| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <1266457062-31150-7-git-send-email-dm@chelsio.com>
Date: Wed, 17 Feb 2010 17:37:41 -0800
From: Dimitris Michailidis <dm@...lsio.com>
To: netdev@...r.kernel.org
Subject: [PATCH net-next 6/7] cxgb4: Add main driver file and driver Makefile
Signed-off-by: Dimitris Michailidis <dm@...lsio.com>
---
drivers/net/cxgb4/Makefile | 7 +
drivers/net/cxgb4/cxgb4_main.c | 4203 ++++++++++++++++++++++++++++++++++++++++
2 files changed, 4210 insertions(+), 0 deletions(-)
create mode 100644 drivers/net/cxgb4/Makefile
create mode 100644 drivers/net/cxgb4/cxgb4_main.c
diff --git a/drivers/net/cxgb4/Makefile b/drivers/net/cxgb4/Makefile
new file mode 100644
index 0000000..4986674
--- /dev/null
+++ b/drivers/net/cxgb4/Makefile
@@ -0,0 +1,7 @@
+#
+# Chelsio T4 driver
+#
+
+obj-$(CONFIG_CHELSIO_T4) += cxgb4.o
+
+cxgb4-objs := cxgb4_main.o l2t.o t4_hw.o sge.o
diff --git a/drivers/net/cxgb4/cxgb4_main.c b/drivers/net/cxgb4/cxgb4_main.c
new file mode 100644
index 0000000..6de7561
--- /dev/null
+++ b/drivers/net/cxgb4/cxgb4_main.c
@@ -0,0 +1,4203 @@
+/*
+ * This file is part of the Chelsio T4 Ethernet driver for Linux.
+ *
+ * Copyright (c) 2003-2010 Chelsio Communications, Inc. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/bitmap.h>
+#include <linux/crc32.h>
+#include <linux/ctype.h>
+#include <linux/debugfs.h>
+#include <linux/etherdevice.h>
+#include <linux/firmware.h>
+#include <linux/if_vlan.h>
+#include <linux/init.h>
+#include <linux/log2.h>
+#include <linux/mdio.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/mutex.h>
+#include <linux/netdevice.h>
+#include <linux/pci.h>
+#include <linux/aer.h>
+#include <linux/proc_fs.h>
+#include <linux/rtnetlink.h>
+#include <linux/seq_file.h>
+#include <linux/sockios.h>
+#include <linux/vmalloc.h>
+#include <linux/workqueue.h>
+#include <net/neighbour.h>
+#include <net/netevent.h>
+#include <asm/uaccess.h>
+
+#include "cxgb4.h"
+#include "t4_regs.h"
+#include "t4_msg.h"
+#include "t4fw_api.h"
+#include "l2t.h"
+
+#define DRV_VERSION "1.0.0-ko"
+#define DRV_DESC "Chelsio T4 Network Driver"
+
+/*
+ * Max interrupt hold-off timer value in us. Queues fall back to this value
+ * under extreme memory pressure so it's largish to give the system time to
+ * recover.
+ */
+#define MAX_SGE_TIMERVAL 200U
+
+enum {
+ MEMWIN0_APERTURE = 65536,
+ MEMWIN0_BASE = 0x30000,
+ MEMWIN1_APERTURE = 32768,
+ MEMWIN1_BASE = 0x28000,
+ MEMWIN2_APERTURE = 2048,
+ MEMWIN2_BASE = 0x1b800,
+};
+
+enum {
+ MAX_TXQ_ENTRIES = 16384,
+ MAX_CTRL_TXQ_ENTRIES = 1024,
+ MAX_RSPQ_ENTRIES = 16384,
+ MAX_RX_BUFFERS = 16384,
+ MIN_TXQ_ENTRIES = 32,
+ MIN_CTRL_TXQ_ENTRIES = 32,
+ MIN_RSPQ_ENTRIES = 128,
+ MIN_FL_ENTRIES = 16
+};
+
+#define DFLT_MSG_ENABLE (NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK | \
+ NETIF_MSG_TIMER | NETIF_MSG_IFDOWN | NETIF_MSG_IFUP |\
+ NETIF_MSG_RX_ERR | NETIF_MSG_TX_ERR)
+
+#define CH_DEVICE(devid) { PCI_VDEVICE(CHELSIO, devid), 0 }
+
+static DEFINE_PCI_DEVICE_TABLE(cxgb4_pci_tbl) = {
+ CH_DEVICE(0xa000), /* PE10K */
+ { 0, }
+};
+
+#define FW_FNAME "cxgb4/t4fw.bin"
+
+MODULE_DESCRIPTION(DRV_DESC);
+MODULE_AUTHOR("Chelsio Communications");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_VERSION(DRV_VERSION);
+MODULE_DEVICE_TABLE(pci, cxgb4_pci_tbl);
+MODULE_FIRMWARE(FW_FNAME);
+
+static int dflt_msg_enable = DFLT_MSG_ENABLE;
+
+module_param(dflt_msg_enable, int, 0644);
+MODULE_PARM_DESC(dflt_msg_enable, "Chelsio T4 default message enable bitmap");
+
+/*
+ * The driver uses the best interrupt scheme available on a platform in the
+ * order MSI-X, MSI, legacy INTx interrupts. This parameter determines which
+ * of these schemes the driver may consider as follows:
+ *
+ * msi = 2: choose from among all three options
+ * msi = 1: only consider MSI and INTx interrupts
+ * msi = 0: force INTx interrupts
+ */
+static int msi = 2;
+
+module_param(msi, int, 0644);
+MODULE_PARM_DESC(msi, "whether to use INTx (0), MSI (1) or MSI-X (2)");
+
+/*
+ * Queue interrupt hold-off timer values. Queues default to the first of these
+ * upon creation.
+ */
+static unsigned int intr_holdoff[SGE_NTIMERS - 1] = { 5, 10, 20, 50, 100 };
+
+module_param_array(intr_holdoff, uint, NULL, 0644);
+MODULE_PARM_DESC(intr_holdoff, "values for queue interrupt hold-off timers "
+ "0..4 in microseconds");
+
+static unsigned int intr_cnt[SGE_NCOUNTERS - 1] = { 4, 8, 16 };
+
+module_param_array(intr_cnt, uint, NULL, 0644);
+MODULE_PARM_DESC(intr_cnt,
+ "thresholds 1..3 for queue interrupt packet counters");
+
+static int vf_acls;
+
+#ifdef CONFIG_PCI_IOV
+module_param(vf_acls, bool, 0644);
+MODULE_PARM_DESC(vf_acls, "if set enable virtualization L2 ACL enforcement");
+
+static unsigned int num_vf[4];
+
+module_param_array(num_vf, uint, NULL, 0644);
+MODULE_PARM_DESC(num_vf, "number of VFs for each of PFs 0-3");
+#endif
+
+static struct dentry *cxgb4_debugfs_root;
+static struct proc_dir_entry *cxgb4_proc_root;
+
+static LIST_HEAD(adapter_list);
+static DEFINE_MUTEX(uld_mutex);
+static struct cxgb4_uld_info ulds[CXGB4_ULD_MAX];
+static const char *uld_str[] = { "RDMA", "iSCSI" };
+
+static void link_report(struct net_device *dev)
+{
+ if (!netif_carrier_ok(dev))
+ printk(KERN_INFO "%s: link down\n", dev->name);
+ else {
+ static const char *fc[] = { "no", "Rx", "Tx", "Tx/Rx" };
+
+ const char *s = "10Mbps";
+ const struct port_info *p = netdev_priv(dev);
+
+ switch (p->link_cfg.speed) {
+ case SPEED_10000: s = "10Gbps"; break;
+ case SPEED_1000: s = "1000Mbps"; break;
+ case SPEED_100: s = "100Mbps"; break;
+ }
+
+ printk(KERN_INFO "%s: link up, %s, full-duplex, %s PAUSE\n",
+ dev->name, s, fc[p->link_cfg.fc]);
+ }
+}
+
+void t4_os_link_changed(struct adapter *adapter, int port_id, int link_stat)
+{
+ struct net_device *dev = adapter->port[port_id];
+
+ /* Skip changes from disabled ports. */
+ if (netif_running(dev) && link_stat != netif_carrier_ok(dev)) {
+ if (link_stat)
+ netif_carrier_on(dev);
+ else
+ netif_carrier_off(dev);
+
+ link_report(dev);
+ }
+}
+
+void t4_os_portmod_changed(const struct adapter *adap, int port_id)
+{
+ static const char *mod_str[] = {
+ NULL, "LR", "SR", "ER", "passive DA", "active DA"
+ };
+
+ const struct net_device *dev = adap->port[port_id];
+ const struct port_info *pi = netdev_priv(dev);
+
+ if (pi->mod_type == FW_PORT_MOD_TYPE_NONE)
+ printk(KERN_INFO "%s: port module unplugged\n", dev->name);
+ else
+ printk(KERN_INFO "%s: %s port module inserted\n", dev->name,
+ mod_str[pi->mod_type]);
+}
+
+/*
+ * Configure the exact and hash address filters to handle a port's multicast
+ * and secondary unicast MAC addresses.
+ */
+static int set_addr_filters(const struct net_device *dev, bool sleep)
+{
+ u64 mhash = 0;
+ u64 uhash = 0;
+ bool free = true;
+ u16 filt_idx[7];
+ const u8 *addr[7];
+ int ret, naddr = 0;
+ const struct dev_addr_list *d;
+ const struct netdev_hw_addr *ha;
+ int uc_cnt = netdev_uc_count(dev);
+ const struct port_info *pi = netdev_priv(dev);
+
+ /* first do the secondary unicast addresses */
+ netdev_for_each_uc_addr(ha, dev) {
+ addr[naddr++] = ha->addr;
+ if (--uc_cnt == 0 || naddr >= ARRAY_SIZE(addr)) {
+ ret = t4_alloc_mac_filt(pi->adapter, 0, pi->viid, free,
+ naddr, addr, filt_idx, &uhash, sleep);
+ if (ret < 0)
+ return ret;
+
+ free = false;
+ naddr = 0;
+ }
+ }
+
+ /* next set up the multicast addresses */
+ netdev_for_each_mc_addr(d, dev) {
+ addr[naddr++] = d->dmi_addr;
+ if (naddr >= ARRAY_SIZE(addr) || d->next == NULL) {
+ ret = t4_alloc_mac_filt(pi->adapter, 0, pi->viid, free,
+ naddr, addr, filt_idx, &mhash, sleep);
+ if (ret < 0)
+ return ret;
+
+ free = false;
+ naddr = 0;
+ }
+ }
+
+ return t4_set_addr_hash(pi->adapter, 0, pi->viid, uhash != 0,
+ uhash | mhash, sleep);
+}
+
+/*
+ * Set Rx properties of a port, such as promiscruity, address filters, and MTU.
+ * If @mtu is -1 it is left unchanged.
+ */
+static int set_rxmode(struct net_device *dev, int mtu, bool sleep_ok)
+{
+ int ret;
+ struct port_info *pi = netdev_priv(dev);
+
+ ret = set_addr_filters(dev, sleep_ok);
+ if (ret == 0)
+ ret = t4_set_rxmode(pi->adapter, 0, pi->viid, mtu,
+ (dev->flags & IFF_PROMISC) ? 1 : 0,
+ (dev->flags & IFF_ALLMULTI) ? 1 : 0, 1,
+ sleep_ok);
+ return ret;
+}
+
+/**
+ * link_start - enable a port
+ * @dev: the port to enable
+ *
+ * Performs the MAC and PHY actions needed to enable a port.
+ */
+static int link_start(struct net_device *dev)
+{
+ int ret;
+ struct port_info *pi = netdev_priv(dev);
+
+ /*
+ * We do not set address filters and promiscuity here, the stack does
+ * that step explicitly.
+ */
+ ret = t4_set_rxmode(pi->adapter, 0, pi->viid, dev->mtu, -1, -1, -1,
+ true);
+ if (ret == 0) {
+ ret = t4_change_mac(pi->adapter, 0, pi->viid,
+ pi->xact_addr_filt, dev->dev_addr, true);
+ if (ret >= 0) {
+ pi->xact_addr_filt = ret;
+ ret = 0;
+ }
+ }
+ if (ret == 0)
+ ret = t4_link_start(pi->adapter, 0, pi->tx_chan, &pi->link_cfg);
+ if (ret == 0)
+ ret = t4_enable_vi(pi->adapter, 0, pi->viid, true, true);
+ return ret;
+}
+
+/*
+ * Response queue handler for the FW event queue.
+ */
+static int fwevtq_handler(struct sge_rspq *q, const __be64 *rsp,
+ const struct pkt_gl *gl)
+{
+ u8 opcode = ((const struct rss_header *)rsp)->opcode;
+
+ rsp++; /* skip RSS header */
+ if (likely(opcode == CPL_SGE_EGR_UPDATE)) {
+ const struct cpl_sge_egr_update *p = (void *)rsp;
+ unsigned int qid = G_EGR_QID(ntohl(p->opcode_qid));
+ struct sge_txq *txq = q->adap->sge.egr_map[qid];
+
+ txq->restarts++;
+ if ((u8 *)txq < (u8 *)q->adap->sge.ethrxq) {
+ struct sge_eth_txq *eq;
+
+ eq = container_of(txq, struct sge_eth_txq, q);
+ netif_tx_wake_queue(eq->txq);
+ } else {
+ struct sge_ofld_txq *oq;
+
+ oq = container_of(txq, struct sge_ofld_txq, q);
+ tasklet_schedule(&oq->qresume_tsk);
+ }
+ } else if (opcode == CPL_FW6_MSG || opcode == CPL_FW4_MSG) {
+ const struct cpl_fw6_msg *p = (void *)rsp;
+
+ if (p->type == 0)
+ t4_handle_fw_rpl(q->adap, p->data);
+ } else if (opcode == CPL_L2T_WRITE_RPL) {
+ const struct cpl_l2t_write_rpl *p = (void *)rsp;
+
+ do_l2t_write_rpl(q->adap, p);
+ } else
+ CH_ERR(q->adap, "unexpected CPL %#x on FW event queue\n",
+ opcode);
+ return 0;
+}
+
+/**
+ * uldrx_handler - response queue handler for ULD queues
+ * @q: the response queue that received the packet
+ * @rsp: the response queue descriptor holding the offload message
+ * @gl: the gather list of packet fragments
+ *
+ * Deliver an ingress offload packet to a ULD. All processing is done by
+ * the ULD, we just maintain statistics.
+ */
+static int uldrx_handler(struct sge_rspq *q, const __be64 *rsp,
+ const struct pkt_gl *gl)
+{
+ struct sge_ofld_rxq *rxq = container_of(q, struct sge_ofld_rxq, rspq);
+
+ if (ulds[q->uld].rx_handler(q->adap->uld_handle[q->uld], rsp, gl)) {
+ rxq->stats.nomem++;
+ return -1;
+ }
+ if (gl == NULL)
+ rxq->stats.imm++;
+ else if (gl == CXGB4_MSG_AN)
+ rxq->stats.an++;
+ else
+ rxq->stats.pkts++;
+ return 0;
+}
+
+static void disable_msi(struct adapter *adapter)
+{
+ if (adapter->flags & USING_MSIX) {
+ pci_disable_msix(adapter->pdev);
+ adapter->flags &= ~USING_MSIX;
+ } else if (adapter->flags & USING_MSI) {
+ pci_disable_msi(adapter->pdev);
+ adapter->flags &= ~USING_MSI;
+ }
+}
+
+/*
+ * Interrupt handler for non-data events used with MSI-X.
+ */
+static irqreturn_t t4_nondata_intr(int irq, void *cookie)
+{
+ struct adapter *adap = cookie;
+
+ u32 v = t4_read_reg(adap, MYPF_REG(A_PL_PF_INT_CAUSE));
+ if (v & F_PFSW) {
+ adap->swintr = 1;
+ t4_write_reg(adap, MYPF_REG(A_PL_PF_INT_CAUSE), v);
+ }
+ t4_slow_intr_handler(adap);
+ return IRQ_HANDLED;
+}
+
+/*
+ * Name the MSI-X interrupts.
+ */
+static void name_msix_vecs(struct adapter *adap)
+{
+ int i, j, msi_idx = 2, n = sizeof(adap->msix_info[0].desc) - 1;
+
+ /* non-data interrupts */
+ snprintf(adap->msix_info[0].desc, n, "%s", adap->name);
+ adap->msix_info[0].desc[n] = 0;
+
+ /* FW events */
+ snprintf(adap->msix_info[1].desc, n, "%s-FWeventq", adap->name);
+ adap->msix_info[1].desc[n] = 0;
+
+ /* Ethernet queues */
+ for_each_port(adap, j) {
+ struct net_device *d = adap->port[j];
+ const struct port_info *pi = netdev_priv(d);
+
+ for (i = 0; i < pi->nqsets; i++, msi_idx++) {
+ snprintf(adap->msix_info[msi_idx].desc, n, "%s-Rx%d",
+ d->name, i);
+ adap->msix_info[msi_idx].desc[n] = 0;
+ }
+ }
+
+ /* offload queues */
+ for_each_ofldrxq(&adap->sge, i) {
+ snprintf(adap->msix_info[msi_idx].desc, n, "%s-ofld%d",
+ adap->name, i);
+ adap->msix_info[msi_idx++].desc[n] = 0;
+ }
+ for_each_rdmarxq(&adap->sge, i) {
+ snprintf(adap->msix_info[msi_idx].desc, n, "%s-rdma%d",
+ adap->name, i);
+ adap->msix_info[msi_idx++].desc[n] = 0;
+ }
+}
+
+static int request_msix_queue_irqs(struct adapter *adap)
+{
+ struct sge *s = &adap->sge;
+ int err, ethqidx, ofldqidx = 0, rdmaqidx = 0, msi = 2;
+
+ err = request_irq(adap->msix_info[1].vec, t4_sge_intr_msix, 0,
+ adap->msix_info[1].desc, &s->fw_evtq);
+ if (err)
+ return err;
+
+ for_each_ethrxq(s, ethqidx) {
+ err = request_irq(adap->msix_info[msi].vec, t4_sge_intr_msix, 0,
+ adap->msix_info[msi].desc,
+ &s->ethrxq[ethqidx].rspq);
+ if (err)
+ goto unwind;
+ msi++;
+ }
+ for_each_ofldrxq(s, ofldqidx) {
+ err = request_irq(adap->msix_info[msi].vec, t4_sge_intr_msix, 0,
+ adap->msix_info[msi].desc,
+ &s->ofldrxq[ofldqidx].rspq);
+ if (err)
+ goto unwind;
+ msi++;
+ }
+ for_each_rdmarxq(s, rdmaqidx) {
+ err = request_irq(adap->msix_info[msi].vec, t4_sge_intr_msix, 0,
+ adap->msix_info[msi].desc,
+ &s->rdmarxq[rdmaqidx].rspq);
+ if (err)
+ goto unwind;
+ msi++;
+ }
+ return 0;
+
+unwind:
+ while (--rdmaqidx >= 0)
+ free_irq(adap->msix_info[--msi].vec,
+ &s->rdmarxq[rdmaqidx].rspq);
+ while (--ofldqidx >= 0)
+ free_irq(adap->msix_info[--msi].vec,
+ &s->ofldrxq[ofldqidx].rspq);
+ while (--ethqidx >= 0)
+ free_irq(adap->msix_info[--msi].vec, &s->ethrxq[ethqidx].rspq);
+ free_irq(adap->msix_info[1].vec, &s->fw_evtq);
+ return err;
+}
+
+static void free_msix_queue_irqs(struct adapter *adap)
+{
+ int i, msi = 2;
+ struct sge *s = &adap->sge;
+
+ free_irq(adap->msix_info[1].vec, &s->fw_evtq);
+ for_each_ethrxq(s, i)
+ free_irq(adap->msix_info[msi++].vec, &s->ethrxq[i].rspq);
+ for_each_ofldrxq(s, i)
+ free_irq(adap->msix_info[msi++].vec, &s->ofldrxq[i].rspq);
+ for_each_rdmarxq(s, i)
+ free_irq(adap->msix_info[msi++].vec, &s->rdmarxq[i].rspq);
+}
+
+/**
+ * setup_rss - configure RSS
+ * @adap: the adapter
+ *
+ * Sets up RSS to distribute packets to multiple receive queues. We
+ * configure the RSS CPU lookup table to distribute to the number of HW
+ * receive queues, and the response queue lookup table to narrow that
+ * down to the response queues actually configured for each port.
