IB/mlx4: Add a driver Mellanox ConnectX InfiniBand adapters

Add an InfiniBand driver for Mellanox ConnectX adapters.  Because
these adapters can also be used as ethernet NICs and Fibre Channel 
HBAs, the driver is split into two modules: 
 
  mlx4_core: Handles low-level things like device initialization and 
    processing firmware commands.  Also controls resource allocation 
    so that the InfiniBand, ethernet and FC functions can share a 
    device without stepping on each other. 
 
  mlx4_ib: Handles InfiniBand-specific things; plugs into the 
    InfiniBand midlayer. 

Signed-off-by: Roland Dreier <rolandd@cisco.com>
This commit is contained in:
Roland Dreier
2007-05-08 18:00:38 -07:00
parent 1bf66a3042
commit 225c7b1fee
43 changed files with 11592 additions and 0 deletions

View File

@@ -45,6 +45,8 @@ source "drivers/infiniband/hw/ehca/Kconfig"
source "drivers/infiniband/hw/amso1100/Kconfig"
source "drivers/infiniband/hw/cxgb3/Kconfig"
source "drivers/infiniband/hw/mlx4/Kconfig"
source "drivers/infiniband/ulp/ipoib/Kconfig"
source "drivers/infiniband/ulp/srp/Kconfig"

View File

@@ -4,6 +4,7 @@ obj-$(CONFIG_INFINIBAND_IPATH) += hw/ipath/
obj-$(CONFIG_INFINIBAND_EHCA) += hw/ehca/
obj-$(CONFIG_INFINIBAND_AMSO1100) += hw/amso1100/
obj-$(CONFIG_INFINIBAND_CXGB3) += hw/cxgb3/
obj-$(CONFIG_MLX4_INFINIBAND) += hw/mlx4/
obj-$(CONFIG_INFINIBAND_IPOIB) += ulp/ipoib/
obj-$(CONFIG_INFINIBAND_SRP) += ulp/srp/
obj-$(CONFIG_INFINIBAND_ISER) += ulp/iser/

View File

@@ -0,0 +1,9 @@
config MLX4_INFINIBAND
tristate "Mellanox ConnectX HCA support"
depends on INFINIBAND
select MLX4_CORE
---help---
This driver provides low-level InfiniBand support for
Mellanox ConnectX PCI Express host channel adapters (HCAs).
This is required to use InfiniBand protocols such as
IP-over-IB or SRP with these devices.

View File

@@ -0,0 +1,3 @@
obj-$(CONFIG_MLX4_INFINIBAND) += mlx4_ib.o
mlx4_ib-y := ah.o cq.o doorbell.o mad.o main.o mr.o qp.o srq.o

View File

@@ -0,0 +1,100 @@
/*
* Copyright (c) 2007 Cisco Systems, 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 "mlx4_ib.h"
struct ib_ah *mlx4_ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr)
{
struct mlx4_dev *dev = to_mdev(pd->device)->dev;
struct mlx4_ib_ah *ah;
ah = kmalloc(sizeof *ah, GFP_ATOMIC);
if (!ah)
return ERR_PTR(-ENOMEM);
memset(&ah->av, 0, sizeof ah->av);
ah->av.port_pd = cpu_to_be32(to_mpd(pd)->pdn | (ah_attr->port_num << 24));
ah->av.g_slid = ah_attr->src_path_bits;
ah->av.dlid = cpu_to_be16(ah_attr->dlid);
if (ah_attr->static_rate) {
ah->av.stat_rate = ah_attr->static_rate + MLX4_STAT_RATE_OFFSET;
while (ah->av.stat_rate > IB_RATE_2_5_GBPS + MLX4_STAT_RATE_OFFSET &&
!(1 << ah->av.stat_rate & dev->caps.stat_rate_support))
--ah->av.stat_rate;
}
ah->av.sl_tclass_flowlabel = cpu_to_be32(ah_attr->sl << 28);
if (ah_attr->ah_flags & IB_AH_GRH) {
ah->av.g_slid |= 0x80;
ah->av.gid_index = ah_attr->grh.sgid_index;
ah->av.hop_limit = ah_attr->grh.hop_limit;
ah->av.sl_tclass_flowlabel |=
cpu_to_be32((ah_attr->grh.traffic_class << 20) |
ah_attr->grh.flow_label);
memcpy(ah->av.dgid, ah_attr->grh.dgid.raw, 16);
}
return &ah->ibah;
}
int mlx4_ib_query_ah(struct ib_ah *ibah, struct ib_ah_attr *ah_attr)
{
struct mlx4_ib_ah *ah = to_mah(ibah);
memset(ah_attr, 0, sizeof *ah_attr);
ah_attr->dlid = be16_to_cpu(ah->av.dlid);
ah_attr->sl = be32_to_cpu(ah->av.sl_tclass_flowlabel) >> 28;
ah_attr->port_num = be32_to_cpu(ah->av.port_pd) >> 24;
if (ah->av.stat_rate)
ah_attr->static_rate = ah->av.stat_rate - MLX4_STAT_RATE_OFFSET;
ah_attr->src_path_bits = ah->av.g_slid & 0x7F;
if (mlx4_ib_ah_grh_present(ah)) {
ah_attr->ah_flags = IB_AH_GRH;
ah_attr->grh.traffic_class =
be32_to_cpu(ah->av.sl_tclass_flowlabel) >> 20;
ah_attr->grh.flow_label =
be32_to_cpu(ah->av.sl_tclass_flowlabel) & 0xfffff;
ah_attr->grh.hop_limit = ah->av.hop_limit;
ah_attr->grh.sgid_index = ah->av.gid_index;
memcpy(ah_attr->grh.dgid.raw, ah->av.dgid, 16);
}
return 0;
}
int mlx4_ib_destroy_ah(struct ib_ah *ah)
{
kfree(to_mah(ah));
return 0;
}

