Spdk/test/unit/lib/idxd/idxd.c/idxd_ut.c

313 lines
8.7 KiB
C
Raw Normal View History

/*-
* BSD LICENSE
*
* Copyright (c) Intel Corporation.
* All rights reserved.
*
* 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.
* * Neither the name of Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* 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 "spdk_cunit.h"
#include "spdk_internal/mock.h"
#include "spdk_internal/idxd.h"
#include "common/lib/test_env.c"
#include "idxd/idxd.h"
#define FAKE_REG_SIZE 0x800
#define GRP_CFG_OFFSET 0x400
#define MAX_TOKENS 0x40
#define MAX_ARRAY_SIZE 0x20
DEFINE_STUB(spdk_pci_idxd_get_driver, struct spdk_pci_driver *, (void), NULL);
int
spdk_pci_enumerate(struct spdk_pci_driver *driver, spdk_pci_enum_cb enum_cb, void *enum_ctx)
{
return -1;
}
int
spdk_pci_device_map_bar(struct spdk_pci_device *dev, uint32_t bar,
void **mapped_addr, uint64_t *phys_addr, uint64_t *size)
{
*mapped_addr = NULL;
*phys_addr = 0;
*size = 0;
return 0;
}
int
spdk_pci_device_unmap_bar(struct spdk_pci_device *dev, uint32_t bar, void *addr)
{
return 0;
}
int
spdk_pci_device_cfg_read32(struct spdk_pci_device *dev, uint32_t *value,
uint32_t offset)
{
*value = 0xFFFFFFFFu;
return 0;
}
int
spdk_pci_device_cfg_write32(struct spdk_pci_device *dev, uint32_t value,
uint32_t offset)
{
return 0;
}
#define movdir64b mock_movdir64b
static inline void
mock_movdir64b(void *dst, const void *src)
{
return;
}
#include "idxd/idxd.c"
#define WQ_CFG_OFFSET 0x500
#define TOTAL_WQE_SIZE 0x40
static int
test_idxd_wq_config(void)
{
struct spdk_idxd_device idxd = {};
union idxd_wqcfg wqcfg = {};
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
uint32_t expected[8] = {0x40, 0, 0x11, 0x9e, 0, 0, 0x40000000, 0};
uint32_t wq_size;
int rc, i, j;
idxd.reg_base = calloc(1, FAKE_REG_SIZE);
SPDK_CU_ASSERT_FATAL(idxd.reg_base != NULL);
SPDK_CU_ASSERT_FATAL(g_dev_cfg->num_groups <= MAX_ARRAY_SIZE);
idxd.groups = calloc(g_dev_cfg->num_groups, sizeof(struct idxd_group));
SPDK_CU_ASSERT_FATAL(idxd.groups != NULL);
idxd.registers.wqcap.total_wq_size = TOTAL_WQE_SIZE;
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
idxd.registers.wqcap.num_wqs = g_dev_cfg->total_wqs;
idxd.registers.gencap.max_batch_shift = LOG2_WQ_MAX_BATCH;
idxd.registers.gencap.max_xfer_shift = LOG2_WQ_MAX_XFER;
idxd.wqcfg_offset = WQ_CFG_OFFSET;
wq_size = idxd.registers.wqcap.total_wq_size / g_dev_cfg->total_wqs;
rc = idxd_wq_config(&idxd);
CU_ASSERT(rc == 0);
for (i = 0; i < g_dev_cfg->total_wqs; i++) {
CU_ASSERT(idxd.queues[i].wqcfg.wq_size == wq_size);
CU_ASSERT(idxd.queues[i].wqcfg.mode == WQ_MODE_DEDICATED);
CU_ASSERT(idxd.queues[i].wqcfg.max_batch_shift == LOG2_WQ_MAX_BATCH);
CU_ASSERT(idxd.queues[i].wqcfg.max_xfer_shift == LOG2_WQ_MAX_XFER);
CU_ASSERT(idxd.queues[i].wqcfg.wq_state == WQ_ENABLED);
CU_ASSERT(idxd.queues[i].wqcfg.priority == WQ_PRIORITY_1);
CU_ASSERT(idxd.queues[i].idxd == &idxd);
CU_ASSERT(idxd.queues[i].group == &idxd.groups[i % g_dev_cfg->num_groups]);
}
for (i = 0 ; i < idxd.registers.wqcap.num_wqs; i++) {
for (j = 0 ; j < WQCFG_NUM_DWORDS; j++) {
wqcfg.raw[j] = spdk_mmio_read_4((uint32_t *)(idxd.reg_base + idxd.wqcfg_offset + i * 32 + j *
4));
CU_ASSERT(wqcfg.raw[j] == expected[j]);
}
}
free(idxd.queues);
free(idxd.reg_base);
free(idxd.groups);
return 0;
}
static int
test_idxd_group_config(void)
{
struct spdk_idxd_device idxd = {};
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
uint64_t wqs[MAX_ARRAY_SIZE] = {};
uint64_t engines[MAX_ARRAY_SIZE] = {};
union idxd_group_flags flags[MAX_ARRAY_SIZE] = {};
int rc, i;
uint64_t base_offset;
idxd.reg_base = calloc(1, FAKE_REG_SIZE);
SPDK_CU_ASSERT_FATAL(idxd.reg_base != NULL);
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
SPDK_CU_ASSERT_FATAL(g_dev_cfg->num_groups <= MAX_ARRAY_SIZE);
idxd.registers.groupcap.num_groups = g_dev_cfg->num_groups;
idxd.registers.enginecap.num_engines = g_dev_cfg->total_engines;
idxd.registers.wqcap.num_wqs = g_dev_cfg->total_wqs;
idxd.registers.groupcap.total_tokens = MAX_TOKENS;
idxd.grpcfg_offset = GRP_CFG_OFFSET;
rc = idxd_group_config(&idxd);
CU_ASSERT(rc == 0);
for (i = 0 ; i < idxd.registers.groupcap.num_groups; i++) {
base_offset = idxd.grpcfg_offset + i * 64;
wqs[i] = spdk_mmio_read_8((uint64_t *)(idxd.reg_base + base_offset));
engines[i] = spdk_mmio_read_8((uint64_t *)(idxd.reg_base + base_offset + CFG_ENGINE_OFFSET));
flags[i].raw = spdk_mmio_read_8((uint64_t *)(idxd.reg_base + base_offset + CFG_FLAG_OFFSET));
}
