summaryrefslogtreecommitdiff
path: root/drivers/scsi/bfa/bfad.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/scsi/bfa/bfad.c
Initial import
Diffstat (limited to 'drivers/scsi/bfa/bfad.c')
-rw-r--r--drivers/scsi/bfa/bfad.c1820
1 files changed, 1820 insertions, 0 deletions
diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c
new file mode 100644
index 000000000..f95adcadb
--- /dev/null
+++ b/drivers/scsi/bfa/bfad.c
@@ -0,0 +1,1820 @@
+/*
+ * Copyright (c) 2005-2010 Brocade Communications Systems, Inc.
+ * All rights reserved
+ * www.brocade.com
+ *
+ * Linux driver for Brocade Fibre Channel Host Bus Adapter.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License (GPL) Version 2 as
+ * published by the Free Software Foundation
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ */
+
+/*
+ * bfad.c Linux driver PCI interface module.
+ */
+#include <linux/module.h>
+#include <linux/kthread.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/pci.h>
+#include <linux/firmware.h>
+#include <asm/uaccess.h>
+#include <asm/fcntl.h>
+
+#include "bfad_drv.h"
+#include "bfad_im.h"
+#include "bfa_fcs.h"
+#include "bfa_defs.h"
+#include "bfa.h"
+
+BFA_TRC_FILE(LDRV, BFAD);
+DEFINE_MUTEX(bfad_mutex);
+LIST_HEAD(bfad_list);
+
+static int bfad_inst;
+static int num_sgpgs_parm;
+int supported_fc4s;
+char *host_name, *os_name, *os_patch;
+int num_rports, num_ios, num_tms;
+int num_fcxps, num_ufbufs;
+int reqq_size, rspq_size, num_sgpgs;
+int rport_del_timeout = BFA_FCS_RPORT_DEF_DEL_TIMEOUT;
+int bfa_lun_queue_depth = BFAD_LUN_QUEUE_DEPTH;
+int bfa_io_max_sge = BFAD_IO_MAX_SGE;
+int bfa_log_level = 3; /* WARNING log level */
+int ioc_auto_recover = BFA_TRUE;
+int bfa_linkup_delay = -1;
+int fdmi_enable = BFA_TRUE;
+int pcie_max_read_reqsz;
+int bfa_debugfs_enable = 1;
+int msix_disable_cb = 0, msix_disable_ct = 0;
+int max_xfer_size = BFAD_MAX_SECTORS >> 1;
+int max_rport_logins = BFA_FCS_MAX_RPORT_LOGINS;
+
+/* Firmware releated */
+u32 bfi_image_cb_size, bfi_image_ct_size, bfi_image_ct2_size;
+u32 *bfi_image_cb, *bfi_image_ct, *bfi_image_ct2;
+
+#define BFAD_FW_FILE_CB "/*(DEBLOBBED)*/"
+#define BFAD_FW_FILE_CT "/*(DEBLOBBED)*/"
+#define BFAD_FW_FILE_CT2 "/*(DEBLOBBED)*/"
+
+static u32 *bfad_load_fwimg(struct pci_dev *pdev);
+static void bfad_free_fwimg(void);
+static void bfad_read_firmware(struct pci_dev *pdev, u32 **bfi_image,
+ u32 *bfi_image_size, char *fw_name);
+
+static const char *msix_name_ct[] = {
+ "ctrl",
+ "cpe0", "cpe1", "cpe2", "cpe3",
+ "rme0", "rme1", "rme2", "rme3" };
+
+static const char *msix_name_cb[] = {
+ "cpe0", "cpe1", "cpe2", "cpe3",
+ "rme0", "rme1", "rme2", "rme3",
+ "eemc", "elpu0", "elpu1", "epss", "mlpu" };
+
+/*(DEBLOBBED)*/
+
+module_param(os_name, charp, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(os_name, "OS name of the hba host machine");
+module_param(os_patch, charp, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(os_patch, "OS patch level of the hba host machine");
+module_param(host_name, charp, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(host_name, "Hostname of the hba host machine");
+module_param(num_rports, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(num_rports, "Max number of rports supported per port "
+ "(physical/logical), default=1024");
+module_param(num_ios, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(num_ios, "Max number of ioim requests, default=2000");
+module_param(num_tms, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(num_tms, "Max number of task im requests, default=128");
+module_param(num_fcxps, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(num_fcxps, "Max number of fcxp requests, default=64");
+module_param(num_ufbufs, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(num_ufbufs, "Max number of unsolicited frame "
+ "buffers, default=64");
+module_param(reqq_size, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(reqq_size, "Max number of request queue elements, "
+ "default=256");
+module_param(rspq_size, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(rspq_size, "Max number of response queue elements, "
+ "default=64");
+module_param(num_sgpgs, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(num_sgpgs, "Number of scatter/gather pages, default=2048");
+module_param(rport_del_timeout, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(rport_del_timeout, "Rport delete timeout, default=90 secs, "
+ "Range[>0]");
+module_param(bfa_lun_queue_depth, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(bfa_lun_queue_depth, "Lun queue depth, default=32, Range[>0]");
+module_param(bfa_io_max_sge, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(bfa_io_max_sge, "Max io scatter/gather elements, default=255");
+module_param(bfa_log_level, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(bfa_log_level, "Driver log level, default=3, "
+ "Range[Critical:1|Error:2|Warning:3|Info:4]");
+module_param(ioc_auto_recover, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(ioc_auto_recover, "IOC auto recovery, default=1, "
+ "Range[off:0|on:1]");
+module_param(bfa_linkup_delay, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(bfa_linkup_delay, "Link up delay, default=30 secs for "
+ "boot port. Otherwise 10 secs in RHEL4 & 0 for "
+ "[RHEL5, SLES10, ESX40] Range[>0]");
+module_param(msix_disable_cb, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(msix_disable_cb, "Disable Message Signaled Interrupts "
+ "for Brocade-415/425/815/825 cards, default=0, "
+ " Range[false:0|true:1]");
+module_param(msix_disable_ct, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(msix_disable_ct, "Disable Message Signaled Interrupts "
+ "if possible for Brocade-1010/1020/804/1007/902/1741 "
+ "cards, default=0, Range[false:0|true:1]");
+module_param(fdmi_enable, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(fdmi_enable, "Enables fdmi registration, default=1, "
+ "Range[false:0|true:1]");
+module_param(pcie_max_read_reqsz, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(pcie_max_read_reqsz, "PCIe max read request size, default=0 "
