[deliverable/linux.git] / drivers / infiniband / hw / cxgb4 / device.c

/*
 * Copyright (c) 2009-2010 Chelsio, 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.
 */
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/debugfs.h>

#include <rdma/ib_verbs.h>

#include "iw_cxgb4.h"

#define DRV_VERSION "0.1"

MODULE_AUTHOR("Steve Wise");
MODULE_DESCRIPTION("Chelsio T4 RDMA Driver");
MODULE_LICENSE("Dual BSD/GPL");
MODULE_VERSION(DRV_VERSION);

static LIST_HEAD(dev_list);
static DEFINE_MUTEX(dev_mutex);

static struct dentry *c4iw_debugfs_root;

struct c4iw_debugfs_data {
	struct c4iw_dev *devp;
	char *buf;
	int bufsize;
	int pos;
};

static int count_idrs(int id, void *p, void *data)
{
	int *countp = data;

	*countp = *countp + 1;
	return 0;
}

static ssize_t debugfs_read(struct file *file, char __user *buf, size_t count,
			    loff_t *ppos)
{
	struct c4iw_debugfs_data *d = file->private_data;

	return simple_read_from_buffer(buf, count, ppos, d->buf, d->pos);
}

static int dump_qp(int id, void *p, void *data)
{
	struct c4iw_qp *qp = p;
	struct c4iw_debugfs_data *qpd = data;
	int space;
	int cc;

	if (id != qp->wq.sq.qid)
		return 0;

	space = qpd->bufsize - qpd->pos - 1;
	if (space == 0)
		return 1;

	if (qp->ep)
		cc = snprintf(qpd->buf + qpd->pos, space, "qp id %u state %u "
			     "ep tid %u state %u %pI4:%u->%pI4:%u\n",
			     qp->wq.sq.qid, (int)qp->attr.state,
			     qp->ep->hwtid, (int)qp->ep->com.state,
			     &qp->ep->com.local_addr.sin_addr.s_addr,
			     ntohs(qp->ep->com.local_addr.sin_port),
			     &qp->ep->com.remote_addr.sin_addr.s_addr,
			     ntohs(qp->ep->com.remote_addr.sin_port));
	else
		cc = snprintf(qpd->buf + qpd->pos, space, "qp id %u state %u\n",
			      qp->wq.sq.qid, (int)qp->attr.state);
	if (cc < space)
		qpd->pos += cc;
	return 0;
}

static int qp_release(struct inode *inode, struct file *file)
{
	struct c4iw_debugfs_data *qpd = file->private_data;
	if (!qpd) {
		printk(KERN_INFO "%s null qpd?\n", __func__);
		return 0;
	}
	kfree(qpd->buf);
	kfree(qpd);
	return 0;
}

static int qp_open(struct inode *inode, struct file *file)
{
	struct c4iw_debugfs_data *qpd;
	int ret = 0;
	int count = 1;

	qpd = kmalloc(sizeof *qpd, GFP_KERNEL);
	if (!qpd) {
		ret = -ENOMEM;
		goto out;
	}
	qpd->devp = inode->i_private;
	qpd->pos = 0;

	spin_lock_irq(&qpd->devp->lock);
	idr_for_each(&qpd->devp->qpidr, count_idrs, &count);
	spin_unlock_irq(&qpd->devp->lock);

	qpd->bufsize = count * 128;
	qpd->buf = kmalloc(qpd->bufsize, GFP_KERNEL);
	if (!qpd->buf) {
		ret = -ENOMEM;
		goto err1;
	}

	spin_lock_irq(&qpd->devp->lock);
	idr_for_each(&qpd->devp->qpidr, dump_qp, qpd);
	spin_unlock_irq(&qpd->devp->lock);

	qpd->buf[qpd->pos++] = 0;
	file->private_data = qpd;
	goto out;
err1:
	kfree(qpd);
out:
	return ret;
}

static const struct file_operations qp_debugfs_fops = {
	.owner   = THIS_MODULE,
	.open    = qp_open,
	.release = qp_release,
	.read    = debugfs_read,
	.llseek  = default_llseek,
};

static int dump_stag(int id, void *p, void *data)
{
	struct c4iw_debugfs_data *stagd = data;
	int space;
	int cc;

	space = stagd->bufsize - stagd->pos - 1;
	if (space == 0)
		return 1;

	cc = snprintf(stagd->buf + stagd->pos, space, "0x%x\n", id<<8);
	if (cc < space)
		stagd->pos += cc;
	return 0;
}

static int stag_release(struct inode *inode, struct file *file)
{
	struct c4iw_debugfs_data *stagd = file->private_data;
	if (!stagd) {
		printk(KERN_INFO "%s null stagd?\n", __func__);
		return 0;
	}
	kfree(stagd->buf);
	kfree(stagd);
	return 0;
}

static int stag_open(struct inode *inode, struct file *file)
{
	struct c4iw_debugfs_data *stagd;
	int ret = 0;
	int count = 1;

	stagd = kmalloc(sizeof *stagd, GFP_KERNEL);
	if (!stagd) {
		ret = -ENOMEM;
		goto out;
	}
	stagd->devp = inode->i_private;
	stagd->pos = 0;

	spin_lock_irq(&stagd->devp->lock);
	idr_for_each(&stagd->devp->mmidr, count_idrs, &count);
	spin_unlock_irq(&stagd->devp->lock);

	stagd->bufsize = count * sizeof("0x12345678\n");
	stagd->buf = kmalloc(stagd->bufsize, GFP_KERNEL);
	if (!stagd->buf) {
		ret = -ENOMEM;
		goto err1;
	}

	spin_lock_irq(&stagd->devp->lock);
	idr_for_each(&stagd->devp->mmidr, dump_stag, stagd);
	spin_unlock_irq(&stagd->devp->lock);