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,216 @@
/*
* Copyright (c) 2007 Cisco Systems, 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/slab.h>
#include "mlx4_ib.h"
struct mlx4_ib_db_pgdir {
struct list_head list;
DECLARE_BITMAP(order0, MLX4_IB_DB_PER_PAGE);
DECLARE_BITMAP(order1, MLX4_IB_DB_PER_PAGE / 2);
unsigned long *bits[2];
__be32 *db_page;
dma_addr_t db_dma;
};
static struct mlx4_ib_db_pgdir *mlx4_ib_alloc_db_pgdir(struct mlx4_ib_dev *dev)
{
struct mlx4_ib_db_pgdir *pgdir;
pgdir = kzalloc(sizeof *pgdir, GFP_KERNEL);
if (!pgdir)
return NULL;
bitmap_fill(pgdir->order1, MLX4_IB_DB_PER_PAGE / 2);
pgdir->bits[0] = pgdir->order0;
pgdir->bits[1] = pgdir->order1;
pgdir->db_page = dma_alloc_coherent(dev->ib_dev.dma_device,
PAGE_SIZE, &pgdir->db_dma,
GFP_KERNEL);
if (!pgdir->db_page) {
kfree(pgdir);
return NULL;
}
return pgdir;
}
static int mlx4_ib_alloc_db_from_pgdir(struct mlx4_ib_db_pgdir *pgdir,
struct mlx4_ib_db *db, int order)
{
int o;
int i;
for (o = order; o <= 1; ++o) {
i = find_first_bit(pgdir->bits[o], MLX4_IB_DB_PER_PAGE >> o);
if (i < MLX4_IB_DB_PER_PAGE >> o)
goto found;
}
return -ENOMEM;
found:
clear_bit(i, pgdir->bits[o]);
i <<= o;
if (o > order)
set_bit(i ^ 1, pgdir->bits[order]);
db->u.pgdir = pgdir;
db->index = i;
db->db = pgdir->db_page + db->index;
db->dma = pgdir->db_dma + db->index * 4;
db->order = order;
return 0;
}
int mlx4_ib_db_alloc(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db, int order)
{
struct mlx4_ib_db_pgdir *pgdir;
int ret = 0;
mutex_lock(&dev->pgdir_mutex);
list_for_each_entry(pgdir, &dev->pgdir_list, list)
if (!mlx4_ib_alloc_db_from_pgdir(pgdir, db, order))
goto out;
pgdir = mlx4_ib_alloc_db_pgdir(dev);
if (!pgdir) {
ret = -ENOMEM;
goto out;
}
list_add(&pgdir->list, &dev->pgdir_list);
/* This should never fail -- we just allocated an empty page: */
WARN_ON(mlx4_ib_alloc_db_from_pgdir(pgdir, db, order));
out:
mutex_unlock(&dev->pgdir_mutex);
return ret;
}
void mlx4_ib_db_free(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db)
{
int o;
int i;
mutex_lock(&dev->pgdir_mutex);
o = db->order;
i = db->index;
if (db->order == 0 && test_bit(i ^ 1, db->u.pgdir->order0)) {
clear_bit(i ^ 1, db->u.pgdir->order0);
++o;
}
i >>= o;
set_bit(i, db->u.pgdir->bits[o]);
if (bitmap_full(db->u.pgdir->order1, MLX4_IB_DB_PER_PAGE / 2)) {
dma_free_coherent(dev->ib_dev.dma_device, PAGE_SIZE,
db->u.pgdir->db_page, db->u.pgdir->db_dma);
list_del(&db->u.pgdir->list);
kfree(db->u.pgdir);
}
mutex_unlock(&dev->pgdir_mutex);
}
struct mlx4_ib_user_db_page {
struct list_head list;
struct ib_umem *umem;
unsigned long user_virt;
int refcnt;
};
int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
struct mlx4_ib_db *db)
{
struct mlx4_ib_user_db_page *page;
struct ib_umem_chunk *chunk;
int err = 0;
mutex_lock(&context->db_page_mutex);
list_for_each_entry(page, &context->db_page_list, list)
if (page->user_virt == (virt & PAGE_MASK))
goto found;
page = kmalloc(sizeof *page, GFP_KERNEL);
if (!page) {
err = -ENOMEM;
goto out;
}
page->user_virt = (virt & PAGE_MASK);
page->refcnt = 0;
page->umem = ib_umem_get(&context->ibucontext, virt & PAGE_MASK,
PAGE_SIZE, 0);
if (IS_ERR(page->umem)) {
err = PTR_ERR(page->umem);
kfree(page);
goto out;
}
list_add(&page->list, &context->db_page_list);
found:
chunk = list_entry(page->umem->chunk_list.next, struct ib_umem_chunk, list);
db->dma = sg_dma_address(chunk->page_list) + (virt & ~PAGE_MASK);
db->u.user_page = page;
++page->refcnt;
out:
mutex_unlock(&context->db_page_mutex);
return err;
}
void mlx4_ib_db_unmap_user(struct mlx4_ib_ucontext *context, struct mlx4_ib_db *db)
{
mutex_lock(&context->db_page_mutex);
if (!--db->u.user_page->refcnt) {
list_del(&db->u.user_page->list);
ib_umem_release(db->u.user_page->umem);
kfree(db->u.user_page);
}
mutex_unlock(&context->db_page_mutex);
}