+ * We always configure the RSS mapping for all ports since the mapping
+ * table has plenty of entries.
+ */
+static int setup_rss(struct adapter *adap)
+{
+ int i, j, err;
+ u16 rss[MAX_ETH_QSETS];
+
+ for_each_port(adap, i) {
+ const struct port_info *pi = adap2pinfo(adap, i);
+ const struct sge_eth_rxq *q = &adap->sge.ethrxq[pi->first_qset];
+
+ for (j = 0; j < pi->nqsets; j++)
+ rss[j] = q[j].rspq.abs_id;
+
+ err = t4_config_rss_range(adap, 0, pi->viid, 0, pi->rss_size,
+ rss, pi->nqsets);
+ if (err)
+ return err;
+ }
+ return 0;
+}
+
+/*
+ * Wait until all NAPI handlers are descheduled.
+ */
+static void quiesce_rx(struct adapter *adap)
+{
+ int i;
+ struct sge *s = &adap->sge;
+
+ for_each_ethrxq(s, i)
+ napi_disable(&s->ethrxq[i].rspq.napi);
+
+ for_each_ofldrxq(s, i)
+ napi_disable(&s->ofldrxq[i].rspq.napi);
+
+ for_each_rdmarxq(s, i)
+ napi_disable(&s->rdmarxq[i].rspq.napi);
+
+ napi_disable(&s->fw_evtq.napi);
+}
+
+static void qenable(struct sge_rspq *q)
+{
+ napi_enable(&q->napi);
+ /* 0-increment GTS to start the timer and enable interrupts */
+ t4_write_reg(q->adap, MYPF_REG(A_SGE_PF_GTS),
+ V_SEINTARM(q->intr_params) | V_INGRESSQID(q->cntxt_id));
+}
+
+/*
+ * Enable NAPI scheduling and interrupt generation for all Rx queues.
+ */
+static void enable_rx(struct adapter *adap)
+{
+ int i;
+ struct sge *s = &adap->sge;
+
+ for_each_ethrxq(s, i)
+ qenable(&s->ethrxq[i].rspq);
+
+ for_each_ofldrxq(s, i)
+ qenable(&s->ofldrxq[i].rspq);
+
+ for_each_rdmarxq(s, i)
+ qenable(&s->rdmarxq[i].rspq);
+
+ qenable(&s->fw_evtq);
+
+ /* the interrupt queue doesn't use NAPI */
+ if (!(adap->flags & USING_MSIX))
+ t4_write_reg(adap, MYPF_REG(A_SGE_PF_GTS),
+ V_SEINTARM(s->intrq.intr_params) |
+ V_INGRESSQID(s->intrq.cntxt_id));
+}
+
+/**
+ * setup_sge_queues - configure SGE Tx/Rx/response queues
+ * @adap: the adapter
+ *
+ * Determines how many sets of SGE queues to use and initializes them.
+ * We support multiple queue sets per port if we have MSI-X, otherwise
+ * just one queue set per port.
+ */
+static int setup_sge_queues(struct adapter *adap)
+{
+ int err, msi_idx, i, j;
+ struct sge *s = &adap->sge;
+
+ bitmap_zero(s->starving_fl, MAX_EGRQ);
+ bitmap_zero(s->txq_maperr, MAX_EGRQ);
+
+ if (adap->flags & USING_MSIX)
+ msi_idx = 1; /* vector 0 is for non-queue interrupts */
+ else {
+ err = t4_sge_alloc_rxq(adap, &s->intrq, false, adap->port[0], 0,
+ NULL, NULL);
+ if (err)
+ return err;
+ msi_idx = -((int)s->intrq.abs_id + 1);
+ }
+
+ err = t4_sge_alloc_rxq(adap, &s->fw_evtq, true, adap->port[0],
+ msi_idx, NULL, fwevtq_handler);
+ if (err) {
+freeout: t4_free_sge_resources(adap);
+ return err;
+ }
+
+ for_each_port(adap, i) {
+ struct net_device *dev = adap->port[i];
+ struct port_info *pi = netdev_priv(dev);
+ struct sge_eth_rxq *q = &s->ethrxq[pi->first_qset];
+ struct sge_eth_txq *t = &s->ethtxq[pi->first_qset];
+
+ for (j = 0; j < pi->nqsets; j++, q++) {
+ if (msi_idx > 0)
+ msi_idx++;
+ err = t4_sge_alloc_rxq(adap, &q->rspq, false, dev,
+ msi_idx, &q->fl,
+ t4_ethrx_handler);
+ if (err)
+ goto freeout;
+ q->rspq.idx = j;
+ memset(&q->stats, 0, sizeof(q->stats));
+ }
+ for (j = 0; j < pi->nqsets; j++, t++) {
+ err = t4_sge_alloc_eth_txq(adap, t, dev,
+ netdev_get_tx_queue(dev, j),
+ s->fw_evtq.cntxt_id);
+ if (err)
+ goto freeout;
+ }
+ }
+
+ j = s->ofldqsets / adap->params.nports; /* ofld queues per channel */
+ for_each_ofldrxq(s, i) {
+ struct sge_ofld_rxq *q = &s->ofldrxq[i];
+ struct net_device *dev = adap->port[i / j];
+
+ if (msi_idx > 0)
+ msi_idx++;
+ err = t4_sge_alloc_rxq(adap, &q->rspq, false, dev, msi_idx,
+ &q->fl, uldrx_handler);
+ if (err)
+ goto freeout;
+ memset(&q->stats, 0, sizeof(q->stats));
+ s->ofld_rxq[i] = q->rspq.abs_id;
+ err = t4_sge_alloc_ofld_txq(adap, &s->ofldtxq[i], dev,
+ s->fw_evtq.cntxt_id);
+ if (err)
+ goto freeout;
+ }
+
+ for_each_rdmarxq(s, i) {
+ struct sge_ofld_rxq *q = &s->rdmarxq[i];
+
+ if (msi_idx > 0)
+ msi_idx++;
+ err = t4_sge_alloc_rxq(adap, &q->rspq, false, adap->port[i],
+ msi_idx, &q->fl, uldrx_handler);
+ if (err)
+ goto freeout;
+ memset(&q->stats, 0, sizeof(q->stats));
+ s->rdma_rxq[i] = q->rspq.abs_id;
+ }
+
+ for_each_port(adap, i) {
+ /*
+ * Note that ->rdmarxq[i].rspq.cntxt_id below is 0 if we don't
+ * have RDMA queues, and that's the right value.
+ */
+ err = t4_sge_alloc_ctrl_txq(adap, &s->ctrlq[i], adap->port[i],
+ s->fw_evtq.cntxt_id,
+ s->rdmarxq[i].rspq.cntxt_id);
+ if (err)
+ goto freeout;
+ }
+
+ t4_write_reg(adap, A_MPS_TRC_RSS_CONTROL,
+ V_RSSCONTROL(netdev2pinfo(adap->port[0])->tx_chan) |
+ V_QUEUENUMBER(s->ethrxq[0].rspq.abs_id));
+ return 0;
+}
+
+static int upgrade_fw(struct adapter *adap)
+{
+ int ret;
+ const struct firmware *fw;
+ struct device *dev = adap->pdev_dev;
+
+ ret = request_firmware(&fw, FW_FNAME, dev);
+ if (ret < 0) {
+ dev_err(dev, "unable to load firmware image " FW_FNAME
+ ", error %d\n", ret);
+ return ret;
+ }
+ ret = t4_load_fw(adap, fw->data, fw->size);
+ release_firmware(fw);
+ if (!ret)
+ dev_info(dev, "firmware upgraded to " FW_FNAME "\n");
+ return ret;
+}
+
+/*
+ * Allocate a chunk of memory using kmalloc or, if that fails, vmalloc.
+ * The allocated memory is cleared.
+ */
+void *t4_alloc_mem(size_t size)
+{
+ void *p = kmalloc(size, GFP_KERNEL);
+
+ if (!p)
+ p = vmalloc(size);
+ if (p)
+ memset(p, 0, size);
+ return p;
+}
+
+/*
+ * Free memory allocated through alloc_mem().
+ */
+void t4_free_mem(void *addr)
+{
+ if (is_vmalloc_addr(addr))
+ vfree(addr);
+ else
+ kfree(addr);
+}
+
+/*
+ * Implementation of ethtool operations.
+ */
+
+static u32 get_msglevel(struct net_device *dev)
+{
+ return netdev2adap(dev)->msg_enable;
+}
+
+static void set_msglevel(struct net_device *dev, u32 val)
+{
+ netdev2adap(dev)->msg_enable = val;
+}
+
+static char stats_strings[][ETH_GSTRING_LEN] = {
+ "TxOctetsOK ",
+ "TxFramesOK ",
+ "TxBroadcastFrames ",
+ "TxMulticastFrames ",
+ "TxUnicastFrames ",
+ "TxErrorFrames ",
+
+ "TxFrames64 ",
+ "TxFrames65To127 ",
+ "TxFrames128To255 ",
+ "TxFrames256To511 ",
+ "TxFrames512To1023 ",
+ "TxFrames1024To1518 ",
+ "TxFrames1519ToMax ",
+
+ "TxFramesDropped ",
+ "TxPauseFrames ",
+ "TxPPP0Frames ",
+ "TxPPP1Frames ",
+ "TxPPP2Frames ",
+ "TxPPP3Frames ",
+ "TxPPP4Frames ",
+ "TxPPP5Frames ",
+ "TxPPP6Frames ",
+ "TxPPP7Frames ",
+
+ "RxOctetsOK ",
+ "RxFramesOK ",
+ "RxBroadcastFrames ",
+ "RxMulticastFrames ",
+ "RxUnicastFrames ",
+
+ "RxFramesTooLong ",
+ "RxJabberErrors ",
+ "RxFCSErrors ",
+ "RxLengthErrors ",
+ "RxSymbolErrors ",
+ "RxRuntFrames ",
+
+ "RxFrames64 ",
+ "RxFrames65To127 ",
+ "RxFrames128To255 ",
+ "RxFrames256To511 ",
+ "RxFrames512To1023 ",
+ "RxFrames1024To1518 ",
+ "RxFrames1519ToMax ",
+
+ "RxPauseFrames ",
+ "RxPPP0Frames ",
+ "RxPPP1Frames ",
+ "RxPPP2Frames ",
+ "RxPPP3Frames ",
+ "RxPPP4Frames ",
+ "RxPPP5Frames ",
+ "RxPPP6Frames ",
+ "RxPPP7Frames ",
+
+ "RxBG0FramesDropped ",
+ "RxBG1FramesDropped ",
+ "RxBG2FramesDropped ",
+ "RxBG3FramesDropped ",
+ "RxBG0FramesTrunc ",
+ "RxBG1FramesTrunc ",
+ "RxBG2FramesTrunc ",
+ "RxBG3FramesTrunc ",
+
+ "TSO ",
+ "TxCsumOffload ",
+ "RxCsumGood ",
+ "VLANextractions ",
+ "VLANinsertions ",
+};
+
+static int get_sset_count(struct net_device *dev, int sset)
+{
+ switch (sset) {
+ case ETH_SS_STATS:
+ return ARRAY_SIZE(stats_strings);
+ default:
+ return -EOPNOTSUPP;
+ }
+}
+
+#define T4_REGMAP_SIZE (160 * 1024)
+
+static int get_regs_len(struct net_device *dev)
+{
+ return T4_REGMAP_SIZE;
+}
+
+static int get_eeprom_len(struct net_device *dev)
+{
+ return EEPROMSIZE;
+}
+
+static void get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
+{
+ struct adapter *adapter = netdev2adap(dev);
+
+ strcpy(info->driver, KBUILD_MODNAME);
+ strcpy(info->version, DRV_VERSION);
+ strcpy(info->bus_info, pci_name(adapter->pdev));
+
+ if (!adapter->params.fw_vers)
+ strcpy(info->fw_version, "N/A");
+ else
+ snprintf(info->fw_version, sizeof(info->fw_version),
+ "%u.%u.%u.%u, TP %u.%u.%u.%u",
+ G_FW_HDR_FW_VER_MAJOR(adapter->params.fw_vers),
+ G_FW_HDR_FW_VER_MINOR(adapter->params.fw_vers),
+ G_FW_HDR_FW_VER_MICRO(adapter->params.fw_vers),
+ G_FW_HDR_FW_VER_BUILD(adapter->params.fw_vers),
+ G_FW_HDR_FW_VER_MAJOR(adapter->params.tp_vers),
+ G_FW_HDR_FW_VER_MINOR(adapter->params.tp_vers),
+ G_FW_HDR_FW_VER_MICRO(adapter->params.tp_vers),
+ G_FW_HDR_FW_VER_BUILD(adapter->params.tp_vers));
+}
+
+static void get_strings(struct net_device *dev, u32 stringset, u8 *data)
+{
+ if (stringset == ETH_SS_STATS)
+ memcpy(data, stats_strings, sizeof(stats_strings));
+}
+
+/*
+ * port stats maintained per queue of the port. They should be in the same
+ * order as in stats_strings above.
+ */
+struct queue_port_stats {
+ u64 tso;
+ u64 tx_csum;
+ u64 rx_csum;
+ u64 vlan_ex;
+ u64 vlan_ins;
+};
+
+static void collect_sge_port_stats(const struct adapter *adap,
+ const struct port_info *p, struct queue_port_stats *s)
+{
+ int i;
+ const struct sge_eth_txq *tx = &adap->sge.ethtxq[p->first_qset];
+ const struct sge_eth_rxq *rx = &adap->sge.ethrxq[p->first_qset];
+
+ memset(s, 0, sizeof(*s));
+ for (i = 0; i < p->nqsets; i++, rx++, tx++) {
+ s->tso += tx->tso;
+ s->tx_csum += tx->tx_cso;
+ s->rx_csum += rx->stats.rx_cso;
+ s->vlan_ex += rx->stats.vlan_ex;
+ s->vlan_ins += tx->vlan_ins;
+ }
+}
+
+static void get_stats(struct net_device *dev, struct ethtool_stats *stats,
+ u64 *data)
+{
+ struct port_info *pi = netdev_priv(dev);
+ struct adapter *adapter = pi->adapter;
+
+ t4_get_port_stats(adapter, pi->tx_chan, (struct port_stats *)data);
+
+ data += sizeof(struct port_stats) / sizeof(u64);
+ collect_sge_port_stats(adapter, pi, (struct queue_port_stats *)data);
+}
+
+/*
+ * Return a version number to identify the type of adapter. The scheme is:
+ * - bits 0..9: chip version
+ * - bits 10..15: chip revision
+ */
+static inline unsigned int mk_adap_vers(const struct adapter *ap)
+{
+ return 4 | (ap->params.rev << 10);
+}
+
+static void reg_block_dump(struct adapter *ap, void *buf, unsigned int start,
+ unsigned int end)
+{
+ u32 *p = buf + start;
+
+ for ( ; start <= end; start += sizeof(u32))
+ *p++ = t4_read_reg(ap, start);
+}
+
+static void get_regs(struct net_device *dev, struct ethtool_regs *regs,
+ void *buf)
+{
+ static const unsigned int reg_ranges[] = {
+ 0x1008, 0x1108,
+ 0x1180, 0x11b4,
+ 0x11fc, 0x123c,
+ 0x1300, 0x173c,
+ 0x1800, 0x18fc,
+ 0x3000, 0x30d8,
+ 0x30e0, 0x5924,
+ 0x5960, 0x59d4,
+ 0x5a00, 0x5af8,
+ 0x6000, 0x6098,
+ 0x6100, 0x6150,
+ 0x6200, 0x6208,
+ 0x6240, 0x6248,
+ 0x6280, 0x6338,
+ 0x6370, 0x638c,
+ 0x6400, 0x643c,
+ 0x6500, 0x6524,
+ 0x6a00, 0x6a38,
+ 0x6a60, 0x6a78,
+ 0x6b00, 0x6b84,
+ 0x6bf0, 0x6c84,
+ 0x6cf0, 0x6d84,
+ 0x6df0, 0x6e84,
+ 0x6ef0, 0x6f84,
+ 0x6ff0, 0x7084,
+ 0x70f0, 0x7184,
+ 0x71f0, 0x7284,
+ 0x72f0, 0x7384,
+ 0x73f0, 0x7450,
+ 0x7500, 0x7530,
+ 0x7600, 0x761c,
+ 0x7680, 0x76cc,
+ 0x7700, 0x7798,
+ 0x77c0, 0x77fc,
+ 0x7900, 0x79fc,
+ 0x7b00, 0x7c38,
+ 0x7d00, 0x7efc,
+ 0x8dc0, 0x8e1c,
+ 0x8e30, 0x8e78,
+ 0x8ea0, 0x8f6c,
+ 0x8fc0, 0x9074,
+ 0x90fc, 0x90fc,
+ 0x9400, 0x9458,
+ 0x9600, 0x96bc,
+ 0x9800, 0x9808,
+ 0x9820, 0x983c,
+ 0x9850, 0x9864,
+ 0x9c00, 0x9c6c,
+ 0x9c80, 0x9cec,
+ 0x9d00, 0x9d6c,
+ 0x9d80, 0x9dec,
+ 0x9e00, 0x9e6c,
+ 0x9e80, 0x9eec,
+ 0x9f00, 0x9f6c,
+ 0x9f80, 0x9fec,
+ 0xd004, 0xd03c,
+ 0xdfc0, 0xdfe0,
+ 0xe000, 0xea7c,
+ 0xf000, 0x11190,
+ 0x19040, 0x19124,
+ 0x19150, 0x191b0,
+ 0x191d0, 0x191e8,
+ 0x19238, 0x1924c,
+ 0x193f8, 0x19474,
+ 0x19490, 0x194f8,
+ 0x19800, 0x19f30,
+ 0x1a000, 0x1a06c,
+ 0x1a0b0, 0x1a120,
+ 0x1a128, 0x1a138,
+ 0x1a190, 0x1a1c4,
+ 0x1a1fc, 0x1a1fc,
+ 0x1e040, 0x1e04c,
+ 0x1e240, 0x1e28c,
+ 0x1e2c0, 0x1e2c0,
+ 0x1e2e0, 0x1e2e0,
+ 0x1e300, 0x1e384,
+ 0x1e3c0, 0x1e3c8,
+ 0x1e440, 0x1e44c,
+ 0x1e640, 0x1e68c,
+ 0x1e6c0, 0x1e6c0,
+ 0x1e6e0, 0x1e6e0,
+ 0x1e700, 0x1e784,
+ 0x1e7c0, 0x1e7c8,
+ 0x1e840, 0x1e84c,
+ 0x1ea40, 0x1ea8c,
+ 0x1eac0, 0x1eac0,
+ 0x1eae0, 0x1eae0,
+ 0x1eb00, 0x1eb84,
+ 0x1ebc0, 0x1ebc8,
+ 0x1ec40, 0x1ec4c,
+ 0x1ee40, 0x1ee8c,
+ 0x1eec0, 0x1eec0,
+ 0x1eee0, 0x1eee0,
+ 0x1ef00, 0x1ef84,
+ 0x1efc0, 0x1efc8,
+ 0x1f040, 0x1f04c,
+ 0x1f240, 0x1f28c,
+ 0x1f2c0, 0x1f2c0,
+ 0x1f2e0, 0x1f2e0,
+ 0x1f300, 0x1f384,
+ 0x1f3c0, 0x1f3c8,
+ 0x1f440, 0x1f44c,
+ 0x1f640, 0x1f68c,
+ 0x1f6c0, 0x1f6c0,
+ 0x1f6e0, 0x1f6e0,
+ 0x1f700, 0x1f784,
+ 0x1f7c0, 0x1f7c8,
+ 0x1f840, 0x1f84c,
+ 0x1fa40, 0x1fa8c,
+ 0x1fac0, 0x1fac0,
