summaryrefslogtreecommitdiff
path: root/drivers/bcma/driver_pci_host.c
diff options
context:
space:
mode:
authorAndré Fabian Silva Delgado <emulatorman@parabola.nu>2015-08-05 17:04:01 -0300
committerAndré Fabian Silva Delgado <emulatorman@parabola.nu>2015-08-05 17:04:01 -0300
commit57f0f512b273f60d52568b8c6b77e17f5636edc0 (patch)
tree5e910f0e82173f4ef4f51111366a3f1299037a7b /drivers/bcma/driver_pci_host.c
Initial import
Diffstat (limited to 'drivers/bcma/driver_pci_host.c')
-rw-r--r--drivers/bcma/driver_pci_host.c623
1 files changed, 623 insertions, 0 deletions
diff --git a/drivers/bcma/driver_pci_host.c b/drivers/bcma/driver_pci_host.c
new file mode 100644
index 000000000..c42cec7c7
--- /dev/null
+++ b/drivers/bcma/driver_pci_host.c
@@ -0,0 +1,623 @@
+/*
+ * Broadcom specific AMBA
+ * PCI Core in hostmode
+ *
+ * Copyright 2005 - 2011, Broadcom Corporation
+ * Copyright 2006, 2007, Michael Buesch <m@bues.ch>
+ * Copyright 2011, 2012, Hauke Mehrtens <hauke@hauke-m.de>
+ *
+ * Licensed under the GNU/GPL. See COPYING for details.
+ */
+
+#include "bcma_private.h"
+#include <linux/pci.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/bcma/bcma.h>
+#include <asm/paccess.h>
+
+/* Probe a 32bit value on the bus and catch bus exceptions.
+ * Returns nonzero on a bus exception.
+ * This is MIPS specific */
+#define mips_busprobe32(val, addr) get_dbe((val), ((u32 *)(addr)))
+
+/* Assume one-hot slot wiring */
+#define BCMA_PCI_SLOT_MAX 16
+#define PCI_CONFIG_SPACE_SIZE 256
+
+bool bcma_core_pci_is_in_hostmode(struct bcma_drv_pci *pc)
+{
+ struct bcma_bus *bus = pc->core->bus;
+ u16 chipid_top;
+ u32 tmp;
+
+ chipid_top = (bus->chipinfo.id & 0xFF00);
+ if (chipid_top != 0x4700 &&
+ chipid_top != 0x5300)
+ return false;
+
+ bcma_core_enable(pc->core, 0);
+
+ return !mips_busprobe32(tmp, pc->core->io_addr);
+}
+
+static u32 bcma_pcie_read_config(struct bcma_drv_pci *pc, u32 address)
+{
+ pcicore_write32(pc, BCMA_CORE_PCI_CONFIG_ADDR, address);
+ pcicore_read32(pc, BCMA_CORE_PCI_CONFIG_ADDR);
+ return pcicore_read32(pc, BCMA_CORE_PCI_CONFIG_DATA);
+}
+
+static void bcma_pcie_write_config(struct bcma_drv_pci *pc, u32 address,
+ u32 data)
+{
+ pcicore_write32(pc, BCMA_CORE_PCI_CONFIG_ADDR, address);
+ pcicore_read32(pc, BCMA_CORE_PCI_CONFIG_ADDR);
+ pcicore_write32(pc, BCMA_CORE_PCI_CONFIG_DATA, data);
+}
+
+static u32 bcma_get_cfgspace_addr(struct bcma_drv_pci *pc, unsigned int dev,
+ unsigned int func, unsigned int off)
+{
+ u32 addr = 0;
+
+ /* Issue config commands only when the data link is up (atleast
+ * one external pcie device is present).