	stagd->buf[stagd->pos++] = 0;
	file->private_data = stagd;
	goto out;
err1:
	kfree(stagd);
out:
	return ret;
}

static const struct file_operations stag_debugfs_fops = {
	.owner   = THIS_MODULE,
	.open    = stag_open,
	.release = stag_release,
	.read    = debugfs_read,
	.llseek  = default_llseek,
};

static int setup_debugfs(struct c4iw_dev *devp)
{
	struct dentry *de;

	if (!devp->debugfs_root)
		return -1;

	de = debugfs_create_file("qps", S_IWUSR, devp->debugfs_root,
				 (void *)devp, &qp_debugfs_fops);
	if (de && de->d_inode)
		de->d_inode->i_size = 4096;

	de = debugfs_create_file("stags", S_IWUSR, devp->debugfs_root,
				 (void *)devp, &stag_debugfs_fops);
	if (de && de->d_inode)
		de->d_inode->i_size = 4096;
	return 0;
}

void c4iw_release_dev_ucontext(struct c4iw_rdev *rdev,
			       struct c4iw_dev_ucontext *uctx)
{
	struct list_head *pos, *nxt;
	struct c4iw_qid_list *entry;

	mutex_lock(&uctx->lock);
	list_for_each_safe(pos, nxt, &uctx->qpids) {
		entry = list_entry(pos, struct c4iw_qid_list, entry);
		list_del_init(&entry->entry);
		if (!(entry->qid & rdev->qpmask))
			c4iw_put_resource(&rdev->resource.qid_fifo, entry->qid,
					  &rdev->resource.qid_fifo_lock);
		kfree(entry);
	}

	list_for_each_safe(pos, nxt, &uctx->qpids) {
		entry = list_entry(pos, struct c4iw_qid_list, entry);
		list_del_init(&entry->entry);
		kfree(entry);
	}
	mutex_unlock(&uctx->lock);
}

void c4iw_init_dev_ucontext(struct c4iw_rdev *rdev,
			    struct c4iw_dev_ucontext *uctx)
{
	INIT_LIST_HEAD(&uctx->qpids);
	INIT_LIST_HEAD(&uctx->cqids);
	mutex_init(&uctx->lock);
}

/* Caller takes care of locking if needed */
static int c4iw_rdev_open(struct c4iw_rdev *rdev)
{
	int err;

	c4iw_init_dev_ucontext(rdev, &rdev->uctx);

	/*
	 * qpshift is the number of bits to shift the qpid left in order
	 * to get the correct address of the doorbell for that qp.
	 */
	rdev->qpshift = PAGE_SHIFT - ilog2(rdev->lldi.udb_density);
	rdev->qpmask = rdev->lldi.udb_density - 1;
	rdev->cqshift = PAGE_SHIFT - ilog2(rdev->lldi.ucq_density);
	rdev->cqmask = rdev->lldi.ucq_density - 1;
	PDBG("%s dev %s stag start 0x%0x size 0x%0x num stags %d "
	     "pbl start 0x%0x size 0x%0x rq start 0x%0x size 0x%0x "
	     "qp qid start %u size %u cq qid start %u size %u\n",
	     __func__, pci_name(rdev->lldi.pdev), rdev->lldi.vr->stag.start,
	     rdev->lldi.vr->stag.size, c4iw_num_stags(rdev),
	     rdev->lldi.vr->pbl.start,
	     rdev->lldi.vr->pbl.size, rdev->lldi.vr->rq.start,
	     rdev->lldi.vr->rq.size,
	     rdev->lldi.vr->qp.start,
	     rdev->lldi.vr->qp.size,
	     rdev->lldi.vr->cq.start,
	     rdev->lldi.vr->cq.size);
	PDBG("udb len 0x%x udb base %p db_reg %p gts_reg %p qpshift %lu "
	     "qpmask 0x%x cqshift %lu cqmask 0x%x\n",
	     (unsigned)pci_resource_len(rdev->lldi.pdev, 2),
	     (void *)pci_resource_start(rdev->lldi.pdev, 2),
	     rdev->lldi.db_reg,
	     rdev->lldi.gts_reg,
	     rdev->qpshift, rdev->qpmask,
	     rdev->cqshift, rdev->cqmask);

	if (c4iw_num_stags(rdev) == 0) {
		err = -EINVAL;
		goto err1;
	}

	err = c4iw_init_resource(rdev, c4iw_num_stags(rdev), T4_MAX_NUM_PD);
	if (err) {
		printk(KERN_ERR MOD "error %d initializing resources\n", err);
		goto err1;
	}
	err = c4iw_pblpool_create(rdev);
	if (err) {
		printk(KERN_ERR MOD "error %d initializing pbl pool\n", err);
		goto err2;
	}
	err = c4iw_rqtpool_create(rdev);
	if (err) {
		printk(KERN_ERR MOD "error %d initializing rqt pool\n", err);
		goto err3;
	}
	err = c4iw_ocqp_pool_create(rdev);
	if (err) {
		printk(KERN_ERR MOD "error %d initializing ocqp pool\n", err);
		goto err4;
	}
	return 0;
err4:
	c4iw_rqtpool_destroy(rdev);
err3:
	c4iw_pblpool_destroy(rdev);
err2:
	c4iw_destroy_resource(&rdev->resource);
err1:
	return err;
}

static void c4iw_rdev_close(struct c4iw_rdev *rdev)
{
	c4iw_pblpool_destroy(rdev);
	c4iw_rqtpool_destroy(rdev);
	c4iw_destroy_resource(&rdev->resource);
}

static void c4iw_remove(struct c4iw_dev *dev)
{
	PDBG("%s c4iw_dev %p\n", __func__,  dev);
	cancel_delayed_work_sync(&dev->db_drop_task);
	list_del(&dev->entry);
	if (dev->registered)
		c4iw_unregister_device(dev);
	c4iw_rdev_close(&dev->rdev);
	idr_destroy(&dev->cqidr);
	idr_destroy(&dev->qpidr);
	idr_destroy(&dev->mmidr);
	iounmap(dev->rdev.oc_mw_kva);
	ib_dealloc_device(&dev->ibdev);
}