+ 0x1fae0, 0x1fae0,
+ 0x1fb00, 0x1fb84,
+ 0x1fbc0, 0x1fbc8,
+ 0x1fc40, 0x1fc4c,
+ 0x1fe40, 0x1fe8c,
+ 0x1fec0, 0x1fec0,
+ 0x1fee0, 0x1fee0,
+ 0x1ff00, 0x1ff84,
+ 0x1ffc0, 0x1ffc8,
+ 0x20000, 0x2002c,
+ 0x20100, 0x2013c,
+ 0x20190, 0x201c4,
+ 0x20200, 0x20318,
+ 0x20400, 0x20528,
+ 0x20540, 0x20614,
+ 0x21000, 0x21040,
+ 0x2104c, 0x21060,
+ 0x210c0, 0x210ec,
+ 0x21200, 0x21268,
+ 0x21270, 0x21284,
+ 0x212fc, 0x21388,
+ 0x21400, 0x21404,
+ 0x21500, 0x21518,
+ 0x2152c, 0x2153c,
+ 0x21550, 0x21554,
+ 0x21600, 0x21600,
+ 0x21608, 0x21628,
+ 0x21630, 0x2163c,
+ 0x21700, 0x2171c,
+ 0x21780, 0x2178c,
+ 0x21800, 0x21c38,
+ 0x21c80, 0x21d7c,
+ 0x21e00, 0x21e04,
+ 0x22000, 0x2202c,
+ 0x22100, 0x2213c,
+ 0x22190, 0x221c4,
+ 0x22200, 0x22318,
+ 0x22400, 0x22528,
+ 0x22540, 0x22614,
+ 0x23000, 0x23040,
+ 0x2304c, 0x23060,
+ 0x230c0, 0x230ec,
+ 0x23200, 0x23268,
+ 0x23270, 0x23284,
+ 0x232fc, 0x23388,
+ 0x23400, 0x23404,
+ 0x23500, 0x23518,
+ 0x2352c, 0x2353c,
+ 0x23550, 0x23554,
+ 0x23600, 0x23600,
+ 0x23608, 0x23628,
+ 0x23630, 0x2363c,
+ 0x23700, 0x2371c,
+ 0x23780, 0x2378c,
+ 0x23800, 0x23c38,
+ 0x23c80, 0x23d7c,
+ 0x23e00, 0x23e04,
+ 0x24000, 0x2402c,
+ 0x24100, 0x2413c,
+ 0x24190, 0x241c4,
+ 0x24200, 0x24318,
+ 0x24400, 0x24528,
+ 0x24540, 0x24614,
+ 0x25000, 0x25040,
+ 0x2504c, 0x25060,
+ 0x250c0, 0x250ec,
+ 0x25200, 0x25268,
+ 0x25270, 0x25284,
+ 0x252fc, 0x25388,
+ 0x25400, 0x25404,
+ 0x25500, 0x25518,
+ 0x2552c, 0x2553c,
+ 0x25550, 0x25554,
+ 0x25600, 0x25600,
+ 0x25608, 0x25628,
+ 0x25630, 0x2563c,
+ 0x25700, 0x2571c,
+ 0x25780, 0x2578c,
+ 0x25800, 0x25c38,
+ 0x25c80, 0x25d7c,
+ 0x25e00, 0x25e04,
+ 0x26000, 0x2602c,
+ 0x26100, 0x2613c,
+ 0x26190, 0x261c4,
+ 0x26200, 0x26318,
+ 0x26400, 0x26528,
+ 0x26540, 0x26614,
+ 0x27000, 0x27040,
+ 0x2704c, 0x27060,
+ 0x270c0, 0x270ec,
+ 0x27200, 0x27268,
+ 0x27270, 0x27284,
+ 0x272fc, 0x27388,
+ 0x27400, 0x27404,
+ 0x27500, 0x27518,
+ 0x2752c, 0x2753c,
+ 0x27550, 0x27554,
+ 0x27600, 0x27600,
+ 0x27608, 0x27628,
+ 0x27630, 0x2763c,
+ 0x27700, 0x2771c,
+ 0x27780, 0x2778c,
+ 0x27800, 0x27c38,
+ 0x27c80, 0x27d7c,
+ 0x27e00, 0x27e04
+ };
+
+ int i;
+ struct adapter *ap = netdev2adap(dev);
+
+ regs->version = mk_adap_vers(ap);
+
+ memset(buf, 0, T4_REGMAP_SIZE);
+ for (i = 0; i < ARRAY_SIZE(reg_ranges); i += 2)
+ reg_block_dump(ap, buf, reg_ranges[i], reg_ranges[i + 1]);
+}
+
+static int restart_autoneg(struct net_device *dev)
+{
+ struct port_info *p = netdev_priv(dev);
+
+ if (!netif_running(dev))
+ return -EAGAIN;
+ if (p->link_cfg.autoneg != AUTONEG_ENABLE)
+ return -EINVAL;
+ t4_restart_aneg(p->adapter, 0, p->tx_chan);
+ return 0;
+}
+
+static int identify_port(struct net_device *dev, u32 data)
+{
+ if (data == 0)
+ data = 2; /* default to 2 seconds */
+
+ return t4_identify_port(netdev2adap(dev), 0, netdev2pinfo(dev)->viid,
+ data * 5);
+}
+
+static unsigned int from_fw_linkcaps(unsigned int type, unsigned int caps)
+{
+ unsigned int v = 0;
+
+ if (type == FW_PORT_TYPE_BT_SGMII || type == FW_PORT_TYPE_BT_XAUI) {
+ v |= SUPPORTED_TP;
+ if (caps & FW_PORT_CAP_SPEED_100M)
+ v |= SUPPORTED_100baseT_Full;
+ if (caps & FW_PORT_CAP_SPEED_1G)
+ v |= SUPPORTED_1000baseT_Full;
+ if (caps & FW_PORT_CAP_SPEED_10G)
+ v |= SUPPORTED_10000baseT_Full;
+ } else if (type == FW_PORT_TYPE_KX4 || type == FW_PORT_TYPE_KX) {
+ v |= SUPPORTED_Backplane;
+ if (caps & FW_PORT_CAP_SPEED_1G)
+ v |= SUPPORTED_1000baseKX_Full;
+ if (caps & FW_PORT_CAP_SPEED_10G)
+ v |= SUPPORTED_10000baseKX4_Full;
+ } else if (type == FW_PORT_TYPE_KR)
+ v |= SUPPORTED_Backplane | SUPPORTED_10000baseKR_Full;
+ else if (type == FW_PORT_TYPE_FIBER)
+ v |= SUPPORTED_FIBRE;
+
+ if (caps & FW_PORT_CAP_ANEG)
+ v |= SUPPORTED_Autoneg;
+ return v;
+}
+
+static unsigned int to_fw_linkcaps(unsigned int caps)
+{
+ unsigned int v = 0;
+
+ if (caps & ADVERTISED_100baseT_Full)
+ v |= FW_PORT_CAP_SPEED_100M;
+ if (caps & ADVERTISED_1000baseT_Full)
+ v |= FW_PORT_CAP_SPEED_1G;
+ if (caps & ADVERTISED_10000baseT_Full)
+ v |= FW_PORT_CAP_SPEED_10G;
+ return v;
+}
+
+static int get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+ const struct port_info *p = netdev_priv(dev);
+
+ if (p->port_type == FW_PORT_TYPE_BT_SGMII ||
+ p->port_type == FW_PORT_TYPE_BT_XAUI)
+ cmd->port = PORT_TP;
+ else if (p->port_type == FW_PORT_TYPE_FIBER)
+ cmd->port = PORT_FIBRE;
+ else if (p->port_type == FW_PORT_TYPE_TWINAX)
+ cmd->port = PORT_DA;
+ else
+ cmd->port = PORT_OTHER;
+
+ if (p->mdio_addr >= 0) {
+ cmd->phy_address = p->mdio_addr;
+ cmd->transceiver = XCVR_EXTERNAL;
+ cmd->mdio_support = p->port_type == FW_PORT_TYPE_BT_SGMII ?
+ MDIO_SUPPORTS_C22 : MDIO_SUPPORTS_C45;
+ } else {
+ cmd->phy_address = 0; /* not really, but no better option */
+ cmd->transceiver = XCVR_INTERNAL;
+ cmd->mdio_support = 0;
+ }
+
+ cmd->supported = from_fw_linkcaps(p->port_type, p->link_cfg.supported);
+ cmd->advertising = from_fw_linkcaps(p->port_type,
+ p->link_cfg.advertising);
+ cmd->speed = netif_carrier_ok(dev) ? p->link_cfg.speed : 0;
+ cmd->duplex = DUPLEX_FULL;
+ cmd->autoneg = p->link_cfg.autoneg;
+ cmd->maxtxpkt = 0;
+ cmd->maxrxpkt = 0;
+ return 0;
+}
+
+static unsigned int speed_to_caps(int speed)
+{
+ if (speed == SPEED_100)
+ return FW_PORT_CAP_SPEED_100M;
+ if (speed == SPEED_1000)
+ return FW_PORT_CAP_SPEED_1G;
+ if (speed == SPEED_10000)
+ return FW_PORT_CAP_SPEED_10G;
+ return 0;
+}
+
+static int set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+ unsigned int cap;
+ struct port_info *p = netdev_priv(dev);
+ struct link_config *lc = &p->link_cfg;
+
+ if (cmd->duplex != DUPLEX_FULL) /* only full-duplex supported */
+ return -EINVAL;
+
+ if (!(lc->supported & FW_PORT_CAP_ANEG)) {
+ /*
+ * PHY offers a single speed. See if that's what's
+ * being requested.
+ */
+ if (cmd->autoneg == AUTONEG_DISABLE &&
+ (lc->supported & speed_to_caps(cmd->speed)))
+ return 0;
+ return -EINVAL;
+ }
+
+ if (cmd->autoneg == AUTONEG_DISABLE) {
+ cap = speed_to_caps(cmd->speed);
+
+ if (!(lc->supported & cap) || cmd->speed == SPEED_1000 ||
+ cmd->speed == SPEED_10000)
+ return -EINVAL;
+ lc->requested_speed = cap;
+ lc->advertising = 0;
+ } else {
+ cap = to_fw_linkcaps(cmd->advertising);
+ if (!(lc->supported & cap))
+ return -EINVAL;
+ lc->requested_speed = 0;
+ lc->advertising = cap | FW_PORT_CAP_ANEG;
+ }
+ lc->autoneg = cmd->autoneg;
+
+ if (netif_running(dev))
+ return t4_link_start(p->adapter, 0, p->tx_chan, lc);
+ return 0;
+}
+
+static void get_pauseparam(struct net_device *dev,
+ struct ethtool_pauseparam *epause)
+{
+ struct port_info *p = netdev_priv(dev);
+
+ epause->autoneg = (p->link_cfg.requested_fc & PAUSE_AUTONEG) != 0;
+ epause->rx_pause = (p->link_cfg.fc & PAUSE_RX) != 0;
+ epause->tx_pause = (p->link_cfg.fc & PAUSE_TX) != 0;
+}
+
+static int set_pauseparam(struct net_device *dev,
+ struct ethtool_pauseparam *epause)
+{
+ struct port_info *p = netdev_priv(dev);
+ struct link_config *lc = &p->link_cfg;
+
+ if (epause->autoneg == AUTONEG_DISABLE)
+ lc->requested_fc = 0;
+ else if (lc->supported & FW_PORT_CAP_ANEG)
+ lc->requested_fc = PAUSE_AUTONEG;
+ else
+ return -EINVAL;
+
+ if (epause->rx_pause)
+ lc->requested_fc |= PAUSE_RX;
+ if (epause->tx_pause)
+ lc->requested_fc |= PAUSE_TX;
+ if (netif_running(dev))
+ return t4_link_start(p->adapter, 0, p->tx_chan, lc);
+ return 0;
+}
+
+static u32 get_rx_csum(struct net_device *dev)
+{
+ struct port_info *p = netdev_priv(dev);
+
+ return p->rx_offload & RX_CSO;
+}
+
+static int set_rx_csum(struct net_device *dev, u32 data)
+{
+ struct port_info *p = netdev_priv(dev);
+
+ if (data)
+ p->rx_offload |= RX_CSO;
+ else
+ p->rx_offload &= ~RX_CSO;
+ return 0;
+}
+
+static void get_sge_param(struct net_device *dev, struct ethtool_ringparam *e)
+{
+ const struct port_info *pi = netdev_priv(dev);
+ const struct sge *s = &pi->adapter->sge;
+
+ e->rx_max_pending = MAX_RX_BUFFERS;
+ e->rx_mini_max_pending = MAX_RSPQ_ENTRIES;
+ e->rx_jumbo_max_pending = 0;
+ e->tx_max_pending = MAX_TXQ_ENTRIES;
+
+ e->rx_pending = s->ethrxq[pi->first_qset].fl.size - 8;
+ e->rx_mini_pending = s->ethrxq[pi->first_qset].rspq.size;
+ e->rx_jumbo_pending = 0;
+ e->tx_pending = s->ethtxq[pi->first_qset].q.size;
+}
+
+static int set_sge_param(struct net_device *dev, struct ethtool_ringparam *e)
+{
+ int i;
+ const struct port_info *pi = netdev_priv(dev);
+ struct adapter *adapter = pi->adapter;
+ struct sge *s = &adapter->sge;
+
+ if (e->rx_pending > MAX_RX_BUFFERS || e->rx_jumbo_pending ||
+ e->tx_pending > MAX_TXQ_ENTRIES ||
+ e->rx_mini_pending > MAX_RSPQ_ENTRIES ||
+ e->rx_mini_pending < MIN_RSPQ_ENTRIES ||
+ e->rx_pending < MIN_FL_ENTRIES || e->tx_pending < MIN_TXQ_ENTRIES)
+ return -EINVAL;
+
+ if (adapter->flags & FULL_INIT_DONE)
+ return -EBUSY;
+
+ for (i = 0; i < pi->nqsets; ++i) {
+ s->ethtxq[pi->first_qset + i].q.size = e->tx_pending;
+ s->ethrxq[pi->first_qset + i].fl.size = e->rx_pending + 8;
+ s->ethrxq[pi->first_qset + i].rspq.size = e->rx_mini_pending;
+ }
+ return 0;
+}
+
+static int closest_timer(const struct sge *s, int time)
+{
+ int i, delta, match = 0, min_delta = INT_MAX;
+
+ for (i = 0; i < ARRAY_SIZE(s->timer_val); i++) {
+ delta = time - s->timer_val[i];
+ if (delta < 0)
+ delta = -delta;
+ if (delta < min_delta) {
+ min_delta = delta;
+ match = i;
+ }
+ }
+ return match;
+}
+
+static int closest_thres(const struct sge *s, int thres)
+{
+ int i, delta, match = 0, min_delta = INT_MAX;
+
+ for (i = 0; i < ARRAY_SIZE(s->counter_val); i++) {
+ delta = thres - s->counter_val[i];
+ if (delta < 0)
+ delta = -delta;
+ if (delta < min_delta) {
+ min_delta = delta;
+ match = i;
+ }
+ }
+ return match;
+}
+
+/*
+ * Return a queue's interrupt hold-off time in us. 0 means no timer.
+ */
+static unsigned int qtimer_val(const struct adapter *adap,
+ const struct sge_rspq *q)
+{
+ unsigned int idx = G_QINTR_TIMER_IDX(q->intr_params);
+
+ return idx < SGE_NTIMERS ? adap->sge.timer_val[idx] : 0;
+}
+
+/**
+ * set_rxq_intr_params - set a queue's interrupt holdoff parameters
+ * @adap: the adapter
+ * @q: the Rx queue
+ * @us: the hold-off time in us, or 0 to disable timer
+ * @cnt: the hold-off packet count, or 0 to disable counter
+ *
+ * Sets an Rx queue's interrupt hold-off time and packet count. At least
+ * one of the two needs to be enabled for the queue to generate interrupts.
+ */
+static int set_rxq_intr_params(struct adapter *adap, struct sge_rspq *q,
+ unsigned int us, unsigned int cnt)
+{
+ if ((us | cnt) == 0)
+ cnt = 1;
+
+ if (cnt) {
+ int err;
+ u32 v, new_idx;
+
+ new_idx = closest_thres(&adap->sge, cnt);
+ if (q->desc && q->pktcnt_idx != new_idx) {
+ /* the queue has already been created, update it */
+ v = V_FW_PARAMS_MNEM(FW_PARAMS_MNEM_DMAQ) |
+ V_FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_DMAQ_IQ_INTCNTTHRESH) |
+ V_FW_PARAMS_PARAM_YZ(q->cntxt_id);
+ err = t4_set_params(adap, 0, 1, &v, &new_idx);
+ if (err)
+ return err;
+ }
+ q->pktcnt_idx = new_idx;
+ }
+
+ us = us == 0 ? E_TIMERIDX_NONE : closest_timer(&adap->sge, us);
+ q->intr_params = V_QINTR_TIMER_IDX(us) | V_QINTR_CNT_EN(cnt > 0);
+ return 0;
+}
+
+static int set_coalesce(struct net_device *dev, struct ethtool_coalesce *c)
+{
+ const struct port_info *pi = netdev_priv(dev);
+ struct adapter *adap = pi->adapter;
+
+ return set_rxq_intr_params(adap, &adap->sge.ethrxq[pi->first_qset].rspq,
+ c->rx_coalesce_usecs, c->rx_max_coalesced_frames);
+}
+
+static int get_coalesce(struct net_device *dev, struct ethtool_coalesce *c)
+{
+ const struct port_info *pi = netdev_priv(dev);
+ const struct adapter *adap = pi->adapter;
+ const struct sge_rspq *rq = &adap->sge.ethrxq[pi->first_qset].rspq;
+
+ c->rx_coalesce_usecs = qtimer_val(adap, rq);
+ c->rx_max_coalesced_frames = (rq->intr_params & F_QINTR_CNT_EN) ?
+ adap->sge.counter_val[rq->pktcnt_idx] : 0;
+ return 0;
+}
+
+/*
+ * Translate a physical EEPROM address to virtual. The first 1K is accessed
+ * through virtual addresses starting at 31K, the rest is accessed through
+ * virtual addresses starting at 0. This mapping is correct only for PF0.
+ */
+static int eeprom_ptov(unsigned int phys_addr)
+{
+ if (phys_addr < 1024)
+ return phys_addr + (31 << 10);
+ if (phys_addr < EEPROMSIZE)
+ return phys_addr - 1024;
+ return -EINVAL;
+}
+
+/*
+ * The next two routines implement eeprom read/write from physical addresses.
+ * The physical->virtual translation is correct only for PF0.
