linux/drivers/infiniband/hw/ehca/ehca_eq.c
Stefan Roscher e2f81daf23 IB/ehca: Support small QP queues
eHCA2 supports QP queues that can be as small as 512 bytes. This
greatly reduces memory overhead for consumers that use lots of QPs
with small queues (e.g. RDMA-only QPs). Apart from dealing with
firmware, this code needs to manage bite-sized chunks of kernel pages,
making sure that no kernel page is shared between different protection
domains.

Signed-off-by: Hoang-Nam Nguyen <hnguyen@de.ibm.com>
2007-07-20 21:19:47 -07:00

188 lines
5.0 KiB
C

/*
* IBM eServer eHCA Infiniband device driver for Linux on POWER
*
* Event queue handling
*
* Authors: Waleri Fomin <fomin@de.ibm.com>
* Khadija Souissi <souissi@de.ibm.com>
* Reinhard Ernst <rernst@de.ibm.com>
* Heiko J Schick <schickhj@de.ibm.com>
* Hoang-Nam Nguyen <hnguyen@de.ibm.com>
*
*
* Copyright (c) 2005 IBM Corporation
*
* All rights reserved.
*
* This source code is distributed under a dual license of GPL v2.0 and OpenIB
* BSD.
*
* OpenIB BSD License
*
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER
* IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#include "ehca_classes.h"
#include "ehca_irq.h"
#include "ehca_iverbs.h"
#include "ehca_qes.h"
#include "hcp_if.h"
#include "ipz_pt_fn.h"
int ehca_create_eq(struct ehca_shca *shca,
struct ehca_eq *eq,
const enum ehca_eq_type type, const u32 length)
{
u64 ret;
u32 nr_pages;
u32 i;
void *vpage;
struct ib_device *ib_dev = &shca->ib_device;
spin_lock_init(&eq->spinlock);
spin_lock_init(&eq->irq_spinlock);
eq->is_initialized = 0;
if (type != EHCA_EQ && type != EHCA_NEQ) {
ehca_err(ib_dev, "Invalid EQ type %x. eq=%p", type, eq);
return -EINVAL;
}
if (!length) {
ehca_err(ib_dev, "EQ length must not be zero. eq=%p", eq);
return -EINVAL;
}
ret = hipz_h_alloc_resource_eq(shca->ipz_hca_handle,
&eq->pf,
type,
length,
&eq->ipz_eq_handle,
&eq->length,
&nr_pages, &eq->ist);
if (ret != H_SUCCESS) {
ehca_err(ib_dev, "Can't allocate EQ/NEQ. eq=%p", eq);
return -EINVAL;
}
ret = ipz_queue_ctor(NULL, &eq->ipz_queue, nr_pages,
EHCA_PAGESIZE, sizeof(struct ehca_eqe), 0, 0);
if (!ret) {
ehca_err(ib_dev, "Can't allocate EQ pages eq=%p", eq);
goto create_eq_exit1;
}
for (i = 0; i < nr_pages; i++) {
u64 rpage;
vpage = ipz_qpageit_get_inc(&eq->ipz_queue);
if (!vpage) {
ret = H_RESOURCE;
goto create_eq_exit2;
}
rpage = virt_to_abs(vpage);
ret = hipz_h_register_rpage_eq(shca->ipz_hca_handle,
eq->ipz_eq_handle,
&eq->pf,
0, 0, rpage, 1);
if (i == (nr_pages - 1)) {
/* last page */
vpage = ipz_qpageit_get_inc(&eq->ipz_queue);
if (ret != H_SUCCESS || vpage)
goto create_eq_exit2;
} else {
if (ret != H_PAGE_REGISTERED || !vpage)
goto create_eq_exit2;
}
}
ipz_qeit_reset(&eq->ipz_queue);
/* register interrupt handlers and initialize work queues */
if (type == EHCA_EQ) {
ret = ibmebus_request_irq(NULL, eq->ist, ehca_interrupt_eq,
IRQF_DISABLED, "ehca_eq",
(void *)shca);
if (ret < 0)
ehca_err(ib_dev, "Can't map interrupt handler.");
tasklet_init(&eq->interrupt_task, ehca_tasklet_eq, (long)shca);
} else if (type == EHCA_NEQ) {
ret = ibmebus_request_irq(NULL, eq->ist, ehca_interrupt_neq,
IRQF_DISABLED, "ehca_neq",
(void *)shca);
if (ret < 0)
ehca_err(ib_dev, "Can't map interrupt handler.");
tasklet_init(&eq->interrupt_task, ehca_tasklet_neq, (long)shca);
}
eq->is_initialized = 1;
return 0;
create_eq_exit2:
ipz_queue_dtor(NULL, &eq->ipz_queue);
create_eq_exit1:
hipz_h_destroy_eq(shca->ipz_hca_handle, eq);
return -EINVAL;
}
void *ehca_poll_eq(struct ehca_shca *shca, struct ehca_eq *eq)
{
unsigned long flags;
void *eqe;
spin_lock_irqsave(&eq->spinlock, flags);
eqe = ipz_eqit_eq_get_inc_valid(&eq->ipz_queue);
spin_unlock_irqrestore(&eq->spinlock, flags);
return eqe;
}
int ehca_destroy_eq(struct ehca_shca *shca, struct ehca_eq *eq)
{
unsigned long flags;
u64 h_ret;
spin_lock_irqsave(&eq->spinlock, flags);
ibmebus_free_irq(NULL, eq->ist, (void *)shca);
h_ret = hipz_h_destroy_eq(shca->ipz_hca_handle, eq);
spin_unlock_irqrestore(&eq->spinlock, flags);
if (h_ret != H_SUCCESS) {
ehca_err(&shca->ib_device, "Can't free EQ resources.");
return -EINVAL;
}
ipz_queue_dtor(NULL, &eq->ipz_queue);
return 0;
}