View File

@@ -0,0 +1,339 @@
/*
* Copyright (c) 2007 Cisco Systems, 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 <rdma/ib_mad.h>
#include <rdma/ib_smi.h>
#include <linux/mlx4/cmd.h>
#include "mlx4_ib.h"
enum {
MLX4_IB_VENDOR_CLASS1 = 0x9,
MLX4_IB_VENDOR_CLASS2 = 0xa
};
int mlx4_MAD_IFC(struct mlx4_ib_dev *dev, int ignore_mkey, int ignore_bkey,
int port, struct ib_wc *in_wc, struct ib_grh *in_grh,
void *in_mad, void *response_mad)
{
struct mlx4_cmd_mailbox *inmailbox, *outmailbox;
void *inbox;
int err;
u32 in_modifier = port;
u8 op_modifier = 0;
inmailbox = mlx4_alloc_cmd_mailbox(dev->dev);
if (IS_ERR(inmailbox))
return PTR_ERR(inmailbox);
inbox = inmailbox->buf;
outmailbox = mlx4_alloc_cmd_mailbox(dev->dev);
if (IS_ERR(outmailbox)) {
mlx4_free_cmd_mailbox(dev->dev, inmailbox);
return PTR_ERR(outmailbox);
}
memcpy(inbox, in_mad, 256);
/*
* Key check traps can't be generated unless we have in_wc to
* tell us where to send the trap.
*/
if (ignore_mkey || !in_wc)
op_modifier |= 0x1;
if (ignore_bkey || !in_wc)
op_modifier |= 0x2;
if (in_wc) {
struct {
__be32 my_qpn;
u32 reserved1;
__be32 rqpn;
u8 sl;
u8 g_path;
u16 reserved2[2];
__be16 pkey;
u32 reserved3[11];
u8 grh[40];
} *ext_info;
memset(inbox + 256, 0, 256);
ext_info = inbox + 256;
ext_info->my_qpn = cpu_to_be32(in_wc->qp->qp_num);
ext_info->rqpn = cpu_to_be32(in_wc->src_qp);
ext_info->sl = in_wc->sl << 4;
ext_info->g_path = in_wc->dlid_path_bits |
(in_wc->wc_flags & IB_WC_GRH ? 0x80 : 0);
ext_info->pkey = cpu_to_be16(in_wc->pkey_index);
if (in_grh)
memcpy(ext_info->grh, in_grh, 40);
op_modifier |= 0x4;
in_modifier |= in_wc->slid << 16;
}
err = mlx4_cmd_box(dev->dev, inmailbox->dma, outmailbox->dma,
in_modifier, op_modifier,
MLX4_CMD_MAD_IFC, MLX4_CMD_TIME_CLASS_C);
if (!err);
memcpy(response_mad, outmailbox->buf, 256);
mlx4_free_cmd_mailbox(dev->dev, inmailbox);
mlx4_free_cmd_mailbox(dev->dev, outmailbox);
return err;
}
static void update_sm_ah(struct mlx4_ib_dev *dev, u8 port_num, u16 lid, u8 sl)
{
struct ib_ah *new_ah;
struct ib_ah_attr ah_attr;
if (!dev->send_agent[port_num - 1][0])
return;
memset(&ah_attr, 0, sizeof ah_attr);
ah_attr.dlid = lid;
ah_attr.sl = sl;
ah_attr.port_num = port_num;
new_ah = ib_create_ah(dev->send_agent[port_num - 1][0]->qp->pd,
&ah_attr);
if (IS_ERR(new_ah))
return;
spin_lock(&dev->sm_lock);
if (dev->sm_ah[port_num - 1])
ib_destroy_ah(dev->sm_ah[port_num - 1]);
dev->sm_ah[port_num - 1] = new_ah;
spin_unlock(&dev->sm_lock);
}
/*
* Snoop SM MADs for port info and P_Key table sets, so we can
* synthesize LID change and P_Key change events.
*/
static void smp_snoop(struct ib_device *ibdev, u8 port_num, struct ib_mad *mad)
{
struct ib_event event;
if ((mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED ||
mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) &&
mad->mad_hdr.method == IB_MGMT_METHOD_SET) {
if (mad->mad_hdr.attr_id == IB_SMP_ATTR_PORT_INFO) {
struct ib_port_info *pinfo =
(struct ib_port_info *) ((struct ib_smp *) mad)->data;
update_sm_ah(to_mdev(ibdev), port_num,
be16_to_cpu(pinfo->sm_lid),
pinfo->neighbormtu_mastersmsl & 0xf);
event.device = ibdev;
event.element.port_num = port_num;
if(pinfo->clientrereg_resv_subnetto & 0x80)
event.event = IB_EVENT_CLIENT_REREGISTER;
else
event.event = IB_EVENT_LID_CHANGE;
ib_dispatch_event(&event);
}
if (mad->mad_hdr.attr_id == IB_SMP_ATTR_PKEY_TABLE) {
event.device = ibdev;
event.event = IB_EVENT_PKEY_CHANGE;
event.element.port_num = port_num;
ib_dispatch_event(&event);
}
}
}
static void node_desc_override(struct ib_device *dev,
struct ib_mad *mad)
{
if ((mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED ||
mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) &&
mad->mad_hdr.method == IB_MGMT_METHOD_GET_RESP &&
mad->mad_hdr.attr_id == IB_SMP_ATTR_NODE_DESC) {
spin_lock(&to_mdev(dev)->sm_lock);
memcpy(((struct ib_smp *) mad)->data, dev->node_desc, 64);
spin_unlock(&to_mdev(dev)->sm_lock);
}
}
static void forward_trap(struct mlx4_ib_dev *dev, u8 port_num, struct ib_mad *mad)
{
int qpn = mad->mad_hdr.mgmt_class != IB_MGMT_CLASS_SUBN_LID_ROUTED;
struct ib_mad_send_buf *send_buf;
struct ib_mad_agent *agent = dev->send_agent[port_num - 1][qpn];
int ret;
if (agent) {
send_buf = ib_create_send_mad(agent, qpn, 0, 0, IB_MGMT_MAD_HDR,
IB_MGMT_MAD_DATA, GFP_ATOMIC);
/*
* We rely here on the fact that MLX QPs don't use the
* address handle after the send is posted (this is
* wrong following the IB spec strictly, but we know
* it's OK for our devices).
*/
spin_lock(&dev->sm_lock);
memcpy(send_buf->mad, mad, sizeof *mad);
if ((send_buf->ah = dev->sm_ah[port_num - 1]))
ret = ib_post_send_mad(send_buf, NULL);
else
ret = -EINVAL;
spin_unlock(&dev->sm_lock);
if (ret)
ib_free_send_mad(send_buf);
}
}
int mlx4_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
struct ib_wc *in_wc, struct ib_grh *in_grh,
struct ib_mad *in_mad, struct ib_mad *out_mad)
{
u16 slid;
int err;
slid = in_wc ? in_wc->slid : be16_to_cpu(IB_LID_PERMISSIVE);
if (in_mad->mad_hdr.method == IB_MGMT_METHOD_TRAP && slid == 0) {
forward_trap(to_mdev(ibdev), port_num, in_mad);
return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED;
}
if (in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED ||
in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) {
if (in_mad->mad_hdr.method != IB_MGMT_METHOD_GET &&
in_mad->mad_hdr.method != IB_MGMT_METHOD_SET &&
in_mad->mad_hdr.method != IB_MGMT_METHOD_TRAP_REPRESS)
return IB_MAD_RESULT_SUCCESS;
/*
* Don't process SMInfo queries or vendor-specific
* MADs -- the SMA can't handle them.
*/
if (in_mad->mad_hdr.attr_id == IB_SMP_ATTR_SM_INFO ||
((in_mad->mad_hdr.attr_id & IB_SMP_ATTR_VENDOR_MASK) ==
IB_SMP_ATTR_VENDOR_MASK))
return IB_MAD_RESULT_SUCCESS;
} else if (in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_PERF_MGMT ||
in_mad->mad_hdr.mgmt_class == MLX4_IB_VENDOR_CLASS1 ||
in_mad->mad_hdr.mgmt_class == MLX4_IB_VENDOR_CLASS2) {
if (in_mad->mad_hdr.method != IB_MGMT_METHOD_GET &&
in_mad->mad_hdr.method != IB_MGMT_METHOD_SET)
return IB_MAD_RESULT_SUCCESS;
} else
return IB_MAD_RESULT_SUCCESS;
err = mlx4_MAD_IFC(to_mdev(ibdev),
mad_flags & IB_MAD_IGNORE_MKEY,
mad_flags & IB_MAD_IGNORE_BKEY,
port_num, in_wc, in_grh, in_mad, out_mad);
if (err)
return IB_MAD_RESULT_FAILURE;
if (!out_mad->mad_hdr.status) {
smp_snoop(ibdev, port_num, in_mad);
node_desc_override(ibdev, out_mad);
}
/* set return bit in status of directed route responses */
if (in_mad->mad_hdr.mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)
out_mad->mad_hdr.status |= cpu_to_be16(1 << 15);
if (in_mad->mad_hdr.method == IB_MGMT_METHOD_TRAP_REPRESS)
/* no response for trap repress */
return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED;
return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_REPLY;
}
static void send_handler(struct ib_mad_agent *agent,
struct ib_mad_send_wc *mad_send_wc)
{
ib_free_send_mad(mad_send_wc->send_buf);
}
int mlx4_ib_mad_init(struct mlx4_ib_dev *dev)
{
struct ib_mad_agent *agent;
int p, q;
int ret;
for (p = 0; p < dev->dev->caps.num_ports; ++p)
for (q = 0; q <= 1; ++q) {
agent = ib_register_mad_agent(&dev->ib_dev, p + 1,
q ? IB_QPT_GSI : IB_QPT_SMI,
NULL, 0, send_handler,
NULL, NULL);
if (IS_ERR(agent)) {
ret = PTR_ERR(agent);
goto err;
}
dev->send_agent[p][q] = agent;
}
return 0;
err:
for (p = 0; p < dev->dev->caps.num_ports; ++p)
for (q = 0; q <= 1; ++q)
if (dev->send_agent[p][q])
ib_unregister_mad_agent(dev->send_agent[p][q]);
return ret;
}
void mlx4_ib_mad_cleanup(struct mlx4_ib_dev *dev)
{
struct ib_mad_agent *agent;
int p, q;
for (p = 0; p < dev->dev->caps.num_ports; ++p) {
for (q = 0; q <= 1; ++q) {
agent = dev->send_agent[p][q];
dev->send_agent[p][q] = NULL;
ib_unregister_mad_agent(agent);
}
if (dev->sm_ah[p])
ib_destroy_ah(dev->sm_ah[p]);
}
}