+ */
+static int eeprom_rd_phys(struct adapter *adap, unsigned int phys_addr, u32 *v)
+{
+ int vaddr = eeprom_ptov(phys_addr);
+
+ if (vaddr >= 0)
+ vaddr = pci_read_vpd(adap->pdev, vaddr, sizeof(u32), v);
+ return vaddr < 0 ? vaddr : 0;
+}
+
+static int eeprom_wr_phys(struct adapter *adap, unsigned int phys_addr, u32 v)
+{
+ int vaddr = eeprom_ptov(phys_addr);
+
+ if (vaddr >= 0)
+ vaddr = pci_write_vpd(adap->pdev, vaddr, sizeof(u32), &v);
+ return vaddr < 0 ? vaddr : 0;
+}
+
+#define EEPROM_MAGIC 0x38E2F10C
+
+static int get_eeprom(struct net_device *dev, struct ethtool_eeprom *e,
+ u8 *data)
+{
+ int i, err = 0;
+ struct adapter *adapter = netdev2adap(dev);
+
+ u8 *buf = kmalloc(EEPROMSIZE, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+
+ e->magic = EEPROM_MAGIC;
+ for (i = e->offset & ~3; !err && i < e->offset + e->len; i += 4)
+ err = eeprom_rd_phys(adapter, i, (u32 *)&buf[i]);
+
+ if (!err)
+ memcpy(data, buf + e->offset, e->len);
+ kfree(buf);
+ return err;
+}
+
+static int set_eeprom(struct net_device *dev, struct ethtool_eeprom *eeprom,
+ u8 *data)
+{
+ u8 *buf;
+ int err = 0;
+ u32 aligned_offset, aligned_len, *p;
+ struct adapter *adapter = netdev2adap(dev);
+
+ if (eeprom->magic != EEPROM_MAGIC)
+ return -EINVAL;
+
+ aligned_offset = eeprom->offset & ~3;
+ aligned_len = (eeprom->len + (eeprom->offset & 3) + 3) & ~3;
+
+ if (aligned_offset != eeprom->offset || aligned_len != eeprom->len) {
+ /*
+ * RMW possibly needed for first or last words.
+ */
+ buf = kmalloc(aligned_len, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ err = eeprom_rd_phys(adapter, aligned_offset, (u32 *)buf);
+ if (!err && aligned_len > 4)
+ err = eeprom_rd_phys(adapter,
+ aligned_offset + aligned_len - 4,
+ (u32 *)&buf[aligned_len - 4]);
+ if (err)
+ goto out;
+ memcpy(buf + (eeprom->offset & 3), data, eeprom->len);
+ } else
+ buf = data;
+
+ err = t4_seeprom_wp(adapter, false);
+ if (err)
+ goto out;
+
+ for (p = (u32 *)buf; !err && aligned_len; aligned_len -= 4, p++) {
+ err = eeprom_wr_phys(adapter, aligned_offset, *p);
+ aligned_offset += 4;
+ }
+
+ if (!err)
+ err = t4_seeprom_wp(adapter, true);
+out:
+ if (buf != data)
+ kfree(buf);
+ return err;
+}
+
+static int set_flash(struct net_device *netdev, struct ethtool_flash *ef)
+{
+ int ret;
+ const struct firmware *fw;
+ struct adapter *adap = netdev2adap(netdev);
+
+ ef->data[sizeof(ef->data) - 1] = '\0';
+ ret = request_firmware(&fw, ef->data, adap->pdev_dev);
+ if (ret < 0)
+ return ret;
+
+ ret = t4_load_fw(adap, fw->data, fw->size);
+ release_firmware(fw);
+ if (!ret)
+ dev_info(adap->pdev_dev, "loaded firmware %s\n", ef->data);
+ return ret;
+}
+
+#define WOL_SUPPORTED (WAKE_BCAST | WAKE_MAGIC)
+#define BCAST_CRC 0xa0ccc1a6
+
+static void get_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
+{
+ wol->supported = WAKE_BCAST | WAKE_MAGIC;
+ wol->wolopts = netdev2adap(dev)->wol;
+ memset(&wol->sopass, 0, sizeof(wol->sopass));
+}
+
+static int set_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
+{
+ int err = 0;
+ struct port_info *pi = netdev_priv(dev);
+
+ if (wol->wolopts & ~WOL_SUPPORTED)
+ return -EINVAL;
+ t4_wol_magic_enable(pi->adapter, pi->tx_chan,
+ (wol->wolopts & WAKE_MAGIC) ? dev->dev_addr : NULL);
+ if (wol->wolopts & WAKE_BCAST) {
+ err = t4_wol_pat_enable(pi->adapter, pi->tx_chan, 0xfe, ~0ULL,
+ ~0ULL, 0, false);
+ if (!err)
+ err = t4_wol_pat_enable(pi->adapter, pi->tx_chan, 1,
+ ~6ULL, ~0ULL, BCAST_CRC, true);
+ } else
+ t4_wol_pat_enable(pi->adapter, pi->tx_chan, 0, 0, 0, 0, false);
+ return err;
+}
+
+static int set_tso(struct net_device *dev, u32 value)
+{
+ if (value)
+ dev->features |= NETIF_F_TSO | NETIF_F_TSO6;
+ else
+ dev->features &= ~(NETIF_F_TSO | NETIF_F_TSO6);
+ return 0;
+}
+
+static struct ethtool_ops cxgb_ethtool_ops = {
+ .get_settings = get_settings,
+ .set_settings = set_settings,
+ .get_drvinfo = get_drvinfo,
+ .get_msglevel = get_msglevel,
+ .set_msglevel = set_msglevel,
+ .get_ringparam = get_sge_param,
+ .set_ringparam = set_sge_param,
+ .get_coalesce = get_coalesce,
+ .set_coalesce = set_coalesce,
+ .get_eeprom_len = get_eeprom_len,
+ .get_eeprom = get_eeprom,
+ .set_eeprom = set_eeprom,
+ .get_pauseparam = get_pauseparam,
+ .set_pauseparam = set_pauseparam,
+ .get_rx_csum = get_rx_csum,
+ .set_rx_csum = set_rx_csum,
+ .set_tx_csum = ethtool_op_set_tx_ipv6_csum,
+ .set_sg = ethtool_op_set_sg,
+ .get_link = ethtool_op_get_link,
+ .get_strings = get_strings,
+ .phys_id = identify_port,
+ .nway_reset = restart_autoneg,
+ .get_sset_count = get_sset_count,
+ .get_ethtool_stats = get_stats,
+ .get_regs_len = get_regs_len,
+ .get_regs = get_regs,
+ .get_wol = get_wol,
+ .set_wol = set_wol,
+ .set_tso = set_tso,
+ .flash_device = set_flash,
+};
+
+/*
+ * Read a vector of numbers from a user space buffer. Each number must be
+ * between min and max inclusive and in the given base.
+ */
+static int rd_usr_int_vec(const char __user *buf, size_t usr_len, int vec_len,
+ unsigned long *vals, unsigned long min,
+ unsigned long max, int base)
+{
+ size_t l;
+ unsigned long v;
+ char c, word[68], *end;
+
+ while (usr_len) {
+ /* skip whitespace to beginning of next word */
+ while (usr_len) {
+ if (get_user(c, buf))
+ return -EFAULT;
+ if (!isspace(c))
+ break;
+ usr_len--;
+ buf++;
+ }
+
+ if (!usr_len)
+ break;
+ if (!vec_len)
+ return -EINVAL; /* too many numbers */
+
+ /* get next word (possibly going beyond its end) */
+ l = min(usr_len, sizeof(word) - 1);
+ if (copy_from_user(word, buf, l))
+ return -EFAULT;
+ word[l] = '\0';
+
+ v = simple_strtoul(word, &end, base);
+ l = end - word;
+ if (!l)
+ return -EINVAL; /* catch embedded '\0's */
+ if (*end && !isspace(*end))
+ return -EINVAL;
+ /*
+ * Complain if we encountered a too long sequence of digits.
+ * The most we can consume in one iteration is for a 64-bit
+ * number in binary. Too bad simple_strtoul doesn't catch
+ * overflows.
+ */
+ if (l > 64)
+ return -EINVAL;
+ if (v < min || v > max)
+ return -ERANGE;
+ *vals++ = v;
+ vec_len--;
+ usr_len -= l;
+ buf += l;
+ }
+ if (vec_len)
+ return -EINVAL; /* not enough numbers */
+ return 0;
+}
+
+static void seq_noop_stop(struct seq_file *seq, void *v)
+{
+}
+
+/*
+ * generic seq_file support for showing a table of size rows x width.
+ */
+struct seq_tab {
+ int (*show)(struct seq_file *seq, void *v, int idx);
+ unsigned int rows; /* # of entries */
+ unsigned char width; /* size in bytes of each entry */
+ unsigned char skip_first; /* whether the first line is a header */
+ char data[0]; /* the table data */
+};
+
+static void *seq_tab_get_idx(struct seq_tab *tb, loff_t pos)
+{
+ pos -= tb->skip_first;
+ return pos >= tb->rows ? NULL : &tb->data[pos * tb->width];
+}
+
+static void *seq_tab_start(struct seq_file *seq, loff_t *pos)
+{
+ struct seq_tab *tb = seq->private;
+
+ if (tb->skip_first && *pos == 0)
+ return SEQ_START_TOKEN;
+
+ return seq_tab_get_idx(tb, *pos);
+}
+
+static void *seq_tab_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ v = seq_tab_get_idx(seq->private, *pos + 1);
+ if (v)
+ ++*pos;
+ return v;
+}
+
+static int seq_tab_show(struct seq_file *seq, void *v)
+{
+ const struct seq_tab *tb = seq->private;
+
+ /*
+ * index is bogus when v isn't within data, eg when it's
+ * SEQ_START_TOKEN, but that's OK
+ */
+ return tb->show(seq, v, ((char *)v - tb->data) / tb->width);
+}
+
+static struct seq_operations seq_tab_ops = {
+ .start = seq_tab_start,
+ .next = seq_tab_next,
+ .stop = seq_noop_stop,
+ .show = seq_tab_show
+};
+
+static struct seq_tab *seq_open_tab(struct file *f, unsigned int rows,
+ unsigned int width, unsigned int have_header,
+ int (*show)(struct seq_file *seq, void *v, int i))
+{
+ struct seq_tab *p;
+
+ p = __seq_open_private(f, &seq_tab_ops, sizeof(*p) + rows * width);
+ if (p) {
+ p->show = show;
+ p->rows = rows;
+ p->width = width;
+ p->skip_first = have_header != 0;
+ }
+ return p;
+}
+
+/*
+ * debugfs support
+ */
+
+static int mem_open(struct inode *inode, struct file *file)
+{
+ file->private_data = inode->i_private;
+ return 0;
+}
+
+static ssize_t mem_read(struct file *file, char __user *buf, size_t count,
+ loff_t *ppos)
+{
+ loff_t pos = *ppos;
+ loff_t avail = file->f_path.dentry->d_inode->i_size;
+ unsigned int mem = (uintptr_t)file->private_data & 3;
+ struct adapter *adap = file->private_data - mem;
+
+ if (pos < 0)
+ return -EINVAL;
+ if (pos >= avail)
+ return 0;
+ if (count > avail - pos)
+ count = avail - pos;
+
+ while (count) {
+ size_t len;
+ int ret, ofst;
+ __be32 data[16];
+
+ if (mem == MEM_MC)
+ ret = t4_mc_read(adap, pos, data, NULL);
+ else
+ ret = t4_edc_read(adap, mem, pos, data, NULL);
+ if (ret)
+ return ret;
+
+ ofst = pos % sizeof(data);
+ len = min(count, sizeof(data) - ofst);
+ if (copy_to_user(buf, (u8 *)data + ofst, len))
+ return -EFAULT;
+
+ buf += len;
+ pos += len;
+ count -= len;
+ }
+ count = pos - *ppos;
+ *ppos = pos;
+ return count;
+}
+
+static const struct file_operations mem_debugfs_fops = {
+ .owner = THIS_MODULE,
+ .open = mem_open,
+ .read = mem_read,
+};
+
+static void __devinit add_debugfs_mem(struct adapter *adap, const char *name,
+ unsigned int idx, unsigned int size_mb)
+{
+ struct dentry *de;
+
+ de = debugfs_create_file(name, S_IRUSR, adap->debugfs_root,
+ (void *)adap + idx, &mem_debugfs_fops);
+ if (de && de->d_inode)
+ de->d_inode->i_size = size_mb << 20;
+}
+
+static int __devinit setup_debugfs(struct adapter *adap)
+{
+ int i;
+
+ if (!adap->debugfs_root)
+ return -1;
+
+ i = t4_read_reg(adap, A_MA_TARGET_MEM_ENABLE);
+ if (i & F_EDRAM0_ENABLE)
+ add_debugfs_mem(adap, "edc0", MEM_EDC0, 5);
+ if (i & F_EDRAM1_ENABLE)
+ add_debugfs_mem(adap, "edc1", MEM_EDC1, 5);
+ if (i & F_EXT_MEM_ENABLE)
+ add_debugfs_mem(adap, "mc", MEM_MC,
+ G_EXT_MEM_SIZE(t4_read_reg(adap, A_MA_EXT_MEMORY_BAR)));
+ return 0;
+}
+
+/*
+ * /proc support
+ */
+
+#define DEFINE_SIMPLE_PROC_FILE(name) \
+static int name##_open(struct inode *inode, struct file *file) \
+{ \
+ return single_open(file, name##_show, PDE(inode)->data); \
+} \
+static const struct file_operations name##_proc_fops = { \
+ .owner = THIS_MODULE, \
+ .open = name##_open, \
+ .read = seq_read, \
+ .llseek = seq_lseek, \
+ .release = single_release \
+}
+
+static int sge_stats_show(struct seq_file *seq, void *v)
+{
+ struct adapter *adap = seq->private;
+ int eth_entries = DIV_ROUND_UP(adap->sge.ethqsets, 4);
+ int ofld_entries = DIV_ROUND_UP(adap->sge.ofldqsets, 4);
+ int rdma_entries = DIV_ROUND_UP(adap->sge.rdmaqs, 4);
+ int ctrl_entries = DIV_ROUND_UP(MAX_CTRL_QUEUES, 4);
+ int i, r = (uintptr_t)v - 1;
+
+ if (r)
+ seq_putc(seq, '\n');
+
+#define S3(fmt_spec, s, v) \
+ seq_printf(seq, "%-12s", s); \
+ for (i = 0; i < n; ++i) \
+ seq_printf(seq, " %16" fmt_spec, v); \
+ seq_putc(seq, '\n');
+#define S(s, v) S3("s", s, v)
+#define T3(fmt_spec, s, v) S3(fmt_spec, s, tx[i].v)
+#define T(s, v) T3("lu", s, v)
+#define R3(fmt_spec, s, v) S3(fmt_spec, s, qs[i].v)
+#define R(s, v) R3("lu", s, v)
+
+ if (r < eth_entries) {
+ const struct sge_eth_rxq *qs = &adap->sge.ethrxq[r * 4];
+ const struct sge_eth_txq *tx = &adap->sge.ethtxq[r * 4];
+ int n = min(4, adap->sge.ethqsets - 4 * r);
+
+ S("QType:", "Ethernet");
+ S("Interface:",
+ qs[i].rspq.netdev ? qs[i].rspq.netdev->name : "N/A");
+ R("RxPackets:", stats.pkts);
+ R("RxCSO:", stats.rx_cso);
+ R("VLANxtract:", stats.vlan_ex);
+ R("LROmerged:", stats.lro_merged);
+ R("LROpackets:", stats.lro_pkts);
+ R("RxDrops:", stats.rx_drops);
+ T("TSO:", tso);
+ T("TxCSO:", tx_cso);
+ T("VLANins:", vlan_ins);
+ T("TxQFull:", q.stops);
+ T("TxQRestarts:", q.restarts);
+ T("TxMapErr:", mapping_err);
+ R("FLAllocErr:", fl.alloc_failed);
+ R("FLLrgAlcErr:", fl.large_alloc_failed);
+ R("FLstarving:", fl.starving);
+ } else if ((r -= eth_entries) < ofld_entries) {
+ const struct sge_ofld_rxq *qs = &adap->sge.ofldrxq[r * 4];
+ const struct sge_ofld_txq *tx = &adap->sge.ofldtxq[r * 4];
+ int n = min(4, adap->sge.ofldqsets - 4 * r);
+
+ S("QType:", "OFLD");
+ R("RxPackets:", stats.pkts);
+ R("RxImmPkts:", stats.imm);
+ R("RxNoMem:", stats.nomem);
+ T("TxQFull:", q.stops);
+ T("TxQRestarts:", q.restarts);
+ T("TxMapErr:", mapping_err);
+ R("FLAllocErr:", fl.alloc_failed);
+ R("FLLrgAlcErr:", fl.large_alloc_failed);
+ R("FLstarving:", fl.starving);
+ } else if ((r -= ofld_entries) < rdma_entries) {
+ const struct sge_ofld_rxq *qs = &adap->sge.rdmarxq[r * 4];
+ int n = min(4, adap->sge.rdmaqs - 4 * r);
+
+ S("QType:", "RDMA");
+ R("RxPackets:", stats.pkts);
+ R("RxImmPkts:", stats.imm);
+ R("RxAN:", stats.an);
+ R("RxNoMem:", stats.nomem);
+ R("FLAllocErr:", fl.alloc_failed);
+ R("FLstarving:", fl.starving);
+ } else if ((r -= rdma_entries) < ctrl_entries) {
+ const struct sge_ctrl_txq *tx = &adap->sge.ctrlq[r * 4];
+ int n = min(4, MAX_CTRL_QUEUES - 4 * r);
+
+ S("QType:", "Control");
+ T("TxQFull:", q.stops);
+ T("TxQRestarts:", q.restarts);
+ } else if ((r -= ctrl_entries) == 0) {
+ seq_printf(seq, "%-12s %16s\n", "QType:", "FW event queue");
+ }
+#undef R
+#undef T
+#undef S
+#undef R3
+#undef T3
+#undef S3
+ return 0;
+}
+
+static int sge_queue_entries(const struct adapter *adap)
+{
+ return DIV_ROUND_UP(adap->sge.ethqsets, 4) +
+ DIV_ROUND_UP(adap->sge.ofldqsets, 4) +
+ DIV_ROUND_UP(adap->sge.rdmaqs, 4) +
+ DIV_ROUND_UP(MAX_CTRL_QUEUES, 4) + 1;
+}
+
+static void *sge_queue_start(struct seq_file *seq, loff_t *pos)
+{
+ int entries = sge_queue_entries(seq->private);
+
+ return *pos < entries ? (void *)((uintptr_t)*pos + 1) : NULL;
+}
+
+static void *sge_queue_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ int entries = sge_queue_entries(seq->private);
+
+ ++*pos;
+ return *pos < entries ? (void *)((uintptr_t)*pos + 1) : NULL;
+}
+
+static struct seq_operations sge_stats_seq_ops = {
+ .start = sge_queue_start,
+ .next = sge_queue_next,
+ .stop = seq_noop_stop,
+ .show = sge_stats_show
+};
+
+static int sge_stats_open(struct inode *inode, struct file *file)
+{
+ int res = seq_open(file, &sge_stats_seq_ops);
+
+ if (!res) {
+ struct seq_file *seq = file->private_data;
+ seq->private = PDE(inode)->data;
+ }
+ return res;
+}
+
+static const struct file_operations sge_stats_proc_fops = {
+ .owner = THIS_MODULE,
+ .open = sge_stats_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release,
+};
+
+static int lb_stats_show(struct seq_file *seq, void *v)
+{
+ static const char *stat_name[] = {
+ "OctetsOK:", "FramesOK:", "BcastFrames:", "McastFrames:",
+ "UcastFrames:", "ErrorFrames:", "Frames64:", "Frames65To127:",
+ "Frames128To255:", "Frames256To511:", "Frames512To1023:",
+ "Frames1024To1518:", "Frames1519ToMax:", "FramesDropped:",
+ "BG0FramesDropped:", "BG1FramesDropped:", "BG2FramesDropped:",
+ "BG3FramesDropped:", "BG0FramesTrunc:", "BG1FramesTrunc:",