static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
{
	struct c4iw_dev *devp;
	int ret;

	devp = (struct c4iw_dev *)ib_alloc_device(sizeof(*devp));
	if (!devp) {
		printk(KERN_ERR MOD "Cannot allocate ib device\n");
		return NULL;
	}
	devp->rdev.lldi = *infop;

	devp->rdev.oc_mw_pa = pci_resource_start(devp->rdev.lldi.pdev, 2) +
		(pci_resource_len(devp->rdev.lldi.pdev, 2) -
		 roundup_pow_of_two(devp->rdev.lldi.vr->ocq.size));
	devp->rdev.oc_mw_kva = ioremap_wc(devp->rdev.oc_mw_pa,
					       devp->rdev.lldi.vr->ocq.size);

	printk(KERN_INFO MOD "ocq memory: "
	       "hw_start 0x%x size %u mw_pa 0x%lx mw_kva %p\n",
	       devp->rdev.lldi.vr->ocq.start, devp->rdev.lldi.vr->ocq.size,
	       devp->rdev.oc_mw_pa, devp->rdev.oc_mw_kva);

	mutex_lock(&dev_mutex);

	ret = c4iw_rdev_open(&devp->rdev);
	if (ret) {
		mutex_unlock(&dev_mutex);
		printk(KERN_ERR MOD "Unable to open CXIO rdev err %d\n", ret);
		ib_dealloc_device(&devp->ibdev);
		return NULL;
	}

	idr_init(&devp->cqidr);
	idr_init(&devp->qpidr);
	idr_init(&devp->mmidr);
	spin_lock_init(&devp->lock);
	list_add_tail(&devp->entry, &dev_list);
	mutex_unlock(&dev_mutex);

	if (c4iw_debugfs_root) {
		devp->debugfs_root = debugfs_create_dir(
					pci_name(devp->rdev.lldi.pdev),
					c4iw_debugfs_root);
		setup_debugfs(devp);
	}
	return devp;
}

static void *c4iw_uld_add(const struct cxgb4_lld_info *infop)
{
	struct c4iw_dev *dev;
	static int vers_printed;
	int i;

	if (!vers_printed++)
		printk(KERN_INFO MOD "Chelsio T4 RDMA Driver - version %s\n",
		       DRV_VERSION);

	dev = c4iw_alloc(infop);
	if (!dev)
		goto out;

	PDBG("%s found device %s nchan %u nrxq %u ntxq %u nports %u\n",
	     __func__, pci_name(dev->rdev.lldi.pdev),
	     dev->rdev.lldi.nchan, dev->rdev.lldi.nrxq,
	     dev->rdev.lldi.ntxq, dev->rdev.lldi.nports);

	for (i = 0; i < dev->rdev.lldi.nrxq; i++)
		PDBG("rxqid[%u] %u\n", i, dev->rdev.lldi.rxq_ids[i]);
out:
	return dev;
}

static struct sk_buff *t4_pktgl_to_skb(const struct pkt_gl *gl,
				       unsigned int skb_len,
				       unsigned int pull_len)
{
	struct sk_buff *skb;
	struct skb_shared_info *ssi;

	if (gl->tot_len <= 512) {
		skb = alloc_skb(gl->tot_len, GFP_ATOMIC);
		if (unlikely(!skb))
			goto out;
		__skb_put(skb, gl->tot_len);
		skb_copy_to_linear_data(skb, gl->va, gl->tot_len);
	} else {
		skb = alloc_skb(skb_len, GFP_ATOMIC);
		if (unlikely(!skb))
			goto out;
		__skb_put(skb, pull_len);
		skb_copy_to_linear_data(skb, gl->va, pull_len);

		ssi = skb_shinfo(skb);
		ssi->frags[0].page = gl->frags[0].page;
		ssi->frags[0].page_offset = gl->frags[0].page_offset + pull_len;
		ssi->frags[0].size = gl->frags[0].size - pull_len;
		if (gl->nfrags > 1)
			memcpy(&ssi->frags[1], &gl->frags[1],
			       (gl->nfrags - 1) * sizeof(skb_frag_t));
		ssi->nr_frags = gl->nfrags;

		skb->len = gl->tot_len;
		skb->data_len = skb->len - pull_len;
		skb->truesize += skb->data_len;

		/* Get a reference for the last page, we don't own it */
		get_page(gl->frags[gl->nfrags - 1].page);
	}
out:
	return skb;
}

static int c4iw_uld_rx_handler(void *handle, const __be64 *rsp,
			const struct pkt_gl *gl)
{
	struct c4iw_dev *dev = handle;
	struct sk_buff *skb;
	const struct cpl_act_establish *rpl;
	unsigned int opcode;

	if (gl == NULL) {
		/* omit RSS and rsp_ctrl at end of descriptor */
		unsigned int len = 64 - sizeof(struct rsp_ctrl) - 8;

		skb = alloc_skb(256, GFP_ATOMIC);
		if (!skb)
			goto nomem;
		__skb_put(skb, len);
		skb_copy_to_linear_data(skb, &rsp[1], len);
	} else if (gl == CXGB4_MSG_AN) {
		const struct rsp_ctrl *rc = (void *)rsp;

		u32 qid = be32_to_cpu(rc->pldbuflen_qid);
		c4iw_ev_handler(dev, qid);
		return 0;
	} else {
		skb = t4_pktgl_to_skb(gl, 128, 128);
		if (unlikely(!skb))
			goto nomem;
	}

	rpl = cplhdr(skb);
	opcode = rpl->ot.opcode;

	if (c4iw_handlers[opcode])
		c4iw_handlers[opcode](dev, skb);
	else
		printk(KERN_INFO "%s no handler opcode 0x%x...\n", __func__,
		       opcode);

	return 0;
nomem:
	return -1;
}

static int c4iw_uld_state_change(void *handle, enum cxgb4_state new_state)
{
	struct c4iw_dev *dev = handle;

