOW_HIGH 0x4038
+#define REG_BRIDGE_MEM_WINDOW_1 0x403c
+#define REG_BRIDGE_GISB_WINDOW 0x402c
+#define REG_BRIDGE_STATE 0x4068
+#define REG_BRIDGE_LINK_STATE 0x00bc
+#define REG_BRIDGE_BUS_WINDOW_LOW 0x400c
+#define REG_BRIDGE_BUS_WINDOW_HIGH 0x4010
+#define REG_BRIDGE_CPU_WINDOW_LOW 0x4070
+#define REG_BRIDGE_CPU_WINDOW_START_HIGH 0x4080
+#define REG_BRIDGE_CPU_WINDOW_END_HIGH 0x4084
+
+#define REG_MSI_ADDR_LOW 0x4044
+#define REG_MSI_ADDR_HIGH 0x4048
+#define REG_MSI_CONFIG 0x404c
+#define REG_MSI_CLR 0x4508
+#define REG_MSI_MASK_CLR 0x4514
+#define REG_MSI_RAISED 0x4500
+#define REG_MSI_EOI 0x4060
+#define NUM_MSI 32
+
+#define REG_EP_CONFIG_CHOICE 0x9000
+#define REG_EP_CONFIG_DATA 0x8000
+
+/*
+ * These values were obtained from runtime inspection of a Linux system using a
+ * JTAG. The very limited documentation I have obtained from Broadcom does not
+ * explain how to compute them.
+ */
+#define REG_VALUE_4GB_WINDOW 0x11
+#define REG_VALUE_4GB_CONFIG 0x88003000
+#define REG_VALUE_MSI_CONFIG 0xffe06540
+
+struct bcm_pcib_irqsrc {
+ struct intr_irqsrc isrc;
+ u_int irq;
+ bool allocated;
+};
+
+struct bcm_pcib_softc {
+ struct generic_pcie_fdt_softc base;
+ device_t dev;
+ struct mtx config_mtx;
+ struct mtx msi_mtx;
+ struct resource *msi_irq_res;
+ void *msi_intr_cookie;
+ struct bcm_pcib_irqsrc *msi_isrcs;
+ pci_addr_t msi_addr;
+};
+
+static struct ofw_compat_data compat_data[] = {
+ {"brcm,bcm2711-pcie", 1},
+ {"brcm,bcm7211-pcie", 1},
+ {"brcm,bcm7445-pcie", 1},
+
+ {NULL, 0}
+};
+
+static int
+bcm_pcib_probe(device_t dev)
+{
+
+ if (!ofw_bus_status_okay(dev))
+ return (ENXIO);
+
+ if (!ofw_bus_search_compatible(dev, compat_data)->ocd_data)
+ return (ENXIO);
+
+ device_set_desc(dev,
+ "BCM2838-compatible PCI-express controller");
+ return (BUS_PROBE_DEFAULT);
+}
+
+static void
+bcm_pcib_set_reg(struct bcm_pcib_softc *sc, uint32_t reg, uint32_t val)
+{
+
+ bus_space_write_4(sc->base.base.bst, sc->base.base.bsh, reg,
+ htole32(val));
+}
+
+static uint32_t
+bcm_pcib_read_reg(struct bcm_pcib_softc *sc, uint32_t reg)
+{
+
+ return (le32toh(bus_space_read_4(sc->base.base.bst, sc->base.base.bsh,
+ reg)));
+}
+
+static void
+bcm_pcib_reset_controller(struct bcm_pcib_softc *sc)
+{
+ uint32_t val;
+
+ val = bcm_pcib_read_reg(sc, REG_BRIDGE_CTRL);
+ val = val | BRIDGE_RESET_FLAG | BRIDGE_DISABLE_FLAG;
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CTRL, val);
+
+ DELAY(100);
+
+ val = bcm_pcib_read_reg(sc, REG_BRIDGE_CTRL);
+ val = val & ~BRIDGE_RESET_FLAG;
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CTRL, val);
+
+ DELAY(100);
+
+ bcm_pcib_set_reg(sc, REG_BRIDGE_SERDES_MODE, 0);
+
+ DELAY(100);
+}
+
+static void
+bcm_pcib_enable_controller(struct bcm_pcib_softc *sc)
+{
+ uint32_t val;
+
+ val = bcm_pcib_read_reg(sc, REG_BRIDGE_CTRL);
+ val = val & ~BRIDGE_DISABLE_FLAG;
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CTRL, val);
+
+ DELAY(100);
+}
+
+static int
+bcm_pcib_check_ranges(device_t dev)
+{
+ struct bcm_pcib_softc *sc;
+ struct pcie_range *ranges;
+ int error = 0, i;
+
+ sc = device_get_softc(dev);
+ ranges = &sc->base.base.ranges[0];
+
+ /* The first range needs to be non-zero. */
+ if (ranges[0].size == 0) {
+ device_printf(dev, "error: first outbound memory range "
+ "(pci addr: 0x%jx, cpu addr: 0x%jx) has zero size.\n",
+ ranges[0].pci_base, ranges[0].phys_base);
+ error = ENXIO;
+ }
+
+ /*
+ * The controller can actually handle three distinct ranges, but we
+ * only implement support for one.