+ "BG2FramesTrunc:", "BG3FramesTrunc:"
+ };
+
+ int i, j;
+ u64 *p0, *p1;
+ struct lb_port_stats s[2];
+
+ memset(s, 0, sizeof(s));
+
+ for (i = 0; i < 4; i += 2) {
+ t4_get_lb_stats(seq->private, i, &s[0]);
+ t4_get_lb_stats(seq->private, i + 1, &s[1]);
+
+ p0 = &s[0].octets;
+ p1 = &s[1].octets;
+ seq_printf(seq, "%s Loopback %u "
+ " Loopback %u\n", i == 0 ? "" : "\n", i, i + 1);
+
+ for (j = 0; j < ARRAY_SIZE(stat_name); j++)
+ seq_printf(seq, "%-17s %20llu %20llu\n", stat_name[j],
+ (unsigned long long)*p0++,
+ (unsigned long long)*p1++);
+ }
+ return 0;
+}
+
+DEFINE_SIMPLE_PROC_FILE(lb_stats);
+
+static int tcp_stats_show(struct seq_file *seq, void *v)
+{
+ struct tp_tcp_stats v4, v6;
+ struct adapter *adap = seq->private;
+
+ spin_lock(&adap->stats_lock);
+ t4_tp_get_tcp_stats(adap, &v4, &v6);
+ spin_unlock(&adap->stats_lock);
+
+ seq_puts(seq,
+ " IP IPv6\n");
+ seq_printf(seq, "OutRsts: %20u %20u\n",
+ v4.tcpOutRsts, v6.tcpOutRsts);
+ seq_printf(seq, "InSegs: %20llu %20llu\n",
+ (unsigned long long)v4.tcpInSegs,
+ (unsigned long long)v6.tcpInSegs);
+ seq_printf(seq, "OutSegs: %20llu %20llu\n",
+ (unsigned long long)v4.tcpOutSegs,
+ (unsigned long long)v6.tcpOutSegs);
+ seq_printf(seq, "RetransSegs: %20llu %20llu\n",
+ (unsigned long long)v4.tcpRetransSegs,
+ (unsigned long long)v6.tcpRetransSegs);
+ return 0;
+}
+
+DEFINE_SIMPLE_PROC_FILE(tcp_stats);
+
+static int tp_err_stats_show(struct seq_file *seq, void *v)
+{
+ struct tp_err_stats stats;
+ struct adapter *adap = seq->private;
+
+ spin_lock(&adap->stats_lock);
+ t4_tp_get_err_stats(adap, &stats);
+ spin_unlock(&adap->stats_lock);
+
+ seq_puts(seq, " channel 0 channel 1 channel 2 "
+ "channel 3\n");
+ seq_printf(seq, "macInErrs: %10u %10u %10u %10u\n",
+ stats.macInErrs[0], stats.macInErrs[1], stats.macInErrs[2],
+ stats.macInErrs[3]);
+ seq_printf(seq, "hdrInErrs: %10u %10u %10u %10u\n",
+ stats.hdrInErrs[0], stats.hdrInErrs[1], stats.hdrInErrs[2],
+ stats.hdrInErrs[3]);
+ seq_printf(seq, "tcpInErrs: %10u %10u %10u %10u\n",
+ stats.tcpInErrs[0], stats.tcpInErrs[1], stats.tcpInErrs[2],
+ stats.tcpInErrs[3]);
+ seq_printf(seq, "tcp6InErrs: %10u %10u %10u %10u\n",
+ stats.tcp6InErrs[0], stats.tcp6InErrs[1],
+ stats.tcp6InErrs[2], stats.tcp6InErrs[3]);
+ seq_printf(seq, "tnlCongDrops: %10u %10u %10u %10u\n",
+ stats.tnlCongDrops[0], stats.tnlCongDrops[1],
+ stats.tnlCongDrops[2], stats.tnlCongDrops[3]);
+ seq_printf(seq, "tnlTxDrops: %10u %10u %10u %10u\n",
+ stats.tnlTxDrops[0], stats.tnlTxDrops[1],
+ stats.tnlTxDrops[2], stats.tnlTxDrops[3]);
+ seq_printf(seq, "ofldVlanDrops: %10u %10u %10u %10u\n",
+ stats.ofldVlanDrops[0], stats.ofldVlanDrops[1],
+ stats.ofldVlanDrops[2], stats.ofldVlanDrops[3]);
+ seq_printf(seq, "ofldChanDrops: %10u %10u %10u %10u\n\n",
+ stats.ofldChanDrops[0], stats.ofldChanDrops[1],
+ stats.ofldChanDrops[2], stats.ofldChanDrops[3]);
+ seq_printf(seq, "ofldNoNeigh: %u\nofldCongDefer: %u\n",
+ stats.ofldNoNeigh, stats.ofldCongDefer);
+ return 0;
+}
+
+DEFINE_SIMPLE_PROC_FILE(tp_err_stats);
+
+static int rss_show(struct seq_file *seq, void *v, int idx)
+{
+ u16 *entry = v;
+
+ seq_printf(seq, "%4d: %4u %4u %4u %4u %4u %4u %4u %4u\n",
+ idx * 8, entry[0], entry[1], entry[2], entry[3], entry[4],
+ entry[5], entry[6], entry[7]);
+ return 0;
+}
+
+static int rss_open(struct inode *inode, struct file *file)
+{
+ int ret;
+ struct seq_tab *p;
+
+ p = seq_open_tab(file, RSS_NENTRIES / 8, 8 * sizeof(u16), 0, rss_show);
+ if (!p)
+ return -ENOMEM;
+ ret = t4_read_rss(PDE(inode)->data, (u16 *)p->data);
+ if (ret)
+ seq_release_private(inode, file);
+ return ret;
+}
+
+static const struct file_operations rss_proc_fops = {
+ .owner = THIS_MODULE,
+ .open = rss_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release_private
+};
+
+static int mtutab_show(struct seq_file *seq, void *v)
+{
+ u16 mtus[NMTUS];
+ struct adapter *adap = seq->private;
+
+ spin_lock(&adap->stats_lock);
+ t4_read_mtu_tbl(adap, mtus, NULL);
+ spin_unlock(&adap->stats_lock);
+
+ seq_printf(seq, "%u %u %u %u %u %u %u %u %u %u %u %u %u %u %u %u\n",
+ mtus[0], mtus[1], mtus[2], mtus[3], mtus[4], mtus[5],
+ mtus[6], mtus[7], mtus[8], mtus[9], mtus[10], mtus[11],
+ mtus[12], mtus[13], mtus[14], mtus[15]);
+ return 0;
+}
+
+static int mtutab_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, mtutab_show, PDE(inode)->data);
+}
+
+static ssize_t mtutab_write(struct file *file, const char __user *buf,
+ size_t count, loff_t *pos)
+{
+ int i;
+ unsigned long mtus[NMTUS];
+ struct adapter *adap = PDE(file->f_path.dentry->d_inode)->data;
+
+ if (adap->flags & FULL_INIT_DONE)
+ return -EBUSY;
+
+ /* Require min MTU of 81 to accommodate SACK */
+ i = rd_usr_int_vec(buf, count, NMTUS, mtus, 81, MAX_MTU, 10);
+ if (i)
+ return i;
+
+ /* MTUs must be in ascending order */
+ for (i = 1; i < NMTUS; ++i)
+ if (mtus[i] < mtus[i - 1])
+ return -EINVAL;
+
+ for (i = 0; i < NMTUS; ++i)
+ adap->params.mtus[i] = mtus[i];
+ t4_load_mtus(adap, adap->params.mtus, adap->params.a_wnd,
+ adap->params.b_wnd);
+ return count;
+}
+
+static const struct file_operations mtutab_proc_fops = {
+ .owner = THIS_MODULE,
+ .open = mtutab_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+ .write = mtutab_write
+};
+
+static int mps_trc_show(struct seq_file *seq, void *v)
+{
+ int enabled, i;
+ struct trace_params tp;
+ struct adapter *adap = seq->private;
+ unsigned int trcidx = (uintptr_t)adap & 3;
+
+ adap = (void *)adap - trcidx;
+ t4_get_trace_filter(adap, &tp, trcidx, &enabled);
+ if (!enabled) {
+ seq_puts(seq, "tracer is disabled\n");
+ return 0;
+ }
+
+ if (tp.skip_ofst * 8 >= TRACE_LEN) {
+ dev_err(adap->pdev_dev, "illegal trace pattern skip offset\n");
+ return -EINVAL;
+ }
+ if (tp.port < 8) {
+ i = adap->chan_map[tp.port & 3];
+ if (i >= MAX_NPORTS) {
+ dev_err(adap->pdev_dev, "tracer %u is assigned "
+ "to non-existing port\n", trcidx);
+ return -EINVAL;
+ }
+ seq_printf(seq, "tracer is capturing %s %s, ",
+ adap->port[i]->name, tp.port < 4 ? "Rx" : "Tx");
+ } else
+ seq_printf(seq, "tracer is capturing loopback %d, ",
+ tp.port - 8);
+ seq_printf(seq, "snap length: %u, min length: %u\n", tp.snap_len,
+ tp.min_len);
+ seq_printf(seq, "packets captured %smatch filter\n",
+ tp.invert ? "do not " : "");
+
+ if (tp.skip_ofst) {
+ seq_puts(seq, "filter pattern: ");
+ for (i = 0; i < tp.skip_ofst * 2; i += 2)
+ seq_printf(seq, "%08x%08x", tp.data[i], tp.data[i + 1]);
+ seq_putc(seq, '/');
+ for (i = 0; i < tp.skip_ofst * 2; i += 2)
+ seq_printf(seq, "%08x%08x", tp.mask[i], tp.mask[i + 1]);
+ seq_puts(seq, "@0\n");
+ }
+
+ seq_puts(seq, "filter pattern: ");
+ for (i = tp.skip_ofst * 2; i < TRACE_LEN / 4; i += 2)
+ seq_printf(seq, "%08x%08x", tp.data[i], tp.data[i + 1]);
+ seq_putc(seq, '/');
+ for (i = tp.skip_ofst * 2; i < TRACE_LEN / 4; i += 2)
+ seq_printf(seq, "%08x%08x", tp.mask[i], tp.mask[i + 1]);
+ seq_printf(seq, "@%u\n", (tp.skip_ofst + tp.skip_len) * 8);
+ return 0;
+}
+
+static int mps_trc_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, mps_trc_show, PDE(inode)->data);
+}
+
+static unsigned int xdigit2int(unsigned char c)
+{
+ return isdigit(c) ? c - '0' : tolower(c) - 'a' + 10;
+}
+
+#define TRC_PORT_NONE 0xff
+
+/*
+ * Set an MPS trace filter. Syntax is:
+ *
+ * disable
+ *
+ * to disable tracing, or
+ *
+ * interface [snaplen=<val>] [minlen=<val>] [not] [<pattern>]...
+ *
+ * where interface is one of rxN, txN, or loopbackN, N = 0..3, and pattern
+ * has the form
+ *
+ * <pattern data>[/<pattern mask>][@<anchor>]
+ *
+ * Up to 2 filter patterns can be specified. If 2 are supplied the first one
+ * must be anchored at 0. An omited mask is taken as a mask of 1s, an omitted
+ * anchor is taken as 0.
+ */
+static ssize_t mps_trc_write(struct file *file, const char __user *buf,
+ size_t count, loff_t *pos)
+{
+ int i, j, enable;
+ u32 *data, *mask;
+ struct trace_params tp;
+ char *s, *p, *word, *end;
+ struct adapter *adap = PDE(file->f_path.dentry->d_inode)->data;
+ unsigned int trcidx = (uintptr_t)adap & 3;
+
+ adap = (void *)adap - trcidx;
+
+ /*
+ * Don't accept input more than 1K, can't be anything valid except lots
+ * of whitespace. Well, use less.
+ */
+ if (count > 1024)
+ return -EFBIG;
+ p = s = kzalloc(count + 1, GFP_USER);
+ if (!s)
+ return -ENOMEM;
+ if (copy_from_user(s, buf, count)) {
+ count = -EFAULT;
+ goto out;
+ }
+
+ if (s[count - 1] == '\n')
+ s[count - 1] = '\0';
+
+ enable = strcmp("disable", s) != 0;
+ if (!enable)
+ goto apply;
+
+ memset(&tp, 0, sizeof(tp));
+ tp.port = TRC_PORT_NONE;
+ i = 0; /* counts pattern nibbles */
+
+ while (p) {
+ while (isspace(*p))
+ p++;
+ word = strsep(&p, " ");
+ if (!*word)
+ break;
+
+ if (!strncmp(word, "snaplen=", 8)) {
+ j = simple_strtoul(word + 8, &end, 10);
+ if (*end || j > 9600) {
+inval: count = -EINVAL;
+ goto out;
+ }
+ tp.snap_len = j;
+ continue;
+ }
+ if (!strncmp(word, "minlen=", 7)) {
+ j = simple_strtoul(word + 7, &end, 10);
+ if (*end || j > M_TFMINPKTSIZE)
+ goto inval;
+ tp.min_len = j;
+ continue;
+ }
+ if (!strcmp(word, "not")) {
+ tp.invert = !tp.invert;
+ continue;
+ }
+ if (!strncmp(word, "loopback", 8) && tp.port == TRC_PORT_NONE) {
+ if (word[8] < '0' || word[8] > '3' || word[9])
+ goto inval;
+ tp.port = word[8] - '0' + 8;
+ continue;
+ }
+ if (!strncmp(word, "tx", 2) && tp.port == TRC_PORT_NONE) {
+ if (word[2] < '0' || word[2] > '3' || word[3])
+ goto inval;
+ tp.port = word[2] - '0' + 4;
+ if (adap->chan_map[tp.port & 3] >= MAX_NPORTS)
+ goto inval;
+ continue;
+ }
+ if (!strncmp(word, "rx", 2) && tp.port == TRC_PORT_NONE) {
+ if (word[2] < '0' || word[2] > '3' || word[3])
+ goto inval;
+ tp.port = word[2] - '0';
+ if (adap->chan_map[tp.port] >= MAX_NPORTS)
+ goto inval;
+ continue;
+ }
+ if (!isxdigit(*word))
+ goto inval;
+
+ /* we have found a trace pattern */
+ if (i) { /* split pattern */
+ if (tp.skip_len) /* too many splits */
+ goto inval;
+ tp.skip_ofst = i / 16;
+ }
+
+ data = &tp.data[i / 8];
+ mask = &tp.mask[i / 8];
+ j = i;
+
+ while (isxdigit(*word)) {
+ if (i >= TRACE_LEN * 2) {
+ count = -EFBIG;
+ goto out;
+ }
+ *data = (*data << 4) + xdigit2int(*word++);
+ if (++i % 8 == 0)
+ data++;
+ }
+ if (*word == '/') {
+ word++;
+ while (isxdigit(*word)) {
+ if (j >= i) /* mask longer than data */
+ goto inval;
+ *mask = (*mask << 4) + xdigit2int(*word++);
+ if (++j % 8 == 0)
+ mask++;
+ }
+ if (i != j) /* mask shorter than data */
+ goto inval;
+ } else { /* no mask, use all 1s */
+ for ( ; i - j >= 8; j += 8)
+ *mask++ = 0xffffffff;
+ if (i % 8)
+ *mask = (1 << (i % 8) * 4) - 1;
+ }
+ if (*word == '@') {
+ j = simple_strtoul(word + 1, &end, 10);
+ if (*end && *end != '\n')
+ goto inval;
+ if (j & 7) /* doesn't start at multiple of 8 */
+ goto inval;
+ j /= 8;
+ if (j < tp.skip_ofst) /* overlaps earlier pattern */
+ goto inval;
+ if (j - tp.skip_ofst > 31) /* skip too big */
+ goto inval;
+ tp.skip_len = j - tp.skip_ofst;
+ }
+ if (i % 8) {
+ *data <<= (8 - i % 8) * 4;
+ *mask <<= (8 - i % 8) * 4;
+ i = (i + 15) & ~15; /* 8-byte align */
+ }
+ }
+
+ if (tp.port == TRC_PORT_NONE)
+ goto inval;
+apply:
+ i = t4_set_trace_filter(adap, &tp, trcidx, enable);
+ if (i)
+ count = i;
+out:
+ kfree(s);
+ return count;
+}
+
+static const struct file_operations mps_trc_proc_fops = {
+ .owner = THIS_MODULE,
+ .open = mps_trc_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+ .write = mps_trc_write
+};
+
+static int tid_info_show(struct seq_file *seq, void *v)
+{
+ struct adapter *adap = seq->private;
+ const struct tid_info *t = &adap->tids;
+
+ if (t4_read_reg(adap, A_LE_DB_CONFIG) & F_HASHEN) {
+ unsigned int sb = t4_read_reg(adap, A_LE_DB_SERVER_INDEX) / 4;
+
+ if (sb)
+ seq_printf(seq, "TID range: 0..%u/%u..%u", sb - 1,
+ t4_read_reg(adap, A_LE_DB_TID_HASHBASE) / 4,
+ t->ntids - 1);
+ else
+ seq_printf(seq, "TID range: %u..%u",
+ t4_read_reg(adap, A_LE_DB_TID_HASHBASE) / 4,
+ t->ntids - 1);
+ } else
+ seq_printf(seq, "TID range: 0..%u", t->ntids - 1);
+ seq_printf(seq, ", in use: %u\n", atomic_read(&t->tids_in_use));
+ seq_printf(seq, "STID range: %u..%u, in use: %u\n", t->stid_base,
+ t->stid_base + t->nstids - 1, t->stids_in_use);
+ seq_printf(seq, "ATID range: 0..%u, in use: %u\n", t->natids - 1,
+ t->atids_in_use);
+ seq_printf(seq, "FTID range: %u..%u\n", t->ftid_base,
+ t->ftid_base + t->nftids - 1);
+ seq_printf(seq, "HW TID usage: %u IP users, %u IPv6 users\n",
+ t4_read_reg(adap, A_LE_DB_ACT_CNT_IPV4),
+ t4_read_reg(adap, A_LE_DB_ACT_CNT_IPV6));
+ return 0;
+}
+
+DEFINE_SIMPLE_PROC_FILE(tid_info);
+
+static int uld_show(struct seq_file *seq, void *v)
+{
+ int i;
+ struct adapter *adap = seq->private;
+
+ for (i = 0; i < CXGB4_ULD_MAX; i++)
+ if (adap->uld_handle[i])
+ seq_printf(seq, "%s: %s\n", uld_str[i], ulds[i].name);
+ return 0;
+}
+
+DEFINE_SIMPLE_PROC_FILE(uld);
+
+enum {
+ ADAP_NEED_L2T = 1 << 0,
+ ADAP_NEED_OFLD = 1 << 1,
+};
+
+struct cxgb4_proc_entry {
+ const char *name;
+ mode_t mode;
+ unsigned int req; /* adapter requirements to create this file */
+ unsigned char data; /* data passed in low bits of adapter pointer */
+ const struct file_operations *fops;
+};
+
+static struct cxgb4_proc_entry proc_files[] = {
+ { "l2t", 0444, ADAP_NEED_L2T, 0, &t4_l2t_proc_fops },
+ { "lb_stats", 0444, 0, 0, &lb_stats_proc_fops },
+ { "path_mtus", 0644, 0, 0, &mtutab_proc_fops },
+ { "qstats", 0444, 0, 0, &sge_stats_proc_fops },
+ { "rss", 0444, 0, 0, &rss_proc_fops },
+ { "tcp_stats", 0444, 0, 0, &tcp_stats_proc_fops },
+ { "tids", 0444, ADAP_NEED_OFLD, 0, &tid_info_proc_fops },
+ { "tp_err_stats", 0444, 0, 0, &tp_err_stats_proc_fops },
+ { "trace0", 0644, 0, 0, &mps_trc_proc_fops },
+ { "trace1", 0644, 0, 1, &mps_trc_proc_fops },
+ { "trace2", 0644, 0, 2, &mps_trc_proc_fops },
+ { "trace3", 0644, 0, 3, &mps_trc_proc_fops },
+ { "uld", 0444, 0, 0, &uld_proc_fops },
+};
+
+static int __devinit setup_proc(struct adapter *adap,
+ struct proc_dir_entry *dir)
+{
+ int i, created;
+ struct proc_dir_entry *p;
+ int ofld = is_offload(adap);
+
+ if (!dir)
+ return -EINVAL;
+
+ /* If we can create any of the entries we do. */
+ for (created = i = 0; i < ARRAY_SIZE(proc_files); ++i) {
+ unsigned int req = proc_files[i].req;
+
+ if ((req & ADAP_NEED_OFLD) && !ofld)
+ continue;
+ if ((req & ADAP_NEED_L2T) && !adap->l2t)
+ continue;
+
+ p = proc_create_data(proc_files[i].name, proc_files[i].mode,
+ dir, proc_files[i].fops,
+ (void *)adap + proc_files[i].data);
+ if (p)
+ created++;
+ }
+
+ return created;
+}
+
+static void cleanup_proc(struct proc_dir_entry *dir)
+{
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(proc_files); ++i)
+ remove_proc_entry(proc_files[i].name, dir);
+}
+
+/*
+ * upper-layer driver support
+ */
+
+/*
+ * Allocate an active-open TID and set it to the supplied value.