	PDBG("%s new_state %u\n", __func__, new_state);
	switch (new_state) {
	case CXGB4_STATE_UP:
		printk(KERN_INFO MOD "%s: Up\n", pci_name(dev->rdev.lldi.pdev));
		if (!dev->registered) {
			int ret;
			ret = c4iw_register_device(dev);
			if (ret)
				printk(KERN_ERR MOD
				       "%s: RDMA registration failed: %d\n",
				       pci_name(dev->rdev.lldi.pdev), ret);
		}
		break;
	case CXGB4_STATE_DOWN:
		printk(KERN_INFO MOD "%s: Down\n",
		       pci_name(dev->rdev.lldi.pdev));
		if (dev->registered)
			c4iw_unregister_device(dev);
		break;
	case CXGB4_STATE_START_RECOVERY:
		printk(KERN_INFO MOD "%s: Fatal Error\n",
		       pci_name(dev->rdev.lldi.pdev));
		if (dev->registered)
			c4iw_unregister_device(dev);
		break;
	case CXGB4_STATE_DETACH:
		printk(KERN_INFO MOD "%s: Detach\n",
		       pci_name(dev->rdev.lldi.pdev));
		mutex_lock(&dev_mutex);
		c4iw_remove(dev);
		mutex_unlock(&dev_mutex);
		break;
	}
	return 0;
}

static struct cxgb4_uld_info c4iw_uld_info = {
	.name = DRV_NAME,
	.add = c4iw_uld_add,
	.rx_handler = c4iw_uld_rx_handler,
	.state_change = c4iw_uld_state_change,
};

static int __init c4iw_init_module(void)
{
	int err;

	err = c4iw_cm_init();
	if (err)
		return err;

	c4iw_debugfs_root = debugfs_create_dir(DRV_NAME, NULL);
	if (!c4iw_debugfs_root)
		printk(KERN_WARNING MOD
		       "could not create debugfs entry, continuing\n");

	cxgb4_register_uld(CXGB4_ULD_RDMA, &c4iw_uld_info);

	return 0;
}

static void __exit c4iw_exit_module(void)
{
	struct c4iw_dev *dev, *tmp;

	mutex_lock(&dev_mutex);
	list_for_each_entry_safe(dev, tmp, &dev_list, entry) {
		c4iw_remove(dev);
	}
	mutex_unlock(&dev_mutex);
	cxgb4_unregister_uld(CXGB4_ULD_RDMA);
	c4iw_cm_term();
	debugfs_remove_recursive(c4iw_debugfs_root);
}