+ "(use system setting), Range[128|256|512|1024|2048|4096]");
+module_param(bfa_debugfs_enable, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(bfa_debugfs_enable, "Enables debugfs feature, default=1,"
+ " Range[false:0|true:1]");
+module_param(max_xfer_size, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(max_xfer_size, "default=32MB,"
+ " Range[64k|128k|256k|512k|1024k|2048k]");
+module_param(max_rport_logins, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(max_rport_logins, "Max number of logins to initiator and target rports on a port (physical/logical), default=1024");
+
+static void
+bfad_sm_uninit(struct bfad_s *bfad, enum bfad_sm_event event);
+static void
+bfad_sm_created(struct bfad_s *bfad, enum bfad_sm_event event);
+static void
+bfad_sm_initializing(struct bfad_s *bfad, enum bfad_sm_event event);
+static void
+bfad_sm_operational(struct bfad_s *bfad, enum bfad_sm_event event);
+static void
+bfad_sm_stopping(struct bfad_s *bfad, enum bfad_sm_event event);
+static void
+bfad_sm_failed(struct bfad_s *bfad, enum bfad_sm_event event);
+static void
+bfad_sm_fcs_exit(struct bfad_s *bfad, enum bfad_sm_event event);
+
+/*
+ * Beginning state for the driver instance, awaiting the pci_probe event
+ */
+static void
+bfad_sm_uninit(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_CREATE:
+ bfa_sm_set_state(bfad, bfad_sm_created);
+ bfad->bfad_tsk = kthread_create(bfad_worker, (void *) bfad,
+ "%s", "bfad_worker");
+ if (IS_ERR(bfad->bfad_tsk)) {
+ printk(KERN_INFO "bfad[%d]: Kernel thread "
+ "creation failed!\n", bfad->inst_no);
+ bfa_sm_send_event(bfad, BFAD_E_KTHREAD_CREATE_FAILED);
+ }
+ bfa_sm_send_event(bfad, BFAD_E_INIT);
+ break;
+
+ case BFAD_E_STOP:
+ /* Ignore stop; already in uninit */
+ break;
+
+ default:
+ bfa_sm_fault(bfad, event);
+ }
+}
+
+/*
+ * Driver Instance is created, awaiting event INIT to initialize the bfad
+ */
+static void
+bfad_sm_created(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ unsigned long flags;
+ bfa_status_t ret;
+
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_INIT:
+ bfa_sm_set_state(bfad, bfad_sm_initializing);
+
+ init_completion(&bfad->comp);
+
+ /* Enable Interrupt and wait bfa_init completion */
+ if (bfad_setup_intr(bfad)) {
+ printk(KERN_WARNING "bfad%d: bfad_setup_intr failed\n",
+ bfad->inst_no);
+ bfa_sm_send_event(bfad, BFAD_E_INIT_FAILED);
+ break;
+ }
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_iocfc_init(&bfad->bfa);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ /* Set up interrupt handler for each vectors */
+ if ((bfad->bfad_flags & BFAD_MSIX_ON) &&
+ bfad_install_msix_handler(bfad)) {
+ printk(KERN_WARNING "%s: install_msix failed, bfad%d\n",
+ __func__, bfad->inst_no);
+ }
+
+ bfad_init_timer(bfad);
+
+ wait_for_completion(&bfad->comp);
+
+ if ((bfad->bfad_flags & BFAD_HAL_INIT_DONE)) {
+ bfa_sm_send_event(bfad, BFAD_E_INIT_SUCCESS);
+ } else {
+ printk(KERN_WARNING
+ "bfa %s: bfa init failed\n",
+ bfad->pci_name);
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_fcs_init(&bfad->bfa_fcs);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ ret = bfad_cfg_pport(bfad, BFA_LPORT_ROLE_FCP_IM);
+ if (ret != BFA_STATUS_OK) {
+ init_completion(&bfad->comp);
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->pport.flags |= BFAD_PORT_DELETE;
+ bfa_fcs_exit(&bfad->bfa_fcs);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ wait_for_completion(&bfad->comp);
+
+ bfa_sm_send_event(bfad, BFAD_E_INIT_FAILED);
+ break;
+ }
+ bfad->bfad_flags |= BFAD_HAL_INIT_FAIL;
+ bfa_sm_send_event(bfad, BFAD_E_HAL_INIT_FAILED);
+ }
+
+ break;
+
+ case BFAD_E_KTHREAD_CREATE_FAILED:
+ bfa_sm_set_state(bfad, bfad_sm_uninit);
+ break;
+
+ default:
+ bfa_sm_fault(bfad, event);
+ }
+}
+
+static void
+bfad_sm_initializing(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ int retval;
+ unsigned long flags;
+
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_INIT_SUCCESS:
+ kthread_stop(bfad->bfad_tsk);
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_tsk = NULL;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ retval = bfad_start_ops(bfad);
+ if (retval != BFA_STATUS_OK) {
+ bfa_sm_set_state(bfad, bfad_sm_failed);
+ break;
+ }
+ bfa_sm_set_state(bfad, bfad_sm_operational);
+ break;
+
+ case BFAD_E_INIT_FAILED:
+ bfa_sm_set_state(bfad, bfad_sm_uninit);
+ kthread_stop(bfad->bfad_tsk);
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_tsk = NULL;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ break;
+
+ case BFAD_E_HAL_INIT_FAILED:
+ bfa_sm_set_state(bfad, bfad_sm_failed);
+ break;
+ default:
+ bfa_sm_fault(bfad, event);
+ }
+}
+
+static void
+bfad_sm_failed(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ int retval;
+
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_INIT_SUCCESS:
+ retval = bfad_start_ops(bfad);
+ if (retval != BFA_STATUS_OK)
+ break;
+ bfa_sm_set_state(bfad, bfad_sm_operational);
+ break;
+
+ case BFAD_E_STOP:
+ bfa_sm_set_state(bfad, bfad_sm_fcs_exit);
+ bfa_sm_send_event(bfad, BFAD_E_FCS_EXIT_COMP);
+ break;
+
+ case BFAD_E_EXIT_COMP:
+ bfa_sm_set_state(bfad, bfad_sm_uninit);
+ bfad_remove_intr(bfad);
+ del_timer_sync(&bfad->hal_tmo);
+ break;
+
+ default:
+ bfa_sm_fault(bfad, event);
+ }
+}
+
+static void
+bfad_sm_operational(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_STOP:
+ bfa_sm_set_state(bfad, bfad_sm_fcs_exit);
+ bfad_fcs_stop(bfad);
+ break;
+
+ default:
+ bfa_sm_fault(bfad, event);
+ }
+}
+
+static void
+bfad_sm_fcs_exit(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_FCS_EXIT_COMP:
+ bfa_sm_set_state(bfad, bfad_sm_stopping);
+ bfad_stop(bfad);
+ break;
+
+ default:
+ bfa_sm_fault(bfad, event);
+ }
+}
+
+static void
+bfad_sm_stopping(struct bfad_s *bfad, enum bfad_sm_event event)
+{
+ bfa_trc(bfad, event);
+
+ switch (event) {
+ case BFAD_E_EXIT_COMP:
+ bfa_sm_set_state(bfad, bfad_sm_uninit);
+ bfad_remove_intr(bfad);
+ del_timer_sync(&bfad->hal_tmo);
+ bfad_im_probe_undo(bfad);
+ bfad->bfad_flags &= ~BFAD_FC4_PROBE_DONE;
+ bfad_uncfg_pport(bfad);
+ break;
+
+ default:
+ bfa_sm_fault(bfad, event);
+ break;
+ }
+}
+
+/*
+ * BFA callbacks
+ */
+void
+bfad_hcb_comp(void *arg, bfa_status_t status)
+{
+ struct bfad_hal_comp *fcomp = (struct bfad_hal_comp *)arg;
+
+ fcomp->status = status;
+ complete(&fcomp->comp);
+}
+
+/*
+ * bfa_init callback
+ */
+void
+bfa_cb_init(void *drv, bfa_status_t init_status)
+{
+ struct bfad_s *bfad = drv;
+
+ if (init_status == BFA_STATUS_OK) {
+ bfad->bfad_flags |= BFAD_HAL_INIT_DONE;
+
+ /*
+ * If BFAD_HAL_INIT_FAIL flag is set:
+ * Wake up the kernel thread to start
+ * the bfad operations after HAL init done
+ */
+ if ((bfad->bfad_flags & BFAD_HAL_INIT_FAIL)) {
+ bfad->bfad_flags &= ~BFAD_HAL_INIT_FAIL;
+ wake_up_process(bfad->bfad_tsk);
+ }
+ }
+
+ complete(&bfad->comp);
+}
+
+/*
+ * BFA_FCS callbacks
+ */
+struct bfad_port_s *
+bfa_fcb_lport_new(struct bfad_s *bfad, struct bfa_fcs_lport_s *port,
+ enum bfa_lport_role roles, struct bfad_vf_s *vf_drv,
+ struct bfad_vport_s *vp_drv)
+{
+ bfa_status_t rc;
+ struct bfad_port_s *port_drv;
+
+ if (!vp_drv && !vf_drv) {
+ port_drv = &bfad->pport;
+ port_drv->pvb_type = BFAD_PORT_PHYS_BASE;
+ } else if (!vp_drv && vf_drv) {
+ port_drv = &vf_drv->base_port;
+ port_drv->pvb_type = BFAD_PORT_VF_BASE;
+ } else if (vp_drv && !vf_drv) {
+ port_drv = &vp_drv->drv_port;
+ port_drv->pvb_type = BFAD_PORT_PHYS_VPORT;
+ } else {
+ port_drv = &vp_drv->drv_port;
+ port_drv->pvb_type = BFAD_PORT_VF_VPORT;
+ }
+
+ port_drv->fcs_port = port;
+ port_drv->roles = roles;
+
+ if (roles & BFA_LPORT_ROLE_FCP_IM) {
+ rc = bfad_im_port_new(bfad, port_drv);
+ if (rc != BFA_STATUS_OK) {
+ bfad_im_port_delete(bfad, port_drv);
+ port_drv = NULL;
+ }
+ }
+
+ return port_drv;
+}
+
+/*
+ * FCS RPORT alloc callback, after successful PLOGI by FCS
+ */
+bfa_status_t
+bfa_fcb_rport_alloc(struct bfad_s *bfad, struct bfa_fcs_rport_s **rport,
+ struct bfad_rport_s **rport_drv)
+{
+ bfa_status_t rc = BFA_STATUS_OK;
+
+ *rport_drv = kzalloc(sizeof(struct bfad_rport_s), GFP_ATOMIC);
+ if (*rport_drv == NULL) {
+ rc = BFA_STATUS_ENOMEM;
+ goto ext;
+ }
+
+ *rport = &(*rport_drv)->fcs_rport;
+
+ext:
+ return rc;
+}
+
+/*
+ * FCS PBC VPORT Create
+ */
+void
+bfa_fcb_pbc_vport_create(struct bfad_s *bfad, struct bfi_pbc_vport_s pbc_vport)
+{
+
+ struct bfa_lport_cfg_s port_cfg = {0};
+ struct bfad_vport_s *vport;
+ int rc;
+
+ vport = kzalloc(sizeof(struct bfad_vport_s), GFP_ATOMIC);
+ if (!vport) {
+ bfa_trc(bfad, 0);
+ return;
+ }
+
+ vport->drv_port.bfad = bfad;
+ port_cfg.roles = BFA_LPORT_ROLE_FCP_IM;
+ port_cfg.pwwn = pbc_vport.vp_pwwn;
+ port_cfg.nwwn = pbc_vport.vp_nwwn;
+ port_cfg.preboot_vp = BFA_TRUE;
+
+ rc = bfa_fcs_pbc_vport_create(&vport->fcs_vport, &bfad->bfa_fcs, 0,
+ &port_cfg, vport);
+
+ if (rc != BFA_STATUS_OK) {
+ bfa_trc(bfad, 0);
+ return;
+ }
+
+ list_add_tail(&vport->list_entry, &bfad->pbc_vport_list);
+}
+
+void
+bfad_hal_mem_release(struct bfad_s *bfad)
+{
+ struct bfa_meminfo_s *hal_meminfo = &bfad->meminfo;
+ struct bfa_mem_dma_s *dma_info, *dma_elem;
+ struct bfa_mem_kva_s *kva_info, *kva_elem;
+ struct list_head *dm_qe, *km_qe;
+
+ dma_info = &hal_meminfo->dma_info;
+ kva_info = &hal_meminfo->kva_info;
+
+ /* Iterate through the KVA meminfo queue */
+ list_for_each(km_qe, &kva_info->qe) {
+ kva_elem = (struct bfa_mem_kva_s *) km_qe;
+ vfree(kva_elem->kva);
+ }
+
+ /* Iterate through the DMA meminfo queue */
+ list_for_each(dm_qe, &dma_info->qe) {
+ dma_elem = (struct bfa_mem_dma_s *) dm_qe;
+ dma_free_coherent(&bfad->pcidev->dev,
+ dma_elem->mem_len, dma_elem->kva,
+ (dma_addr_t) dma_elem->dma);
+ }
+
+ memset(hal_meminfo, 0, sizeof(struct bfa_meminfo_s));
+}
+
+void
+bfad_update_hal_cfg(struct bfa_iocfc_cfg_s *bfa_cfg)
+{
+ if (num_rports > 0)
+ bfa_cfg->fwcfg.num_rports = num_rports;
+ if (num_ios > 0)
+ bfa_cfg->fwcfg.num_ioim_reqs = num_ios;
+ if (num_tms > 0)
+ bfa_cfg->fwcfg.num_tskim_reqs = num_tms;
+ if (num_fcxps > 0 && num_fcxps <= BFA_FCXP_MAX)
+ bfa_cfg->fwcfg.num_fcxp_reqs = num_fcxps;
+ if (num_ufbufs > 0 && num_ufbufs <= BFA_UF_MAX)
+ bfa_cfg->fwcfg.num_uf_bufs = num_ufbufs;
+ if (reqq_size > 0)
+ bfa_cfg->drvcfg.num_reqq_elems = reqq_size;
+ if (rspq_size > 0)
+ bfa_cfg->drvcfg.num_rspq_elems = rspq_size;
+ if (num_sgpgs > 0 && num_sgpgs <= BFA_SGPG_MAX)
+ bfa_cfg->drvcfg.num_sgpgs = num_sgpgs;
+
+ /*
+ * populate the hal values back to the driver for sysfs use.
+ * otherwise, the default values will be shown as 0 in sysfs
+ */
+ num_rports = bfa_cfg->fwcfg.num_rports;
+ num_ios = bfa_cfg->fwcfg.num_ioim_reqs;
+ num_tms = bfa_cfg->fwcfg.num_tskim_reqs;
+ num_fcxps = bfa_cfg->fwcfg.num_fcxp_reqs;
+ num_ufbufs = bfa_cfg->fwcfg.num_uf_bufs;
+ reqq_size = bfa_cfg->drvcfg.num_reqq_elems;
+ rspq_size = bfa_cfg->drvcfg.num_rspq_elems;
+ num_sgpgs = bfa_cfg->drvcfg.num_sgpgs;
+}
+
+bfa_status_t
+bfad_hal_mem_alloc(struct bfad_s *bfad)
+{
+ struct bfa_meminfo_s *hal_meminfo = &bfad->meminfo;
+ struct bfa_mem_dma_s *dma_info, *dma_elem;
+ struct bfa_mem_kva_s *kva_info, *kva_elem;
+ struct list_head *dm_qe, *km_qe;
+ bfa_status_t rc = BFA_STATUS_OK;
+ dma_addr_t phys_addr;
+
+ bfa_cfg_get_default(&bfad->ioc_cfg);
+ bfad_update_hal_cfg(&bfad->ioc_cfg);
+ bfad->cfg_data.ioc_queue_depth = bfad->ioc_cfg.fwcfg.num_ioim_reqs;
+ bfa_cfg_get_meminfo(&bfad->ioc_cfg, hal_meminfo, &bfad->bfa);
+
+ dma_info = &hal_meminfo->dma_info;
+ kva_info = &hal_meminfo->kva_info;
+
+ /* Iterate through the KVA meminfo queue */
+ list_for_each(km_qe, &kva_info->qe) {
+ kva_elem = (struct bfa_mem_kva_s *) km_qe;
+ kva_elem->kva = vmalloc(kva_elem->mem_len);
+ if (kva_elem->kva == NULL) {
+ bfad_hal_mem_release(bfad);
+ rc = BFA_STATUS_ENOMEM;
+ goto ext;
+ }
+ memset(kva_elem->kva, 0, kva_elem->mem_len);
+ }
+
+ /* Iterate through the DMA meminfo queue */
+ list_for_each(dm_qe, &dma_info->qe) {
+ dma_elem = (struct bfa_mem_dma_s *) dm_qe;
+ dma_elem->kva = dma_alloc_coherent(&bfad->pcidev->dev,
+ dma_elem->mem_len,
+ &phys_addr, GFP_KERNEL);
+ if (dma_elem->kva == NULL) {
+ bfad_hal_mem_release(bfad);
+ rc = BFA_STATUS_ENOMEM;
+ goto ext;
+ }
+ dma_elem->dma = phys_addr;
+ memset(dma_elem->kva, 0, dma_elem->mem_len);
+ }
+ext:
+ return rc;
+}
+
+/*
+ * Create a vport under a vf.
+ */
+bfa_status_t
+bfad_vport_create(struct bfad_s *bfad, u16 vf_id,
+ struct bfa_lport_cfg_s *port_cfg, struct device *dev)
+{
+ struct bfad_vport_s *vport;
+ int rc = BFA_STATUS_OK;
+ unsigned long flags;
+ struct completion fcomp;
+
+ vport = kzalloc(sizeof(struct bfad_vport_s), GFP_KERNEL);
+ if (!vport) {
+ rc = BFA_STATUS_ENOMEM;
+ goto ext;
+ }
+
+ vport->drv_port.bfad = bfad;
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ rc = bfa_fcs_vport_create(&vport->fcs_vport, &bfad->bfa_fcs, vf_id,
+ port_cfg, vport);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ if (rc != BFA_STATUS_OK)
+ goto ext_free_vport;
+
+ if (port_cfg->roles & BFA_LPORT_ROLE_FCP_IM) {
+ rc = bfad_im_scsi_host_alloc(bfad, vport->drv_port.im_port,
+ dev);
+ if (rc != BFA_STATUS_OK)
+ goto ext_free_fcs_vport;
+ }
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_fcs_vport_start(&vport->fcs_vport);
+ list_add_tail(&vport->list_entry, &bfad->vport_list);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ return BFA_STATUS_OK;
+
+ext_free_fcs_vport:
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ vport->comp_del = &fcomp;
+ init_completion(vport->comp_del);
+ bfa_fcs_vport_delete(&vport->fcs_vport);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ wait_for_completion(vport->comp_del);
+ext_free_vport:
+ kfree(vport);
+ext:
+ return rc;
+}
+
+void
+bfad_bfa_tmo(unsigned long data)
+{
+ struct bfad_s *bfad = (struct bfad_s *) data;
+ unsigned long flags;
+ struct list_head doneq;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+
+ bfa_timer_beat(&bfad->bfa.timer_mod);
+
+ bfa_comp_deq(&bfad->bfa, &doneq);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ if (!list_empty(&doneq)) {
+ bfa_comp_process(&bfad->bfa, &doneq);
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_comp_free(&bfad->bfa, &doneq);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ }
+
+ mod_timer(&bfad->hal_tmo,
+ jiffies + msecs_to_jiffies(BFA_TIMER_FREQ));
+}
+
+void
+bfad_init_timer(struct bfad_s *bfad)
+{
+ init_timer(&bfad->hal_tmo);
+ bfad->hal_tmo.function = bfad_bfa_tmo;
+ bfad->hal_tmo.data = (unsigned long)bfad;
+
+ mod_timer(&bfad->hal_tmo,
+ jiffies + msecs_to_jiffies(BFA_TIMER_FREQ));
+}
+
+int
+bfad_pci_init(struct pci_dev *pdev, struct bfad_s *bfad)
+{
+ int rc = -ENODEV;
+
+ if (pci_enable_device(pdev)) {
+ printk(KERN_ERR "pci_enable_device fail %p\n", pdev);
+ goto out;
+ }
+
+ if (pci_request_regions(pdev, BFAD_DRIVER_NAME))
+ goto out_disable_device;
+
+ pci_set_master(pdev);
+
+
+ if ((pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) != 0) ||
+ (pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)) != 0)) {
+ if ((pci_set_dma_mask(pdev, DMA_BIT_MASK(32)) != 0) ||
+ (pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32)) != 0)) {
+ printk(KERN_ERR "pci_set_dma_mask fail %p\n", pdev);
+ goto out_release_region;
+ }
+ }
+
+ /* Enable PCIE Advanced Error Recovery (AER) if kernel supports */
+ pci_enable_pcie_error_reporting(pdev);
+
+ bfad->pci_bar0_kva = pci_iomap(pdev, 0, pci_resource_len(pdev, 0));
+ bfad->pci_bar2_kva = pci_iomap(pdev, 2, pci_resource_len(pdev, 2));
+
+ if (bfad->pci_bar0_kva == NULL) {
+ printk(KERN_ERR "Fail to map bar0\n");
+ goto out_release_region;
+ }
+
+ bfad->hal_pcidev.pci_slot = PCI_SLOT(pdev->devfn);
+ bfad->hal_pcidev.pci_func = PCI_FUNC(pdev->devfn);
+ bfad->hal_pcidev.pci_bar_kva = bfad->pci_bar0_kva;
+ bfad->hal_pcidev.device_id = pdev->device;
+ bfad->hal_pcidev.ssid = pdev->subsystem_device;
+ bfad->pci_name = pci_name(pdev);
+
+ bfad->pci_attr.vendor_id = pdev->vendor;
+ bfad->pci_attr.device_id = pdev->device;
+ bfad->pci_attr.ssid = pdev->subsystem_device;
+ bfad->pci_attr.ssvid = pdev->subsystem_vendor;
+ bfad->pci_attr.pcifn = PCI_FUNC(pdev->devfn);
+
+ bfad->pcidev = pdev;
+
+ /* Adjust PCIe Maximum Read Request Size */
+ if (pci_is_pcie(pdev) && pcie_max_read_reqsz) {
+ if (pcie_max_read_reqsz >= 128 &&
+ pcie_max_read_reqsz <= 4096 &&
+ is_power_of_2(pcie_max_read_reqsz)) {
+ int max_rq = pcie_get_readrq(pdev);
+ printk(KERN_WARNING "BFA[%s]: "
+ "pcie_max_read_request_size is %d, "
+ "reset to %d\n", bfad->pci_name, max_rq,
+ pcie_max_read_reqsz);
+ pcie_set_readrq(pdev, pcie_max_read_reqsz);
+ } else {
+ printk(KERN_WARNING "BFA[%s]: invalid "
+ "pcie_max_read_request_size %d ignored\n",
+ bfad->pci_name, pcie_max_read_reqsz);
+ }
+ }
+
+ pci_save_state(pdev);
+
+ return 0;
+
+out_release_region:
+ pci_release_regions(pdev);
+out_disable_device:
+ pci_disable_device(pdev);
+out:
+ return rc;
+}
+
+void
+bfad_pci_uninit(struct pci_dev *pdev, struct bfad_s *bfad)
+{
+ pci_iounmap(pdev, bfad->pci_bar0_kva);
+ pci_iounmap(pdev, bfad->pci_bar2_kva);
+ pci_release_regions(pdev);
+ /* Disable PCIE Advanced Error Recovery (AER) */
+ pci_disable_pcie_error_reporting(pdev);
+ pci_disable_device(pdev);
+}
+
+bfa_status_t
+bfad_drv_init(struct bfad_s *bfad)
+{
+ bfa_status_t rc;
+ unsigned long flags;
+
+ bfad->cfg_data.rport_del_timeout = rport_del_timeout;
+ bfad->cfg_data.lun_queue_depth = bfa_lun_queue_depth;
+ bfad->cfg_data.io_max_sge = bfa_io_max_sge;
+ bfad->cfg_data.binding_method = FCP_PWWN_BINDING;
+
+ rc = bfad_hal_mem_alloc(bfad);
+ if (rc != BFA_STATUS_OK) {
+ printk(KERN_WARNING "bfad%d bfad_hal_mem_alloc failure\n",
+ bfad->inst_no);
+ printk(KERN_WARNING
+ "Not enough memory to attach all Brocade HBA ports, %s",
+ "System may need more memory.\n");
+ return BFA_STATUS_FAILED;
+ }
+
+ bfad->bfa.trcmod = bfad->trcmod;
+ bfad->bfa.plog = &bfad->plog_buf;
+ bfa_plog_init(&bfad->plog_buf);
+ bfa_plog_str(&bfad->plog_buf, BFA_PL_MID_DRVR, BFA_PL_EID_DRIVER_START,
+ 0, "Driver Attach");
+
+ bfa_attach(&bfad->bfa, bfad, &bfad->ioc_cfg, &bfad->meminfo,
+ &bfad->hal_pcidev);
+
+ /* FCS INIT */
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfa_fcs.trcmod = bfad->trcmod;
+ bfa_fcs_attach(&bfad->bfa_fcs, &bfad->bfa, bfad, BFA_FALSE);
+ bfad->bfa_fcs.fdmi_enabled = fdmi_enable;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ bfad->bfad_flags |= BFAD_DRV_INIT_DONE;
+
+ return BFA_STATUS_OK;
+}
+
+void
+bfad_drv_uninit(struct bfad_s *bfad)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ init_completion(&bfad->comp);
+ bfa_iocfc_stop(&bfad->bfa);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ wait_for_completion(&bfad->comp);
+
+ del_timer_sync(&bfad->hal_tmo);
+ bfa_isr_disable(&bfad->bfa);
+ bfa_detach(&bfad->bfa);
+ bfad_remove_intr(bfad);
+ bfad_hal_mem_release(bfad);
+
+ bfad->bfad_flags &= ~BFAD_DRV_INIT_DONE;
+}
+
+void
+bfad_drv_start(struct bfad_s *bfad)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_iocfc_start(&bfad->bfa);
+ bfa_fcs_pbc_vport_init(&bfad->bfa_fcs);
+ bfa_fcs_fabric_modstart(&bfad->bfa_fcs);
+ bfad->bfad_flags |= BFAD_HAL_START_DONE;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ if (bfad->im)
+ flush_workqueue(bfad->im->drv_workq);
+}
+
+void
+bfad_fcs_stop(struct bfad_s *bfad)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ init_completion(&bfad->comp);
+ bfad->pport.flags |= BFAD_PORT_DELETE;
+ bfa_fcs_exit(&bfad->bfa_fcs);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ wait_for_completion(&bfad->comp);
+
+ bfa_sm_send_event(bfad, BFAD_E_FCS_EXIT_COMP);
+}
+
+void
+bfad_stop(struct bfad_s *bfad)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ init_completion(&bfad->comp);
+ bfa_iocfc_stop(&bfad->bfa);
+ bfad->bfad_flags &= ~BFAD_HAL_START_DONE;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ wait_for_completion(&bfad->comp);
+
+ bfa_sm_send_event(bfad, BFAD_E_EXIT_COMP);
+}
+
+bfa_status_t
+bfad_cfg_pport(struct bfad_s *bfad, enum bfa_lport_role role)
+{
+ int rc = BFA_STATUS_OK;
+
+ /* Allocate scsi_host for the physical port */
+ if ((supported_fc4s & BFA_LPORT_ROLE_FCP_IM) &&
+ (role & BFA_LPORT_ROLE_FCP_IM)) {
+ if (bfad->pport.im_port == NULL) {
+ rc = BFA_STATUS_FAILED;
+ goto out;
+ }
+
+ rc = bfad_im_scsi_host_alloc(bfad, bfad->pport.im_port,
+ &bfad->pcidev->dev);
+ if (rc != BFA_STATUS_OK)
+ goto out;
+
+ bfad->pport.roles |= BFA_LPORT_ROLE_FCP_IM;
+ }
+
+ bfad->bfad_flags |= BFAD_CFG_PPORT_DONE;
+
+out:
+ return rc;
+}
+
+void
+bfad_uncfg_pport(struct bfad_s *bfad)
+{
+ if ((supported_fc4s & BFA_LPORT_ROLE_FCP_IM) &&
+ (bfad->pport.roles & BFA_LPORT_ROLE_FCP_IM)) {
+ bfad_im_scsi_host_free(bfad, bfad->pport.im_port);
+ bfad_im_port_clean(bfad->pport.im_port);
+ kfree(bfad->pport.im_port);
+ bfad->pport.roles &= ~BFA_LPORT_ROLE_FCP_IM;
+ }
+
+ bfad->bfad_flags &= ~BFAD_CFG_PPORT_DONE;
+}
+
+bfa_status_t
+bfad_start_ops(struct bfad_s *bfad) {
+
+ int retval;
+ unsigned long flags;
+ struct bfad_vport_s *vport, *vport_new;
+ struct bfa_fcs_driver_info_s driver_info;
+
+ /* Limit min/max. xfer size to [64k-32MB] */
+ if (max_xfer_size < BFAD_MIN_SECTORS >> 1)
+ max_xfer_size = BFAD_MIN_SECTORS >> 1;
+ if (max_xfer_size > BFAD_MAX_SECTORS >> 1)
+ max_xfer_size = BFAD_MAX_SECTORS >> 1;
+
+ /* Fill the driver_info info to fcs*/
+ memset(&driver_info, 0, sizeof(driver_info));
+ strncpy(driver_info.version, BFAD_DRIVER_VERSION,
+ sizeof(driver_info.version) - 1);
+ if (host_name)
+ strncpy(driver_info.host_machine_name, host_name,
+ sizeof(driver_info.host_machine_name) - 1);
+ if (os_name)
+ strncpy(driver_info.host_os_name, os_name,
+ sizeof(driver_info.host_os_name) - 1);
+ if (os_patch)
+ strncpy(driver_info.host_os_patch, os_patch,
+ sizeof(driver_info.host_os_patch) - 1);
+
+ strncpy(driver_info.os_device_name, bfad->pci_name,
+ sizeof(driver_info.os_device_name) - 1);
+
+ /* FCS driver info init */
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_fcs_driver_info_init(&bfad->bfa_fcs, &driver_info);
+
+ if (bfad->bfad_flags & BFAD_CFG_PPORT_DONE)
+ bfa_fcs_update_cfg(&bfad->bfa_fcs);
+ else
+ bfa_fcs_init(&bfad->bfa_fcs);
+
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ if (!(bfad->bfad_flags & BFAD_CFG_PPORT_DONE)) {
+ retval = bfad_cfg_pport(bfad, BFA_LPORT_ROLE_FCP_IM);
+ if (retval != BFA_STATUS_OK)
+ return BFA_STATUS_FAILED;
+ }
+
+ /* Setup fc host fixed attribute if the lk supports */
+ bfad_fc_host_init(bfad->pport.im_port);
+
+ /* BFAD level FC4 IM specific resource allocation */
+ retval = bfad_im_probe(bfad);
+ if (retval != BFA_STATUS_OK) {
+ printk(KERN_WARNING "bfad_im_probe failed\n");
+ if (bfa_sm_cmp_state(bfad, bfad_sm_initializing))
+ bfa_sm_set_state(bfad, bfad_sm_failed);
+ return BFA_STATUS_FAILED;
+ } else
+ bfad->bfad_flags |= BFAD_FC4_PROBE_DONE;
+
+ bfad_drv_start(bfad);
+
+ /* Complete pbc vport create */
+ list_for_each_entry_safe(vport, vport_new, &bfad->pbc_vport_list,
+ list_entry) {
+ struct fc_vport_identifiers vid;
+ struct fc_vport *fc_vport;
+ char pwwn_buf[BFA_STRING_32];
+
+ memset(&vid, 0, sizeof(vid));
+ vid.roles = FC_PORT_ROLE_FCP_INITIATOR;
+ vid.vport_type = FC_PORTTYPE_NPIV;
+ vid.disable = false;
+ vid.node_name = wwn_to_u64((u8 *)
+ (&((vport->fcs_vport).lport.port_cfg.nwwn)));
+ vid.port_name = wwn_to_u64((u8 *)
+ (&((vport->fcs_vport).lport.port_cfg.pwwn)));
+ fc_vport = fc_vport_create(bfad->pport.im_port->shost, 0, &vid);
+ if (!fc_vport) {
+ wwn2str(pwwn_buf, vid.port_name);
+ printk(KERN_WARNING "bfad%d: failed to create pbc vport"
+ " %s\n", bfad->inst_no, pwwn_buf);
+ }
+ list_del(&vport->list_entry);
+ kfree(vport);
+ }
+
+ /*
+ * If bfa_linkup_delay is set to -1 default; try to retrive the
+ * value using the bfad_get_linkup_delay(); else use the
+ * passed in module param value as the bfa_linkup_delay.
+ */
+ if (bfa_linkup_delay < 0) {
+ bfa_linkup_delay = bfad_get_linkup_delay(bfad);
+ bfad_rport_online_wait(bfad);
+ bfa_linkup_delay = -1;
+ } else
+ bfad_rport_online_wait(bfad);
+
+ BFA_LOG(KERN_INFO, bfad, bfa_log_level, "bfa device claimed\n");
+
+ return BFA_STATUS_OK;
+}
+
+int
+bfad_worker(void *ptr)
+{
+ struct bfad_s *bfad = ptr;
+ unsigned long flags;
+
+ if (kthread_should_stop())
+ return 0;
+
+ /* Send event BFAD_E_INIT_SUCCESS */
+ bfa_sm_send_event(bfad, BFAD_E_INIT_SUCCESS);
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_tsk = NULL;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ return 0;
+}
+
+/*
+ * BFA driver interrupt functions
+ */
+irqreturn_t
+bfad_intx(int irq, void *dev_id)
+{
+ struct bfad_s *bfad = dev_id;
+ struct list_head doneq;
+ unsigned long flags;
+ bfa_boolean_t rc;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ rc = bfa_intx(&bfad->bfa);
+ if (!rc) {
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ return IRQ_NONE;
+ }
+
+ bfa_comp_deq(&bfad->bfa, &doneq);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ if (!list_empty(&doneq)) {
+ bfa_comp_process(&bfad->bfa, &doneq);
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_comp_free(&bfad->bfa, &doneq);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ }
+
+ return IRQ_HANDLED;
+
+}
+
+static irqreturn_t
+bfad_msix(int irq, void *dev_id)
+{
+ struct bfad_msix_s *vec = dev_id;
+ struct bfad_s *bfad = vec->bfad;
+ struct list_head doneq;
+ unsigned long flags;
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+
+ bfa_msix(&bfad->bfa, vec->msix.entry);
+ bfa_comp_deq(&bfad->bfa, &doneq);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ if (!list_empty(&doneq)) {
+ bfa_comp_process(&bfad->bfa, &doneq);
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_comp_free(&bfad->bfa, &doneq);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ }
+
+ return IRQ_HANDLED;
+}
+
+/*
+ * Initialize the MSIX entry table.
+ */
+static void
+bfad_init_msix_entry(struct bfad_s *bfad, struct msix_entry *msix_entries,
+ int mask, int max_bit)
+{
+ int i;
+ int match = 0x00000001;
+
+ for (i = 0, bfad->nvec = 0; i < MAX_MSIX_ENTRY; i++) {
+ if (mask & match) {
+ bfad->msix_tab[bfad->nvec].msix.entry = i;
+ bfad->msix_tab[bfad->nvec].bfad = bfad;
+ msix_entries[bfad->nvec].entry = i;
+ bfad->nvec++;
+ }
+
+ match <<= 1;
+ }
+
+}
+
+int
+bfad_install_msix_handler(struct bfad_s *bfad)
+{
+ int i, error = 0;
+
+ for (i = 0; i < bfad->nvec; i++) {
+ sprintf(bfad->msix_tab[i].name, "bfa-%s-%s",
+ bfad->pci_name,
+ ((bfa_asic_id_cb(bfad->hal_pcidev.device_id)) ?
+ msix_name_cb[i] : msix_name_ct[i]));
+
+ error = request_irq(bfad->msix_tab[i].msix.vector,
+ (irq_handler_t) bfad_msix, 0,
+ bfad->msix_tab[i].name, &bfad->msix_tab[i]);
+ bfa_trc(bfad, i);
+ bfa_trc(bfad, bfad->msix_tab[i].msix.vector);
+ if (error) {
+ int j;
+
+ for (j = 0; j < i; j++)
+ free_irq(bfad->msix_tab[j].msix.vector,
+ &bfad->msix_tab[j]);
+
+ bfad->bfad_flags &= ~BFAD_MSIX_ON;
+ pci_disable_msix(bfad->pcidev);
+
+ return 1;
+ }
+ }
+
+ return 0;
+}
+
+/*
+ * Setup MSIX based interrupt.
+ */
+int
+bfad_setup_intr(struct bfad_s *bfad)
+{
+ int error;
+ u32 mask = 0, i, num_bit = 0, max_bit = 0;
+ struct msix_entry msix_entries[MAX_MSIX_ENTRY];
+ struct pci_dev *pdev = bfad->pcidev;
+ u16 reg;
+
+ /* Call BFA to get the msix map for this PCI function. */
+ bfa_msix_getvecs(&bfad->bfa, &mask, &num_bit, &max_bit);
+
+ /* Set up the msix entry table */
+ bfad_init_msix_entry(bfad, msix_entries, mask, max_bit);
+
+ if ((bfa_asic_id_ctc(pdev->device) && !msix_disable_ct) ||
+ (bfa_asic_id_cb(pdev->device) && !msix_disable_cb)) {
+
+ error = pci_enable_msix_exact(bfad->pcidev,
+ msix_entries, bfad->nvec);
+ /* In CT1 & CT2, try to allocate just one vector */
+ if (error == -ENOSPC && bfa_asic_id_ctc(pdev->device)) {
+ printk(KERN_WARNING "bfa %s: trying one msix "
+ "vector failed to allocate %d[%d]\n",
+ bfad->pci_name, bfad->nvec, error);
+ bfad->nvec = 1;
+ error = pci_enable_msix_exact(bfad->pcidev,
+ msix_entries, 1);
+ }
+
+ if (error) {
+ printk(KERN_WARNING "bfad%d: "
+ "pci_enable_msix_exact failed (%d), "
+ "use line based.\n",
+ bfad->inst_no, error);
+ goto line_based;
+ }
+
+ /* Disable INTX in MSI-X mode */
+ pci_read_config_word(pdev, PCI_COMMAND, &reg);
+
+ if (!(reg & PCI_COMMAND_INTX_DISABLE))
+ pci_write_config_word(pdev, PCI_COMMAND,
+ reg | PCI_COMMAND_INTX_DISABLE);
+
+ /* Save the vectors */
+ for (i = 0; i < bfad->nvec; i++) {
+ bfa_trc(bfad, msix_entries[i].vector);
+ bfad->msix_tab[i].msix.vector = msix_entries[i].vector;
+ }
+
+ bfa_msix_init(&bfad->bfa, bfad->nvec);
+
+ bfad->bfad_flags |= BFAD_MSIX_ON;
+
+ return 0;
+ }
+
+line_based:
+ error = request_irq(bfad->pcidev->irq, (irq_handler_t)bfad_intx,
+ BFAD_IRQ_FLAGS, BFAD_DRIVER_NAME, bfad);
+ if (error)
+ return error;
+
+ bfad->bfad_flags |= BFAD_INTX_ON;
+
+ return 0;
+}
+
+void
+bfad_remove_intr(struct bfad_s *bfad)
+{
+ int i;
+
+ if (bfad->bfad_flags & BFAD_MSIX_ON) {
+ for (i = 0; i < bfad->nvec; i++)
+ free_irq(bfad->msix_tab[i].msix.vector,
+ &bfad->msix_tab[i]);
+
+ pci_disable_msix(bfad->pcidev);
+ bfad->bfad_flags &= ~BFAD_MSIX_ON;
+ } else if (bfad->bfad_flags & BFAD_INTX_ON) {
+ free_irq(bfad->pcidev->irq, bfad);
+ }
+}
+
+/*
+ * PCI probe entry.
+ */
+int
+bfad_pci_probe(struct pci_dev *pdev, const struct pci_device_id *pid)
+{
+ struct bfad_s *bfad;
+ int error = -ENODEV, retval, i;
+
+ /* For single port cards - only claim function 0 */
+ if ((pdev->device == BFA_PCI_DEVICE_ID_FC_8G1P) &&
+ (PCI_FUNC(pdev->devfn) != 0))
+ return -ENODEV;
+
+ bfad = kzalloc(sizeof(struct bfad_s), GFP_KERNEL);
+ if (!bfad) {
+ error = -ENOMEM;
+ goto out;
+ }
+
+ bfad->trcmod = kzalloc(sizeof(struct bfa_trc_mod_s), GFP_KERNEL);
+ if (!bfad->trcmod) {
+ printk(KERN_WARNING "Error alloc trace buffer!\n");
+ error = -ENOMEM;
+ goto out_alloc_trace_failure;
+ }
+
+ /* TRACE INIT */
+ bfa_trc_init(bfad->trcmod);
+ bfa_trc(bfad, bfad_inst);
+
+ /* AEN INIT */
+ INIT_LIST_HEAD(&bfad->free_aen_q);
+ INIT_LIST_HEAD(&bfad->active_aen_q);
+ for (i = 0; i < BFA_AEN_MAX_ENTRY; i++)
+ list_add_tail(&bfad->aen_list[i].qe, &bfad->free_aen_q);
+
+ if (!(bfad_load_fwimg(pdev))) {
+ kfree(bfad->trcmod);
+ goto out_alloc_trace_failure;
+ }
+
+ retval = bfad_pci_init(pdev, bfad);
+ if (retval) {
+ printk(KERN_WARNING "bfad_pci_init failure!\n");
+ error = retval;
+ goto out_pci_init_failure;
+ }
+
+ mutex_lock(&bfad_mutex);
+ bfad->inst_no = bfad_inst++;
+ list_add_tail(&bfad->list_entry, &bfad_list);
+ mutex_unlock(&bfad_mutex);
+
+ /* Initializing the state machine: State set to uninit */
+ bfa_sm_set_state(bfad, bfad_sm_uninit);
+
+ spin_lock_init(&bfad->bfad_lock);
+ spin_lock_init(&bfad->bfad_aen_spinlock);
+
+ pci_set_drvdata(pdev, bfad);
+
+ bfad->ref_count = 0;
+ bfad->pport.bfad = bfad;
+ INIT_LIST_HEAD(&bfad->pbc_vport_list);
+ INIT_LIST_HEAD(&bfad->vport_list);
+
+ /* Setup the debugfs node for this bfad */
+ if (bfa_debugfs_enable)
+ bfad_debugfs_init(&bfad->pport);
+
+ retval = bfad_drv_init(bfad);
+ if (retval != BFA_STATUS_OK)
+ goto out_drv_init_failure;
+
+ bfa_sm_send_event(bfad, BFAD_E_CREATE);
+
+ if (bfa_sm_cmp_state(bfad, bfad_sm_uninit))
+ goto out_bfad_sm_failure;
+
+ return 0;
+
+out_bfad_sm_failure:
+ bfad_hal_mem_release(bfad);
+out_drv_init_failure:
+ /* Remove the debugfs node for this bfad */
+ kfree(bfad->regdata);
+ bfad_debugfs_exit(&bfad->pport);
+ mutex_lock(&bfad_mutex);
+ bfad_inst--;
+ list_del(&bfad->list_entry);
+ mutex_unlock(&bfad_mutex);
+ bfad_pci_uninit(pdev, bfad);
+out_pci_init_failure:
+ kfree(bfad->trcmod);
+out_alloc_trace_failure:
+ kfree(bfad);
+out:
+ return error;
+}
+
+/*
+ * PCI remove entry.
+ */
+void
+bfad_pci_remove(struct pci_dev *pdev)
+{
+ struct bfad_s *bfad = pci_get_drvdata(pdev);
+ unsigned long flags;
+
+ bfa_trc(bfad, bfad->inst_no);
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ if (bfad->bfad_tsk != NULL) {
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ kthread_stop(bfad->bfad_tsk);
+ } else {
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ }
+
+ /* Send Event BFAD_E_STOP */
+ bfa_sm_send_event(bfad, BFAD_E_STOP);
+
+ /* Driver detach and dealloc mem */
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_detach(&bfad->bfa);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ bfad_hal_mem_release(bfad);
+
+ /* Remove the debugfs node for this bfad */
+ kfree(bfad->regdata);
+ bfad_debugfs_exit(&bfad->pport);
+
+ /* Cleaning the BFAD instance */
+ mutex_lock(&bfad_mutex);
+ bfad_inst--;
+ list_del(&bfad->list_entry);
+ mutex_unlock(&bfad_mutex);
+ bfad_pci_uninit(pdev, bfad);
+
+ kfree(bfad->trcmod);
+ kfree(bfad);
+}
+
+/*
+ * PCI Error Recovery entry, error detected.
+ */
+static pci_ers_result_t
+bfad_pci_error_detected(struct pci_dev *pdev, pci_channel_state_t state)
+{
+ struct bfad_s *bfad = pci_get_drvdata(pdev);
+ unsigned long flags;
+ pci_ers_result_t ret = PCI_ERS_RESULT_NONE;
+
+ dev_printk(KERN_ERR, &pdev->dev,
+ "error detected state: %d - flags: 0x%x\n",
+ state, bfad->bfad_flags);
+
+ switch (state) {
+ case pci_channel_io_normal: /* non-fatal error */
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_flags &= ~BFAD_EEH_BUSY;
+ /* Suspend/fail all bfa operations */
+ bfa_ioc_suspend(&bfad->bfa.ioc);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ del_timer_sync(&bfad->hal_tmo);
+ ret = PCI_ERS_RESULT_CAN_RECOVER;
+ break;
+ case pci_channel_io_frozen: /* fatal error */
+ init_completion(&bfad->comp);
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_flags |= BFAD_EEH_BUSY;
+ /* Suspend/fail all bfa operations */
+ bfa_ioc_suspend(&bfad->bfa.ioc);
+ bfa_fcs_stop(&bfad->bfa_fcs);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ wait_for_completion(&bfad->comp);
+
+ bfad_remove_intr(bfad);
+ del_timer_sync(&bfad->hal_tmo);
+ pci_disable_device(pdev);
+ ret = PCI_ERS_RESULT_NEED_RESET;
+ break;
+ case pci_channel_io_perm_failure: /* PCI Card is DEAD */
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_flags |= BFAD_EEH_BUSY |
+ BFAD_EEH_PCI_CHANNEL_IO_PERM_FAILURE;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ /* If the error_detected handler is called with the reason
+ * pci_channel_io_perm_failure - it will subsequently call
+ * pci_remove() entry point to remove the pci device from the
+ * system - So defer the cleanup to pci_remove(); cleaning up
+ * here causes inconsistent state during pci_remove().
+ */
+ ret = PCI_ERS_RESULT_DISCONNECT;
+ break;
+ default:
+ WARN_ON(1);
+ }
+
+ return ret;
+}
+
+int
+restart_bfa(struct bfad_s *bfad)
+{
+ unsigned long flags;
+ struct pci_dev *pdev = bfad->pcidev;
+
+ bfa_attach(&bfad->bfa, bfad, &bfad->ioc_cfg,
+ &bfad->meminfo, &bfad->hal_pcidev);
+
+ /* Enable Interrupt and wait bfa_init completion */
+ if (bfad_setup_intr(bfad)) {
+ dev_printk(KERN_WARNING, &pdev->dev,
+ "%s: bfad_setup_intr failed\n", bfad->pci_name);
+ bfa_sm_send_event(bfad, BFAD_E_INIT_FAILED);
+ return -1;
+ }
+
+ init_completion(&bfad->comp);
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfa_iocfc_init(&bfad->bfa);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+
+ /* Set up interrupt handler for each vectors */
+ if ((bfad->bfad_flags & BFAD_MSIX_ON) &&
+ bfad_install_msix_handler(bfad))
+ dev_printk(KERN_WARNING, &pdev->dev,
+ "%s: install_msix failed.\n", bfad->pci_name);
+
+ bfad_init_timer(bfad);
+ wait_for_completion(&bfad->comp);
+ bfad_drv_start(bfad);
+
+ return 0;
+}
+
+/*
+ * PCI Error Recovery entry, re-initialize the chip.
+ */
+static pci_ers_result_t
+bfad_pci_slot_reset(struct pci_dev *pdev)
+{
+ struct bfad_s *bfad = pci_get_drvdata(pdev);
+ u8 byte;
+
+ dev_printk(KERN_ERR, &pdev->dev,
+ "bfad_pci_slot_reset flags: 0x%x\n", bfad->bfad_flags);
+
+ if (pci_enable_device(pdev)) {
+ dev_printk(KERN_ERR, &pdev->dev, "Cannot re-enable "
+ "PCI device after reset.\n");
+ return PCI_ERS_RESULT_DISCONNECT;
+ }
+
+ pci_restore_state(pdev);
+
+ /*
+ * Read some byte (e.g. DMA max. payload size which can't
+ * be 0xff any time) to make sure - we did not hit another PCI error
+ * in the middle of recovery. If we did, then declare permanent failure.
+ */
+ pci_read_config_byte(pdev, 0x68, &byte);
+ if (byte == 0xff) {
+ dev_printk(KERN_ERR, &pdev->dev,
+ "slot_reset failed ... got another PCI error !\n");
+ goto out_disable_device;
+ }
+
+ pci_save_state(pdev);
+ pci_set_master(pdev);
+
+ if (pci_set_dma_mask(bfad->pcidev, DMA_BIT_MASK(64)) != 0)
+ if (pci_set_dma_mask(bfad->pcidev, DMA_BIT_MASK(32)) != 0)
+ goto out_disable_device;
+
+ pci_cleanup_aer_uncorrect_error_status(pdev);
+
+ if (restart_bfa(bfad) == -1)
+ goto out_disable_device;
+
+ pci_enable_pcie_error_reporting(pdev);
+ dev_printk(KERN_WARNING, &pdev->dev,
+ "slot_reset completed flags: 0x%x!\n", bfad->bfad_flags);
+
+ return PCI_ERS_RESULT_RECOVERED;
+
+out_disable_device:
+ pci_disable_device(pdev);
+ return PCI_ERS_RESULT_DISCONNECT;
+}
+
+static pci_ers_result_t
+bfad_pci_mmio_enabled(struct pci_dev *pdev)
+{
+ unsigned long flags;
+ struct bfad_s *bfad = pci_get_drvdata(pdev);
+
+ dev_printk(KERN_INFO, &pdev->dev, "mmio_enabled\n");
+
+ /* Fetch FW diagnostic information */
+ bfa_ioc_debug_save_ftrc(&bfad->bfa.ioc);
+
+ /* Cancel all pending IOs */
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ init_completion(&bfad->comp);
+ bfa_fcs_stop(&bfad->bfa_fcs);
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+ wait_for_completion(&bfad->comp);
+
+ bfad_remove_intr(bfad);
+ del_timer_sync(&bfad->hal_tmo);
+ pci_disable_device(pdev);
+
+ return PCI_ERS_RESULT_NEED_RESET;
+}
+
+static void
+bfad_pci_resume(struct pci_dev *pdev)
+{
+ unsigned long flags;
+ struct bfad_s *bfad = pci_get_drvdata(pdev);
+
+ dev_printk(KERN_WARNING, &pdev->dev, "resume\n");
+
+ /* wait until the link is online */
+ bfad_rport_online_wait(bfad);
+
+ spin_lock_irqsave(&bfad->bfad_lock, flags);
+ bfad->bfad_flags &= ~BFAD_EEH_BUSY;
+ spin_unlock_irqrestore(&bfad->bfad_lock, flags);
+}
+
+struct pci_device_id bfad_id_table[] = {
+ {
+ .vendor = BFA_PCI_VENDOR_ID_BROCADE,
+ .device = BFA_PCI_DEVICE_ID_FC_8G2P,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ },
+ {
+ .vendor = BFA_PCI_VENDOR_ID_BROCADE,
+ .device = BFA_PCI_DEVICE_ID_FC_8G1P,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ },
+ {
+ .vendor = BFA_PCI_VENDOR_ID_BROCADE,
+ .device = BFA_PCI_DEVICE_ID_CT,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ .class = (PCI_CLASS_SERIAL_FIBER << 8),
+ .class_mask = ~0,
+ },
+ {
+ .vendor = BFA_PCI_VENDOR_ID_BROCADE,
+ .device = BFA_PCI_DEVICE_ID_CT_FC,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ .class = (PCI_CLASS_SERIAL_FIBER << 8),
+ .class_mask = ~0,
+ },
+ {
+ .vendor = BFA_PCI_VENDOR_ID_BROCADE,
+ .device = BFA_PCI_DEVICE_ID_CT2,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ .class = (PCI_CLASS_SERIAL_FIBER << 8),
+ .class_mask = ~0,
+ },
+
+ {
+ .vendor = BFA_PCI_VENDOR_ID_BROCADE,
+ .device = BFA_PCI_DEVICE_ID_CT2_QUAD,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ .class = (PCI_CLASS_SERIAL_FIBER << 8),
+ .class_mask = ~0,
+ },
+ {0, 0},
+};
+
+MODULE_DEVICE_TABLE(pci, bfad_id_table);
+
+/*
+ * PCI error recovery handlers.
+ */
+static struct pci_error_handlers bfad_err_handler = {
+ .error_detected = bfad_pci_error_detected,
+ .slot_reset = bfad_pci_slot_reset,
+ .mmio_enabled = bfad_pci_mmio_enabled,
+ .resume = bfad_pci_resume,
+};
+
+static struct pci_driver bfad_pci_driver = {
+ .name = BFAD_DRIVER_NAME,
+ .id_table = bfad_id_table,
+ .probe = bfad_pci_probe,
+ .remove = bfad_pci_remove,
+ .err_handler = &bfad_err_handler,
+};
+
+/*
+ * Driver module init.
+ */
+static int __init
+bfad_init(void)
+{
+ int error = 0;
+
+ printk(KERN_INFO "Brocade BFA FC/FCOE SCSI driver - version: %s\n",
+ BFAD_DRIVER_VERSION);
+
+ if (num_sgpgs > 0)
+ num_sgpgs_parm = num_sgpgs;
+
+ error = bfad_im_module_init();
+ if (error) {
+ error = -ENOMEM;
+ printk(KERN_WARNING "bfad_im_module_init failure\n");
+ goto ext;
+ }
+
+ if (strcmp(FCPI_NAME, " fcpim") == 0)
+ supported_fc4s |= BFA_LPORT_ROLE_FCP_IM;
+
+ bfa_auto_recover = ioc_auto_recover;
+ bfa_fcs_rport_set_del_timeout(rport_del_timeout);
+ bfa_fcs_rport_set_max_logins(max_rport_logins);
+
+ error = pci_register_driver(&bfad_pci_driver);
+ if (error) {
+ printk(KERN_WARNING "pci_register_driver failure\n");
+ goto ext;
+ }
+
+ return 0;
+
+ext:
+ bfad_im_module_exit();
+ return error;
+}
+
+/*
+ * Driver module exit.
+ */
+static void __exit
+bfad_exit(void)
+{
+ pci_unregister_driver(&bfad_pci_driver);
+ bfad_im_module_exit();
+ bfad_free_fwimg();
+}
+
+/* Firmware handling */
+static void
+bfad_read_firmware(struct pci_dev *pdev, u32 **bfi_image,
+ u32 *bfi_image_size, char *fw_name)
+{
+ const struct firmware *fw;
+
+ if (reject_firmware(&fw, fw_name, &pdev->dev)) {
+ printk(KERN_ALERT "Can't locate firmware %s\n", fw_name);
+ *bfi_image = NULL;
+ goto out;
+ }
+
+ *bfi_image = vmalloc(fw->size);
+ if (NULL == *bfi_image) {
+ printk(KERN_ALERT "Fail to allocate buffer for fw image "
+ "size=%x!\n", (u32) fw->size);
+ goto out;
+ }
+
+ memcpy(*bfi_image, fw->data, fw->size);
+ *bfi_image_size = fw->size/sizeof(u32);
+out:
+ release_firmware(fw);
+}
+
+static u32 *
+bfad_load_fwimg(struct pci_dev *pdev)
+{
+ if (bfa_asic_id_ct2(pdev->device)) {
+ if (bfi_image_ct2_size == 0)
+ bfad_read_firmware(pdev, &bfi_image_ct2,
+ &bfi_image_ct2_size, BFAD_FW_FILE_CT2);
+ return bfi_image_ct2;
+ } else if (bfa_asic_id_ct(pdev->device)) {
+ if (bfi_image_ct_size == 0)
+ bfad_read_firmware(pdev, &bfi_image_ct,
+ &bfi_image_ct_size, BFAD_FW_FILE_CT);
+ return bfi_image_ct;
+ } else if (bfa_asic_id_cb(pdev->device)) {
+ if (bfi_image_cb_size == 0)
+ bfad_read_firmware(pdev, &bfi_image_cb,
+ &bfi_image_cb_size, BFAD_FW_FILE_CB);
+ return bfi_image_cb;
+ }
+
+ return NULL;
+}
+
+static void
+bfad_free_fwimg(void)
+{
+ if (bfi_image_ct2_size && bfi_image_ct2)
+ vfree(bfi_image_ct2);
+ if (bfi_image_ct_size && bfi_image_ct)
+ vfree(bfi_image_ct);
+ if (bfi_image_cb_size && bfi_image_cb)
+ vfree(bfi_image_cb);
+}
+
+module_init(bfad_init);
+module_exit(bfad_exit);
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Brocade Fibre Channel HBA Driver" BFAD_PROTO_NAME);
+MODULE_AUTHOR("Brocade Communications Systems, Inc.");
+MODULE_VERSION(BFAD_DRIVER_VERSION);