+ */
+int cxgb4_alloc_atid(struct tid_info *t, void *data)
+{
+ int atid = -1;
+
+ spin_lock_bh(&t->atid_lock);
+ if (t->afree) {
+ union aopen_entry *p = t->afree;
+
+ atid = p - t->atid_tab;
+ t->afree = p->next;
+ p->data = data;
+ t->atids_in_use++;
+ }
+ spin_unlock_bh(&t->atid_lock);
+ return atid;
+}
+EXPORT_SYMBOL(cxgb4_alloc_atid);
+
+/*
+ * Release an active-open TID.
+ */
+void cxgb4_free_atid(struct tid_info *t, unsigned int atid)
+{
+ union aopen_entry *p = &t->atid_tab[atid];
+
+ spin_lock_bh(&t->atid_lock);
+ p->next = t->afree;
+ t->afree = p;
+ t->atids_in_use--;
+ spin_unlock_bh(&t->atid_lock);
+}
+EXPORT_SYMBOL(cxgb4_free_atid);
+
+/*
+ * Allocate a server TID and set it to the supplied value.
+ */
+int cxgb4_alloc_stid(struct tid_info *t, int family, void *data)
+{
+ int stid;
+
+ spin_lock_bh(&t->stid_lock);
+ if (family == PF_INET) {
+ stid = find_first_zero_bit(t->stid_bmap, t->nstids);
+ if (stid < t->nstids)
+ __set_bit(stid, t->stid_bmap);
+ else
+ stid = -1;
+ } else {
+ stid = bitmap_find_free_region(t->stid_bmap, t->nstids, 2);
+ if (stid < 0)
+ stid = -1;
+ }
+ if (stid >= 0) {
+ t->stid_tab[stid].data = data;
+ stid += t->stid_base;
+ t->stids_in_use++;
+ }
+ spin_unlock_bh(&t->stid_lock);
+ return stid;
+}
+EXPORT_SYMBOL(cxgb4_alloc_stid);
+
+/*
+ * Release a server TID.
+ */
+void cxgb4_free_stid(struct tid_info *t, unsigned int stid, int family)
+{
+ stid -= t->stid_base;
+ spin_lock_bh(&t->stid_lock);
+ if (family == PF_INET)
+ __clear_bit(stid, t->stid_bmap);
+ else
+ bitmap_release_region(t->stid_bmap, stid, 2);
+ t->stid_tab[stid].data = NULL;
+ t->stids_in_use--;
+ spin_unlock_bh(&t->stid_lock);
+}
+EXPORT_SYMBOL(cxgb4_free_stid);
+
+/*
+ * Populate a TID_RELEASE WR. Caller must properly size the skb.
+ */
+static void mk_tid_release(struct sk_buff *skb, unsigned int chan,
+ unsigned int tid)
+{
+ struct cpl_tid_release *req;
+
+ set_wr_txq(skb, CPL_PRIORITY_SETUP, chan);
+ req = (struct cpl_tid_release *)__skb_put(skb, sizeof(*req));
+ INIT_TP_WR(req, tid);
+ OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_TID_RELEASE, tid));
+}
+
+/*
+ * Queue a TID release request and if necessary schedule a work queue to
+ * process it.
+ */
+void cxgb4_queue_tid_release(struct tid_info *t, unsigned int chan,
+ unsigned int tid)
+{
+ void **p = &t->tid_tab[tid];
+ struct adapter *adap = container_of(t, struct adapter, tids);
+
+ spin_lock_bh(&adap->tid_release_lock);
+ *p = adap->tid_release_head;
+ /* Low 2 bits encode the Tx channel number */
+ adap->tid_release_head = (void **)((uintptr_t)p | chan);
+ if (!*p)
+ schedule_work(&adap->tid_release_task);
+ spin_unlock_bh(&adap->tid_release_lock);
+}
+EXPORT_SYMBOL(cxgb4_queue_tid_release);
+
+/*
+ * Process the list of pending TID release requests.
+ */
+static void process_tid_release_list(struct work_struct *work)
+{
+ struct sk_buff *skb;
+ struct adapter *adap;
+
+ adap = container_of(work, struct adapter, tid_release_task);
+
+ spin_lock_bh(&adap->tid_release_lock);
+ while (adap->tid_release_head) {
+ void **p = adap->tid_release_head;
+ unsigned int chan = (uintptr_t)p & 3;
+ p = (void *)p - chan;
+
+ adap->tid_release_head = *p;
+ *p = NULL;
+ spin_unlock_bh(&adap->tid_release_lock);
+
+ while (!(skb = alloc_skb(sizeof(struct cpl_tid_release),
+ GFP_KERNEL)))
+ yield();
+
+ mk_tid_release(skb, chan, p - adap->tids.tid_tab);
+ t4_ofld_send(adap, skb);
+ spin_lock_bh(&adap->tid_release_lock);
+ }
+ spin_unlock_bh(&adap->tid_release_lock);
+}
+
+/*
+ * Release a TID and inform HW. If we are unable to allocate the release
+ * message we defer to a work queue.
+ */
+void cxgb4_remove_tid(struct tid_info *t, unsigned int chan, unsigned int tid)
+{
+ struct sk_buff *skb;
+ struct adapter *adap = container_of(t, struct adapter, tids);
+
+ WARN_ON(tid >= t->ntids);
+
+ skb = alloc_skb(sizeof(struct cpl_tid_release), GFP_ATOMIC);
+ if (likely(skb)) {
+ t->tid_tab[tid] = NULL;
+ mk_tid_release(skb, chan, tid);
+ t4_ofld_send(adap, skb);
+ } else
+ cxgb4_queue_tid_release(t, chan, tid);
+ atomic_dec(&t->tids_in_use);
+}
+EXPORT_SYMBOL(cxgb4_remove_tid);
+
+/*
+ * Allocate and initialize the TID tables. Returns 0 on success.
+ */
+static int tid_init(struct tid_info *t)
+{
+ size_t size;
+ unsigned int natids = t->natids;
+
+ size = t->ntids * sizeof(*t->tid_tab) + natids * sizeof(*t->atid_tab) +
+ t->nstids * sizeof(*t->stid_tab) +
+ BITS_TO_LONGS(t->nstids) * sizeof(long);
+ t->tid_tab = t4_alloc_mem(size);
+ if (!t->tid_tab)
+ return -ENOMEM;
+
+ t->atid_tab = (union aopen_entry *)&t->tid_tab[t->ntids];
+ t->stid_tab = (struct serv_entry *)&t->atid_tab[natids];
+ t->stid_bmap = (unsigned long *)&t->stid_tab[t->nstids];
+ spin_lock_init(&t->stid_lock);
+ spin_lock_init(&t->atid_lock);
+
+ t->stids_in_use = 0;
+ t->afree = NULL;
+ t->atids_in_use = 0;
+ atomic_set(&t->tids_in_use, 0);
+
+ /* Setup the free list for atid_tab and clear the stid bitmap. */
+ if (natids) {
+ while (--natids)
+ t->atid_tab[natids - 1].next = &t->atid_tab[natids];
+ t->afree = t->atid_tab;
+ }
+ bitmap_zero(t->stid_bmap, t->nstids);
+ return 0;
+}
+
+/**
+ * cxgb4_create_server - create an IP server
+ * @dev: the device
+ * @stid: the server TID
+ * @sip: local IP address to bind server to
+ * @sport: the server's TCP port
+ * @queue: queue to direct messages from this server to
+ *
+ * Create an IP server for the given port and address.
+ * Returns <0 on error and one of the %NET_XMIT_* values on success.
+ */
+int cxgb4_create_server(const struct net_device *dev, unsigned int stid,
+ __be32 sip, __be16 sport, unsigned int queue)
+{
+ unsigned int chan;
+ struct sk_buff *skb;
+ struct adapter *adap;
+ struct cpl_pass_open_req *req;
+
+ skb = alloc_skb(sizeof(*req), GFP_KERNEL);
+ if (!skb)
+ return -ENOMEM;
+
+ adap = netdev2adap(dev);
+ req = (struct cpl_pass_open_req *)__skb_put(skb, sizeof(*req));
+ INIT_TP_WR(req, 0);
+ OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_PASS_OPEN_REQ, stid));
+ req->local_port = sport;
+ req->peer_port = htons(0);
+ req->local_ip = sip;
+ req->peer_ip = htonl(0);
+ chan = netdev2pinfo(adap->sge.ingr_map[queue]->netdev)->tx_chan;
+ req->opt0 = cpu_to_be64(V_TX_CHAN(chan));
+ req->opt1 = cpu_to_be64(V_CONN_POLICY(CPL_CONN_POLICY_ASK) |
+ F_SYN_RSS_ENABLE | V_SYN_RSS_QUEUE(queue));
+ return t4_mgmt_tx(adap, skb);
+}
+EXPORT_SYMBOL(cxgb4_create_server);
+
+/**
+ * cxgb4_create_server6 - create an IPv6 server
+ * @dev: the device
+ * @stid: the server TID
+ * @sip: local IPv6 address to bind server to
+ * @sport: the server's TCP port
+ * @queue: queue to direct messages from this server to
+ *
+ * Create an IPv6 server for the given port and address.
+ * Returns <0 on error and one of the %NET_XMIT_* values on success.
+ */
+int cxgb4_create_server6(const struct net_device *dev, unsigned int stid,
+ const struct in6_addr *sip, __be16 sport,
+ unsigned int queue)
+{
+ unsigned int chan;
+ struct sk_buff *skb;
+ struct adapter *adap;
+ struct cpl_pass_open_req6 *req;
+
+ skb = alloc_skb(sizeof(*req), GFP_KERNEL);
+ if (!skb)
+ return -ENOMEM;
+
+ adap = netdev2adap(dev);
+ req = (struct cpl_pass_open_req6 *)__skb_put(skb, sizeof(*req));
+ INIT_TP_WR(req, 0);
+ OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_PASS_OPEN_REQ6, stid));
+ req->local_port = sport;
+ req->peer_port = htons(0);
+ req->local_ip_hi = *(__be64 *)(sip->s6_addr);
+ req->local_ip_lo = *(__be64 *)(sip->s6_addr + 8);
+ req->peer_ip_hi = cpu_to_be64(0);
+ req->peer_ip_lo = cpu_to_be64(0);
+ chan = netdev2pinfo(adap->sge.ingr_map[queue]->netdev)->tx_chan;
+ req->opt0 = cpu_to_be64(V_TX_CHAN(chan));
+ req->opt1 = cpu_to_be64(V_CONN_POLICY(CPL_CONN_POLICY_ASK) |
+ F_SYN_RSS_ENABLE | V_SYN_RSS_QUEUE(queue));
+ return t4_mgmt_tx(adap, skb);
+}
+EXPORT_SYMBOL(cxgb4_create_server6);
+
+/**
+ * cxgb4_best_mtu - find the entry in the MTU table closest to an MTU
+ * @mtus: the HW MTU table
+ * @mtu: the target MTU
+ * @idx: index of selected entry in the MTU table
+ *
+ * Returns the index and the value in the HW MTU table that is closest to
+ * but does not exceed @mtu, unless @mtu is smaller than any value in the
+ * table, in which case that smallest available value is selected.
+ */
+unsigned int cxgb4_best_mtu(const unsigned short *mtus, unsigned short mtu,
+ unsigned int *idx)
+{
+ unsigned int i = 0;
+
+ while (i < NMTUS - 1 && mtus[i + 1] <= mtu)
+ ++i;
+ if (idx)
+ *idx = i;
+ return mtus[i];
+}
+EXPORT_SYMBOL(cxgb4_best_mtu);
+
+/**
+ * cxgb4_port_chan - get the HW channel of a port
+ * @dev: the net device for the port
+ *
+ * Return the HW Tx channel of the given port.
+ */
+unsigned int cxgb4_port_chan(const struct net_device *dev)
+{
+ return netdev2pinfo(dev)->tx_chan;
+}
+EXPORT_SYMBOL(cxgb4_port_chan);
+
+/**
+ * cxgb4_netdev_by_hwid - return the net device of a HW port
+ * @pdev: identifies the adapter
+ * @id: the HW port id
+ *
+ * Return the net device associated with the interface with the given HW
+ * id.
+ */
+struct net_device *cxgb4_netdev_by_hwid(struct pci_dev *pdev, unsigned int id)
+{
+ const struct adapter *adap = pci_get_drvdata(pdev);
+
+ if (!adap || id >= NCHAN)
+ return NULL;
+ id = adap->chan_map[id];
+ return id < MAX_NPORTS ? adap->port[id] : NULL;
+}
+EXPORT_SYMBOL(cxgb4_netdev_by_hwid);
+
+void cxgb4_iscsi_init(struct net_device *dev, unsigned int tag_mask,
+ const unsigned int *pgsz_order)
+{
+ struct adapter *adap = netdev2adap(dev);
+
+ t4_write_reg(adap, A_ULP_RX_ISCSI_TAGMASK, tag_mask);
+ t4_write_reg(adap, A_ULP_RX_ISCSI_PSZ, V_HPZ0(pgsz_order[0]) |
+ V_HPZ1(pgsz_order[1]) | V_HPZ2(pgsz_order[2]) |
+ V_HPZ3(pgsz_order[3]));
+}
+EXPORT_SYMBOL(cxgb4_iscsi_init);
+
+static struct pci_driver cxgb4_driver;
+
+static void check_neigh_update(struct neighbour *neigh)
+{
+ const struct device *parent;
+ const struct net_device *netdev = neigh->dev;
+
+ if (netdev->priv_flags & IFF_802_1Q_VLAN)
+ netdev = vlan_dev_real_dev(netdev);
+ parent = netdev->dev.parent;
+ if (parent && parent->driver == &cxgb4_driver.driver)
+ t4_l2t_update(dev_get_drvdata(parent), neigh);
+}
+
+static int netevent_cb(struct notifier_block *nb, unsigned long event,
+ void *data)
+{
+ switch (event) {
+ case NETEVENT_NEIGH_UPDATE :
+ check_neigh_update(data);
+ break;
+ case NETEVENT_PMTU_UPDATE:
+ case NETEVENT_REDIRECT:
+ default:
+ break;
+ }
+ return 0;
+}
+
+static bool netevent_registered;
+static struct notifier_block cxgb4_netevent_nb = {
+ .notifier_call = netevent_cb
+};
+
+static void uld_attach(struct adapter *adap, unsigned int uld)
+{
+ void *handle;
+ struct cxgb4_lld_info lli;
+
+ lli.pdev = adap->pdev;
+ lli.l2t = adap->l2t;
+ lli.tids = &adap->tids;
+ lli.ports = adap->port;
+ lli.vr = &adap->vres;
+ lli.mtus = adap->params.mtus;
+ if (uld == CXGB4_ULD_RDMA) {
+ lli.rxq_ids = adap->sge.rdma_rxq;
+ lli.nrxq = adap->sge.rdmaqs;
+ } else if (uld == CXGB4_ULD_ISCSI) {
+ lli.rxq_ids = adap->sge.ofld_rxq;
+ lli.nrxq = adap->sge.ofldqsets;
+ }
+ lli.ntxq = adap->sge.ofldqsets;
+ lli.nchan = adap->params.nports;
+ lli.nports = adap->params.nports;
+ lli.wr_cred = adap->params.ofldq_wr_cred;
+ lli.adapter_type = CXGB4_ADAP_T4;
+ lli.udb_density = 1 << G_QUEUESPERPAGEPF0(
+ t4_read_reg(adap, A_SGE_EGRESS_QUEUES_PER_PAGE_PF));
+ lli.ucq_density = 1 << G_QUEUESPERPAGEPF0(
+ t4_read_reg(adap, A_SGE_INGRESS_QUEUES_PER_PAGE_PF));
+ lli.gts_reg = adap->regs + MYPF_REG(A_SGE_PF_GTS);
+ lli.db_reg = adap->regs + MYPF_REG(A_SGE_PF_KDOORBELL);
+ lli.fw_vers = adap->params.fw_vers;
+
+ handle = ulds[uld].add(&lli);
+ if (IS_ERR(handle)) {
+ CH_WARN(adap, "could not attach to the %s driver, error %ld\n",
+ uld_str[uld], PTR_ERR(handle));
+ return;
+ }
+
+ adap->uld_handle[uld] = handle;
+
+ if (!netevent_registered) {
+ register_netevent_notifier(&cxgb4_netevent_nb);
+ netevent_registered = true;
+ }
+}
+
+static void attach_ulds(struct adapter *adap)
+{
+ unsigned int i;
+
+ mutex_lock(&uld_mutex);
+ list_add_tail(&adap->list_node, &adapter_list);
+ for (i = 0; i < CXGB4_ULD_MAX; i++)
+ if (ulds[i].add)
+ uld_attach(adap, i);
+ mutex_unlock(&uld_mutex);
+}
+
+static void detach_ulds(struct adapter *adap)
+{
+ unsigned int i;
+
+ mutex_lock(&uld_mutex);
+ list_del(&adap->list_node);
+ for (i = 0; i < CXGB4_ULD_MAX; i++)
+ if (adap->uld_handle[i]) {
+ ulds[i].state_change(adap->uld_handle[i],
+ CXGB4_STATE_DETACH);
+ adap->uld_handle[i] = NULL;
+ }
+ if (netevent_registered && list_empty(&adapter_list)) {
+ unregister_netevent_notifier(&cxgb4_netevent_nb);
+ netevent_registered = false;
+ }
+ mutex_unlock(&uld_mutex);
+}
+
+static void notify_ulds(struct adapter *adap, enum cxgb4_state new_state)
+{
+ unsigned int i;
+
+ mutex_lock(&uld_mutex);
+ for (i = 0; i < CXGB4_ULD_MAX; i++)
+ if (adap->uld_handle[i])
+ ulds[i].state_change(adap->uld_handle[i], new_state);
+ mutex_unlock(&uld_mutex);
+}
+
+/**
+ * cxgb4_register_uld - register an upper-layer driver
+ * @type: the ULD type
+ * @p: the ULD methods
+ *
+ * Registers an upper-layer driver with this driver and notifies the ULD
+ * about any presently available devices that support its type. Returns
+ * %-EBUSY if a ULD of the same type is already registered.
+ */
+int cxgb4_register_uld(enum cxgb4_uld type, const struct cxgb4_uld_info *p)
+{
+ int ret = 0;
+ struct adapter *adap;
+
+ if (type >= CXGB4_ULD_MAX)
+ return -EINVAL;
+ mutex_lock(&uld_mutex);
+ if (ulds[type].add) {
+ ret = -EBUSY;
+ goto out;
+ }
+ ulds[type] = *p;
+ list_for_each_entry(adap, &adapter_list, list_node)
+ uld_attach(adap, type);
+out: mutex_unlock(&uld_mutex);
+ return ret;
+}
+EXPORT_SYMBOL(cxgb4_register_uld);
+
+/**
+ * cxgb4_unregister_uld - unregister an upper-layer driver
+ * @type: the ULD type
+ *
+ * Unregisters an existing upper-layer driver.