module_init(c4iw_init_module);
module_exit(c4iw_exit_module);
Commit	Line	Data
cfdda9d7 SW	1	/*
	2	* Copyright (c) 2009-2010 Chelsio, Inc. All rights reserved.
	3	*
	4	* This software is available to you under a choice of one of two
	5	* licenses. You may choose to be licensed under the terms of the GNU
	6	* General Public License (GPL) Version 2, available from the file
	7	* COPYING in the main directory of this source tree, or the
	8	* OpenIB.org BSD license below:
	9	*
	10	* Redistribution and use in source and binary forms, with or
	11	* without modification, are permitted provided that the following
	12	* conditions are met:
	13	*
	14	* - Redistributions of source code must retain the above
	15	* copyright notice, this list of conditions and the following
	16	* disclaimer.
	17	*
	18	* - Redistributions in binary form must reproduce the above
	19	* copyright notice, this list of conditions and the following
	20	* disclaimer in the documentation and/or other materials
	21	* provided with the distribution.
	22	*
	23	* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
	24	* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
	25	* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
	26	* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
	27	* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
	28	* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
	29	* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
	30	* SOFTWARE.
	31	*/
	32	#include <linux/module.h>
	33	#include <linux/moduleparam.h>
	34	#include <linux/debugfs.h>
	35
	36	#include <rdma/ib_verbs.h>
	37
	38	#include "iw_cxgb4.h"
	39
	40	#define DRV_VERSION "0.1"
	41
	42	MODULE_AUTHOR("Steve Wise");
	43	MODULE_DESCRIPTION("Chelsio T4 RDMA Driver");
	44	MODULE_LICENSE("Dual BSD/GPL");
	45	MODULE_VERSION(DRV_VERSION);
	46
	47	static LIST_HEAD(dev_list);
	48	static DEFINE_MUTEX(dev_mutex);
	49
	50	static struct dentry *c4iw_debugfs_root;
	51
9e8d1fa3	52	struct c4iw_debugfs_data {
cfdda9d7 SW	53	struct c4iw_dev *devp;
	54	char *buf;
	55	int bufsize;
	56	int pos;
	57	};
	58
9e8d1fa3	59	static int count_idrs(int id, void p, void data)
cfdda9d7	60	{
cfdda9d7 SW	61	int *countp = data;
cfdda9d7 SW	62
cfdda9d7 SW	63	countp = countp + 1;
	64	return 0;
	65	}
	66
9e8d1fa3 SW	67	static ssize_t debugfs_read(struct file file, char __user buf, size_t count,
	68	loff_t *ppos)
	69	{
	70	struct c4iw_debugfs_data *d = file->private_data;
9e8d1fa3	71
3160977a	72	return simple_read_from_buffer(buf, count, ppos, d->buf, d->pos);
9e8d1fa3 SW	73	}
	74
	75	static int dump_qp(int id, void p, void data)
cfdda9d7 SW	76	{
cfdda9d7 SW	77	struct c4iw_qp *qp = p;
9e8d1fa3	78	struct c4iw_debugfs_data *qpd = data;
cfdda9d7 SW	79	int space;
	80	int cc;
	81
	82	if (id != qp->wq.sq.qid)
	83	return 0;
	84
	85	space = qpd->bufsize - qpd->pos - 1;
	86	if (space == 0)
	87	return 1;
	88
	89	if (qp->ep)
	90	cc = snprintf(qpd->buf + qpd->pos, space, "qp id %u state %u "
	91	"ep tid %u state %u %pI4:%u->%pI4:%u\n",
	92	qp->wq.sq.qid, (int)qp->attr.state,
	93	qp->ep->hwtid, (int)qp->ep->com.state,
	94	&qp->ep->com.local_addr.sin_addr.s_addr,
	95	ntohs(qp->ep->com.local_addr.sin_port),
	96	&qp->ep->com.remote_addr.sin_addr.s_addr,
	97	ntohs(qp->ep->com.remote_addr.sin_port));
	98	else
	99	cc = snprintf(qpd->buf + qpd->pos, space, "qp id %u state %u\n",
	100	qp->wq.sq.qid, (int)qp->attr.state);
	101	if (cc < space)
	102	qpd->pos += cc;
	103	return 0;
	104	}
	105
	106	static int qp_release(struct inode inode, struct file file)
	107	{
9e8d1fa3	108	struct c4iw_debugfs_data *qpd = file->private_data;
cfdda9d7 SW	109	if (!qpd) {
	110	printk(KERN_INFO "%s null qpd?\n", __func__);
	111	return 0;
	112	}
	113	kfree(qpd->buf);
	114	kfree(qpd);
	115	return 0;
	116	}
	117
	118	static int qp_open(struct inode inode, struct file file)
	119	{
9e8d1fa3	120	struct c4iw_debugfs_data *qpd;
cfdda9d7 SW	121	int ret = 0;
	122	int count = 1;
	123
	124	qpd = kmalloc(sizeof *qpd, GFP_KERNEL);
	125	if (!qpd) {
	126	ret = -ENOMEM;
	127	goto out;
	128	}
	129	qpd->devp = inode->i_private;
	130	qpd->pos = 0;
	131
	132	spin_lock_irq(&qpd->devp->lock);
9e8d1fa3	133	idr_for_each(&qpd->devp->qpidr, count_idrs, &count);
cfdda9d7 SW	134	spin_unlock_irq(&qpd->devp->lock);
	135
	136	qpd->bufsize = count * 128;
	137	qpd->buf = kmalloc(qpd->bufsize, GFP_KERNEL);
	138	if (!qpd->buf) {
	139	ret = -ENOMEM;
	140	goto err1;
	141	}
	142
	143	spin_lock_irq(&qpd->devp->lock);
9e8d1fa3	144	idr_for_each(&qpd->devp->qpidr, dump_qp, qpd);
cfdda9d7 SW	145	spin_unlock_irq(&qpd->devp->lock);
	146
	147	qpd->buf[qpd->pos++] = 0;