+ */
+ if (dev >= 2 || !(bcma_pcie_read(pc, BCMA_CORE_PCI_DLLP_LSREG)
+ & BCMA_CORE_PCI_DLLP_LSREG_LINKUP))
+ goto out;
+
+ /* Type 0 transaction */
+ /* Slide the PCI window to the appropriate slot */
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI1, BCMA_CORE_PCI_SBTOPCI_CFG0);
+ /* Calculate the address */
+ addr = pc->host_controller->host_cfg_addr;
+ addr |= (dev << BCMA_CORE_PCI_CFG_SLOT_SHIFT);
+ addr |= (func << BCMA_CORE_PCI_CFG_FUN_SHIFT);
+ addr |= (off & ~3);
+
+out:
+ return addr;
+}
+
+static int bcma_extpci_read_config(struct bcma_drv_pci *pc, unsigned int dev,
+ unsigned int func, unsigned int off,
+ void *buf, int len)
+{
+ int err = -EINVAL;
+ u32 addr, val;
+ void __iomem *mmio = 0;
+
+ WARN_ON(!pc->hostmode);
+ if (unlikely(len != 1 && len != 2 && len != 4))
+ goto out;
+ if (dev == 0) {
+ /* we support only two functions on device 0 */
+ if (func > 1)
+ goto out;
+
+ /* accesses to config registers with offsets >= 256
+ * requires indirect access.
+ */
+ if (off >= PCI_CONFIG_SPACE_SIZE) {
+ addr = (func << 12);
+ addr |= (off & 0x0FFC);
+ val = bcma_pcie_read_config(pc, addr);
+ } else {
+ addr = BCMA_CORE_PCI_PCICFG0;
+ addr |= (func << 8);
+ addr |= (off & 0xFC);
+ val = pcicore_read32(pc, addr);
+ }
+ } else {
+ addr = bcma_get_cfgspace_addr(pc, dev, func, off);
+ if (unlikely(!addr))
+ goto out;
+ err = -ENOMEM;
+ mmio = ioremap_nocache(addr, sizeof(val));
+ if (!mmio)
+ goto out;
+
+ if (mips_busprobe32(val, mmio)) {
+ val = 0xFFFFFFFF;
+ goto unmap;
+ }
+ }
+ val >>= (8 * (off & 3));
+
+ switch (len) {
+ case 1:
+ *((u8 *)buf) = (u8)val;
+ break;
+ case 2:
+ *((u16 *)buf) = (u16)val;
+ break;
+ case 4:
+ *((u32 *)buf) = (u32)val;
+ break;
+ }
+ err = 0;
+unmap:
+ if (mmio)
+ iounmap(mmio);
+out:
+ return err;
+}
+
+static int bcma_extpci_write_config(struct bcma_drv_pci *pc, unsigned int dev,
+ unsigned int func, unsigned int off,
+ const void *buf, int len)
+{
+ int err = -EINVAL;
+ u32 addr, val;
+ void __iomem *mmio = 0;
+ u16 chipid = pc->core->bus->chipinfo.id;
+
+ WARN_ON(!pc->hostmode);
+ if (unlikely(len != 1 && len != 2 && len != 4))
+ goto out;
+ if (dev == 0) {
+ /* we support only two functions on device 0 */
+ if (func > 1)
+ goto out;
+
+ /* accesses to config registers with offsets >= 256
+ * requires indirect access.