+ */
+int cxgb4_unregister_uld(enum cxgb4_uld type)
+{
+ struct adapter *adap;
+
+ if (type >= CXGB4_ULD_MAX)
+ return -EINVAL;
+ mutex_lock(&uld_mutex);
+ list_for_each_entry(adap, &adapter_list, list_node)
+ adap->uld_handle[type] = NULL;
+ ulds[type].add = NULL;
+ mutex_unlock(&uld_mutex);
+ return 0;
+}
+EXPORT_SYMBOL(cxgb4_unregister_uld);
+
+/**
+ * cxgb_up - enable the adapter
+ * @adap: adapter being enabled
+ *
+ * Called when the first port is enabled, this function performs the
+ * actions necessary to make an adapter operational, such as completing
+ * the initialization of HW modules, and enabling interrupts.
+ *
+ * Must be called with the rtnl lock held.
+ */
+static int cxgb_up(struct adapter *adap)
+{
+ int err = 0;
+
+ if (!(adap->flags & FULL_INIT_DONE)) {
+ err = setup_sge_queues(adap);
+ if (err)
+ goto out;
+ err = setup_rss(adap);
+ if (err) {
+ t4_free_sge_resources(adap);
+ goto out;
+ }
+ if (adap->flags & USING_MSIX)
+ name_msix_vecs(adap);
+ adap->flags |= FULL_INIT_DONE;
+ }
+
+ if (adap->flags & USING_MSIX) {
+ err = request_irq(adap->msix_info[0].vec, t4_nondata_intr, 0,
+ adap->msix_info[0].desc, adap);
+ if (err)
+ goto irq_err;
+
+ err = request_msix_queue_irqs(adap);
+ if (err) {
+ free_irq(adap->msix_info[0].vec, adap);
+ goto irq_err;
+ }
+ } else if ((err = request_irq(adap->pdev->irq, t4_intr_handler(adap),
+ (adap->flags & USING_MSI) ? 0 : IRQF_SHARED,
+ adap->name, adap)))
+ goto irq_err;
+
+ enable_rx(adap);
+ t4_sge_start(adap);
+ t4_intr_enable(adap);
+ notify_ulds(adap, CXGB4_STATE_UP);
+ out:
+ return err;
+ irq_err:
+ CH_ERR(adap, "request_irq failed, err %d\n", err);
+ goto out;
+}
+
+static void cxgb_down(struct adapter *adapter)
+{
+ t4_intr_disable(adapter);
+ cancel_work_sync(&adapter->tid_release_task);
+
+ if (adapter->flags & USING_MSIX) {
+ free_msix_queue_irqs(adapter);
+ free_irq(adapter->msix_info[0].vec, adapter);
+ } else
+ free_irq(adapter->pdev->irq, adapter);
+ quiesce_rx(adapter);
+}
+
+/*
+ * net_device operations
+ */
+static int cxgb_open(struct net_device *dev)
+{
+ int err;
+ struct port_info *pi = netdev_priv(dev);
+ struct adapter *adapter = pi->adapter;
+
+ if (!adapter->open_device_map && (err = cxgb_up(adapter)) < 0)
+ return err;
+
+ dev->real_num_tx_queues = pi->nqsets;
+ set_bit(pi->tx_chan, &adapter->open_device_map);
+ link_start(dev);
+ netif_tx_start_all_queues(dev);
+ return 0;
+}
+
+static int cxgb_close(struct net_device *dev)
+{
+ int ret;
+ struct port_info *pi = netdev_priv(dev);
+ struct adapter *adapter = pi->adapter;
+
+ netif_tx_stop_all_queues(dev);
+ netif_carrier_off(dev);
+ ret = t4_enable_vi(adapter, 0, pi->viid, false, false);
+
+ clear_bit(pi->tx_chan, &adapter->open_device_map);
+
+ if (!adapter->open_device_map)
+ cxgb_down(adapter);
+ return 0;
+}
+
+static struct net_device_stats *cxgb_get_stats(struct net_device *dev)
+{
+ struct port_stats stats;
+ struct port_info *p = netdev_priv(dev);
+ struct adapter *adapter = p->adapter;
+ struct net_device_stats *ns = &dev->stats;
+
+ spin_lock(&adapter->stats_lock);
+ t4_get_port_stats(adapter, p->tx_chan, &stats);
+ spin_unlock(&adapter->stats_lock);
+
+ ns->tx_bytes = stats.tx_octets;
+ ns->tx_packets = stats.tx_frames;
+ ns->rx_bytes = stats.rx_octets;
+ ns->rx_packets = stats.rx_frames;
+ ns->multicast = stats.rx_mcast_frames;
+
+ /* detailed rx_errors */
+ ns->rx_length_errors = stats.rx_jabber + stats.rx_too_long +
+ stats.rx_runt;
+ ns->rx_over_errors = 0;
+ ns->rx_crc_errors = stats.rx_fcs_err;
+ ns->rx_frame_errors = stats.rx_symbol_err;
+ ns->rx_fifo_errors = stats.rx_ovflow0 + stats.rx_ovflow1 +
+ stats.rx_ovflow2 + stats.rx_ovflow3 +
+ stats.rx_trunc0 + stats.rx_trunc1 +
+ stats.rx_trunc2 + stats.rx_trunc3;
+ ns->rx_missed_errors = 0;
+
+ /* detailed tx_errors */
+ ns->tx_aborted_errors = 0;
+ ns->tx_carrier_errors = 0;
+ ns->tx_fifo_errors = 0;
+ ns->tx_heartbeat_errors = 0;
+ ns->tx_window_errors = 0;
+
+ ns->tx_errors = stats.tx_error_frames;
+ ns->rx_errors = stats.rx_symbol_err + stats.rx_fcs_err +
+ ns->rx_length_errors + stats.rx_len_err + ns->rx_fifo_errors;
+ return ns;
+}
+
+static int cxgb_ioctl(struct net_device *dev, struct ifreq *req, int cmd)
+{
+ int ret = 0, prtad, devad;
+ struct port_info *pi = netdev_priv(dev);
+ struct mii_ioctl_data *data = (struct mii_ioctl_data *)&req->ifr_data;
+
+ switch (cmd) {
+ case SIOCGMIIPHY:
+ if (pi->mdio_addr < 0)
+ return -EOPNOTSUPP;
+ data->phy_id = pi->mdio_addr;
+ break;
+ case SIOCGMIIREG:
+ case SIOCSMIIREG:
+ if (mdio_phy_id_is_c45(data->phy_id)) {
+ prtad = mdio_phy_id_prtad(data->phy_id);
+ devad = mdio_phy_id_devad(data->phy_id);
+ } else if (data->phy_id < 32) {
+ prtad = data->phy_id;
+ devad = 0;
+ data->reg_num &= 0x1f;
+ } else
+ return -EINVAL;
+
+ if (cmd == SIOCGMIIREG)
+ ret = t4_mdio_rd(pi->adapter, 0, prtad, devad,
+ data->reg_num, &data->val_out);
+ else
+ ret = t4_mdio_wr(pi->adapter, 0, prtad, devad,
+ data->reg_num, data->val_in);
+ break;
+ default:
+ return -EOPNOTSUPP;
+ }
+ return ret;
+}
+
+static void cxgb_set_rxmode(struct net_device *dev)
+{
+ /* unfortunately we can't return errors to the stack */
+ set_rxmode(dev, -1, false);
+}
+
+static int cxgb_change_mtu(struct net_device *dev, int new_mtu)
+{
+ int ret;
+ struct port_info *pi = netdev_priv(dev);
+
+ if (new_mtu < 81 || new_mtu > MAX_MTU) /* accommodate SACK */
+ return -EINVAL;
+ ret = t4_set_rxmode(pi->adapter, 0, pi->viid, new_mtu, -1, -1, -1,
+ true);
+ if (!ret)
+ dev->mtu = new_mtu;
+ return ret;
+}
+
+static int cxgb_set_mac_addr(struct net_device *dev, void *p)
+{
+ int ret;
+ struct sockaddr *addr = p;
+ struct port_info *pi = netdev_priv(dev);
+
+ if (!is_valid_ether_addr(addr->sa_data))
+ return -EINVAL;
+
+ ret = t4_change_mac(pi->adapter, 0, pi->viid, pi->xact_addr_filt,
+ addr->sa_data, true);
+ if (ret < 0)
+ return ret;
+
+ memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
+ pi->xact_addr_filt = ret;
+ return 0;
+}
+
+static void vlan_rx_register(struct net_device *dev, struct vlan_group *grp)
+{
+ struct port_info *pi = netdev_priv(dev);
+
+ pi->vlan_grp = grp;
+ t4_set_vlan_accel(pi->adapter, 1 << pi->tx_chan, grp != NULL);
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+static void cxgb_netpoll(struct net_device *dev)
+{
+ int i;
+ struct port_info *pi = netdev_priv(dev);
+ struct adapter *adap = pi->adapter;
+ struct sge_eth_rxq *rx = &adap->sge.ethrxq[pi->first_qset];
+
+ for (i = 0; i < pi->nqsets; i++, rx++)
+ t4_intr_handler(adap)((adap->flags & USING_MSIX) ? &rx->rspq :
+ adap);
+}
+#endif
+
+static const struct net_device_ops cxgb4_netdev_ops = {
+ .ndo_open = cxgb_open,
+ .ndo_stop = cxgb_close,
+ .ndo_start_xmit = t4_eth_xmit,
+ .ndo_get_stats = cxgb_get_stats,
+ .ndo_set_rx_mode = cxgb_set_rxmode,
+ .ndo_set_mac_address = cxgb_set_mac_addr,
+ .ndo_validate_addr = eth_validate_addr,
+ .ndo_do_ioctl = cxgb_ioctl,
+ .ndo_change_mtu = cxgb_change_mtu,
+ .ndo_vlan_rx_register = vlan_rx_register,
+#ifdef CONFIG_NET_POLL_CONTROLLER
+ .ndo_poll_controller = cxgb_netpoll,
+#endif
+};
+
+void t4_fatal_err(struct adapter *adapter)
+{
+ t4_set_reg_field(adapter, A_SGE_CONTROL, F_GLOBALENABLE, 0);
+ CH_ALERT(adapter, "encountered fatal error, operation suspended\n");
+}
+
+static void setup_memwin(struct adapter *adap)
+{
+ u32 bar0;
+
+ bar0 = pci_resource_start(adap->pdev, 0); /* truncation intentional */
+ t4_write_reg(adap, PCIE_MEM_ACCESS_REG(A_PCIE_MEM_ACCESS_BASE_WIN, 0),
+ (bar0 + MEMWIN0_BASE) | V_BIR(0) |
+ V_WINDOW(ilog2(MEMWIN0_APERTURE) - 10));
+ t4_write_reg(adap, PCIE_MEM_ACCESS_REG(A_PCIE_MEM_ACCESS_BASE_WIN, 1),
+ (bar0 + MEMWIN1_BASE) | V_BIR(0) |
+ V_WINDOW(ilog2(MEMWIN1_APERTURE) - 10));
+ t4_write_reg(adap, PCIE_MEM_ACCESS_REG(A_PCIE_MEM_ACCESS_BASE_WIN, 2),
+ (bar0 + MEMWIN2_BASE) | V_BIR(0) |
+ V_WINDOW(ilog2(MEMWIN2_APERTURE) - 10));
+}
+
+/*
+ * Max # of ATIDs. The absolute HW max is 16K but we keep it lower.
+ */
+#define MAX_ATIDS 8192U
+
+/*
+ * Phase 0 of initialization: contact FW, obtain config, perform basic init.
+ */
+static int adap_init0(struct adapter *adap)
+{
+ int ret;
+ u32 v, port_vec;
+ enum dev_state state;
+ u32 params[7], val[7];
+ struct fw_caps_config_cmd c;
+
+ ret = t4_check_fw_version(adap);
+ if (ret == -EINVAL) {
+ ret = upgrade_fw(adap);
+ if (ret == 0) /* recache FW version */
+ ret = t4_check_fw_version(adap);
+ }
+ if (ret < 0)
+ return ret;
+
+ /* contact FW, request master */
+ ret = t4_fw_hello(adap, 0, 0, MASTER_MUST, &state);
+ if (ret < 0) {
+ CH_ERR(adap, "could not connect to FW, error %d\n", ret);
+ return ret;
+ }
+
+ /* reset device */
+ ret = t4_fw_reset(adap, 0, F_PIORSTMODE | F_PIORST);
+ if (ret < 0)
+ goto bye;
+
+ /* get device capabilities */
+ memset(&c, 0, sizeof(c));
+ c.op_to_write = htonl(V_FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+ F_FW_CMD_REQUEST | F_FW_CMD_READ);
+ c.retval_len16 = htonl(FW_LEN16(c));
+ ret = t4_wr_mbox(adap, 0, &c, sizeof(c), &c);
+ if (ret < 0)
+ goto bye;
+
+ /* select capabilities we'll be using */
+ if (c.niccaps & htons(FW_CAPS_CONFIG_NIC_VM)) {
+ if (!vf_acls)
+ c.niccaps ^= htons(FW_CAPS_CONFIG_NIC_VM);
+ else
+ c.niccaps = htons(FW_CAPS_CONFIG_NIC_VM);
+ } else if (vf_acls) {
+ CH_ERR(adap, "virtualization ACLs not supported");
+ goto bye;
+ }
+ c.op_to_write = htonl(V_FW_CMD_OP(FW_CAPS_CONFIG_CMD) |
+ F_FW_CMD_REQUEST | F_FW_CMD_WRITE);
+ ret = t4_wr_mbox(adap, 0, &c, sizeof(c), NULL);
+ if (ret < 0)
+ goto bye;
+
+ ret = t4_config_glbl_rss(adap, 0,
+ FW_RSS_GLB_CONFIG_CMD_MODE_BASICVIRTUAL,
+ F_FW_RSS_GLB_CONFIG_CMD_TNLMAPEN |
+ F_FW_RSS_GLB_CONFIG_CMD_TNLALLLKP);
+ if (ret < 0)
+ goto bye;
+
+ ret = t4_cfg_pfvf(adap, 0, 0, 0, 64, 64, 64, 0, 0, 4,
+ M_FW_PFVF_CMD_CMASK, M_FW_PFVF_CMD_PMASK, 16,
+ FW_CMD_CAP_PF, FW_CMD_CAP_PF);
+ if (ret < 0)
+ goto bye;
+
+ for (v = 0; v < SGE_NTIMERS - 1; v++)
+ adap->sge.timer_val[v] = min(intr_holdoff[v], MAX_SGE_TIMERVAL);
+ adap->sge.timer_val[SGE_NTIMERS - 1] = MAX_SGE_TIMERVAL;
+ adap->sge.counter_val[0] = 1;
+ for (v = 1; v < SGE_NCOUNTERS; v++)
+ adap->sge.counter_val[v] = min(intr_cnt[v - 1], M_THRESHOLD_0);
+ t4_sge_init(adap);
+
+ /* get basic stuff going */
+ ret = t4_early_init(adap, 0);
+ if (ret < 0)
+ goto bye;
+
+#define FW_PARAM_DEV(param) \
+ V_FW_PARAMS_MNEM(FW_PARAMS_MNEM_DEV) | \
+ V_FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_DEV_##param)
+
+#define FW_PARAM_PFVF(param) \
+ V_FW_PARAMS_MNEM(FW_PARAMS_MNEM_PFVF) | \
+ V_FW_PARAMS_PARAM_X(FW_PARAMS_PARAM_PFVF_##param)
+
+ params[0] = FW_PARAM_DEV(PORTVEC);
+ params[1] = FW_PARAM_PFVF(L2T_START);
+ params[2] = FW_PARAM_PFVF(L2T_END);
+ params[3] = FW_PARAM_PFVF(FILTER_START);
+ params[4] = FW_PARAM_PFVF(FILTER_END);
+ ret = t4_query_params(adap, 0, 5, params, val);
+ if (ret < 0)
+ goto bye;
+ port_vec = val[0];
+ adap->tids.ftid_base = val[3];
+ adap->tids.nftids = val[4] - val[3] + 1;
+
+ if (c.ofldcaps) {
+ /* query offload-related parameters */
+ params[0] = FW_PARAM_DEV(NTID);
+ params[1] = FW_PARAM_PFVF(SERVER_START);
+ params[2] = FW_PARAM_PFVF(SERVER_END);
+ params[3] = FW_PARAM_PFVF(TDDP_START);
+ params[4] = FW_PARAM_PFVF(TDDP_END);
+ params[5] = FW_PARAM_DEV(FLOWC_BUFFIFO_SZ);
+ ret = t4_query_params(adap, 0, 6, params, val);
+ if (ret < 0)
+ goto bye;
+ adap->tids.ntids = val[0];
+ adap->tids.natids = min(adap->tids.ntids / 2, MAX_ATIDS);
+ adap->tids.stid_base = val[1];
+ adap->tids.nstids = val[2] - val[1] + 1;
+ adap->vres.ddp.start = val[3];
+ adap->vres.ddp.size = val[4] - val[3] + 1;
+ adap->params.ofldq_wr_cred = val[5];
+ adap->params.offload = 1;
+ }
+ if (c.rdmacaps) {
+ params[0] = FW_PARAM_PFVF(STAG_START);
+ params[1] = FW_PARAM_PFVF(STAG_END);
+ params[2] = FW_PARAM_PFVF(RQ_START);
+ params[3] = FW_PARAM_PFVF(RQ_END);
+ params[4] = FW_PARAM_PFVF(PBL_START);
+ params[5] = FW_PARAM_PFVF(PBL_END);
+ ret = t4_query_params(adap, 0, 6, params, val);
+ if (ret < 0)
+ goto bye;
+ adap->vres.stag.start = val[0];
+ adap->vres.stag.size = val[1] - val[0] + 1;
+ adap->vres.rq.start = val[2];
+ adap->vres.rq.size = val[3] - val[2] + 1;
+ adap->vres.pbl.start = val[4];
+ adap->vres.pbl.size = val[5] - val[4] + 1;
+ }
+ if (c.iscsicaps) {
+ params[0] = FW_PARAM_PFVF(ISCSI_START);
+ params[1] = FW_PARAM_PFVF(ISCSI_END);
+ ret = t4_query_params(adap, 0, 2, params, val);
+ if (ret < 0)
+ goto bye;
+ adap->vres.iscsi.start = val[0];
+ adap->vres.iscsi.size = val[1] - val[0] + 1;
+ }
+#undef FW_PARAM_PFVF
+#undef FW_PARAM_DEV
+
+ adap->params.nports = hweight32(port_vec);
+ adap->params.portvec = port_vec;
+ adap->flags |= FW_OK;
+
+ /* These are finalized by FW initialization, load their values now */
+ v = t4_read_reg(adap, A_TP_TIMER_RESOLUTION);
+ adap->params.tp.tre = G_TIMERRESOLUTION(v);
+ t4_read_mtu_tbl(adap, adap->params.mtus, NULL);
+ t4_load_mtus(adap, adap->params.mtus, adap->params.a_wnd,
+ adap->params.b_wnd);
+
+ /* tweak some settings */
+ t4_write_reg(adap, A_TP_SHIFT_CNT, V_SYNSHIFTMAX(6) |
+ V_RXTSHIFTMAXR1(4) | V_RXTSHIFTMAXR2(15) |
+ V_PERSHIFTBACKOFFMAX(8) | V_PERSHIFTMAX(8) |
+ V_KEEPALIVEMAXR1(4) | V_KEEPALIVEMAXR2(9));
+ t4_write_reg(adap, A_ULP_RX_TDDP_PSZ, V_HPZ0(PAGE_SHIFT - 12));
+ t4_write_reg(adap, A_TP_PIO_ADDR, A_TP_INGRESS_CONFIG);
+ v = t4_read_reg(adap, A_TP_PIO_DATA);
+ t4_write_reg(adap, A_TP_PIO_DATA, v & ~F_CSUM_HAS_PSEUDO_HDR);
+ setup_memwin(adap);
+ return 0;
+
+ /*
+ * If a command timed out or failed with EIO FW does not operate within
+ * its spec or something catastrophic happened to HW/FW, stop issuing
+ * commands.