	148	file->private_data = qpd;
	149	goto out;
	150	err1:
	151	kfree(qpd);
	152	out:
	153	return ret;
	154	}
	155
9e8d1fa3 SW	156	static const struct file_operations qp_debugfs_fops = {
	157	.owner = THIS_MODULE,
	158	.open = qp_open,
	159	.release = qp_release,
	160	.read = debugfs_read,
8bbac892	161	.llseek = default_llseek,
9e8d1fa3 SW	162	};
	163
	164	static int dump_stag(int id, void p, void data)
cfdda9d7	165	{
9e8d1fa3 SW	166	struct c4iw_debugfs_data *stagd = data;
	167	int space;
	168	int cc;
cfdda9d7	169
9e8d1fa3 SW	170	space = stagd->bufsize - stagd->pos - 1;
	171	if (space == 0)
	172	return 1;
	173
	174	cc = snprintf(stagd->buf + stagd->pos, space, "0x%x\n", id<<8);
	175	if (cc < space)
	176	stagd->pos += cc;
	177	return 0;
	178	}
	179
	180	static int stag_release(struct inode inode, struct file file)
	181	{
	182	struct c4iw_debugfs_data *stagd = file->private_data;
	183	if (!stagd) {
	184	printk(KERN_INFO "%s null stagd?\n", __func__);
cfdda9d7	185	return 0;
9e8d1fa3 SW	186	}
	187	kfree(stagd->buf);
	188	kfree(stagd);
	189	return 0;
	190	}
cfdda9d7	191
9e8d1fa3 SW	192	static int stag_open(struct inode inode, struct file file)
	193	{
	194	struct c4iw_debugfs_data *stagd;
	195	int ret = 0;
	196	int count = 1;
cfdda9d7	197
9e8d1fa3 SW	198	stagd = kmalloc(sizeof *stagd, GFP_KERNEL);
	199	if (!stagd) {
	200	ret = -ENOMEM;
	201	goto out;
	202	}
	203	stagd->devp = inode->i_private;
	204	stagd->pos = 0;
cfdda9d7	205
9e8d1fa3 SW	206	spin_lock_irq(&stagd->devp->lock);
	207	idr_for_each(&stagd->devp->mmidr, count_idrs, &count);
	208	spin_unlock_irq(&stagd->devp->lock);
	209
	210	stagd->bufsize = count * sizeof("0x12345678\n");
	211	stagd->buf = kmalloc(stagd->bufsize, GFP_KERNEL);
	212	if (!stagd->buf) {
	213	ret = -ENOMEM;
	214	goto err1;
cfdda9d7	215	}
9e8d1fa3 SW	216
	217	spin_lock_irq(&stagd->devp->lock);
	218	idr_for_each(&stagd->devp->mmidr, dump_stag, stagd);
	219	spin_unlock_irq(&stagd->devp->lock);
	220
	221	stagd->buf[stagd->pos++] = 0;
	222	file->private_data = stagd;
	223	goto out;
	224	err1:
	225	kfree(stagd);
	226	out:
	227	return ret;
cfdda9d7 SW	228	}
cfdda9d7 SW	229
9e8d1fa3	230	static const struct file_operations stag_debugfs_fops = {
cfdda9d7	231	.owner = THIS_MODULE,
9e8d1fa3 SW	232	.open = stag_open,
	233	.release = stag_release,
	234	.read = debugfs_read,
8bbac892	235	.llseek = default_llseek,
cfdda9d7 SW	236	};
	237
	238	static int setup_debugfs(struct c4iw_dev *devp)
	239	{
	240	struct dentry *de;
	241
	242	if (!devp->debugfs_root)
	243	return -1;
	244
	245	de = debugfs_create_file("qps", S_IWUSR, devp->debugfs_root,
	246	(void *)devp, &qp_debugfs_fops);
	247	if (de && de->d_inode)
	248	de->d_inode->i_size = 4096;
9e8d1fa3 SW	249
	250	de = debugfs_create_file("stags", S_IWUSR, devp->debugfs_root,
	251	(void *)devp, &stag_debugfs_fops);
	252	if (de && de->d_inode)
	253	de->d_inode->i_size = 4096;
cfdda9d7 SW	254	return 0;
	255	}
	256
	257	void c4iw_release_dev_ucontext(struct c4iw_rdev *rdev,
	258	struct c4iw_dev_ucontext *uctx)
	259	{
	260	struct list_head pos, nxt;
	261	struct c4iw_qid_list *entry;
	262
	263	mutex_lock(&uctx->lock);
	264	list_for_each_safe(pos, nxt, &uctx->qpids) {
	265	entry = list_entry(pos, struct c4iw_qid_list, entry);
	266	list_del_init(&entry->entry);
	267	if (!(entry->qid & rdev->qpmask))
	268	c4iw_put_resource(&rdev->resource.qid_fifo, entry->qid,
	269	&rdev->resource.qid_fifo_lock);
	270	kfree(entry);
	271	}
	272
	273	list_for_each_safe(pos, nxt, &uctx->qpids) {
	274	entry = list_entry(pos, struct c4iw_qid_list, entry);
	275	list_del_init(&entry->entry);
	276	kfree(entry);
	277	}
	278	mutex_unlock(&uctx->lock);
	279	}
	280
	281	void c4iw_init_dev_ucontext(struct c4iw_rdev *rdev,
	282	struct c4iw_dev_ucontext *uctx)
	283	{
	284	INIT_LIST_HEAD(&uctx->qpids);
	285	INIT_LIST_HEAD(&uctx->cqids);
	286	mutex_init(&uctx->lock);
	287	}
	288
	289	/* Caller takes care of locking if needed */
	290	static int c4iw_rdev_open(struct c4iw_rdev *rdev)
	291	{
	292	int err;
	293
	294	c4iw_init_dev_ucontext(rdev, &rdev->uctx);
	295
	296	/*
	297	* qpshift is the number of bits to shift the qpid left in order
	298	* to get the correct address of the doorbell for that qp.
	299	*/
	300	rdev->qpshift = PAGE_SHIFT - ilog2(rdev->lldi.udb_density);
	301	rdev->qpmask = rdev->lldi.udb_density - 1;
	302	rdev->cqshift = PAGE_SHIFT - ilog2(rdev->lldi.ucq_density);
	303	rdev->cqmask = rdev->lldi.ucq_density - 1;
	304	PDBG("%s dev %s stag start 0x%0x size 0x%0x num stags %d "
93fb72e4 SW	305	"pbl start 0x%0x size 0x%0x rq start 0x%0x size 0x%0x "
93fb72e4 SW	306	"qp qid start %u size %u cq qid start %u size %u\n",