+ */
+ if (off >= PCI_CONFIG_SPACE_SIZE) {
+ addr = (func << 12);
+ addr |= (off & 0x0FFC);
+ val = bcma_pcie_read_config(pc, addr);
+ } else {
+ addr = BCMA_CORE_PCI_PCICFG0;
+ addr |= (func << 8);
+ addr |= (off & 0xFC);
+ val = pcicore_read32(pc, addr);
+ }
+ } else {
+ addr = bcma_get_cfgspace_addr(pc, dev, func, off);
+ if (unlikely(!addr))
+ goto out;
+ err = -ENOMEM;
+ mmio = ioremap_nocache(addr, sizeof(val));
+ if (!mmio)
+ goto out;
+
+ if (mips_busprobe32(val, mmio)) {
+ val = 0xFFFFFFFF;
+ goto unmap;
+ }
+ }
+
+ switch (len) {
+ case 1:
+ val &= ~(0xFF << (8 * (off & 3)));
+ val |= *((const u8 *)buf) << (8 * (off & 3));
+ break;
+ case 2:
+ val &= ~(0xFFFF << (8 * (off & 3)));
+ val |= *((const u16 *)buf) << (8 * (off & 3));
+ break;
+ case 4:
+ val = *((const u32 *)buf);
+ break;
+ }
+ if (dev == 0) {
+ /* accesses to config registers with offsets >= 256
+ * requires indirect access.
+ */
+ if (off >= PCI_CONFIG_SPACE_SIZE)
+ bcma_pcie_write_config(pc, addr, val);
+ else
+ pcicore_write32(pc, addr, val);
+ } else {
+ writel(val, mmio);
+
+ if (chipid == BCMA_CHIP_ID_BCM4716 ||
+ chipid == BCMA_CHIP_ID_BCM4748)
+ readl(mmio);
+ }
+
+ err = 0;
+unmap:
+ if (mmio)
+ iounmap(mmio);
+out:
+ return err;
+}
+
+static int bcma_core_pci_hostmode_read_config(struct pci_bus *bus,
+ unsigned int devfn,
+ int reg, int size, u32 *val)
+{
+ unsigned long flags;
+ int err;
+ struct bcma_drv_pci *pc;
+ struct bcma_drv_pci_host *pc_host;
+
+ pc_host = container_of(bus->ops, struct bcma_drv_pci_host, pci_ops);
+ pc = pc_host->pdev;
+
+ spin_lock_irqsave(&pc_host->cfgspace_lock, flags);
+ err = bcma_extpci_read_config(pc, PCI_SLOT(devfn),
+ PCI_FUNC(devfn), reg, val, size);
+ spin_unlock_irqrestore(&pc_host->cfgspace_lock, flags);
+
+ return err ? PCIBIOS_DEVICE_NOT_FOUND : PCIBIOS_SUCCESSFUL;
+}
+
+static int bcma_core_pci_hostmode_write_config(struct pci_bus *bus,
+ unsigned int devfn,
+ int reg, int size, u32 val)
+{
+ unsigned long flags;
+ int err;
+ struct bcma_drv_pci *pc;
+ struct bcma_drv_pci_host *pc_host;
+
+ pc_host = container_of(bus->ops, struct bcma_drv_pci_host, pci_ops);
+ pc = pc_host->pdev;
+
+ spin_lock_irqsave(&pc_host->cfgspace_lock, flags);
+ err = bcma_extpci_write_config(pc, PCI_SLOT(devfn),
+ PCI_FUNC(devfn), reg, &val, size);
+ spin_unlock_irqrestore(&pc_host->cfgspace_lock, flags);
+
+ return err ? PCIBIOS_DEVICE_NOT_FOUND : PCIBIOS_SUCCESSFUL;
+}
+
+/* return cap_offset if requested capability exists in the PCI config space */
+static u8 bcma_find_pci_capability(struct bcma_drv_pci *pc, unsigned int dev,
+ unsigned int func, u8 req_cap_id,
+ unsigned char *buf, u32 *buflen)
+{
+ u8 cap_id;
+ u8 cap_ptr = 0;
+ u32 bufsize;
+ u8 byte_val;
+
+ /* check for Header type 0 */
+ bcma_extpci_read_config(pc, dev, func, PCI_HEADER_TYPE, &byte_val,
+ sizeof(u8));
+ if ((byte_val & 0x7F) != PCI_HEADER_TYPE_NORMAL)
+ return cap_ptr;
+
+ /* check if the capability pointer field exists */
+ bcma_extpci_read_config(pc, dev, func, PCI_STATUS, &byte_val,
+ sizeof(u8));
+ if (!(byte_val & PCI_STATUS_CAP_LIST))
+ return cap_ptr;
+
+ /* check if the capability pointer is 0x00 */
+ bcma_extpci_read_config(pc, dev, func, PCI_CAPABILITY_LIST, &cap_ptr,
+ sizeof(u8));
+ if (cap_ptr == 0x00)
+ return cap_ptr;
+
+ /* loop thr'u the capability list and see if the requested capabilty
+ * exists */
+ bcma_extpci_read_config(pc, dev, func, cap_ptr, &cap_id, sizeof(u8));
+ while (cap_id != req_cap_id) {
+ bcma_extpci_read_config(pc, dev, func, cap_ptr + 1, &cap_ptr,
+ sizeof(u8));
+ if (cap_ptr == 0x00)
+ return cap_ptr;
+ bcma_extpci_read_config(pc, dev, func, cap_ptr, &cap_id,
+ sizeof(u8));
+ }
+
+ /* found the caller requested capability */
+ if ((buf != NULL) && (buflen != NULL)) {
+ u8 cap_data;
+
+ bufsize = *buflen;
+ if (!bufsize)
+ return cap_ptr;
+
+ *buflen = 0;
+
+ /* copy the cpability data excluding cap ID and next ptr */
+ cap_data = cap_ptr + 2;
+ if ((bufsize + cap_data) > PCI_CONFIG_SPACE_SIZE)
+ bufsize = PCI_CONFIG_SPACE_SIZE - cap_data;
+ *buflen = bufsize;
+ while (bufsize--) {
+ bcma_extpci_read_config(pc, dev, func, cap_data, buf,
+ sizeof(u8));
+ cap_data++;
+ buf++;
+ }
+ }
+
+ return cap_ptr;
+}
+
+/* If the root port is capable of returning Config Request
+ * Retry Status (CRS) Completion Status to software then
+ * enable the feature.
+ */
+static void bcma_core_pci_enable_crs(struct bcma_drv_pci *pc)
+{
+ struct bcma_bus *bus = pc->core->bus;
+ u8 cap_ptr, root_ctrl, root_cap, dev;
+ u16 val16;
+ int i;
+
+ cap_ptr = bcma_find_pci_capability(pc, 0, 0, PCI_CAP_ID_EXP, NULL,
+ NULL);
+ root_cap = cap_ptr + PCI_EXP_RTCAP;
+ bcma_extpci_read_config(pc, 0, 0, root_cap, &val16, sizeof(u16));
+ if (val16 & BCMA_CORE_PCI_RC_CRS_VISIBILITY) {
+ /* Enable CRS software visibility */
+ root_ctrl = cap_ptr + PCI_EXP_RTCTL;
+ val16 = PCI_EXP_RTCTL_CRSSVE;
+ bcma_extpci_read_config(pc, 0, 0, root_ctrl, &val16,
+ sizeof(u16));
+
+ /* Initiate a configuration request to read the vendor id
+ * field of the device function's config space header after
+ * 100 ms wait time from the end of Reset. If the device is
+ * not done with its internal initialization, it must at
+ * least return a completion TLP, with a completion status
+ * of "Configuration Request Retry Status (CRS)". The root
+ * complex must complete the request to the host by returning
+ * a read-data value of 0001h for the Vendor ID field and
+ * all 1s for any additional bytes included in the request.
+ * Poll using the config reads for max wait time of 1 sec or
+ * until we receive the successful completion status. Repeat
+ * the procedure for all the devices.
+ */
+ for (dev = 1; dev < BCMA_PCI_SLOT_MAX; dev++) {
+ for (i = 0; i < 100000; i++) {
+ bcma_extpci_read_config(pc, dev, 0,
+ PCI_VENDOR_ID, &val16,
+ sizeof(val16));
+ if (val16 != 0x1)
+ break;
+ udelay(10);
+ }
+ if (val16 == 0x1)
+ bcma_err(bus, "PCI: Broken device in slot %d\n",
+ dev);
+ }
+ }
+}
+
+void bcma_core_pci_hostmode_init(struct bcma_drv_pci *pc)
+{
+ struct bcma_bus *bus = pc->core->bus;
+ struct bcma_drv_pci_host *pc_host;
+ u32 tmp;
+ u32 pci_membase_1G;
+ unsigned long io_map_base;
+
+ bcma_info(bus, "PCIEcore in host mode found\n");
+
+ if (bus->sprom.boardflags_lo & BCMA_CORE_PCI_BFL_NOPCI) {
+ bcma_info(bus, "This PCIE core is disabled and not working\n");
+ return;
+ }
+
+ pc_host = kzalloc(sizeof(*pc_host), GFP_KERNEL);
+ if (!pc_host) {
+ bcma_err(bus, "can not allocate memory");
+ return;
+ }
+
+ spin_lock_init(&pc_host->cfgspace_lock);
+
+ pc->host_controller = pc_host;
+ pc_host->pci_controller.io_resource = &pc_host->io_resource;
+ pc_host->pci_controller.mem_resource = &pc_host->mem_resource;
+ pc_host->pci_controller.pci_ops = &pc_host->pci_ops;
+ pc_host->pdev = pc;
+
+ pci_membase_1G = BCMA_SOC_PCI_DMA;
+ pc_host->host_cfg_addr = BCMA_SOC_PCI_CFG;
+
+ pc_host->pci_ops.read = bcma_core_pci_hostmode_read_config;
+ pc_host->pci_ops.write = bcma_core_pci_hostmode_write_config;
+
+ pc_host->mem_resource.name = "BCMA PCIcore external memory",
+ pc_host->mem_resource.start = BCMA_SOC_PCI_DMA;
+ pc_host->mem_resource.end = BCMA_SOC_PCI_DMA + BCMA_SOC_PCI_DMA_SZ - 1;
+ pc_host->mem_resource.flags = IORESOURCE_MEM | IORESOURCE_PCI_FIXED;
+
+ pc_host->io_resource.name = "BCMA PCIcore external I/O",
+ pc_host->io_resource.start = 0x100;
+ pc_host->io_resource.end = 0x7FF;
+ pc_host->io_resource.flags = IORESOURCE_IO | IORESOURCE_PCI_FIXED;
+
+ /* Reset RC */
+ usleep_range(3000, 5000);
+ pcicore_write32(pc, BCMA_CORE_PCI_CTL, BCMA_CORE_PCI_CTL_RST_OE);
+ msleep(50);
+ pcicore_write32(pc, BCMA_CORE_PCI_CTL, BCMA_CORE_PCI_CTL_RST |
+ BCMA_CORE_PCI_CTL_RST_OE);
+
+ /* 64 MB I/O access window. On 4716, use
+ * sbtopcie0 to access the device registers. We
+ * can't use address match 2 (1 GB window) region
+ * as mips can't generate 64-bit address on the
+ * backplane.
+ */
+ if (bus->chipinfo.id == BCMA_CHIP_ID_BCM4716 ||
+ bus->chipinfo.id == BCMA_CHIP_ID_BCM4748) {
+ pc_host->mem_resource.start = BCMA_SOC_PCI_MEM;
+ pc_host->mem_resource.end = BCMA_SOC_PCI_MEM +
+ BCMA_SOC_PCI_MEM_SZ - 1;
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI0,
+ BCMA_CORE_PCI_SBTOPCI_MEM | BCMA_SOC_PCI_MEM);
+ } else if (bus->chipinfo.id == BCMA_CHIP_ID_BCM4706) {
+ tmp = BCMA_CORE_PCI_SBTOPCI_MEM;
+ tmp |= BCMA_CORE_PCI_SBTOPCI_PREF;
+ tmp |= BCMA_CORE_PCI_SBTOPCI_BURST;
+ if (pc->core->core_unit == 0) {
+ pc_host->mem_resource.start = BCMA_SOC_PCI_MEM;
+ pc_host->mem_resource.end = BCMA_SOC_PCI_MEM +
+ BCMA_SOC_PCI_MEM_SZ - 1;
+ pc_host->io_resource.start = 0x100;
+ pc_host->io_resource.end = 0x47F;
+ pci_membase_1G = BCMA_SOC_PCIE_DMA_H32;
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI0,
+ tmp | BCMA_SOC_PCI_MEM);
+ } else if (pc->core->core_unit == 1) {
+ pc_host->mem_resource.start = BCMA_SOC_PCI1_MEM;
+ pc_host->mem_resource.end = BCMA_SOC_PCI1_MEM +
+ BCMA_SOC_PCI_MEM_SZ - 1;
+ pc_host->io_resource.start = 0x480;
+ pc_host->io_resource.end = 0x7FF;
+ pci_membase_1G = BCMA_SOC_PCIE1_DMA_H32;
+ pc_host->host_cfg_addr = BCMA_SOC_PCI1_CFG;
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI0,
+ tmp | BCMA_SOC_PCI1_MEM);
+ }
+ } else
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI0,
+ BCMA_CORE_PCI_SBTOPCI_IO);
+
+ /* 64 MB configuration access window */
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI1, BCMA_CORE_PCI_SBTOPCI_CFG0);
+
+ /* 1 GB memory access window */
+ pcicore_write32(pc, BCMA_CORE_PCI_SBTOPCI2,
+ BCMA_CORE_PCI_SBTOPCI_MEM | pci_membase_1G);
+
+
+ /* As per PCI Express Base Spec 1.1 we need to wait for
+ * at least 100 ms from the end of a reset (cold/warm/hot)
+ * before issuing configuration requests to PCI Express
+ * devices.
+ */
+ msleep(100);
+
+ bcma_core_pci_enable_crs(pc);
+
+ if (bus->chipinfo.id == BCMA_CHIP_ID_BCM4706 ||
+ bus->chipinfo.id == BCMA_CHIP_ID_BCM4716) {
+ u16 val16;
+ bcma_extpci_read_config(pc, 0, 0, BCMA_CORE_PCI_CFG_DEVCTRL,
+ &val16, sizeof(val16));
+ val16 |= (2 << 5); /* Max payload size of 512 */
+ val16 |= (2 << 12); /* MRRS 512 */
+ bcma_extpci_write_config(pc, 0, 0, BCMA_CORE_PCI_CFG_DEVCTRL,
+ &val16, sizeof(val16));
+ }
+
+ /* Enable PCI bridge BAR0 memory & master access */
+ tmp = PCI_COMMAND_MASTER | PCI_COMMAND_MEMORY;
+ bcma_extpci_write_config(pc, 0, 0, PCI_COMMAND, &tmp, sizeof(tmp));
+
+ /* Enable PCI interrupts */
+ pcicore_write32(pc, BCMA_CORE_PCI_IMASK, BCMA_CORE_PCI_IMASK_INTA);
+
+ /* Ok, ready to run, register it to the system.
+ * The following needs change, if we want to port hostmode
+ * to non-MIPS platform. */
+ io_map_base = (unsigned long)ioremap_nocache(pc_host->mem_resource.start,
+ resource_size(&pc_host->mem_resource));
+ pc_host->pci_controller.io_map_base = io_map_base;
+ set_io_port_base(pc_host->pci_controller.io_map_base);
+ /* Give some time to the PCI controller to configure itself with the new
+ * values. Not waiting at this point causes crashes of the machine. */
+ usleep_range(10000, 15000);
+ register_pci_controller(&pc_host->pci_controller);
+ return;
+}
+
+/* Early PCI fixup for a device on the PCI-core bridge. */
+static void bcma_core_pci_fixup_pcibridge(struct pci_dev *dev)
+{
+ if (dev->bus->ops->read != bcma_core_pci_hostmode_read_config) {
+ /* This is not a device on the PCI-core bridge. */
+ return;
+ }
+ if (PCI_SLOT(dev->devfn) != 0)
+ return;
+
+ pr_info("PCI: Fixing up bridge %s\n", pci_name(dev));
+
+ /* Enable PCI bridge bus mastering and memory space */
+ pci_set_master(dev);
+ if (pcibios_enable_device(dev, ~0) < 0) {
+ pr_err("PCI: BCMA bridge enable failed\n");
+ return;
+ }
+
+ /* Enable PCI bridge BAR1 prefetch and burst */
+ pci_write_config_dword(dev, BCMA_PCI_BAR1_CONTROL, 3);
+}
+DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, bcma_core_pci_fixup_pcibridge);
+
+/* Early PCI fixup for all PCI-cores to set the correct memory address. */
+static void bcma_core_pci_fixup_addresses(struct pci_dev *dev)
+{
+ struct resource *res;
+ int pos, err;
+
+ if (dev->bus->ops->read != bcma_core_pci_hostmode_read_config) {
+ /* This is not a device on the PCI-core bridge. */
+ return;
+ }
+ if (PCI_SLOT(dev->devfn) == 0)
+ return;
+
+ pr_info("PCI: Fixing up addresses %s\n", pci_name(dev));
+
+ for (pos = 0; pos < 6; pos++) {
+ res = &dev->resource[pos];
+ if (res->flags & (IORESOURCE_IO | IORESOURCE_MEM)) {
+ err = pci_assign_resource(dev, pos);
+ if (err)
+ pr_err("PCI: Problem fixing up the addresses on %s\n",
+ pci_name(dev));
+ }
+ }
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, bcma_core_pci_fixup_addresses);
+
+/* This function is called when doing a pci_enable_device().
+ * We must first check if the device is a device on the PCI-core bridge. */
+int bcma_core_pci_plat_dev_init(struct pci_dev *dev)
+{
+ struct bcma_drv_pci_host *pc_host;
+ int readrq;
+
+ if (dev->bus->ops->read != bcma_core_pci_hostmode_read_config) {
+ /* This is not a device on the PCI-core bridge. */
+ return -ENODEV;
+ }
+ pc_host = container_of(dev->bus->ops, struct bcma_drv_pci_host,
+ pci_ops);
+
+ pr_info("PCI: Fixing up device %s\n", pci_name(dev));
+
+ /* Fix up interrupt lines */
+ dev->irq = bcma_core_irq(pc_host->pdev->core, 0);
+ pci_write_config_byte(dev, PCI_INTERRUPT_LINE, dev->irq);
+
+ readrq = pcie_get_readrq(dev);
+ if (readrq > 128) {
+ pr_info("change PCIe max read request size from %i to 128\n", readrq);
+ pcie_set_readrq(dev, 128);
+ }
+ return 0;
+}
+EXPORT_SYMBOL(bcma_core_pci_plat_dev_init);
+
+/* PCI device IRQ mapping. */
+int bcma_core_pci_pcibios_map_irq(const struct pci_dev *dev)
+{
+ struct bcma_drv_pci_host *pc_host;
+
+ if (dev->bus->ops->read != bcma_core_pci_hostmode_read_config) {
+ /* This is not a device on the PCI-core bridge. */
+ return -ENODEV;
+ }
+
+ pc_host = container_of(dev->bus->ops, struct bcma_drv_pci_host,
+ pci_ops);
+ return bcma_core_irq(pc_host->pdev->core, 0);
+}
+EXPORT_SYMBOL(bcma_core_pci_pcibios_map_irq);