/* wqe and engine arrays are indexed by group id and are bitmaps of assigned elements. */
CU_ASSERT(wqs[0] == 0x1);
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
CU_ASSERT(engines[0] == 0xf);
CU_ASSERT(flags[0].tokens_allowed == MAX_TOKENS / g_dev_cfg->num_groups);
/* groups allocated by code under test. */
free(idxd.groups);
free(idxd.reg_base);
return 0;
}
static int
test_idxd_reset_dev(void)
{
struct spdk_idxd_device idxd = {};
union idxd_cmdsts_reg *fake_cmd_status_reg;
int rc;
idxd.reg_base = calloc(1, FAKE_REG_SIZE);
SPDK_CU_ASSERT_FATAL(idxd.reg_base != NULL);
fake_cmd_status_reg = idxd.reg_base + IDXD_CMDSTS_OFFSET;
/* Test happy path */
rc = idxd_reset_dev(&idxd);
CU_ASSERT(rc == 0);
/* Test error reported path */
fake_cmd_status_reg->err = 1;
rc = idxd_reset_dev(&idxd);
CU_ASSERT(rc == -EINVAL);
free(idxd.reg_base);
return 0;
}
static int
test_idxd_wait_cmd(void)
{
struct spdk_idxd_device idxd = {};
int timeout = 1;
union idxd_cmdsts_reg *fake_cmd_status_reg;
int rc;
idxd.reg_base = calloc(1, FAKE_REG_SIZE);
SPDK_CU_ASSERT_FATAL(idxd.reg_base != NULL);
fake_cmd_status_reg = idxd.reg_base + IDXD_CMDSTS_OFFSET;
/* Test happy path. */
rc = idxd_wait_cmd(&idxd, timeout);
CU_ASSERT(rc == 0);
/* Setup up our fake register to set the error bit. */
fake_cmd_status_reg->err = 1;
rc = idxd_wait_cmd(&idxd, timeout);
CU_ASSERT(rc == -EINVAL);
fake_cmd_status_reg->err = 0;
/* Setup up our fake register to set the active bit. */
fake_cmd_status_reg->active = 1;
rc = idxd_wait_cmd(&idxd, timeout);
CU_ASSERT(rc == -EBUSY);
free(idxd.reg_base);
return 0;
}
static int
test_spdk_idxd_set_config(void)
{
g_dev_cfg = NULL;
spdk_idxd_set_config(0);
SPDK_CU_ASSERT_FATAL(g_dev_cfg != NULL);
CU_ASSERT(memcmp(&g_dev_cfg0, g_dev_cfg, sizeof(struct device_config)) == 0);
return 0;
}
static int
test_spdk_idxd_reconfigure_chan(void)
{
struct spdk_idxd_io_channel chan = {};
struct spdk_idxd_device idxd = {};
int rc;
uint32_t test_ring_size = 8;
uint32_t num_channels = 2;
lib/idxd: refactor batching for increased performance And to eliminate an artificial constraint on # of user descriptors. The main idea here was to move from a single ring that covered all user descriptors to a pre-allocated ring per pre-allocated batch. In addition, the other major change here is in how we poll for completions. We used to poll the batch rings then the main ring. Now when commands are prepared their completion address is added to a per channel list and the poller simply runs through that list not caring which ring the completion address belongs too. This simplifies the completion logic considerably and will avoid polling locations that can't potentially have a completion. Some minor rework was included as well, mainly getting rid of the ring_ctrl struct as it didn't serve much of a purpose anyway and with how things are setup now its easier to read with all the elements in the channel struct. Also, a change that came in while this was WIP needed a few fixes to function correctly. Addressed those and moved them to a helper function so we have one point of control for xlations. Added support for NOP in cases where a batch is submitted with only 1 descriptor. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie201b28118823100e908e0d1b08e7c10bb8fa9e7 Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3654 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2020-08-03 15:54:55 +00:00
chan.ring_slots = spdk_bit_array_create(test_ring_size);
chan.ring_size = test_ring_size;
chan.completions = spdk_zmalloc(test_ring_size * sizeof(struct idxd_hw_desc), 0, NULL,
SPDK_ENV_LCORE_ID_ANY, SPDK_MALLOC_DMA);
SPDK_CU_ASSERT_FATAL(chan.completions != NULL);
chan.idxd = &idxd;
chan.idxd->num_channels = num_channels;
rc = spdk_idxd_reconfigure_chan(&chan);
CU_ASSERT(rc == 0);
lib/idxd: refactor batching for increased performance And to eliminate an artificial constraint on # of user descriptors. The main idea here was to move from a single ring that covered all user descriptors to a pre-allocated ring per pre-allocated batch. In addition, the other major change here is in how we poll for completions. We used to poll the batch rings then the main ring. Now when commands are prepared their completion address is added to a per channel list and the poller simply runs through that list not caring which ring the completion address belongs too. This simplifies the completion logic considerably and will avoid polling locations that can't potentially have a completion. Some minor rework was included as well, mainly getting rid of the ring_ctrl struct as it didn't serve much of a purpose anyway and with how things are setup now its easier to read with all the elements in the channel struct. Also, a change that came in while this was WIP needed a few fixes to function correctly. Addressed those and moved them to a helper function so we have one point of control for xlations. Added support for NOP in cases where a batch is submitted with only 1 descriptor. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie201b28118823100e908e0d1b08e7c10bb8fa9e7 Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3654 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2020-08-03 15:54:55 +00:00
CU_ASSERT(chan.max_ring_slots == test_ring_size / num_channels);
lib/idxd: refactor batching for increased performance And to eliminate an artificial constraint on # of user descriptors. The main idea here was to move from a single ring that covered all user descriptors to a pre-allocated ring per pre-allocated batch. In addition, the other major change here is in how we poll for completions. We used to poll the batch rings then the main ring. Now when commands are prepared their completion address is added to a per channel list and the poller simply runs through that list not caring which ring the completion address belongs too. This simplifies the completion logic considerably and will avoid polling locations that can't potentially have a completion. Some minor rework was included as well, mainly getting rid of the ring_ctrl struct as it didn't serve much of a purpose anyway and with how things are setup now its easier to read with all the elements in the channel struct. Also, a change that came in while this was WIP needed a few fixes to function correctly. Addressed those and moved them to a helper function so we have one point of control for xlations. Added support for NOP in cases where a batch is submitted with only 1 descriptor. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie201b28118823100e908e0d1b08e7c10bb8fa9e7 Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3654 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2020-08-03 15:54:55 +00:00
spdk_bit_array_free(&chan.ring_slots);
spdk_free(chan.completions);
return 0;
}
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
static int
test_setup(void)
{
g_dev_cfg = &g_dev_cfg0;
return 0;
}
int main(int argc, char **argv)
{
CU_pSuite suite = NULL;
unsigned int num_failures;
CU_set_error_action(CUEA_ABORT);
CU_initialize_registry();
lib/idxd: change config #0 to something more sane Config #1 remains what is shown as an example in the spec. Change config #0 to just have 1 work group and 1 work queue all backed by 4 engines. As the majority of initial use cases will not be implementing separate priorities and/or different back end targets (mem, pmem, etc) having just 1 group and work queue makes the most sense as it allows the silicon to decide which engine to use. Also, having multiple work queues spreads out the available entires such that if we're not using all of the work queues then we're not using all of the resources. As channels are created they are assigned the next available device. As a channel is assigned a device that is already in use it will round robin work queues. If then, for example, we have 16 devices then only the first work queue will ever be used for the first 16 threads which seems and if there are even just 2 work queues per device it would take 32 threads to use all of the resources at the device. By haing just one work queue per device we always have the max number of work queue entries available regardless of how many threads are being used. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: Ie15ff6bdea12525fe3bfc769613084ddd2de50bf Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5845 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot Reviewed-by: Ziye Yang <ziye.yang@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2021-01-09 21:29:34 +00:00
suite = CU_add_suite("idxd", test_setup, NULL);
CU_ADD_TEST(suite, test_spdk_idxd_reconfigure_chan);
CU_ADD_TEST(suite, test_spdk_idxd_set_config);
CU_ADD_TEST(suite, test_idxd_wait_cmd);
CU_ADD_TEST(suite, test_idxd_reset_dev);
CU_ADD_TEST(suite, test_idxd_group_config);
CU_ADD_TEST(suite, test_idxd_wq_config);
CU_basic_set_mode(CU_BRM_VERBOSE);
CU_basic_run_tests();
num_failures = CU_get_number_of_failures();
CU_cleanup_registry();
return num_failures;
}