cfdda9d7 SW	307	__func__, pci_name(rdev->lldi.pdev), rdev->lldi.vr->stag.start,
	308	rdev->lldi.vr->stag.size, c4iw_num_stags(rdev),
	309	rdev->lldi.vr->pbl.start,
	310	rdev->lldi.vr->pbl.size, rdev->lldi.vr->rq.start,
93fb72e4 SW	311	rdev->lldi.vr->rq.size,
	312	rdev->lldi.vr->qp.start,
	313	rdev->lldi.vr->qp.size,
	314	rdev->lldi.vr->cq.start,
	315	rdev->lldi.vr->cq.size);
cfdda9d7 SW	316	PDBG("udb len 0x%x udb base %p db_reg %p gts_reg %p qpshift %lu "
	317	"qpmask 0x%x cqshift %lu cqmask 0x%x\n",
	318	(unsigned)pci_resource_len(rdev->lldi.pdev, 2),
	319	(void *)pci_resource_start(rdev->lldi.pdev, 2),
	320	rdev->lldi.db_reg,
	321	rdev->lldi.gts_reg,
	322	rdev->qpshift, rdev->qpmask,
	323	rdev->cqshift, rdev->cqmask);
	324
	325	if (c4iw_num_stags(rdev) == 0) {
	326	err = -EINVAL;
	327	goto err1;
	328	}
	329
	330	err = c4iw_init_resource(rdev, c4iw_num_stags(rdev), T4_MAX_NUM_PD);
	331	if (err) {
	332	printk(KERN_ERR MOD "error %d initializing resources\n", err);
	333	goto err1;
	334	}
	335	err = c4iw_pblpool_create(rdev);
	336	if (err) {
	337	printk(KERN_ERR MOD "error %d initializing pbl pool\n", err);
	338	goto err2;
	339	}
	340	err = c4iw_rqtpool_create(rdev);
	341	if (err) {
	342	printk(KERN_ERR MOD "error %d initializing rqt pool\n", err);
	343	goto err3;
	344	}
c6d7b267 SW	345	err = c4iw_ocqp_pool_create(rdev);
	346	if (err) {
	347	printk(KERN_ERR MOD "error %d initializing ocqp pool\n", err);
	348	goto err4;
	349	}
cfdda9d7	350	return 0;
c6d7b267 SW	351	err4:
c6d7b267 SW	352	c4iw_rqtpool_destroy(rdev);
cfdda9d7 SW	353	err3:
	354	c4iw_pblpool_destroy(rdev);
	355	err2:
	356	c4iw_destroy_resource(&rdev->resource);
	357	err1:
	358	return err;
	359	}
	360
	361	static void c4iw_rdev_close(struct c4iw_rdev *rdev)
	362	{
	363	c4iw_pblpool_destroy(rdev);
	364	c4iw_rqtpool_destroy(rdev);
	365	c4iw_destroy_resource(&rdev->resource);
	366	}
	367
	368	static void c4iw_remove(struct c4iw_dev *dev)
	369	{
	370	PDBG("%s c4iw_dev %p\n", __func__, dev);
	371	cancel_delayed_work_sync(&dev->db_drop_task);
	372	list_del(&dev->entry);
1c01c538 SW	373	if (dev->registered)
1c01c538 SW	374	c4iw_unregister_device(dev);
cfdda9d7 SW	375	c4iw_rdev_close(&dev->rdev);
	376	idr_destroy(&dev->cqidr);
	377	idr_destroy(&dev->qpidr);
	378	idr_destroy(&dev->mmidr);
c6d7b267	379	iounmap(dev->rdev.oc_mw_kva);
cfdda9d7 SW	380	ib_dealloc_device(&dev->ibdev);
	381	}
	382
	383	static struct c4iw_dev c4iw_alloc(const struct cxgb4_lld_info infop)
	384	{
	385	struct c4iw_dev *devp;
	386	int ret;
	387
	388	devp = (struct c4iw_dev )ib_alloc_device(sizeof(devp));
	389	if (!devp) {
	390	printk(KERN_ERR MOD "Cannot allocate ib device\n");
	391	return NULL;
	392	}
	393	devp->rdev.lldi = *infop;
	394
c6d7b267 SW	395	devp->rdev.oc_mw_pa = pci_resource_start(devp->rdev.lldi.pdev, 2) +
	396	(pci_resource_len(devp->rdev.lldi.pdev, 2) -
	397	roundup_pow_of_two(devp->rdev.lldi.vr->ocq.size));
	398	devp->rdev.oc_mw_kva = ioremap_wc(devp->rdev.oc_mw_pa,
	399	devp->rdev.lldi.vr->ocq.size);
	400
	401	printk(KERN_INFO MOD "ocq memory: "
	402	"hw_start 0x%x size %u mw_pa 0x%lx mw_kva %p\n",
	403	devp->rdev.lldi.vr->ocq.start, devp->rdev.lldi.vr->ocq.size,
	404	devp->rdev.oc_mw_pa, devp->rdev.oc_mw_kva);
	405
cfdda9d7 SW	406	mutex_lock(&dev_mutex);
	407
	408	ret = c4iw_rdev_open(&devp->rdev);
	409	if (ret) {
	410	mutex_unlock(&dev_mutex);
	411	printk(KERN_ERR MOD "Unable to open CXIO rdev err %d\n", ret);
	412	ib_dealloc_device(&devp->ibdev);
	413	return NULL;
	414	}
	415
	416	idr_init(&devp->cqidr);
	417	idr_init(&devp->qpidr);
	418	idr_init(&devp->mmidr);
	419	spin_lock_init(&devp->lock);
	420	list_add_tail(&devp->entry, &dev_list);
	421	mutex_unlock(&dev_mutex);
	422
cfdda9d7 SW	423	if (c4iw_debugfs_root) {
	424	devp->debugfs_root = debugfs_create_dir(
	425	pci_name(devp->rdev.lldi.pdev),
	426	c4iw_debugfs_root);
	427	setup_debugfs(devp);
	428	}
	429	return devp;
	430	}
	431
	432	static void c4iw_uld_add(const struct cxgb4_lld_info infop)
	433	{
	434	struct c4iw_dev *dev;
	435	static int vers_printed;
	436	int i;
	437
	438	if (!vers_printed++)
	439	printk(KERN_INFO MOD "Chelsio T4 RDMA Driver - version %s\n",
	440	DRV_VERSION);
	441
	442	dev = c4iw_alloc(infop);
	443	if (!dev)
	444	goto out;
	445
	446	PDBG("%s found device %s nchan %u nrxq %u ntxq %u nports %u\n",
	447	__func__, pci_name(dev->rdev.lldi.pdev),
	448	dev->rdev.lldi.nchan, dev->rdev.lldi.nrxq,
	449	dev->rdev.lldi.ntxq, dev->rdev.lldi.nports);
	450
	451	for (i = 0; i < dev->rdev.lldi.nrxq; i++)
	452	PDBG("rxqid[%u] %u\n", i, dev->rdev.lldi.rxq_ids[i]);