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,285 @@
/*
* Copyright (c) 2006, 2007 Cisco Systems. 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.
*/
#ifndef MLX4_IB_H
#define MLX4_IB_H
#include <linux/compiler.h>
#include <linux/list.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_umem.h>
#include <linux/mlx4/device.h>
#include <linux/mlx4/doorbell.h>
enum {
MLX4_IB_DB_PER_PAGE = PAGE_SIZE / 4
};
struct mlx4_ib_db_pgdir;
struct mlx4_ib_user_db_page;
struct mlx4_ib_db {
__be32 *db;
union {
struct mlx4_ib_db_pgdir *pgdir;
struct mlx4_ib_user_db_page *user_page;
} u;
dma_addr_t dma;
int index;
int order;
};
struct mlx4_ib_ucontext {
struct ib_ucontext ibucontext;
struct mlx4_uar uar;
struct list_head db_page_list;
struct mutex db_page_mutex;
};
struct mlx4_ib_pd {
struct ib_pd ibpd;
u32 pdn;
};
struct mlx4_ib_cq_buf {
struct mlx4_buf buf;
struct mlx4_mtt mtt;
};
struct mlx4_ib_cq {
struct ib_cq ibcq;
struct mlx4_cq mcq;
struct mlx4_ib_cq_buf buf;
struct mlx4_ib_db db;
spinlock_t lock;
struct ib_umem *umem;
};
struct mlx4_ib_mr {
struct ib_mr ibmr;
struct mlx4_mr mmr;
struct ib_umem *umem;
};
struct mlx4_ib_wq {
u64 *wrid;
spinlock_t lock;
int max;
int max_gs;
int offset;
int wqe_shift;
unsigned head;
unsigned tail;
};
struct mlx4_ib_qp {
struct ib_qp ibqp;
struct mlx4_qp mqp;
struct mlx4_buf buf;
struct mlx4_ib_db db;
struct mlx4_ib_wq rq;
u32 doorbell_qpn;
__be32 sq_signal_bits;
struct mlx4_ib_wq sq;
struct ib_umem *umem;
struct mlx4_mtt mtt;
int buf_size;
struct mutex mutex;
u8 port;
u8 alt_port;
u8 atomic_rd_en;
u8 resp_depth;
u8 state;
};
struct mlx4_ib_srq {
struct ib_srq ibsrq;
struct mlx4_srq msrq;
struct mlx4_buf buf;
struct mlx4_ib_db db;
u64 *wrid;
spinlock_t lock;
int head;
int tail;
u16 wqe_ctr;
struct ib_umem *umem;
struct mlx4_mtt mtt;
struct mutex mutex;
};
struct mlx4_ib_ah {
struct ib_ah ibah;
struct mlx4_av av;
};
struct mlx4_ib_dev {
struct ib_device ib_dev;
struct mlx4_dev *dev;
void __iomem *uar_map;
struct list_head pgdir_list;
struct mutex pgdir_mutex;
struct mlx4_uar priv_uar;
u32 priv_pdn;
MLX4_DECLARE_DOORBELL_LOCK(uar_lock);
struct ib_mad_agent *send_agent[MLX4_MAX_PORTS][2];
struct ib_ah *sm_ah[MLX4_MAX_PORTS];
spinlock_t sm_lock;
struct mutex cap_mask_mutex;
};
static inline struct mlx4_ib_dev *to_mdev(struct ib_device *ibdev)
{
return container_of(ibdev, struct mlx4_ib_dev, ib_dev);
}
static inline struct mlx4_ib_ucontext *to_mucontext(struct ib_ucontext *ibucontext)
{
return container_of(ibucontext, struct mlx4_ib_ucontext, ibucontext);
}
static inline struct mlx4_ib_pd *to_mpd(struct ib_pd *ibpd)
{
return container_of(ibpd, struct mlx4_ib_pd, ibpd);
}
static inline struct mlx4_ib_cq *to_mcq(struct ib_cq *ibcq)
{
return container_of(ibcq, struct mlx4_ib_cq, ibcq);
}
static inline struct mlx4_ib_cq *to_mibcq(struct mlx4_cq *mcq)
{
return container_of(mcq, struct mlx4_ib_cq, mcq);
}
static inline struct mlx4_ib_mr *to_mmr(struct ib_mr *ibmr)
{
return container_of(ibmr, struct mlx4_ib_mr, ibmr);
}
static inline struct mlx4_ib_qp *to_mqp(struct ib_qp *ibqp)
{
return container_of(ibqp, struct mlx4_ib_qp, ibqp);
}
static inline struct mlx4_ib_qp *to_mibqp(struct mlx4_qp *mqp)
{
return container_of(mqp, struct mlx4_ib_qp, mqp);
}
static inline struct mlx4_ib_srq *to_msrq(struct ib_srq *ibsrq)
{
return container_of(ibsrq, struct mlx4_ib_srq, ibsrq);
}
static inline struct mlx4_ib_srq *to_mibsrq(struct mlx4_srq *msrq)
{
return container_of(msrq, struct mlx4_ib_srq, msrq);
}
static inline struct mlx4_ib_ah *to_mah(struct ib_ah *ibah)
{
return container_of(ibah, struct mlx4_ib_ah, ibah);
}
int mlx4_ib_db_alloc(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db, int order);
void mlx4_ib_db_free(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db);
int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
struct mlx4_ib_db *db);
void mlx4_ib_db_unmap_user(struct mlx4_ib_ucontext *context, struct mlx4_ib_db *db);
struct ib_mr *mlx4_ib_get_dma_mr(struct ib_pd *pd, int acc);
int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
struct ib_umem *umem);
struct ib_mr *mlx4_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u64 virt_addr, int access_flags,
struct ib_udata *udata);
int mlx4_ib_dereg_mr(struct ib_mr *mr);
struct ib_cq *mlx4_ib_create_cq(struct ib_device *ibdev, int entries, int vector,
struct ib_ucontext *context,
struct ib_udata *udata);
int mlx4_ib_destroy_cq(struct ib_cq *cq);
int mlx4_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc);
int mlx4_ib_arm_cq(struct ib_cq *cq, enum ib_cq_notify_flags flags);
void __mlx4_ib_cq_clean(struct mlx4_ib_cq *cq, u32 qpn, struct mlx4_ib_srq *srq);
void mlx4_ib_cq_clean(struct mlx4_ib_cq *cq, u32 qpn, struct mlx4_ib_srq *srq);
struct ib_ah *mlx4_ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr);
int mlx4_ib_query_ah(struct ib_ah *ibah, struct ib_ah_attr *ah_attr);
int mlx4_ib_destroy_ah(struct ib_ah *ah);
struct ib_srq *mlx4_ib_create_srq(struct ib_pd *pd,
struct ib_srq_init_attr *init_attr,
struct ib_udata *udata);
int mlx4_ib_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
enum ib_srq_attr_mask attr_mask, struct ib_udata *udata);
int mlx4_ib_destroy_srq(struct ib_srq *srq);
void mlx4_ib_free_srq_wqe(struct mlx4_ib_srq *srq, int wqe_index);
int mlx4_ib_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
struct ib_recv_wr **bad_wr);
struct ib_qp *mlx4_ib_create_qp(struct ib_pd *pd,
struct ib_qp_init_attr *init_attr,
struct ib_udata *udata);
int mlx4_ib_destroy_qp(struct ib_qp *qp);
int mlx4_ib_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
int attr_mask, struct ib_udata *udata);
int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
struct ib_send_wr **bad_wr);
int mlx4_ib_post_recv(struct ib_qp *ibqp, struct ib_recv_wr *wr,
struct ib_recv_wr **bad_wr);
int mlx4_MAD_IFC(struct mlx4_ib_dev *dev, int ignore_mkey, int ignore_bkey,
int port, struct ib_wc *in_wc, struct ib_grh *in_grh,
void *in_mad, void *response_mad);
int mlx4_ib_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
struct ib_wc *in_wc, struct ib_grh *in_grh,
struct ib_mad *in_mad, struct ib_mad *out_mad);
int mlx4_ib_mad_init(struct mlx4_ib_dev *dev);
void mlx4_ib_mad_cleanup(struct mlx4_ib_dev *dev);
static inline int mlx4_ib_ah_grh_present(struct mlx4_ib_ah *ah)
{
return !!(ah->av.g_slid & 0x80);
}
#endif /* MLX4_IB_H */

View File

@@ -0,0 +1,184 @@
/*
* Copyright (c) 2007 Cisco Systems, 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 "mlx4_ib.h"
static u32 convert_access(int acc)
{
return (acc & IB_ACCESS_REMOTE_ATOMIC ? MLX4_PERM_ATOMIC : 0) |
(acc & IB_ACCESS_REMOTE_WRITE ? MLX4_PERM_REMOTE_WRITE : 0) |
(acc & IB_ACCESS_REMOTE_READ ? MLX4_PERM_REMOTE_READ : 0) |
(acc & IB_ACCESS_LOCAL_WRITE ? MLX4_PERM_LOCAL_WRITE : 0) |
MLX4_PERM_LOCAL_READ;
}
struct ib_mr *mlx4_ib_get_dma_mr(struct ib_pd *pd, int acc)
{
struct mlx4_ib_mr *mr;
int err;
mr = kmalloc(sizeof *mr, GFP_KERNEL);
if (!mr)
return ERR_PTR(-ENOMEM);
err = mlx4_mr_alloc(to_mdev(pd->device)->dev, to_mpd(pd)->pdn, 0,
~0ull, convert_access(acc), 0, 0, &mr->mmr);
if (err)
goto err_free;
err = mlx4_mr_enable(to_mdev(pd->device)->dev, &mr->mmr);
if (err)
goto err_mr;
mr->ibmr.rkey = mr->ibmr.lkey = mr->mmr.key;
mr->umem = NULL;
return &mr->ibmr;
err_mr:
mlx4_mr_free(to_mdev(pd->device)->dev, &mr->mmr);
err_free:
kfree(mr);
return ERR_PTR(err);
}
int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
struct ib_umem *umem)
{
u64 *pages;
struct ib_umem_chunk *chunk;
int i, j, k;
int n;
int len;
int err = 0;
pages = (u64 *) __get_free_page(GFP_KERNEL);
if (!pages)
return -ENOMEM;
i = n = 0;
list_for_each_entry(chunk, &umem->chunk_list, list)
for (j = 0; j < chunk->nmap; ++j) {
len = sg_dma_len(&chunk->page_list[j]) >> mtt->page_shift;
for (k = 0; k < len; ++k) {
pages[i++] = sg_dma_address(&chunk->page_list[j]) +
umem->page_size * k;
/*
* Be friendly to WRITE_MTT firmware
* command, and pass it chunks of
* appropriate size.
*/
if (i == PAGE_SIZE / sizeof (u64) - 2) {
err = mlx4_write_mtt(dev->dev, mtt, n,
i, pages);
if (err)
goto out;
n += i;
i = 0;
}
}
}
if (i)
err = mlx4_write_mtt(dev->dev, mtt, n, i, pages);
out:
free_page((unsigned long) pages);
return err;
}
struct ib_mr *mlx4_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
u64 virt_addr, int access_flags,
struct ib_udata *udata)
{
struct mlx4_ib_dev *dev = to_mdev(pd->device);
struct mlx4_ib_mr *mr;
int shift;
int err;
int n;
mr = kmalloc(sizeof *mr, GFP_KERNEL);
if (!mr)
return ERR_PTR(-ENOMEM);
mr->umem = ib_umem_get(pd->uobject->context, start, length, access_flags);
if (IS_ERR(mr->umem)) {
err = PTR_ERR(mr->umem);
goto err_free;
}
n = ib_umem_page_count(mr->umem);
shift = ilog2(mr->umem->page_size);
err = mlx4_mr_alloc(dev->dev, to_mpd(pd)->pdn, virt_addr, length,
convert_access(access_flags), n, shift, &mr->mmr);
if (err)
goto err_umem;
err = mlx4_ib_umem_write_mtt(dev, &mr->mmr.mtt, mr->umem);
if (err)
goto err_mr;
err = mlx4_mr_enable(dev->dev, &mr->mmr);
if (err)
goto err_mr;
mr->ibmr.rkey = mr->ibmr.lkey = mr->mmr.key;
return &mr->ibmr;
err_mr:
mlx4_mr_free(to_mdev(pd->device)->dev, &mr->mmr);
err_umem:
ib_umem_release(mr->umem);
err_free:
kfree(mr);
return ERR_PTR(err);
}
int mlx4_ib_dereg_mr(struct ib_mr *ibmr)
{
struct mlx4_ib_mr *mr = to_mmr(ibmr);
mlx4_mr_free(to_mdev(ibmr->device)->dev, &mr->mmr);
if (mr->umem)
ib_umem_release(mr->umem);
kfree(mr);
return 0;
}

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,334 @@
/*
* Copyright (c) 2007 Cisco Systems, 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/mlx4/qp.h>
#include <linux/mlx4/srq.h>
#include "mlx4_ib.h"
#include "user.h"
static void *get_wqe(struct mlx4_ib_srq *srq, int n)
{
int offset = n << srq->msrq.wqe_shift;
if (srq->buf.nbufs == 1)
return srq->buf.u.direct.buf + offset;
else
return srq->buf.u.page_list[offset >> PAGE_SHIFT].buf +
(offset & (PAGE_SIZE - 1));
}
static void mlx4_ib_srq_event(struct mlx4_srq *srq, enum mlx4_event type)
{
struct ib_event event;
struct ib_srq *ibsrq = &to_mibsrq(srq)->ibsrq;
if (ibsrq->event_handler) {
event.device = ibsrq->device;
event.element.srq = ibsrq;
switch (type) {
case MLX4_EVENT_TYPE_SRQ_LIMIT:
event.event = IB_EVENT_SRQ_LIMIT_REACHED;
break;
case MLX4_EVENT_TYPE_SRQ_CATAS_ERROR:
event.event = IB_EVENT_SRQ_ERR;
break;
default:
printk(KERN_WARNING "mlx4_ib: Unexpected event type %d "
"on SRQ %06x\n", type, srq->srqn);
return;
}
ibsrq->event_handler(&event, ibsrq->srq_context);
}
}
struct ib_srq *mlx4_ib_create_srq(struct ib_pd *pd,
struct ib_srq_init_attr *init_attr,
struct ib_udata *udata)
{
struct mlx4_ib_dev *dev = to_mdev(pd->device);
struct mlx4_ib_srq *srq;
struct mlx4_wqe_srq_next_seg *next;
int desc_size;
int buf_size;
int err;
int i;
/* Sanity check SRQ size before proceeding */
if (init_attr->attr.max_wr >= dev->dev->caps.max_srq_wqes ||
init_attr->attr.max_sge > dev->dev->caps.max_srq_sge)
return ERR_PTR(-EINVAL);
srq = kmalloc(sizeof *srq, GFP_KERNEL);
if (!srq)
return ERR_PTR(-ENOMEM);
mutex_init(&srq->mutex);
spin_lock_init(&srq->lock);
srq->msrq.max = roundup_pow_of_two(init_attr->attr.max_wr + 1);
srq->msrq.max_gs = init_attr->attr.max_sge;
desc_size = max(32UL,
roundup_pow_of_two(sizeof (struct mlx4_wqe_srq_next_seg) +
srq->msrq.max_gs *
sizeof (struct mlx4_wqe_data_seg)));
srq->msrq.wqe_shift = ilog2(desc_size);
buf_size = srq->msrq.max * desc_size;
if (pd->uobject) {
struct mlx4_ib_create_srq ucmd;
if (ib_copy_from_udata(&ucmd, udata, sizeof ucmd)) {
err = -EFAULT;
goto err_srq;
}
srq->umem = ib_umem_get(pd->uobject->context, ucmd.buf_addr,
buf_size, 0);
if (IS_ERR(srq->umem)) {
err = PTR_ERR(srq->umem);
goto err_srq;
}
err = mlx4_mtt_init(dev->dev, ib_umem_page_count(srq->umem),
ilog2(srq->umem->page_size), &srq->mtt);
if (err)
goto err_buf;
err = mlx4_ib_umem_write_mtt(dev, &srq->mtt, srq->umem);
if (err)
goto err_mtt;
err = mlx4_ib_db_map_user(to_mucontext(pd->uobject->context),
ucmd.db_addr, &srq->db);
if (err)
goto err_mtt;
} else {
err = mlx4_ib_db_alloc(dev, &srq->db, 0);
if (err)
goto err_srq;
*srq->db.db = 0;
if (mlx4_buf_alloc(dev->dev, buf_size, PAGE_SIZE * 2, &srq->buf)) {
err = -ENOMEM;
goto err_db;
}
srq->head = 0;
srq->tail = srq->msrq.max - 1;
srq->wqe_ctr = 0;
for (i = 0; i < srq->msrq.max; ++i) {
next = get_wqe(srq, i);
next->next_wqe_index =
cpu_to_be16((i + 1) & (srq->msrq.max - 1));
}
err = mlx4_mtt_init(dev->dev, srq->buf.npages, srq->buf.page_shift,
&srq->mtt);
if (err)
goto err_buf;
err = mlx4_buf_write_mtt(dev->dev, &srq->mtt, &srq->buf);
if (err)
goto err_mtt;
srq->wrid = kmalloc(srq->msrq.max * sizeof (u64), GFP_KERNEL);
if (!srq->wrid) {
err = -ENOMEM;
goto err_mtt;
}
}
err = mlx4_srq_alloc(dev->dev, to_mpd(pd)->pdn, &srq->mtt,
srq->db.dma, &srq->msrq);
if (err)
goto err_wrid;
srq->msrq.event = mlx4_ib_srq_event;
if (pd->uobject)
if (ib_copy_to_udata(udata, &srq->msrq.srqn, sizeof (__u32))) {
err = -EFAULT;
goto err_wrid;
}
init_attr->attr.max_wr = srq->msrq.max - 1;
return &srq->ibsrq;
err_wrid:
if (pd->uobject)
mlx4_ib_db_unmap_user(to_mucontext(pd->uobject->context), &srq->db);
else
kfree(srq->wrid);
err_mtt:
mlx4_mtt_cleanup(dev->dev, &srq->mtt);
err_buf:
if (pd->uobject)
ib_umem_release(srq->umem);
else
mlx4_buf_free(dev->dev, buf_size, &srq->buf);
err_db:
if (!pd->uobject)
mlx4_ib_db_free(dev, &srq->db);
err_srq:
kfree(srq);
return ERR_PTR(err);
}
int mlx4_ib_modify_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr,
enum ib_srq_attr_mask attr_mask, struct ib_udata *udata)
{
struct mlx4_ib_dev *dev = to_mdev(ibsrq->device);
struct mlx4_ib_srq *srq = to_msrq(ibsrq);
int ret;
/* We don't support resizing SRQs (yet?) */
if (attr_mask & IB_SRQ_MAX_WR)
return -EINVAL;
if (attr_mask & IB_SRQ_LIMIT) {
if (attr->srq_limit >= srq->msrq.max)
return -EINVAL;
mutex_lock(&srq->mutex);
ret = mlx4_srq_arm(dev->dev, &srq->msrq, attr->srq_limit);
mutex_unlock(&srq->mutex);
if (ret)
return ret;
}
return 0;
}
int mlx4_ib_destroy_srq(struct ib_srq *srq)
{
struct mlx4_ib_dev *dev = to_mdev(srq->device);
struct mlx4_ib_srq *msrq = to_msrq(srq);
mlx4_srq_free(dev->dev, &msrq->msrq);
mlx4_mtt_cleanup(dev->dev, &msrq->mtt);
if (srq->uobject) {
mlx4_ib_db_unmap_user(to_mucontext(srq->uobject->context), &msrq->db);
ib_umem_release(msrq->umem);
} else {
kfree(msrq->wrid);
mlx4_buf_free(dev->dev, msrq->msrq.max << msrq->msrq.wqe_shift,
&msrq->buf);
mlx4_ib_db_free(dev, &msrq->db);
}
kfree(msrq);
return 0;
}
void mlx4_ib_free_srq_wqe(struct mlx4_ib_srq *srq, int wqe_index)
{
struct mlx4_wqe_srq_next_seg *next;
/* always called with interrupts disabled. */
spin_lock(&srq->lock);
next = get_wqe(srq, srq->tail);
next->next_wqe_index = cpu_to_be16(wqe_index);
srq->tail = wqe_index;
spin_unlock(&srq->lock);
}
int mlx4_ib_post_srq_recv(struct ib_srq *ibsrq, struct ib_recv_wr *wr,
struct ib_recv_wr **bad_wr)
{
struct mlx4_ib_srq *srq = to_msrq(ibsrq);
struct mlx4_wqe_srq_next_seg *next;
struct mlx4_wqe_data_seg *scat;
unsigned long flags;
int err = 0;
int nreq;
int i;
spin_lock_irqsave(&srq->lock, flags);
for (nreq = 0; wr; ++nreq, wr = wr->next) {
if (unlikely(wr->num_sge > srq->msrq.max_gs)) {
err = -EINVAL;
*bad_wr = wr;
break;
}
srq->wrid[srq->head] = wr->wr_id;
next = get_wqe(srq, srq->head);
srq->head = be16_to_cpu(next->next_wqe_index);
scat = (struct mlx4_wqe_data_seg *) (next + 1);
for (i = 0; i < wr->num_sge; ++i) {
scat[i].byte_count = cpu_to_be32(wr->sg_list[i].length);
scat[i].lkey = cpu_to_be32(wr->sg_list[i].lkey);
scat[i].addr = cpu_to_be64(wr->sg_list[i].addr);
}
if (i < srq->msrq.max_gs) {
scat[i].byte_count = 0;
scat[i].lkey = cpu_to_be32(MLX4_INVALID_LKEY);
scat[i].addr = 0;
}
}
if (likely(nreq)) {
srq->wqe_ctr += nreq;
/*
* Make sure that descriptors are written before
* doorbell record.
*/
wmb();
*srq->db.db = cpu_to_be32(srq->wqe_ctr);
}
spin_unlock_irqrestore(&srq->lock, flags);
return err;
}

View File

@@ -0,0 +1,92 @@
/*
* Copyright (c) 2007 Cisco Systems, 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.
*/
#ifndef MLX4_IB_USER_H
#define MLX4_IB_USER_H
#include <linux/types.h>
/*
* Increment this value if any changes that break userspace ABI
* compatibility are made.
*/
#define MLX4_IB_UVERBS_ABI_VERSION 1
/*
* Make sure that all structs defined in this file remain laid out so
* that they pack the same way on 32-bit and 64-bit architectures (to
* avoid incompatibility between 32-bit userspace and 64-bit kernels).
* In particular do not use pointer types -- pass pointers in __u64
* instead.
*/
struct mlx4_ib_alloc_ucontext_resp {
__u32 qp_tab_size;
__u16 bf_reg_size;
__u16 bf_regs_per_page;
};
struct mlx4_ib_alloc_pd_resp {
__u32 pdn;
__u32 reserved;
};
struct mlx4_ib_create_cq {
__u64 buf_addr;
__u64 db_addr;
};
struct mlx4_ib_create_cq_resp {
__u32 cqn;
__u32 reserved;
};
struct mlx4_ib_resize_cq {
__u64 buf_addr;
};
struct mlx4_ib_create_srq {
__u64 buf_addr;
__u64 db_addr;
};
struct mlx4_ib_create_srq_resp {
__u32 srqn;
__u32 reserved;
};
struct mlx4_ib_create_qp {
__u64 buf_addr;
__u64 db_addr;
};
#endif /* MLX4_IB_USER_H */

View File

@@ -2493,6 +2493,20 @@ config PASEMI_MAC
This driver supports the on-chip 1/10Gbit Ethernet controller on
PA Semi's PWRficient line of chips.
config MLX4_CORE
tristate
depends on PCI
default n
config MLX4_DEBUG
bool "Verbose debugging output" if (MLX4_CORE && EMBEDDED)
default y
---help---
This option causes debugging code to be compiled into the
mlx4_core driver. The output can be turned on via the
debug_level module parameter (which can also be set after
the driver is loaded through sysfs).
endmenu
source "drivers/net/tokenring/Kconfig"

View File

@@ -197,6 +197,7 @@ obj-$(CONFIG_SMC911X) += smc911x.o
obj-$(CONFIG_DM9000) += dm9000.o
obj-$(CONFIG_FEC_8XX) += fec_8xx/
obj-$(CONFIG_PASEMI_MAC) += pasemi_mac.o
obj-$(CONFIG_MLX4_CORE) += mlx4/
obj-$(CONFIG_MACB) += macb.o

View File

@@ -0,0 +1,4 @@
obj-$(CONFIG_MLX4_CORE) += mlx4_core.o
mlx4_core-y := alloc.o catas.o cmd.o cq.o eq.o fw.o icm.o intf.o main.o mcg.o \
mr.o pd.o profile.o qp.o reset.o srq.o