+ */
+ for (i = 1; (bootverbose || error) && i < MAX_RANGES_TUPLES; ++i) {
+ if (ranges[i].size > 0)
+ device_printf(dev,
+ "note: outbound memory range %d (pci addr: 0x%jx, "
+ "cpu addr: 0x%jx, size: 0x%jx) will be ignored.\n",
+ i, ranges[i].pci_base, ranges[i].phys_base,
+ ranges[i].size);
+ }
+
+ return (error);
+}
+
+static const char *
+bcm_pcib_link_state_string(uint32_t mode)
+{
+
+ switch(mode & PCIEM_LINK_STA_SPEED) {
+ case 0:
+ return ("not up");
+ case 1:
+ return ("2.5 GT/s");
+ case 2:
+ return ("5.0 GT/s");
+ case 4:
+ return ("8.0 GT/s");
+ default:
+ return ("unknown");
+ }
+}
+
+static bus_addr_t
+bcm_get_offset_and_prepare_config(struct bcm_pcib_softc *sc, u_int bus,
+ u_int slot, u_int func, u_int reg)
+{
+ /*
+ * Config for an end point is only available through a narrow window for
+ * one end point at a time. We first tell the controller which end point
+ * we want, then access it through the window.
+ */
+ uint32_t func_index;
+
+ if (bus == 0 && slot == 0 && func == 0)
+ /*
+ * Special case for root device; its config is always available
+ * through the zero-offset.
+ */
+ return (reg);
+
+ /* Tell the controller to show us the config in question. */
+ func_index = PCIE_ADDR_OFFSET(bus, slot, func, 0);
+ bcm_pcib_set_reg(sc, REG_EP_CONFIG_CHOICE, func_index);
+
+ return (REG_EP_CONFIG_DATA + reg);
+}
+
+static bool
+bcm_pcib_is_valid_quad(struct bcm_pcib_softc *sc, u_int bus, u_int slot,
+ u_int func, u_int reg)
+{
+
+ if ((bus < sc->base.base.bus_start) || (bus > sc->base.base.bus_end))
+ return (false);
+ if ((slot > PCI_SLOTMAX) || (func > PCI_FUNCMAX) || (reg > PCIE_REGMAX))
+ return (false);
+
+ if (bus == 0 && slot == 0 && func == 0)
+ return (true);
+ if (bus == 0)
+ /*
+ * Probing other slots and funcs on bus 0 will lock up the
+ * memory controller.
+ */
+ return (false);
+
+ return (true);
+}
+
+static uint32_t
+bcm_pcib_read_config(device_t dev, u_int bus, u_int slot, u_int func, u_int reg,
+ int bytes)
+{
+ struct bcm_pcib_softc *sc;
+ bus_space_handle_t h;
+ bus_space_tag_t t;
+ bus_addr_t offset;
+ uint32_t data;
+
+ sc = device_get_softc(dev);
+ if (!bcm_pcib_is_valid_quad(sc, bus, slot, func, reg))
+ return (~0U);
+
+ mtx_lock(&sc->config_mtx);
+ offset = bcm_get_offset_and_prepare_config(sc, bus, slot, func, reg);
+
+ t = sc->base.base.bst;
+ h = sc->base.base.bsh;
+
+ switch (bytes) {
+ case 1:
+ data = bus_space_read_1(t, h, offset);
+ break;
+ case 2:
+ data = le16toh(bus_space_read_2(t, h, offset));
+ break;
+ case 4:
+ data = le32toh(bus_space_read_4(t, h, offset));
+ break;
+ default:
+ data = ~0U;
+ break;
+ }
+
+ mtx_unlock(&sc->config_mtx);
+ return (data);
+}
+
+static void
+bcm_pcib_write_config(device_t dev, u_int bus, u_int slot,
+ u_int func, u_int reg, uint32_t val, int bytes)
+{
+ struct bcm_pcib_softc *sc;
+ bus_space_handle_t h;
+ bus_space_tag_t t;
+ uint32_t offset;
+
+ sc = device_get_softc(dev);
+ if (!bcm_pcib_is_valid_quad(sc, bus, slot, func, reg))
+ return;
+
+ mtx_lock(&sc->config_mtx);
+ offset = bcm_get_offset_and_prepare_config(sc, bus, slot, func, reg);
+
+ t = sc->base.base.bst;
+ h = sc->base.base.bsh;
+
+ switch (bytes) {
+ case 1:
+ bus_space_write_1(t, h, offset, val);
+ break;
+ case 2:
+ bus_space_write_2(t, h, offset, htole16(val));
+ break;
+ case 4:
+ bus_space_write_4(t, h, offset, htole32(val));
+ break;
+ default:
+ break;
+ }
+
+ mtx_unlock(&sc->config_mtx);
+}
+
+static void
+bcm_pcib_msi_intr_process(struct bcm_pcib_softc *sc, uint32_t interrupt_bitmap,
+ struct trapframe *tf)
+{
+ struct bcm_pcib_irqsrc *irqsrc;
+ uint32_t bit, irq;
+
+ while ((bit = ffs(interrupt_bitmap))) {
+ irq = bit - 1;
+
+ /* Acknowledge interrupt. */
+ bcm_pcib_set_reg(sc, REG_MSI_CLR, 1 << irq);
+
+ /* Send EOI. */
+ bcm_pcib_set_reg(sc, REG_MSI_EOI, 1);
+
+ /* Despatch to handler. */
+ irqsrc = &sc->msi_isrcs[irq];
+ if (intr_isrc_dispatch(&irqsrc->isrc, tf))
+ device_printf(sc->dev,
+ "note: unexpected interrupt (%d) triggered.\n",
+ irq);
+
+ /* Done with this interrupt. */
+ interrupt_bitmap = interrupt_bitmap & ~(1 << irq);
+ }
+}
+
+static int
+bcm_pcib_msi_intr(void *arg)
+{
+ struct bcm_pcib_softc *sc;
+ struct trapframe *tf;
+ uint32_t interrupt_bitmap;
+
+ sc = (struct bcm_pcib_softc *) arg;
+ tf = curthread->td_intr_frame;
+
+ while ((interrupt_bitmap = bcm_pcib_read_reg(sc, REG_MSI_RAISED)))
+ bcm_pcib_msi_intr_process(sc, interrupt_bitmap, tf);
+
+ return (FILTER_HANDLED);
+}
+
+static int
+bcm_pcib_alloc_msi(device_t dev, device_t child, int count, int maxcount,
+ device_t *pic, struct intr_irqsrc **srcs)
+{
+ struct bcm_pcib_softc *sc;
+ int first_int, i;
+
+ sc = device_get_softc(dev);
+ mtx_lock(&sc->msi_mtx);
+
+ /* Find a continguous region of free message-signalled interrupts. */
+ for (first_int = 0; first_int + count < NUM_MSI; ) {
+ for (i = first_int; i < first_int + count; ++i) {
+ if (sc->msi_isrcs[i].allocated)
+ goto next;
+ }
+ goto found;
+next:
+ first_int = i + 1;
+ }
+
+ /* No appropriate region available. */
+ mtx_unlock(&sc->msi_mtx);
+ device_printf(dev, "warning: failed to allocate %d MSI messages.\n",
+ count);
+ return (ENXIO);
+
+found:
+ /* Mark the messages as in use. */
+ for (i = 0; i < count; ++i) {
+ sc->msi_isrcs[i + first_int].allocated = true;
+ srcs[i] = &(sc->msi_isrcs[i + first_int].isrc);
+ }
+
+ mtx_unlock(&sc->msi_mtx);
+ *pic = device_get_parent(dev);
+
+ return (0);
+}
+
+static int
+bcm_pcib_map_msi(device_t dev, device_t child, struct intr_irqsrc *isrc,
+ uint64_t *addr, uint32_t *data)
+{
+ struct bcm_pcib_softc *sc;
+ struct bcm_pcib_irqsrc *msi_msg;
+
+ sc = device_get_softc(dev);
+ msi_msg = (struct bcm_pcib_irqsrc *) isrc;
+
+ *addr = sc->msi_addr;
+ *data = (REG_VALUE_MSI_CONFIG & 0xffff) | msi_msg->irq;
+ return (0);
+}
+
+static int
+bcm_pcib_release_msi(device_t dev, device_t child, int count,
+ struct intr_irqsrc **isrc)
+{
+ struct bcm_pcib_softc *sc;
+ struct bcm_pcib_irqsrc *msi_isrc;
+ int i;
+
+ sc = device_get_softc(dev);
+ mtx_lock(&sc->msi_mtx);
+
+ for (i = 0; i < count; i++) {
+ msi_isrc = (struct bcm_pcib_irqsrc *) isrc[i];
+ msi_isrc->allocated = false;
+ }
+
+ mtx_unlock(&sc->msi_mtx);
+ return (0);
+}
+
+static int
+bcm_pcib_msi_attach(device_t dev)
+{
+ struct bcm_pcib_softc *sc;
+ phandle_t node, xref;
+ char const *bcm_name;
+ int i, rid;
+
+ sc = device_get_softc(dev);
+ sc->msi_addr = 0xffffffffc;
+
+ /* Clear any pending interrupts. */
+ bcm_pcib_set_reg(sc, REG_MSI_CLR, 0xffffffff);
+
+ rid = 1;
+ sc->msi_irq_res = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
+ RF_ACTIVE);
+ if (sc->msi_irq_res == NULL) {
+ device_printf(dev, "could not allocate MSI irq resource.\n");
+ return (ENXIO);
+ }
+
+ sc->msi_isrcs = malloc(sizeof(*sc->msi_isrcs) * NUM_MSI, M_DEVBUF,
+ M_WAITOK | M_ZERO);
+
+ int error = bus_setup_intr(dev, sc->msi_irq_res, INTR_TYPE_BIO |
+ INTR_MPSAFE, bcm_pcib_msi_intr, NULL, sc, &sc->msi_intr_cookie);
+ if (error) {
+ device_printf(dev, "error: failed to setup MSI handler.\n");
+ return (ENXIO);
+ }
+
+ bcm_name = device_get_nameunit(dev);
+ for (i = 0; i < NUM_MSI; i++) {
+ sc->msi_isrcs[i].irq = i;
+ error = intr_isrc_register(&sc->msi_isrcs[i].isrc, dev, 0,
+ "%s,%u", bcm_name, i);
+ if (error) {
+ device_printf(dev,
+ "error: failed to register interrupt %d.\n", i);
+ return (ENXIO);
+ }
+ }
+
+ node = ofw_bus_get_node(dev);
+ xref = OF_xref_from_node(node);
+ OF_device_register_xref(xref, dev);
+
+ error = intr_msi_register(dev, xref);
+ if (error)
+ return (ENXIO);
+
+ mtx_init(&sc->msi_mtx, "bcm_pcib: msi_mtx", NULL, MTX_DEF);
+
+ bcm_pcib_set_reg(sc, REG_MSI_MASK_CLR, 0xffffffff);
+ bcm_pcib_set_reg(sc, REG_MSI_ADDR_LOW, (sc->msi_addr & 0xffffffff) | 1);
+ bcm_pcib_set_reg(sc, REG_MSI_ADDR_HIGH, (sc->msi_addr >> 32));
+ bcm_pcib_set_reg(sc, REG_MSI_CONFIG, REG_VALUE_MSI_CONFIG);
+
+ return (0);
+}
+
+static void
+bcm_pcib_relocate_bridge_window(device_t dev)
+{
+ /*
+ * In principle an out-of-bounds bridge window could be automatically
+ * adjusted at resource-activation time to lie within the bus address
+ * space by pcib_grow_window(), but that is not possible because the
+ * out-of-bounds resource allocation fails at allocation time. Instead,
+ * we will just fix up the window on the controller here, before it is
+ * re-discovered by pcib_probe_windows().
+ */
+
+ struct bcm_pcib_softc *sc;
+ pci_addr_t base, size, new_base, new_limit;
+ uint16_t val;
+
+ sc = device_get_softc(dev);
+
+ val = bcm_pcib_read_config(dev, 0, 0, 0, PCIR_MEMBASE_1, 2);
+ base = PCI_PPBMEMBASE(0, val);
+
+ val = bcm_pcib_read_config(dev, 0, 0, 0, PCIR_MEMLIMIT_1, 2);
+ size = PCI_PPBMEMLIMIT(0, val) - base;
+
+ new_base = sc->base.base.ranges[0].pci_base;
+ val = (uint16_t) (new_base >> 16);
+ bcm_pcib_write_config(dev, 0, 0, 0, PCIR_MEMBASE_1, val, 2);
+
+ new_limit = new_base + size;
+ val = (uint16_t) (new_limit >> 16);
+ bcm_pcib_write_config(dev, 0, 0, 0, PCIR_MEMLIMIT_1, val, 2);
+}
+
+static uint32_t
+encode_cpu_window_low(pci_addr_t phys_base, bus_size_t size)
+{
+
+ return (((phys_base >> 0x10) & 0xfff0) |
+ ((phys_base + size - 1) & 0xfff00000));
+}
+
+static uint32_t
+encode_cpu_window_start_high(pci_addr_t phys_base)
+{
+
+ return ((phys_base >> 0x20) & 0xff);
+}
+
+static uint32_t
+encode_cpu_window_end_high(pci_addr_t phys_base, bus_size_t size)
+{
+
+ return (((phys_base + size - 1) >> 0x20) & 0xff);
+}
+
+static int
+bcm_pcib_attach(device_t dev)
+{
+ struct bcm_pcib_softc *sc;
+ pci_addr_t phys_base, pci_base;
+ bus_size_t size;
+ uint32_t hardware_rev, bridge_state, link_state;
+ int error, tries;
+
+ sc = device_get_softc(dev);
+ sc->dev = dev;
+
+ error = pci_host_generic_setup_fdt(dev);
+ if (error)
+ return (error);
+
+ error = bcm_pcib_check_ranges(dev);
+ if (error)
+ return (error);
+
+ mtx_init(&sc->config_mtx, "bcm_pcib: config_mtx", NULL, MTX_DEF);
+
+ bcm_pcib_reset_controller(sc);
+
+ hardware_rev = bcm_pcib_read_reg(sc, REG_CONTROLLER_HW_REV) & 0xffff;
+ device_printf(dev, "hardware identifies as revision 0x%x.\n",
+ hardware_rev);
+
+ /*
+ * Set PCI->CPU memory window. This encodes the inbound window showing
+ * up to 4 GiB of system memory to the controller, with zero offset.
+ * Thus, from the perspective of a device on the PCI-E bus, there is a
+ * 1:1 map from PCI-E bus addresses to system memory addresses. However,
+ * a hardware limitation means that the controller can only perform DMA
+ * on the lower 3 GiB of system memory.
+ */
+ bcm_pcib_set_reg(sc, REG_BRIDGE_MEM_WINDOW_LOW, REG_VALUE_4GB_WINDOW);
+ bcm_pcib_set_reg(sc, REG_BRIDGE_MEM_WINDOW_HIGH, 0);
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CONFIG, REG_VALUE_4GB_CONFIG);
+ bcm_pcib_set_reg(sc, REG_BRIDGE_GISB_WINDOW, 0);
+ bcm_pcib_set_reg(sc, REG_BRIDGE_MEM_WINDOW_1, 0);
+
+ bcm_pcib_enable_controller(sc);
+
+ /* Wait for controller to start. */
+ for(tries = 0; ; ++tries) {
+ bridge_state = bcm_pcib_read_reg(sc, REG_BRIDGE_STATE);
+
+ if ((bridge_state & 0x30) == 0x30)
+ /* Controller ready. */
+ break;
+
+ if (tries > 100) {
+ device_printf(dev,
+ "error: controller failed to start.\n");
+ return (ENXIO);
+ }
+
+ DELAY(1000);
+ }
+
+ link_state = bcm_pcib_read_reg(sc, REG_BRIDGE_LINK_STATE) >> 0x10;
+ if (!link_state) {
+ device_printf(dev, "error: controller started but link is not "
+ "up.\n");
+ return (ENXIO);
+ }
+ if (bootverbose)
+ device_printf(dev, "note: reported link speed is %s.\n",
+ bcm_pcib_link_state_string(link_state));
+
+ /*
+ * Set the CPU->PCI memory window. The map in this direction is not 1:1.
+ * Addresses seen by the CPU need to be adjusted to make sense to the
+ * controller as they pass through the window.
+ */
+ pci_base = sc->base.base.ranges[0].pci_base;
+ phys_base = sc->base.base.ranges[0].phys_base;
+ size = sc->base.base.ranges[0].size;
+
+ bcm_pcib_set_reg(sc, REG_BRIDGE_BUS_WINDOW_LOW, pci_base & 0xffffffff);
+ bcm_pcib_set_reg(sc, REG_BRIDGE_BUS_WINDOW_HIGH, pci_base >> 32);
+
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CPU_WINDOW_LOW,
+ encode_cpu_window_low(phys_base, size));
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CPU_WINDOW_START_HIGH,
+ encode_cpu_window_start_high(phys_base));
+ bcm_pcib_set_reg(sc, REG_BRIDGE_CPU_WINDOW_END_HIGH,
+ encode_cpu_window_end_high(phys_base, size));
+
+ /*
+ * The controller starts up declaring itself an endpoint; readvertise it
+ * as a bridge.
+ */
+ bcm_pcib_set_reg(sc, PCI_ID_VAL3,
+ PCIC_BRIDGE << CLASS_SHIFT | PCIS_BRIDGE_PCI << SUBCLASS_SHIFT);
+
+ bcm_pcib_set_reg(sc, REG_BRIDGE_SERDES_MODE, 0x2);
+ DELAY(100);
+
+ bcm_pcib_relocate_bridge_window(dev);
+
+ /* Configure interrupts. */
+ error = bcm_pcib_msi_attach(dev);
+ if (error)
+ return (error);
+
+ /* Done. */
+ device_add_child(dev, "pci", -1);
+ return (bus_generic_attach(dev));
+}
+
+/*
+ * Device method table.
+ */
+static device_method_t bcm_pcib_methods[] = {
+ /* Device interface. */
+ DEVMETHOD(device_probe, bcm_pcib_probe),
+ DEVMETHOD(device_attach, bcm_pcib_attach),
+
+ /* PCIB interface. */
+ DEVMETHOD(pcib_read_config, bcm_pcib_read_config),
+ DEVMETHOD(pcib_write_config, bcm_pcib_write_config),
+
+ /* MSI interface. */
+ DEVMETHOD(msi_alloc_msi, bcm_pcib_alloc_msi),
+ DEVMETHOD(msi_release_msi, bcm_pcib_release_msi),
+ DEVMETHOD(msi_map_msi, bcm_pcib_map_msi),
+
+ DEVMETHOD_END
+};
+
+DEFINE_CLASS_1(pcib, bcm_pcib_driver, bcm_pcib_methods,
+ sizeof(struct bcm_pcib_softc), generic_pcie_fdt_driver);
+
+static devclass_t bcm_pcib_devclass;
+DRIVER_MODULE(pcib, simplebus, bcm_pcib_driver, bcm_pcib_devclass, 0, 0);
+
diff --git a/sys/arm/broadcom/bcm2835/files.bcm283x b/sys/arm/broadcom/bcm2835/files.bcm283x
index 09cea7b4a4c3..f367d46a018f 100644
--- a/sys/arm/broadcom/bcm2835/files.bcm283x
+++ b/sys/arm/broadcom/bcm2835/files.bcm283x
@@ -17,6 +17,7 @@ arm/broadcom/bcm2835/bcm2835_spi.c optional bcm2835_spi
arm/broadcom/bcm2835/bcm2835_vcbus.c standard
arm/broadcom/bcm2835/bcm2835_vcio.c standard
arm/broadcom/bcm2835/bcm2835_wdog.c standard
+arm/broadcom/bcm2835/bcm2838_pci.c optional pci
arm/broadcom/bcm2835/bcm283x_dwc_fdt.c optional dwcotg fdt
dev/mbox/mbox_if.m standard
diff --git a/sys/conf/files.arm64 b/sys/conf/files.arm64
index 2667fc784ee8..bbb19b75084d 100644
--- a/sys/conf/files.arm64
+++ b/sys/conf/files.arm64
@@ -105,6 +105,7 @@ arm/broadcom/bcm2835/bcm2835_vcio.c optional soc_brcm_bcm2837 fdt | soc_brcm_bc
arm/broadcom/bcm2835/bcm2835_wdog.c optional soc_brcm_bcm2837 fdt | soc_brcm_bcm2838 fdt
arm/broadcom/bcm2835/bcm2836.c optional soc_brcm_bcm2837 fdt | soc_brcm_bcm2838 fdt
arm/broadcom/bcm2835/bcm283x_dwc_fdt.c optional dwcotg fdt soc_brcm_bcm2837 | dwcotg fdt soc_brcm_bcm2838
+arm/broadcom/bcm2835/bcm2838_pci.c optional soc_brcm_bcm2838 fdt pci
arm/freescale/vybrid/vf_i2c.c optional vf_i2c iicbus SOC_NXP_LS
arm/mv/a37x0_gpio.c optional a37x0_gpio gpio fdt
arm/mv/a37x0_iic.c optional a37x0_iic iicbus fdt
diff --git a/sys/dev/pci/pci_host_generic_fdt.c b/sys/dev/pci/pci_host_generic_fdt.c
index 471ddfca6e2a..57434f526ef1 100644
--- a/sys/dev/pci/pci_host_generic_fdt.c
+++ b/sys/dev/pci/pci_host_generic_fdt.c
@@ -120,7 +120,7 @@ generic_pcie_fdt_probe(device_t dev)
}
int
-pci_host_generic_attach(device_t dev)
+pci_host_generic_setup_fdt(device_t dev)
{
struct generic_pcie_fdt_softc *sc;
phandle_t node;
@@ -149,12 +149,28 @@ pci_host_generic_attach(device_t dev)
/* TODO parse FDT bus ranges */
sc->base.bus_start = 0;
sc->base.bus_end = 0xFF;
+
error = pci_host_generic_core_attach(dev);
if (error != 0)
return (error);
ofw_bus_setup_iinfo(node, &sc->pci_iinfo, sizeof(cell_t));
+ return (0);
+}
+
+int
+pci_host_generic_attach(device_t dev)
+{
+ struct generic_pcie_fdt_softc *sc;
+ int error;
+
+ sc = device_get_softc(dev);
+
+ error = pci_host_generic_setup_fdt(dev);
+ if (error != 0)
+ return (error);
+
device_add_child(dev, "pci", -1);
return (bus_generic_attach(dev));
}
diff --git a/sys/dev/pci/pci_host_generic_fdt.h b/sys/dev/pci/pci_host_generic_fdt.h
index fba3ddcd7f84..5c3e18e9ad35 100644
--- a/sys/dev/pci/pci_host_generic_fdt.h
+++ b/sys/dev/pci/pci_host_generic_fdt.h
@@ -43,6 +43,7 @@ DECLARE_CLASS(generic_pcie_fdt_driver);
struct resource *pci_host_generic_alloc_resource(device_t,
device_t, int, int *, rman_res_t, rman_res_t, rman_res_t, u_int);
+int pci_host_generic_setup_fdt(device_t);
int pci_host_generic_attach(device_t);
int generic_pcie_get_id(device_t, device_t, enum pci_id_type, uintptr_t *);