cfdda9d7 SW	453	out:
	454	return dev;
	455	}
	456
	457	static struct sk_buff t4_pktgl_to_skb(const struct pkt_gl gl,
	458	unsigned int skb_len,
	459	unsigned int pull_len)
	460	{
	461	struct sk_buff *skb;
	462	struct skb_shared_info *ssi;
	463
	464	if (gl->tot_len <= 512) {
	465	skb = alloc_skb(gl->tot_len, GFP_ATOMIC);
	466	if (unlikely(!skb))
	467	goto out;
	468	__skb_put(skb, gl->tot_len);
	469	skb_copy_to_linear_data(skb, gl->va, gl->tot_len);
	470	} else {
	471	skb = alloc_skb(skb_len, GFP_ATOMIC);
	472	if (unlikely(!skb))
	473	goto out;
	474	__skb_put(skb, pull_len);
	475	skb_copy_to_linear_data(skb, gl->va, pull_len);
	476
	477	ssi = skb_shinfo(skb);
	478	ssi->frags[0].page = gl->frags[0].page;
	479	ssi->frags[0].page_offset = gl->frags[0].page_offset + pull_len;
	480	ssi->frags[0].size = gl->frags[0].size - pull_len;
	481	if (gl->nfrags > 1)
	482	memcpy(&ssi->frags[1], &gl->frags[1],
	483	(gl->nfrags - 1) * sizeof(skb_frag_t));
	484	ssi->nr_frags = gl->nfrags;
	485
	486	skb->len = gl->tot_len;
	487	skb->data_len = skb->len - pull_len;
	488	skb->truesize += skb->data_len;
	489
	490	/* Get a reference for the last page, we don't own it */
	491	get_page(gl->frags[gl->nfrags - 1].page);
	492	}
	493	out:
	494	return skb;
	495	}
	496
	497	static int c4iw_uld_rx_handler(void handle, const __be64 rsp,
	498	const struct pkt_gl *gl)
	499	{
	500	struct c4iw_dev *dev = handle;
	501	struct sk_buff *skb;
	502	const struct cpl_act_establish *rpl;
	503	unsigned int opcode;
	504
	505	if (gl == NULL) {
	506	/* omit RSS and rsp_ctrl at end of descriptor */
	507	unsigned int len = 64 - sizeof(struct rsp_ctrl) - 8;
	508
	509	skb = alloc_skb(256, GFP_ATOMIC);
	510	if (!skb)
	511	goto nomem;
	512	__skb_put(skb, len);
	513	skb_copy_to_linear_data(skb, &rsp[1], len);
	514	} else if (gl == CXGB4_MSG_AN) {
	515	const struct rsp_ctrl rc = (void )rsp;
	516
517	u32 qid = be32_to_cpu(rc->pldbuflen_qid);
518	c4iw_ev_handler(dev, qid);
519	return 0;
520	} else {
521	skb = t4_pktgl_to_skb(gl, 128, 128);
522	if (unlikely(!skb))
523	goto nomem;
524	}
525
526	rpl = cplhdr(skb);
527	opcode = rpl->ot.opcode;
528
529	if (c4iw_handlers[opcode])
530	c4iw_handlers[opcode](dev, skb);
531	else
532	printk(KERN_INFO "%s no handler opcode 0x%x...\n", __func__,
533	opcode);
534
535	return 0;
536	nomem:
537	return -1;
538	}
539
540	static int c4iw_uld_state_change(void *handle, enum cxgb4_state new_state)
541	{
1c01c538 SW	542	struct c4iw_dev *dev = handle;
1c01c538 SW	543
cfdda9d7	544	PDBG("%s new_state %u\n", __func__, new_state);
1c01c538 SW	545	switch (new_state) {
	546	case CXGB4_STATE_UP:
	547	printk(KERN_INFO MOD "%s: Up\n", pci_name(dev->rdev.lldi.pdev));
	548	if (!dev->registered) {
	549	int ret;
	550	ret = c4iw_register_device(dev);
	551	if (ret)
	552	printk(KERN_ERR MOD
	553	"%s: RDMA registration failed: %d\n",
	554	pci_name(dev->rdev.lldi.pdev), ret);
	555	}
	556	break;
	557	case CXGB4_STATE_DOWN:
	558	printk(KERN_INFO MOD "%s: Down\n",
	559	pci_name(dev->rdev.lldi.pdev));
	560	if (dev->registered)
	561	c4iw_unregister_device(dev);
	562	break;
	563	case CXGB4_STATE_START_RECOVERY:
	564	printk(KERN_INFO MOD "%s: Fatal Error\n",
	565	pci_name(dev->rdev.lldi.pdev));
	566	if (dev->registered)
	567	c4iw_unregister_device(dev);
	568	break;
	569	case CXGB4_STATE_DETACH:
	570	printk(KERN_INFO MOD "%s: Detach\n",
	571	pci_name(dev->rdev.lldi.pdev));
	572	mutex_lock(&dev_mutex);
	573	c4iw_remove(dev);
	574	mutex_unlock(&dev_mutex);
	575	break;
	576	}
cfdda9d7 SW	577	return 0;
	578	}
	579
	580	static struct cxgb4_uld_info c4iw_uld_info = {
	581	.name = DRV_NAME,
	582	.add = c4iw_uld_add,
	583	.rx_handler = c4iw_uld_rx_handler,
	584	.state_change = c4iw_uld_state_change,
	585	};
	586
	587	static int __init c4iw_init_module(void)
	588	{
	589	int err;
	590
	591	err = c4iw_cm_init();
	592	if (err)
	593	return err;
	594
	595	c4iw_debugfs_root = debugfs_create_dir(DRV_NAME, NULL);
	596	if (!c4iw_debugfs_root)
	597	printk(KERN_WARNING MOD
	598	"could not create debugfs entry, continuing\n");
	599
	600	cxgb4_register_uld(CXGB4_ULD_RDMA, &c4iw_uld_info);
	601
	602	return 0;
	603	}
	604
	605	static void __exit c4iw_exit_module(void)
	606	{
	607	struct c4iw_dev dev, tmp;
	608
cfdda9d7 SW	609	mutex_lock(&dev_mutex);
	610	list_for_each_entry_safe(dev, tmp, &dev_list, entry) {
	611	c4iw_remove(dev);
	612	}
	613	mutex_unlock(&dev_mutex);
fd388ce6	614	cxgb4_unregister_uld(CXGB4_ULD_RDMA);
cfdda9d7 SW	615	c4iw_cm_term();
	616	debugfs_remove_recursive(c4iw_debugfs_root);
	617	}
	618
	619	module_init(c4iw_init_module);
	620	module_exit(c4iw_exit_module);