179
drivers/net/mlx4/alloc.c Normal file
View File

@@ -0,0 +1,179 @@
/*
* Copyright (c) 2006, 2007 Cisco Systems, 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/errno.h>
#include <linux/slab.h>
#include <linux/bitmap.h>
#include "mlx4.h"
u32 mlx4_bitmap_alloc(struct mlx4_bitmap *bitmap)
{
u32 obj;
spin_lock(&bitmap->lock);
obj = find_next_zero_bit(bitmap->table, bitmap->max, bitmap->last);
if (obj >= bitmap->max) {
bitmap->top = (bitmap->top + bitmap->max) & bitmap->mask;
obj = find_first_zero_bit(bitmap->table, bitmap->max);
}
if (obj < bitmap->max) {
set_bit(obj, bitmap->table);
obj |= bitmap->top;
bitmap->last = obj + 1;
} else
obj = -1;
spin_unlock(&bitmap->lock);
return obj;
}
void mlx4_bitmap_free(struct mlx4_bitmap *bitmap, u32 obj)
{
obj &= bitmap->max - 1;
spin_lock(&bitmap->lock);
clear_bit(obj, bitmap->table);
bitmap->last = min(bitmap->last, obj);
bitmap->top = (bitmap->top + bitmap->max) & bitmap->mask;
spin_unlock(&bitmap->lock);
}
int mlx4_bitmap_init(struct mlx4_bitmap *bitmap, u32 num, u32 mask, u32 reserved)
{
int i;
/* num must be a power of 2 */
if (num != roundup_pow_of_two(num))
return -EINVAL;
bitmap->last = 0;
bitmap->top = 0;
bitmap->max = num;
bitmap->mask = mask;
spin_lock_init(&bitmap->lock);
bitmap->table = kzalloc(BITS_TO_LONGS(num) * sizeof (long), GFP_KERNEL);
if (!bitmap->table)
return -ENOMEM;
for (i = 0; i < reserved; ++i)
set_bit(i, bitmap->table);
return 0;
}
void mlx4_bitmap_cleanup(struct mlx4_bitmap *bitmap)
{
kfree(bitmap->table);
}
/*
* Handling for queue buffers -- we allocate a bunch of memory and
* register it in a memory region at HCA virtual address 0. If the
* requested size is > max_direct, we split the allocation into
* multiple pages, so we don't require too much contiguous memory.
*/
int mlx4_buf_alloc(struct mlx4_dev *dev, int size, int max_direct,
struct mlx4_buf *buf)
{
dma_addr_t t;
if (size <= max_direct) {
buf->nbufs = 1;
buf->npages = 1;
buf->page_shift = get_order(size) + PAGE_SHIFT;
buf->u.direct.buf = dma_alloc_coherent(&dev->pdev->dev,
size, &t, GFP_KERNEL);
if (!buf->u.direct.buf)
return -ENOMEM;
buf->u.direct.map = t;
while (t & ((1 << buf->page_shift) - 1)) {
--buf->page_shift;
buf->npages *= 2;
}
memset(buf->u.direct.buf, 0, size);
} else {
int i;
buf->nbufs = (size + PAGE_SIZE - 1) / PAGE_SIZE;
buf->npages = buf->nbufs;
buf->page_shift = PAGE_SHIFT;
buf->u.page_list = kzalloc(buf->nbufs * sizeof *buf->u.page_list,
GFP_KERNEL);
if (!buf->u.page_list)
return -ENOMEM;
for (i = 0; i < buf->nbufs; ++i) {
buf->u.page_list[i].buf =
dma_alloc_coherent(&dev->pdev->dev, PAGE_SIZE,
&t, GFP_KERNEL);
if (!buf->u.page_list[i].buf)
goto err_free;
buf->u.page_list[i].map = t;
memset(buf->u.page_list[i].buf, 0, PAGE_SIZE);
}
}
return 0;
err_free:
mlx4_buf_free(dev, size, buf);
return -ENOMEM;
}
EXPORT_SYMBOL_GPL(mlx4_buf_alloc);
void mlx4_buf_free(struct mlx4_dev *dev, int size, struct mlx4_buf *buf)
{
int i;
if (buf->nbufs == 1)
dma_free_coherent(&dev->pdev->dev, size, buf->u.direct.buf,
buf->u.direct.map);
else {
for (i = 0; i < buf->nbufs; ++i)
dma_free_coherent(&dev->pdev->dev, PAGE_SIZE,
buf->u.page_list[i].buf,
buf->u.page_list[i].map);
kfree(buf->u.page_list);
}
}
EXPORT_SYMBOL_GPL(mlx4_buf_free);

70
drivers/net/mlx4/catas.c Normal file
View File

@@ -0,0 +1,70 @@
/*
* Copyright (c) 2007 Cisco Systems, 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 "mlx4.h"
void mlx4_handle_catas_err(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
int i;
mlx4_err(dev, "Catastrophic error detected:\n");
for (i = 0; i < priv->fw.catas_size; ++i)
mlx4_err(dev, " buf[%02x]: %08x\n",
i, swab32(readl(priv->catas_err.map + i)));
mlx4_dispatch_event(dev, MLX4_EVENT_TYPE_LOCAL_CATAS_ERROR, 0, 0);
}
void mlx4_map_catas_buf(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
unsigned long addr;
addr = pci_resource_start(dev->pdev, priv->fw.catas_bar) +
priv->fw.catas_offset;
priv->catas_err.map = ioremap(addr, priv->fw.catas_size * 4);
if (!priv->catas_err.map)
mlx4_warn(dev, "Failed to map catastrophic error buffer at 0x%lx\n",
addr);
}
void mlx4_unmap_catas_buf(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
if (priv->catas_err.map)
iounmap(priv->catas_err.map);
}

429
drivers/net/mlx4/cmd.c Normal file
View File

@@ -0,0 +1,429 @@
/*
* Copyright (c) 2004, 2005 Topspin Communications. All rights reserved.
* Copyright (c) 2005 Mellanox Technologies. All rights reserved.
* Copyright (c) 2005, 2006, 2007 Cisco Systems, 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/sched.h>
#include <linux/pci.h>
#include <linux/errno.h>
#include <linux/mlx4/cmd.h>
#include <asm/io.h>
#include "mlx4.h"
#define CMD_POLL_TOKEN 0xffff
enum {
/* command completed successfully: */
CMD_STAT_OK = 0x00,
/* Internal error (such as a bus error) occurred while processing command: */
CMD_STAT_INTERNAL_ERR = 0x01,
/* Operation/command not supported or opcode modifier not supported: */
CMD_STAT_BAD_OP = 0x02,
/* Parameter not supported or parameter out of range: */
CMD_STAT_BAD_PARAM = 0x03,
/* System not enabled or bad system state: */
CMD_STAT_BAD_SYS_STATE = 0x04,
/* Attempt to access reserved or unallocaterd resource: */
CMD_STAT_BAD_RESOURCE = 0x05,
/* Requested resource is currently executing a command, or is otherwise busy: */
CMD_STAT_RESOURCE_BUSY = 0x06,
/* Required capability exceeds device limits: */
CMD_STAT_EXCEED_LIM = 0x08,
/* Resource is not in the appropriate state or ownership: */
CMD_STAT_BAD_RES_STATE = 0x09,
/* Index out of range: */
CMD_STAT_BAD_INDEX = 0x0a,
/* FW image corrupted: */
CMD_STAT_BAD_NVMEM = 0x0b,
/* Attempt to modify a QP/EE which is not in the presumed state: */
CMD_STAT_BAD_QP_STATE = 0x10,
/* Bad segment parameters (Address/Size): */
CMD_STAT_BAD_SEG_PARAM = 0x20,
/* Memory Region has Memory Windows bound to: */
CMD_STAT_REG_BOUND = 0x21,
/* HCA local attached memory not present: */
CMD_STAT_LAM_NOT_PRE = 0x22,
/* Bad management packet (silently discarded): */
CMD_STAT_BAD_PKT = 0x30,
/* More outstanding CQEs in CQ than new CQ size: */
CMD_STAT_BAD_SIZE = 0x40
};
enum {
HCR_IN_PARAM_OFFSET = 0x00,
HCR_IN_MODIFIER_OFFSET = 0x08,
HCR_OUT_PARAM_OFFSET = 0x0c,
HCR_TOKEN_OFFSET = 0x14,
HCR_STATUS_OFFSET = 0x18,
HCR_OPMOD_SHIFT = 12,
HCR_T_BIT = 21,
HCR_E_BIT = 22,
HCR_GO_BIT = 23
};
enum {
GO_BIT_TIMEOUT = 10000
};
struct mlx4_cmd_context {
struct completion done;
int result;
int next;
u64 out_param;
u16 token;
};
static int mlx4_status_to_errno(u8 status) {
static const int trans_table[] = {
[CMD_STAT_INTERNAL_ERR] = -EIO,
[CMD_STAT_BAD_OP] = -EPERM,
[CMD_STAT_BAD_PARAM] = -EINVAL,
[CMD_STAT_BAD_SYS_STATE] = -ENXIO,
[CMD_STAT_BAD_RESOURCE] = -EBADF,
[CMD_STAT_RESOURCE_BUSY] = -EBUSY,
[CMD_STAT_EXCEED_LIM] = -ENOMEM,
[CMD_STAT_BAD_RES_STATE] = -EBADF,
[CMD_STAT_BAD_INDEX] = -EBADF,
[CMD_STAT_BAD_NVMEM] = -EFAULT,
[CMD_STAT_BAD_QP_STATE] = -EINVAL,
[CMD_STAT_BAD_SEG_PARAM] = -EFAULT,
[CMD_STAT_REG_BOUND] = -EBUSY,
[CMD_STAT_LAM_NOT_PRE] = -EAGAIN,
[CMD_STAT_BAD_PKT] = -EINVAL,
[CMD_STAT_BAD_SIZE] = -ENOMEM,
};
if (status >= ARRAY_SIZE(trans_table) ||
(status != CMD_STAT_OK && trans_table[status] == 0))
return -EIO;
return trans_table[status];
}
static int cmd_pending(struct mlx4_dev *dev)
{
u32 status = readl(mlx4_priv(dev)->cmd.hcr + HCR_STATUS_OFFSET);
return (status & swab32(1 << HCR_GO_BIT)) ||
(mlx4_priv(dev)->cmd.toggle ==
!!(status & swab32(1 << HCR_T_BIT)));
}
static int mlx4_cmd_post(struct mlx4_dev *dev, u64 in_param, u64 out_param,
u32 in_modifier, u8 op_modifier, u16 op, u16 token,
int event)
{
struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
u32 __iomem *hcr = cmd->hcr;
int ret = -EAGAIN;
unsigned long end;
mutex_lock(&cmd->hcr_mutex);
end = jiffies;
if (event)
end += HZ * 10;
while (cmd_pending(dev)) {
if (time_after_eq(jiffies, end))
goto out;
cond_resched();
}
/*
* We use writel (instead of something like memcpy_toio)
* because writes of less than 32 bits to the HCR don't work
* (and some architectures such as ia64 implement memcpy_toio
* in terms of writeb).
*/
__raw_writel((__force u32) cpu_to_be32(in_param >> 32), hcr + 0);
__raw_writel((__force u32) cpu_to_be32(in_param & 0xfffffffful), hcr + 1);
__raw_writel((__force u32) cpu_to_be32(in_modifier), hcr + 2);
__raw_writel((__force u32) cpu_to_be32(out_param >> 32), hcr + 3);
__raw_writel((__force u32) cpu_to_be32(out_param & 0xfffffffful), hcr + 4);
__raw_writel((__force u32) cpu_to_be32(token << 16), hcr + 5);
/* __raw_writel may not order writes. */
wmb();
__raw_writel((__force u32) cpu_to_be32((1 << HCR_GO_BIT) |
(cmd->toggle << HCR_T_BIT) |
(event ? (1 << HCR_E_BIT) : 0) |
(op_modifier << HCR_OPMOD_SHIFT) |
op), hcr + 6);
cmd->toggle = cmd->toggle ^ 1;
ret = 0;
out:
mutex_unlock(&cmd->hcr_mutex);
return ret;
}
static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
int out_is_imm, u32 in_modifier, u8 op_modifier,
u16 op, unsigned long timeout)
{
struct mlx4_priv *priv = mlx4_priv(dev);
void __iomem *hcr = priv->cmd.hcr;
int err = 0;
unsigned long end;
down(&priv->cmd.poll_sem);
err = mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
in_modifier, op_modifier, op, CMD_POLL_TOKEN, 0);
if (err)
goto out;
end = msecs_to_jiffies(timeout) + jiffies;
while (cmd_pending(dev) && time_before(jiffies, end))
cond_resched();
if (cmd_pending(dev)) {
err = -ETIMEDOUT;
goto out;
}
if (out_is_imm)
*out_param =
(u64) be32_to_cpu((__force __be32)
__raw_readl(hcr + HCR_OUT_PARAM_OFFSET)) << 32 |
(u64) be32_to_cpu((__force __be32)
__raw_readl(hcr + HCR_OUT_PARAM_OFFSET + 4));
err = mlx4_status_to_errno(be32_to_cpu((__force __be32)
__raw_readl(hcr + HCR_STATUS_OFFSET)) >> 24);
out:
up(&priv->cmd.poll_sem);
return err;
}
void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param)
{
struct mlx4_priv *priv = mlx4_priv(dev);
struct mlx4_cmd_context *context =
&priv->cmd.context[token & priv->cmd.token_mask];
/* previously timed out command completing at long last */
if (token != context->token)
return;
context->result = mlx4_status_to_errno(status);
context->out_param = out_param;
context->token += priv->cmd.token_mask + 1;
complete(&context->done);
}
static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
int out_is_imm, u32 in_modifier, u8 op_modifier,
u16 op, unsigned long timeout)
{
struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
struct mlx4_cmd_context *context;
int err = 0;
down(&cmd->event_sem);
spin_lock(&cmd->context_lock);
BUG_ON(cmd->free_head < 0);
context = &cmd->context[cmd->free_head];
cmd->free_head = context->next;
spin_unlock(&cmd->context_lock);
init_completion(&context->done);
mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
in_modifier, op_modifier, op, context->token, 1);
if (!wait_for_completion_timeout(&context->done, msecs_to_jiffies(timeout))) {
err = -EBUSY;
goto out;
}
err = context->result;
if (err)
goto out;
if (out_is_imm)
*out_param = context->out_param;
out:
spin_lock(&cmd->context_lock);
context->next = cmd->free_head;
cmd->free_head = context - cmd->context;
spin_unlock(&cmd->context_lock);
up(&cmd->event_sem);
return err;
}
int __mlx4_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
int out_is_imm, u32 in_modifier, u8 op_modifier,
u16 op, unsigned long timeout)
{
if (mlx4_priv(dev)->cmd.use_events)
return mlx4_cmd_wait(dev, in_param, out_param, out_is_imm,
in_modifier, op_modifier, op, timeout);
else
return mlx4_cmd_poll(dev, in_param, out_param, out_is_imm,
in_modifier, op_modifier, op, timeout);
}
EXPORT_SYMBOL_GPL(__mlx4_cmd);
int mlx4_cmd_init(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
mutex_init(&priv->cmd.hcr_mutex);
sema_init(&priv->cmd.poll_sem, 1);
priv->cmd.use_events = 0;
priv->cmd.toggle = 1;
priv->cmd.hcr = ioremap(pci_resource_start(dev->pdev, 0) + MLX4_HCR_BASE,
MLX4_HCR_SIZE);
if (!priv->cmd.hcr) {
mlx4_err(dev, "Couldn't map command register.");
return -ENOMEM;
}
priv->cmd.pool = pci_pool_create("mlx4_cmd", dev->pdev,
MLX4_MAILBOX_SIZE,
MLX4_MAILBOX_SIZE, 0);
if (!priv->cmd.pool) {
iounmap(priv->cmd.hcr);
return -ENOMEM;
}
return 0;
}
void mlx4_cmd_cleanup(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
pci_pool_destroy(priv->cmd.pool);
iounmap(priv->cmd.hcr);
}
/*
* Switch to using events to issue FW commands (can only be called
* after event queue for command events has been initialized).
*/
int mlx4_cmd_use_events(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
int i;
priv->cmd.context = kmalloc(priv->cmd.max_cmds *
sizeof (struct mlx4_cmd_context),
GFP_KERNEL);
if (!priv->cmd.context)
return -ENOMEM;
for (i = 0; i < priv->cmd.max_cmds; ++i) {
priv->cmd.context[i].token = i;
priv->cmd.context[i].next = i + 1;
}
priv->cmd.context[priv->cmd.max_cmds - 1].next = -1;
priv->cmd.free_head = 0;
sema_init(&priv->cmd.event_sem, priv->cmd.max_cmds);
spin_lock_init(&priv->cmd.context_lock);
for (priv->cmd.token_mask = 1;
priv->cmd.token_mask < priv->cmd.max_cmds;
priv->cmd.token_mask <<= 1)
; /* nothing */
--priv->cmd.token_mask;
priv->cmd.use_events = 1;
down(&priv->cmd.poll_sem);
return 0;
}
/*
* Switch back to polling (used when shutting down the device)
*/
void mlx4_cmd_use_polling(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
int i;
priv->cmd.use_events = 0;
for (i = 0; i < priv->cmd.max_cmds; ++i)
down(&priv->cmd.event_sem);
kfree(priv->cmd.context);
up(&priv->cmd.poll_sem);
}
struct mlx4_cmd_mailbox *mlx4_alloc_cmd_mailbox(struct mlx4_dev *dev)
{
struct mlx4_cmd_mailbox *mailbox;
mailbox = kmalloc(sizeof *mailbox, GFP_KERNEL);
if (!mailbox)
return ERR_PTR(-ENOMEM);
mailbox->buf = pci_pool_alloc(mlx4_priv(dev)->cmd.pool, GFP_KERNEL,
&mailbox->dma);
if (!mailbox->buf) {
kfree(mailbox);
return ERR_PTR(-ENOMEM);
}
return mailbox;
}
EXPORT_SYMBOL_GPL(mlx4_alloc_cmd_mailbox);
void mlx4_free_cmd_mailbox(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox)
{
if (!mailbox)
return;
pci_pool_free(mlx4_priv(dev)->cmd.pool, mailbox->buf, mailbox->dma);
kfree(mailbox);
}
EXPORT_SYMBOL_GPL(mlx4_free_cmd_mailbox);

Some files were not shown because too many files have changed in this diff Show More