+ */
+bye: if (ret != -ETIMEDOUT && ret != -EIO)
+ t4_fw_bye(adap, 0);
+ return ret;
+}
+
+static inline bool is_10g_port(const struct link_config *lc)
+{
+ return (lc->supported & FW_PORT_CAP_SPEED_10G) != 0;
+}
+
+static inline void init_rspq(struct sge_rspq *q, u8 timer_idx, u8 pkt_cnt_idx,
+ unsigned int size, unsigned int iqe_size)
+{
+ q->intr_params = V_QINTR_TIMER_IDX(timer_idx) |
+ V_QINTR_CNT_EN(pkt_cnt_idx < SGE_NCOUNTERS);
+ q->pktcnt_idx = pkt_cnt_idx < SGE_NCOUNTERS ? pkt_cnt_idx : 0;
+ q->iqe_len = iqe_size;
+ q->size = size;
+}
+
+/*
+ * Perform default configuration of DMA queues depending on the number and type
+ * of ports we found and the number of available CPUs. Most settings can be
+ * modified by the admin prior to actual use.
+ */
+static void __devinit cfg_queues(struct adapter *adap)
+{
+ struct sge *s = &adap->sge;
+ int i, q10g = 0, n10g = 0, qidx = 0;
+
+ for_each_port(adap, i)
+ n10g += is_10g_port(&adap2pinfo(adap, i)->link_cfg);
+
+ /*
+ * We default to 1 queue per non-10G port and up to # of cores queues
+ * per 10G port.
+ */
+ if (n10g)
+ q10g = (MAX_ETH_QSETS - (adap->params.nports - n10g)) / n10g;
+ if (q10g > num_online_cpus())
+ q10g = num_online_cpus();
+
+ for_each_port(adap, i) {
+ struct port_info *pi = adap2pinfo(adap, i);
+
+ pi->first_qset = qidx;
+ pi->nqsets = is_10g_port(&pi->link_cfg) ? q10g : 1;
+ qidx += pi->nqsets;
+ }
+
+ s->ethqsets = qidx;
+ s->max_ethqsets = qidx; /* MSI-X may lower it later */
+
+ if (is_offload(adap)) {
+ /*
+ * For offload we use 1 queue/channel if all ports are up to 1G,
+ * otherwise we divide all available queues amongst the channels
+ * capped by the number of available cores.
+ */
+ if (n10g) {
+ i = min_t(int, ARRAY_SIZE(s->ofldrxq),
+ num_online_cpus());
+ s->ofldqsets = roundup(i, adap->params.nports);
+ } else
+ s->ofldqsets = adap->params.nports;
+ /* For RDMA one Rx queue per channel suffices */
+ s->rdmaqs = adap->params.nports;
+ }
+
+ for (i = 0; i < ARRAY_SIZE(s->ethrxq); i++) {
+ struct sge_eth_rxq *r = &s->ethrxq[i];
+
+ init_rspq(&r->rspq, 0, 0, 1024, 64);
+ r->fl.size = 72;
+ }
+
+ for (i = 0; i < ARRAY_SIZE(s->ethtxq); i++)
+ s->ethtxq[i].q.size = 1024;
+
+ for (i = 0; i < ARRAY_SIZE(s->ctrlq); i++)
+ s->ctrlq[i].q.size = 512;
+
+ for (i = 0; i < ARRAY_SIZE(s->ofldtxq); i++)
+ s->ofldtxq[i].q.size = 1024;
+
+ for (i = 0; i < ARRAY_SIZE(s->ofldrxq); i++) {
+ struct sge_ofld_rxq *r = &s->ofldrxq[i];
+
+ init_rspq(&r->rspq, 0, 0, 1024, 64);
+ r->rspq.uld = CXGB4_ULD_ISCSI;
+ r->fl.size = 72;
+ }
+
+ for (i = 0; i < ARRAY_SIZE(s->rdmarxq); i++) {
+ struct sge_ofld_rxq *r = &s->rdmarxq[i];
+
+ init_rspq(&r->rspq, 0, 0, 511, 64);
+ r->rspq.uld = CXGB4_ULD_RDMA;
+ r->fl.size = 72;
+ }
+
+ init_rspq(&s->fw_evtq, E_TIMERIDX_NONE, 0, 512, 64);
+ init_rspq(&s->intrq, E_TIMERIDX_NONE, 0, 2 * MAX_INGQ, 64);
+}
+
+/*
+ * Reduce the number of Ethernet queues across all ports to at most n.
+ * n provides at least one queue per port.
+ */
+static void __devinit reduce_ethqs(struct adapter *adap, int n)
+{
+ int i;
+ struct port_info *pi;
+
+ while (n < adap->sge.ethqsets)
+ for_each_port(adap, i) {
+ pi = adap2pinfo(adap, i);
+ if (pi->nqsets > 1) {
+ pi->nqsets--;
+ adap->sge.ethqsets--;
+ if (adap->sge.ethqsets <= n)
+ break;
+ }
+ }
+
+ n = 0;
+ for_each_port(adap, i) {
+ pi = adap2pinfo(adap, i);
+ pi->first_qset = n;
+ n += pi->nqsets;
+ }
+}
+
+/* 2 MSI-X vectors needed for the FW queue and non-data interrupts */
+#define EXTRA_VECS 2
+
+static int __devinit enable_msix(struct adapter *adap)
+{
+ int ofld_need = 0;
+ int i, err, want, need;
+ struct sge *s = &adap->sge;
+ unsigned int nchan = adap->params.nports;
+ struct msix_entry entries[MAX_INGQ + 1];
+
+ for (i = 0; i < ARRAY_SIZE(entries); ++i)
+ entries[i].entry = i;
+
+ want = s->max_ethqsets + EXTRA_VECS;
+ if (is_offload(adap)) {
+ want += s->rdmaqs + s->ofldqsets;
+ /* need nchan for each possible ULD */
+ ofld_need = 2 * nchan;
+ }
+ need = adap->params.nports + EXTRA_VECS + ofld_need;
+
+ while ((err = pci_enable_msix(adap->pdev, entries, want)) >= need)
+ want = err;
+
+ if (!err) {
+ /*
+ * Distribute available vectors to the various queue groups.
+ * Every group gets its minimum requirement and NIC gets top
+ * priority for leftovers.
+ */
+ i = want - EXTRA_VECS - ofld_need;
+ if (i < s->max_ethqsets) {
+ s->max_ethqsets = i;
+ if (i < s->ethqsets)
+ reduce_ethqs(adap, i);
+ }
+ if (is_offload(adap)) {
+ i = want - EXTRA_VECS - s->max_ethqsets;
+ i -= ofld_need - nchan;
+ s->ofldqsets = (i / nchan) * nchan; /* round down */
+ }
+ for (i = 0; i < want; ++i)
+ adap->msix_info[i].vec = entries[i].vector;
+ } else if (err > 0)
+ dev_info(adap->pdev_dev,
+ "only %d MSI-X vectors left, not using MSI-X\n", err);
+ return err;
+}
+
+#undef EXTRA_VECS
+
+static void __devinit print_port_info(struct adapter *adap)
+{
+ static const char *base[] = {
+ "R", "KX4", "T", "KX", "T", "KR", "CX4"
+ };
+
+ int i;
+ char buf[80];
+
+ for_each_port(adap, i) {
+ struct net_device *dev = adap->port[i];
+ const struct port_info *pi = netdev_priv(dev);
+ char *bufp = buf;
+
+ if (!test_bit(i, &adap->registered_device_map))
+ continue;
+
+ if (pi->link_cfg.supported & FW_PORT_CAP_SPEED_100M)
+ bufp += sprintf(bufp, "100/");
+ if (pi->link_cfg.supported & FW_PORT_CAP_SPEED_1G)
+ bufp += sprintf(bufp, "1000/");
+ if (pi->link_cfg.supported & FW_PORT_CAP_SPEED_10G)
+ bufp += sprintf(bufp, "10G/");
+ if (bufp != buf)
+ --bufp;
+ sprintf(bufp, "BASE-%s", base[pi->port_type]);
+
+ printk(KERN_INFO "%s: Chelsio %s rev %d %s %sNIC PCIe x%d%s\n",
+ dev->name, adap->params.vpd.id, adap->params.rev,
+ buf, is_offload(adap) ? "R" : "", adap->params.pci.width,
+ (adap->flags & USING_MSIX) ? " MSI-X" :
+ (adap->flags & USING_MSI) ? " MSI" : "");
+ if (adap->name == dev->name)
+ printk(KERN_INFO "%s: S/N: %s, E/C: %s\n", adap->name,
+ adap->params.vpd.sn, adap->params.vpd.ec);
+ }
+}
+
+#define VLAN_FEAT (NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_TSO | NETIF_F_TSO6 |\
+ NETIF_F_IPV6_CSUM | NETIF_F_HIGHDMA)
+
+static int __devinit init_one(struct pci_dev *pdev,
+ const struct pci_device_id *ent)
+{
+ int func;
+ struct port_info *pi;
+ int i, err, pci_using_dac = 0;
+ struct adapter *adapter = NULL;
+
+ printk_once(KERN_INFO "%s - version %s\n", DRV_DESC, DRV_VERSION);
+
+ err = pci_request_regions(pdev, KBUILD_MODNAME);
+ if (err) {
+ /* Just info, some other driver may have claimed the device. */
+ dev_info(&pdev->dev, "cannot obtain PCI resources\n");
+ return err;
+ }
+
+ /* We control everything through PF 0 */
+ func = PCI_FUNC(pdev->devfn);
+ if (func > 0)
+ goto sriov;
+
+ err = pci_enable_device(pdev);
+ if (err) {
+ dev_err(&pdev->dev, "cannot enable PCI device\n");
+ goto out_release_regions;
+ }
+
+ if (!pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) {
+ pci_using_dac = 1;
+ err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+ if (err) {
+ dev_err(&pdev->dev, "unable to obtain 64-bit DMA for "
+ "coherent allocations\n");
+ goto out_disable_device;
+ }
+ } else if ((err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32))) != 0) {
+ dev_err(&pdev->dev, "no usable DMA configuration\n");
+ goto out_disable_device;
+ }
+
+ pci_enable_pcie_error_reporting(pdev);
+ pci_set_master(pdev);
+ pci_save_state(pdev);
+
+ adapter = kzalloc(sizeof(*adapter), GFP_KERNEL);
+ if (!adapter) {
+ err = -ENOMEM;
+ goto out_disable_device;
+ }
+
+ adapter->regs = pci_ioremap_bar(pdev, 0);
+ if (!adapter->regs) {
+ dev_err(&pdev->dev, "cannot map device registers\n");
+ err = -ENOMEM;
+ goto out_free_adapter;
+ }
+
+ adapter->pdev = pdev;
+ adapter->pdev_dev = &pdev->dev;
+ adapter->name = pci_name(pdev);
+ adapter->msg_enable = dflt_msg_enable;
+ memset(adapter->chan_map, 0xff, sizeof(adapter->chan_map));
+
+ spin_lock_init(&adapter->stats_lock);
+ spin_lock_init(&adapter->tid_release_lock);
+
+ INIT_WORK(&adapter->tid_release_task, process_tid_release_list);
+
+ err = t4_prep_adapter(adapter);
+ if (err)
+ goto out_unmap_bar;
+ err = adap_init0(adapter);
+ if (err)
+ goto out_unmap_bar;
+
+ for_each_port(adapter, i) {
+ struct net_device *netdev;
+
+ netdev = alloc_etherdev_mq(sizeof(struct port_info),
+ MAX_ETH_QSETS);
+ if (!netdev) {
+ err = -ENOMEM;
+ goto out_free_dev;
+ }
+
+ SET_NETDEV_DEV(netdev, &pdev->dev);
+
+ adapter->port[i] = netdev;
+ pi = netdev_priv(netdev);
+ pi->adapter = adapter;
+ pi->xact_addr_filt = -1;
+ pi->rx_offload = RX_CSO;
+ pi->port_id = i;
+ netif_carrier_off(netdev);
+ netif_tx_stop_all_queues(netdev);
+ netdev->irq = pdev->irq;
+
+ netdev->features |= NETIF_F_SG | NETIF_F_TSO | NETIF_F_TSO6;
+ netdev->features |= NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM;
+ netdev->features |= NETIF_F_GRO;
+ if (pci_using_dac)
+ netdev->features |= NETIF_F_HIGHDMA;
+
+ netdev->features |= NETIF_F_HW_VLAN_TX | NETIF_F_HW_VLAN_RX;
+ netdev->vlan_features = netdev->features & VLAN_FEAT;
+
+ netdev->netdev_ops = &cxgb4_netdev_ops;
+ SET_ETHTOOL_OPS(netdev, &cxgb_ethtool_ops);
+ }
+
+ pci_set_drvdata(pdev, adapter);
+
+ if (adapter->flags & FW_OK) {
+ err = t4_port_init(adapter, 0, 0, 0);
+ if (err)
+ goto out_free_dev;
+ }
+
+ /*
+ * Configure queues and allocate tables now, they can be needed as
+ * soon as the first register_netdev completes.
+ */
+ cfg_queues(adapter);
+
+ adapter->l2t = t4_init_l2t();
+ if (!adapter->l2t) {
+ /* We tolerate a lack of L2T, giving up some functionality */
+ dev_warn(&pdev->dev, "could not allocate L2T, continuing\n");
+ adapter->params.offload = 0;
+ }
+
+ if (is_offload(adapter) && tid_init(&adapter->tids) < 0) {
+ dev_warn(&pdev->dev, "could not allocate TID table, "
+ "continuing\n");
+ adapter->params.offload = 0;
+ }
+
+ /*
+ * The card is now ready to go. If any errors occur during device
+ * registration we do not fail the whole card but rather proceed only
+ * with the ports we manage to register successfully. However we must
+ * register at least one net device.
+ */
+ for_each_port(adapter, i) {
+ err = register_netdev(adapter->port[i]);
+ if (err)
+ dev_warn(&pdev->dev,
+ "cannot register net device %s, skipping\n",
+ adapter->port[i]->name);
+ else {
+ /*
+ * Change the name we use for messages to the name of
+ * the first successfully registered interface.
+ */
+ if (!adapter->registered_device_map)
+ adapter->name = adapter->port[i]->name;
+
+ __set_bit(i, &adapter->registered_device_map);
+ adapter->chan_map[adap2pinfo(adapter, i)->tx_chan] = i;
+ }
+ }
+ if (!adapter->registered_device_map) {
+ dev_err(&pdev->dev, "could not register any net devices\n");
+ goto out_free_dev;
+ }
+
+ if (cxgb4_proc_root) {
+ adapter->proc_root = proc_mkdir(pci_name(pdev),
+ cxgb4_proc_root);
+ if (!adapter->proc_root)
+ dev_warn(&pdev->dev,
+ "could not create /proc directory");
+ else
+ setup_proc(adapter, adapter->proc_root);
+ }
+
+ if (cxgb4_debugfs_root) {
+ adapter->debugfs_root = debugfs_create_dir(pci_name(pdev),
+ cxgb4_debugfs_root);
+ setup_debugfs(adapter);
+ }
+
+ /* See what interrupts we'll be using */
+ if (msi > 1 && enable_msix(adapter) == 0)
+ adapter->flags |= USING_MSIX;
+ else if (msi > 0 && pci_enable_msi(pdev) == 0)
+ adapter->flags |= USING_MSI;
+
+ if (is_offload(adapter)) {
+ attach_ulds(adapter);
+ }
+
+ print_port_info(adapter);
+
+sriov:
+#ifdef CONFIG_PCI_IOV
+ if (func < ARRAY_SIZE(num_vf) && num_vf[func] > 0)
+ if (pci_enable_sriov(pdev, num_vf[func]) == 0)
+ dev_info(&pdev->dev,
+ "instantiated %u virtual functions\n",
+ num_vf[func]);
+#endif
+ return 0;
+
+ out_free_dev:
+ t4_free_mem(adapter->tids.tid_tab);
+ t4_free_mem(adapter->l2t);
+ for_each_port(adapter, i)
+ if (adapter->port[i])
+ free_netdev(adapter->port[i]);
+ if (adapter->flags & FW_OK)
+ t4_fw_bye(adapter, 0);
+ out_unmap_bar:
+ iounmap(adapter->regs);
+ out_free_adapter:
+ kfree(adapter);
+ out_disable_device:
+ pci_disable_pcie_error_reporting(pdev);
+ pci_disable_device(pdev);
+ out_release_regions:
+ pci_release_regions(pdev);
+ pci_set_drvdata(pdev, NULL);
+ return err;
+}
+
+static void __devexit remove_one(struct pci_dev *pdev)
+{
+ struct adapter *adapter = pci_get_drvdata(pdev);
+
+ pci_disable_sriov(pdev);
+
+ if (adapter) {
+ int i;
+
+ if (is_offload(adapter))
+ detach_ulds(adapter);
+
+ t4_sge_stop(adapter);
+
+ for_each_port(adapter, i)
+ if (test_bit(i, &adapter->registered_device_map))
+ unregister_netdev(adapter->port[i]);
+
+ if (adapter->proc_root) {
+ cleanup_proc(adapter->proc_root);
+ remove_proc_entry(pci_name(pdev), cxgb4_proc_root);
+ }
+
+ if (adapter->debugfs_root) {
+ debugfs_remove_recursive(adapter->debugfs_root);
+ }
+
+ t4_free_sge_resources(adapter);
+ t4_free_mem(adapter->l2t);
+ t4_free_mem(adapter->tids.tid_tab);
+ disable_msi(adapter);
+
+ for_each_port(adapter, i)
+ if (adapter->port[i])
+ free_netdev(adapter->port[i]);
+
+ if (adapter->flags & FW_OK)
+ t4_fw_bye(adapter, 0);
+ iounmap(adapter->regs);
+ kfree(adapter);
+ pci_disable_pcie_error_reporting(pdev);
+ pci_disable_device(pdev);
+ pci_release_regions(pdev);
+ pci_set_drvdata(pdev, NULL);
+ } else if (PCI_FUNC(pdev->devfn) > 0)
+ pci_release_regions(pdev);
+}
+
+static struct pci_driver cxgb4_driver = {
+ .name = KBUILD_MODNAME,
+ .id_table = cxgb4_pci_tbl,
+ .probe = init_one,
+ .remove = __devexit_p(remove_one),
+};
+
+#define DRV_PROC_NAME "driver/" KBUILD_MODNAME
+
+static int __init cxgb4_init_module(void)
+{
+ int ret;
+
+ /* Debugfs support is optional, just warn if this fails */
+ cxgb4_debugfs_root = debugfs_create_dir(KBUILD_MODNAME, NULL);
+ if (!cxgb4_debugfs_root)
+ pr_warning("could not create debugfs entry, continuing\n");
+
+#ifdef CONFIG_PROC_FS
+ cxgb4_proc_root = proc_mkdir(DRV_PROC_NAME, NULL);
+ if (!cxgb4_proc_root)
+ pr_warning("could not create /proc driver directory, "
+ "continuing\n");
+#endif
+
+ ret = pci_register_driver(&cxgb4_driver);
+ if (ret < 0) {
+ remove_proc_entry(DRV_PROC_NAME, NULL);
+ debugfs_remove(cxgb4_debugfs_root);
+ }
+ return ret;
+}
+
+static void __exit cxgb4_cleanup_module(void)
+{
+ pci_unregister_driver(&cxgb4_driver);
+ remove_proc_entry(DRV_PROC_NAME, NULL);
+ debugfs_remove(cxgb4_debugfs_root); /* NULL ok */
+}
+
+module_init(cxgb4_init_module);
+module_exit(cxgb4_cleanup_module);
--
1.5.4
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists