[dpdk-dev] [PATCH v1 0/8] use compiler atomic builtins for examples
Since atomic operations have been adopted in DPDK now[1], change rte_atomicNN_xxx APIs to compiler's atomic built-ins in examples module[2]. [1] https://www.dpdk.org/blog/2021/03/26/dpdk-adopts-the-c11-memory-model/ [2] https://doc.dpdk.org/guides/rel_notes/deprecation.html Joyce Kong (8): examples/bbdev_app: use compiler atomics for flag sync examples/multi_process: use compiler atomics for sync examples/kni: use compiler atomics for status sync examples/performance-thread: use compiler atomics for sync examples/l2fwd-jobstats: use compiler atomics for stats sync examples/vm_power_manager: use compiler atomics for sync examples/server_node_efd: use compiler atomics for sync examples: remove unnecessary include of atomic examples/bbdev_app/main.c | 13 --- examples/bond/main.c | 1 - examples/ip_fragmentation/main.c | 1 - examples/ip_reassembly/main.c | 1 - examples/ipsec-secgw/ipsec-secgw.c| 1 - examples/ipv4_multicast/main.c| 1 - examples/kni/main.c | 27 +++ examples/l2fwd-crypto/main.c | 1 - examples/l2fwd-event/l2fwd_common.h | 1 - examples/l2fwd-event/l2fwd_event.c| 1 - examples/l2fwd-jobstats/main.c| 11 +++--- examples/l2fwd-keepalive/main.c | 1 - examples/l2fwd/main.c | 1 - examples/l3fwd-acl/main.c | 1 - examples/l3fwd-power/main.c | 1 - examples/l3fwd/main.c | 1 - examples/link_status_interrupt/main.c | 1 - .../client_server_mp/mp_client/client.c | 1 - .../client_server_mp/mp_server/init.c | 1 - .../client_server_mp/mp_server/main.c | 7 ++-- examples/multi_process/simple_mp/main.c | 1 - .../multi_process/simple_mp/mp_commands.c | 1 - examples/multi_process/symmetric_mp/main.c| 1 - examples/performance-thread/common/lthread.c | 10 +++--- .../performance-thread/common/lthread_diag.h | 10 +++--- .../performance-thread/common/lthread_int.h | 1 - .../performance-thread/common/lthread_mutex.c | 26 +++--- .../performance-thread/common/lthread_mutex.h | 2 +- .../performance-thread/common/lthread_sched.c | 34 --- .../performance-thread/common/lthread_tls.c | 5 +-- .../performance-thread/l3fwd-thread/main.c| 22 +--- examples/server_node_efd/node/node.c | 1 - examples/server_node_efd/server/init.c| 1 - examples/server_node_efd/server/main.c| 7 ++-- examples/vhost_blk/blk.c | 1 - examples/vhost_blk/vhost_blk.c| 1 - examples/vm_power_manager/channel_manager.c | 1 - examples/vm_power_manager/channel_manager.h | 1 - examples/vm_power_manager/channel_monitor.c | 11 +++--- examples/vmdq/main.c | 1 - examples/vmdq_dcb/main.c | 1 - 41 files changed, 91 insertions(+), 122 deletions(-) -- 2.17.1
[dpdk-dev] [PATCH v1 1/8] examples/bbdev_app: use compiler atomics for flag sync
Convert rte_atomic usages to compiler atomic built-ins
for global_exit_flag sync.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/bbdev_app/main.c | 13 ++---
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/examples/bbdev_app/main.c b/examples/bbdev_app/main.c
index 5251db0b16..75c620ea75 100644
--- a/examples/bbdev_app/main.c
+++ b/examples/bbdev_app/main.c
@@ -18,7 +18,6 @@
#include
#include
-#include
#include
#include
#include
@@ -167,7 +166,7 @@ static const struct app_config_params def_app_config = {
.num_dec_cores = 1,
};
-static rte_atomic16_t global_exit_flag;
+static uint16_t global_exit_flag;
/* display usage */
static inline void
@@ -279,7 +278,7 @@ static void
signal_handler(int signum)
{
printf("\nSignal %d received\n", signum);
- rte_atomic16_set(&global_exit_flag, 1);
+ __atomic_store_n(&global_exit_flag, 1, __ATOMIC_RELAXED);
}
static void
@@ -328,7 +327,7 @@ check_port_link_status(uint16_t port_id)
fflush(stdout);
for (count = 0; count <= MAX_CHECK_TIME &&
- !rte_atomic16_read(&global_exit_flag); count++) {
+ !__atomic_load_n(&global_exit_flag, __ATOMIC_RELAXED);
count++) {
memset(&link, 0, sizeof(link));
link_get_err = rte_eth_link_get_nowait(port_id, &link);
@@ -682,7 +681,7 @@ stats_loop(void *arg)
{
struct stats_lcore_params *stats_lcore = arg;
- while (!rte_atomic16_read(&global_exit_flag)) {
+ while (!__atomic_load_n(&global_exit_flag, __ATOMIC_RELAXED)) {
print_stats(stats_lcore);
rte_delay_ms(500);
}
@@ -928,7 +927,7 @@ processing_loop(void *arg)
const bool run_decoder = (lcore_conf->core_type &
(1 << RTE_BBDEV_OP_TURBO_DEC));
- while (!rte_atomic16_read(&global_exit_flag)) {
+ while (!__atomic_load_n(&global_exit_flag, __ATOMIC_RELAXED)) {
if (run_encoder)
run_encoding(lcore_conf);
if (run_decoder)
@@ -1062,7 +1061,7 @@ main(int argc, char **argv)
.align = __alignof__(struct rte_mbuf *),
};
- rte_atomic16_init(&global_exit_flag);
+ __atomic_store_n(&global_exit_flag, 0, __ATOMIC_RELAXED);
sigret = signal(SIGTERM, signal_handler);
if (sigret == SIG_ERR)
--
2.17.1
Re: [dpdk-dev] [PATCH] examples/performance-thread: fix build issue with clang 12.0.1
> -Original Message-
> From: jer...@marvell.com
> Sent: Monday, August 16, 2021 4:19 PM
> To: John McNamara ; Ian Betts
> ; Tomasz Kulasek
> Cc: dev@dpdk.org; NBU-Contact-Thomas Monjalon
> ; bruce.richard...@intel.com;
> david.march...@redhat.com; Jerin Jacob ; Ali
> Alnubani
> Subject: [PATCH] examples/performance-thread: fix build issue with clang
> 12.0.1
>
> From: Jerin Jacob
>
> In clang 12.0.1 version, the use of pthread_yield() deprecated, use
> sched_yield() instead.
>
> log:
>
> Compiling C object
> examples/dpdk-pthread_shim.p/performance-
> thread_pthread_shim_main.c.o
> ../examples/performance-thread/pthread_shim/main.c: In function
> 'helloworld_pthread':
> ../examples/performance-thread/pthread_shim/main.c:75:9: warning:
> 'pthread_yield' is deprecated: pthread_yield is deprecated, use sched_yield
> instead [-Wdeprecated-declarations]
>
> Fixes: 433ba6228f9a ("examples/performance-thread: add pthread_shim
> app")
>
> Bugzilla ID: 745
>
> Cc: alia...@oss.nvidia.com
>
> Signed-off-by: Jerin Jacob
> ---
I think this issue might be specific to Fedora Rawhide's build of clang 12.0.1,
see my comment in the ticket:
https://bugs.dpdk.org/show_bug.cgi?id=745#c3
Tested-by: Ali Alnubani
Thanks Jerin.
Ali
[dpdk-dev] [PATCH v1 2/8] examples/multi_process: use compiler atomics for sync
Convert rte_atomic32_test_and_set usage to compiler atomic
CAS operation for display_stats sync.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/multi_process/client_server_mp/mp_server/main.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/examples/multi_process/client_server_mp/mp_server/main.c
b/examples/multi_process/client_server_mp/mp_server/main.c
index 9bcee460fd..b4761ebc7b 100644
--- a/examples/multi_process/client_server_mp/mp_server/main.c
+++ b/examples/multi_process/client_server_mp/mp_server/main.c
@@ -20,7 +20,6 @@
#include
#include
#include
-#include
#include
#include
#include
@@ -158,10 +157,12 @@ static int
sleep_lcore(__rte_unused void *dummy)
{
/* Used to pick a display thread - static, so zero-initialised */
- static rte_atomic32_t display_stats;
+ static uint32_t display_stats;
+ uint32_t status = 0;
/* Only one core should display stats */
- if (rte_atomic32_test_and_set(&display_stats)) {
+ if (__atomic_compare_exchange_n(&display_stats, &status, 1, 0,
+ __ATOMIC_RELAXED, __ATOMIC_RELAXED)) {
const unsigned sleeptime = 1;
printf("Core %u displaying statistics\n", rte_lcore_id());
--
2.17.1
[dpdk-dev] [PATCH v1 3/8] examples/kni: use compiler atomics for status sync
Convert rte_atomic usages to compiler atomic builit-ins
for kni_stop and kni_pause sync.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/kni/main.c | 27 +--
1 file changed, 13 insertions(+), 14 deletions(-)
diff --git a/examples/kni/main.c b/examples/kni/main.c
index beabb3c848..ad1f569e18 100644
--- a/examples/kni/main.c
+++ b/examples/kni/main.c
@@ -27,7 +27,6 @@
#include
#include
#include
-#include
#include
#include
#include
@@ -131,8 +130,8 @@ static int kni_change_mtu(uint16_t port_id, unsigned int
new_mtu);
static int kni_config_network_interface(uint16_t port_id, uint8_t if_up);
static int kni_config_mac_address(uint16_t port_id, uint8_t mac_addr[]);
-static rte_atomic32_t kni_stop = RTE_ATOMIC32_INIT(0);
-static rte_atomic32_t kni_pause = RTE_ATOMIC32_INIT(0);
+static uint32_t kni_stop;
+static uint32_t kni_pause;
/* Print out statistics on packets handled */
static void
@@ -185,7 +184,7 @@ signal_handler(int signum)
if (signum == SIGRTMIN || signum == SIGINT || signum == SIGTERM) {
printf("\nSIGRTMIN/SIGINT/SIGTERM received. "
"KNI processing stopping.\n");
- rte_atomic32_inc(&kni_stop);
+ __atomic_fetch_add(&kni_stop, 1, __ATOMIC_RELAXED);
return;
}
}
@@ -311,8 +310,8 @@ main_loop(__rte_unused void *arg)
kni_port_params_array[i]->lcore_rx,
kni_port_params_array[i]->port_id);
while (1) {
- f_stop = rte_atomic32_read(&kni_stop);
- f_pause = rte_atomic32_read(&kni_pause);
+ f_stop = __atomic_load_n(&kni_stop, __ATOMIC_RELAXED);
+ f_pause = __atomic_load_n(&kni_pause, __ATOMIC_RELAXED);
if (f_stop)
break;
if (f_pause)
@@ -324,8 +323,8 @@ main_loop(__rte_unused void *arg)
kni_port_params_array[i]->lcore_tx,
kni_port_params_array[i]->port_id);
while (1) {
- f_stop = rte_atomic32_read(&kni_stop);
- f_pause = rte_atomic32_read(&kni_pause);
+ f_stop = __atomic_load_n(&kni_stop, __ATOMIC_RELAXED);
+ f_pause = __atomic_load_n(&kni_pause, __ATOMIC_RELAXED);
if (f_stop)
break;
if (f_pause)
@@ -856,9 +855,9 @@ kni_change_mtu(uint16_t port_id, unsigned int new_mtu)
{
int ret;
- rte_atomic32_inc(&kni_pause);
+ __atomic_fetch_add(&kni_pause, 1, __ATOMIC_RELAXED);
ret = kni_change_mtu_(port_id, new_mtu);
- rte_atomic32_dec(&kni_pause);
+ __atomic_fetch_sub(&kni_pause, 1, __ATOMIC_RELAXED);
return ret;
}
@@ -877,14 +876,14 @@ kni_config_network_interface(uint16_t port_id, uint8_t
if_up)
RTE_LOG(INFO, APP, "Configure network interface of %d %s\n",
port_id, if_up ? "up" : "down");
- rte_atomic32_inc(&kni_pause);
+ __atomic_fetch_add(&kni_pause, 1, __ATOMIC_RELAXED);
if (if_up != 0) { /* Configure network interface up */
ret = rte_eth_dev_stop(port_id);
if (ret != 0) {
RTE_LOG(ERR, APP, "Failed to stop port %d: %s\n",
port_id, rte_strerror(-ret));
- rte_atomic32_dec(&kni_pause);
+ __atomic_fetch_sub(&kni_pause, 1, __ATOMIC_RELAXED);
return ret;
}
ret = rte_eth_dev_start(port_id);
@@ -893,12 +892,12 @@ kni_config_network_interface(uint16_t port_id, uint8_t
if_up)
if (ret != 0) {
RTE_LOG(ERR, APP, "Failed to stop port %d: %s\n",
port_id, rte_strerror(-ret));
- rte_atomic32_dec(&kni_pause);
+ __atomic_fetch_sub(&kni_pause, 1, __ATOMIC_RELAXED);
return ret;
}
}
- rte_atomic32_dec(&kni_pause);
+ __atomic_fetch_sub(&kni_pause, 1, __ATOMIC_RELAXED);
if (ret < 0)
RTE_LOG(ERR, APP, "Failed to start port %d\n", port_id);
--
2.17.1
[dpdk-dev] [PATCH v1 4/8] examples/performance-thread: use compiler atomics for sync
Convert rte_atomic usages to compiler atomic built-ins
for thread sync.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/performance-thread/common/lthread.c | 10 +++---
.../performance-thread/common/lthread_diag.h | 10 +++---
.../performance-thread/common/lthread_int.h | 1 -
.../performance-thread/common/lthread_mutex.c | 26 +++---
.../performance-thread/common/lthread_mutex.h | 2 +-
.../performance-thread/common/lthread_sched.c | 34 ---
.../performance-thread/common/lthread_tls.c | 5 +--
.../performance-thread/l3fwd-thread/main.c| 22 +---
8 files changed, 53 insertions(+), 57 deletions(-)
diff --git a/examples/performance-thread/common/lthread.c
b/examples/performance-thread/common/lthread.c
index 3f1f48db43..98123f34f8 100644
--- a/examples/performance-thread/common/lthread.c
+++ b/examples/performance-thread/common/lthread.c
@@ -357,9 +357,10 @@ void lthread_exit(void *ptr)
* - if exit before join then we suspend and resume on join
* - if join before exit then we resume the joining thread
*/
+ uint64_t join_initial = LT_JOIN_INITIAL;
if ((lt->join == LT_JOIN_INITIAL)
- && rte_atomic64_cmpset(<->join, LT_JOIN_INITIAL,
- LT_JOIN_EXITING)) {
+ && __atomic_compare_exchange_n(<->join, &join_initial,
+ LT_JOIN_EXITING, 0, __ATOMIC_RELAXED, __ATOMIC_RELAXED)) {
DIAG_EVENT(lt, LT_DIAG_LTHREAD_EXIT, 1, 0);
_suspend();
@@ -415,9 +416,10 @@ int lthread_join(struct lthread *lt, void **ptr)
* - if join before exit we suspend and will resume when exit is called
* - if exit before join we resume the exiting thread
*/
+ uint64_t join_initial = LT_JOIN_INITIAL;
if ((lt->join == LT_JOIN_INITIAL)
- && rte_atomic64_cmpset(<->join, LT_JOIN_INITIAL,
- LT_JOIN_THREAD_SET)) {
+ && __atomic_compare_exchange_n(<->join, &join_initial,
+ LT_JOIN_THREAD_SET, 0, __ATOMIC_RELAXED, __ATOMIC_RELAXED)) {
DIAG_EVENT(current, LT_DIAG_LTHREAD_JOIN, lt, 1);
_suspend();
diff --git a/examples/performance-thread/common/lthread_diag.h
b/examples/performance-thread/common/lthread_diag.h
index e876dda6da..7ee89eef38 100644
--- a/examples/performance-thread/common/lthread_diag.h
+++ b/examples/performance-thread/common/lthread_diag.h
@@ -78,11 +78,11 @@ extern uint64_t diag_mask;
} \
} while (0)
-#define DIAG_COUNT_DEFINE(x) rte_atomic64_t count_##x
-#define DIAG_COUNT_INIT(o, x) rte_atomic64_init(&((o)->count_##x))
-#define DIAG_COUNT_INC(o, x) rte_atomic64_inc(&((o)->count_##x))
-#define DIAG_COUNT_DEC(o, x) rte_atomic64_dec(&((o)->count_##x))
-#define DIAG_COUNT(o, x) rte_atomic64_read(&((o)->count_##x))
+#define DIAG_COUNT_DEFINE(x) uint64_t count_##x
+#define DIAG_COUNT_INIT(o, x) __atomic_store_n(&((o)->count_##x), 0,
__ATOMIC_RELAXED)
+#define DIAG_COUNT_INC(o, x) __atomic_fetch_add(&((o)->count_##x), 1,
__ATOMIC_RELAXED)
+#define DIAG_COUNT_DEC(o, x) __atomic_fetch_sub(&((o)->count_##x), 1,
__ATOMIC_RELAXED)
+#define DIAG_COUNT(o, x) __atomic_load_n(&((o)->count_##x), __ATOMIC_RELAXED)
#define DIAG_USED
diff --git a/examples/performance-thread/common/lthread_int.h
b/examples/performance-thread/common/lthread_int.h
index a352f13b75..d010126f16 100644
--- a/examples/performance-thread/common/lthread_int.h
+++ b/examples/performance-thread/common/lthread_int.h
@@ -21,7 +21,6 @@ extern "C" {
#include
#include
#include
-#include
#include
#include
diff --git a/examples/performance-thread/common/lthread_mutex.c
b/examples/performance-thread/common/lthread_mutex.c
index 01da6cad4f..43cc9bbfb9 100644
--- a/examples/performance-thread/common/lthread_mutex.c
+++ b/examples/performance-thread/common/lthread_mutex.c
@@ -60,7 +60,7 @@ lthread_mutex_init(char *name, struct lthread_mutex **mutex,
m->root_sched = THIS_SCHED;
m->owner = NULL;
- rte_atomic64_init(&m->count);
+ __atomic_store_n(&m->count, 0, __ATOMIC_RELAXED);
DIAG_CREATE_EVENT(m, LT_DIAG_MUTEX_CREATE);
/* success */
@@ -115,10 +115,11 @@ int lthread_mutex_lock(struct lthread_mutex *m)
}
for (;;) {
- rte_atomic64_inc(&m->count);
+ __atomic_fetch_add(&m->count, 1, __ATOMIC_RELAXED);
do {
- if (rte_atomic64_cmpset
- ((uint64_t *) &m->owner, 0, (uint64_t) lt)) {
+ uint64_t lt_init = 0;
+ if (__atomic_compare_exchange_n(&m->owner, <_init, lt,
+ 0, __ATOMIC_RELAXED, __ATOMIC_RELAXED))
{
/* happy days, we got the lock */
DIAG_EVENT(m, LT_DIAG_MUTEX
[dpdk-dev] [PATCH v1 5/8] examples/l2fwd-jobstats: use compiler atomics for stats sync
Convert rte_atomic usages to compiler atomic built-ins
for stats_read_pending sync in l2fwd_jobstats module.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/l2fwd-jobstats/main.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/examples/l2fwd-jobstats/main.c b/examples/l2fwd-jobstats/main.c
index bbb4a27a6d..99a17de181 100644
--- a/examples/l2fwd-jobstats/main.c
+++ b/examples/l2fwd-jobstats/main.c
@@ -16,7 +16,6 @@
#include
#include
#include
-#include
#include
#include
#include
@@ -81,7 +80,7 @@ struct lcore_queue_conf {
struct rte_jobstats idle_job;
struct rte_jobstats_context jobs_context;
- rte_atomic16_t stats_read_pending;
+ uint16_t stats_read_pending;
rte_spinlock_t lock;
} __rte_cache_aligned;
/* >8 End of list of queues to be polled for given lcore. */
@@ -155,9 +154,9 @@ show_lcore_stats(unsigned lcore_id)
uint64_t collection_time = rte_get_timer_cycles();
/* Ask forwarding thread to give us stats. */
- rte_atomic16_set(&qconf->stats_read_pending, 1);
+ __atomic_store_n(&qconf->stats_read_pending, 1, __ATOMIC_RELAXED);
rte_spinlock_lock(&qconf->lock);
- rte_atomic16_set(&qconf->stats_read_pending, 0);
+ __atomic_store_n(&qconf->stats_read_pending, 0, __ATOMIC_RELAXED);
/* Collect context statistics. */
stats_period = ctx->state_time - ctx->start_time;
@@ -526,8 +525,8 @@ l2fwd_main_loop(void)
repeats++;
need_manage = qconf->flush_timer.expire < now;
/* Check if we was esked to give a stats. */
- stats_read_pending =
-
rte_atomic16_read(&qconf->stats_read_pending);
+ stats_read_pending =
__atomic_load_n(&qconf->stats_read_pending,
+ __ATOMIC_RELAXED);
need_manage |= stats_read_pending;
for (i = 0; i < qconf->n_rx_port &&
!need_manage; i++)
--
2.17.1
[dpdk-dev] [PATCH v1 6/8] examples/vm_power_manager: use compiler atomics for sync
Convert rte_atomic32_cmpset to compiler atomic CAS
operation for channel status sync.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/vm_power_manager/channel_monitor.c | 11 ++-
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/examples/vm_power_manager/channel_monitor.c
b/examples/vm_power_manager/channel_monitor.c
index 99f81544d7..3c20406f7c 100644
--- a/examples/vm_power_manager/channel_monitor.c
+++ b/examples/vm_power_manager/channel_monitor.c
@@ -25,7 +25,6 @@
#include
#include
#include
-#include
#include
#include
#ifdef RTE_NET_I40E
@@ -829,8 +828,9 @@ process_request(struct rte_power_channel_packet *pkt,
if (chan_info == NULL)
return -1;
- if (rte_atomic32_cmpset(&(chan_info->status),
CHANNEL_MGR_CHANNEL_CONNECTED,
- CHANNEL_MGR_CHANNEL_PROCESSING) == 0)
+ uint32_t channel_connected = CHANNEL_MGR_CHANNEL_CONNECTED;
+ if (__atomic_compare_exchange_n(&(chan_info->status),
&channel_connected,
+ CHANNEL_MGR_CHANNEL_PROCESSING, 0, __ATOMIC_RELAXED,
__ATOMIC_RELAXED) == 0)
return -1;
if (pkt->command == RTE_POWER_CPU_POWER) {
@@ -934,8 +934,9 @@ process_request(struct rte_power_channel_packet *pkt,
* Return is not checked as channel status may have been set to DISABLED
* from management thread
*/
- rte_atomic32_cmpset(&(chan_info->status),
CHANNEL_MGR_CHANNEL_PROCESSING,
- CHANNEL_MGR_CHANNEL_CONNECTED);
+ uint32_t channel_processing = CHANNEL_MGR_CHANNEL_PROCESSING;
+ __atomic_compare_exchange_n(&(chan_info->status), &channel_processing,
+ CHANNEL_MGR_CHANNEL_CONNECTED, 0, __ATOMIC_RELAXED,
__ATOMIC_RELAXED);
return 0;
}
--
2.17.1
[dpdk-dev] [PATCH v1 7/8] examples/server_node_efd: use compiler atomics for sync
Convert rte_atomic32_test_and_set to compiler CAS atomic
operation for display_stats sync.
Signed-off-by: Joyce Kong
Reviewed-by: Ruifeng Wang
---
examples/server_node_efd/server/main.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/examples/server_node_efd/server/main.c
b/examples/server_node_efd/server/main.c
index 7d07131dbf..b69beb012c 100644
--- a/examples/server_node_efd/server/main.c
+++ b/examples/server_node_efd/server/main.c
@@ -22,7 +22,6 @@
#include
#include
#include
-#include
#include
#include
#include
@@ -180,10 +179,12 @@ static int
sleep_lcore(__rte_unused void *dummy)
{
/* Used to pick a display thread - static, so zero-initialised */
- static rte_atomic32_t display_stats;
+ static uint32_t display_stats;
/* Only one core should display stats */
- if (rte_atomic32_test_and_set(&display_stats)) {
+ uint32_t display_init = 0;
+ if (__atomic_compare_exchange_n(&display_stats, &display_init, 1, 0,
+ __ATOMIC_RELAXED, __ATOMIC_RELAXED)) {
const unsigned int sleeptime = 1;
printf("Core %u displaying statistics\n", rte_lcore_id());
--
2.17.1
[dpdk-dev] [PATCH v1 8/8] examples: remove unnecessary include of atomic
Remove the unnecessary header file rte_atomic.h included in example module. Signed-off-by: Joyce Kong Reviewed-by: Ruifeng Wang --- examples/bond/main.c | 1 - examples/ip_fragmentation/main.c | 1 - examples/ip_reassembly/main.c | 1 - examples/ipsec-secgw/ipsec-secgw.c | 1 - examples/ipv4_multicast/main.c | 1 - examples/l2fwd-crypto/main.c | 1 - examples/l2fwd-event/l2fwd_common.h| 1 - examples/l2fwd-event/l2fwd_event.c | 1 - examples/l2fwd-keepalive/main.c| 1 - examples/l2fwd/main.c | 1 - examples/l3fwd-acl/main.c | 1 - examples/l3fwd-power/main.c| 1 - examples/l3fwd/main.c | 1 - examples/link_status_interrupt/main.c | 1 - examples/multi_process/client_server_mp/mp_client/client.c | 1 - examples/multi_process/client_server_mp/mp_server/init.c | 1 - examples/multi_process/simple_mp/main.c| 1 - examples/multi_process/simple_mp/mp_commands.c | 1 - examples/multi_process/symmetric_mp/main.c | 1 - examples/server_node_efd/node/node.c | 1 - examples/server_node_efd/server/init.c | 1 - examples/vhost_blk/blk.c | 1 - examples/vhost_blk/vhost_blk.c | 1 - examples/vm_power_manager/channel_manager.c| 1 - examples/vm_power_manager/channel_manager.h| 1 - examples/vmdq/main.c | 1 - examples/vmdq_dcb/main.c | 1 - 27 files changed, 27 deletions(-) diff --git a/examples/bond/main.c b/examples/bond/main.c index f48400e211..de1e995a8a 100644 --- a/examples/bond/main.c +++ b/examples/bond/main.c @@ -24,7 +24,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/ip_fragmentation/main.c b/examples/ip_fragmentation/main.c index f245369720..f965624c8b 100644 --- a/examples/ip_fragmentation/main.c +++ b/examples/ip_fragmentation/main.c @@ -21,7 +21,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/ip_reassembly/main.c b/examples/ip_reassembly/main.c index 8645ac790b..6a8e91673d 100644 --- a/examples/ip_reassembly/main.c +++ b/examples/ip_reassembly/main.c @@ -22,7 +22,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/ipsec-secgw/ipsec-secgw.c b/examples/ipsec-secgw/ipsec-secgw.c index f252d34985..012b908410 100644 --- a/examples/ipsec-secgw/ipsec-secgw.c +++ b/examples/ipsec-secgw/ipsec-secgw.c @@ -24,7 +24,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/ipv4_multicast/main.c b/examples/ipv4_multicast/main.c index cc527d7f6b..5830ad61c1 100644 --- a/examples/ipv4_multicast/main.c +++ b/examples/ipv4_multicast/main.c @@ -20,7 +20,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/l2fwd-crypto/main.c b/examples/l2fwd-crypto/main.c index 5f539c458c..e302ad022c 100644 --- a/examples/l2fwd-crypto/main.c +++ b/examples/l2fwd-crypto/main.c @@ -20,7 +20,6 @@ #include #include -#include #include #include #include diff --git a/examples/l2fwd-event/l2fwd_common.h b/examples/l2fwd-event/l2fwd_common.h index 939221d45a..1a418e6a22 100644 --- a/examples/l2fwd-event/l2fwd_common.h +++ b/examples/l2fwd-event/l2fwd_common.h @@ -27,7 +27,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/l2fwd-event/l2fwd_event.c b/examples/l2fwd-event/l2fwd_event.c index 7ba5311d66..acfacfa4fb 100644 --- a/examples/l2fwd-event/l2fwd_event.c +++ b/examples/l2fwd-event/l2fwd_event.c @@ -5,7 +5,6 @@ #include #include -#include #include #include #include diff --git a/examples/l2fwd-keepalive/main.c b/examples/l2fwd-keepalive/main.c index 4e1a17cfe4..cb7f267259 100644 --- a/examples/l2fwd-keepalive/main.c +++ b/examples/l2fwd-keepalive/main.c @@ -24,7 +24,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/l2fwd/main.c b/examples/l2fwd/main.c index 911e40c66e..56a0663111 100644 --- a/examples/l2fwd/main.c +++ b/examples/l2fwd/main.c @@ -24,7 +24,6 @@ #include #include #include -#include #include #include #include diff --git a/examples/l3fwd-acl/main.c b/examples/l3fwd-acl/main.c index a1f457b564..2a942731bc 100644 --- a/examples/l3fwd-acl/main.c +++ b/examples/l3fwd-acl/main.c @@ -20,7 +20,6 @@ #include #include #include -#include #include #include #include diff
[dpdk-dev] [PATCH v4 1/2] drivers/raw: remove octeontx2-dma driver
Removing the rawdev based octeontx2-dma driver as the dependent common/octeontx2 will be soon be going away. Also a new DMA driver will be coming in this place once the rte_dmadev library is in. Signed-off-by: Radha Mohan Chintakuntla --- Changes from v3: Fixed patch application failure on main due to conflict. Changes from v2: Fixed DPDK CI reported issues for more documentation failure. Changes from v1: Fixed compilation issues in documentation MAINTAINERS | 6 - doc/guides/platform/octeontx2.rst | 3 - doc/guides/rawdevs/index.rst| 1 - doc/guides/rawdevs/octeontx2_dma.rst| 103 - drivers/raw/meson.build | 1 - drivers/raw/octeontx2_dma/meson.build | 18 - drivers/raw/octeontx2_dma/otx2_dpi_msg.c| 105 - drivers/raw/octeontx2_dma/otx2_dpi_rawdev.c | 441 drivers/raw/octeontx2_dma/otx2_dpi_rawdev.h | 197 - drivers/raw/octeontx2_dma/otx2_dpi_test.c | 218 -- drivers/raw/octeontx2_dma/version.map | 3 - 11 files changed, 1096 deletions(-) delete mode 100644 doc/guides/rawdevs/octeontx2_dma.rst delete mode 100644 drivers/raw/octeontx2_dma/meson.build delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_msg.c delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_rawdev.c delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_rawdev.h delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_test.c delete mode 100644 drivers/raw/octeontx2_dma/version.map diff --git a/MAINTAINERS b/MAINTAINERS index 266f5ac1da..9f9aa37c68 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1321,12 +1321,6 @@ M: Tomasz Duszynski F: doc/guides/rawdevs/cnxk_bphy.rst F: drivers/raw/cnxk_bphy/ -Marvell OCTEON TX2 DMA -M: Radha Mohan Chintakuntla -M: Veerasenareddy Burru -F: drivers/raw/octeontx2_dma/ -F: doc/guides/rawdevs/octeontx2_dma.rst - Marvell OCTEON TX2 EP M: Radha Mohan Chintakuntla M: Veerasenareddy Burru diff --git a/doc/guides/platform/octeontx2.rst b/doc/guides/platform/octeontx2.rst index 8b5991f03b..3a3d28571c 100644 --- a/doc/guides/platform/octeontx2.rst +++ b/doc/guides/platform/octeontx2.rst @@ -152,9 +152,6 @@ This section lists dataplane H/W block(s) available in OCTEON TX2 SoC. #. **Event Device Driver** See :doc:`../eventdevs/octeontx2` for SSO event device driver information. -#. **DMA Rawdev Driver** - See :doc:`../rawdevs/octeontx2_dma` for DMA driver information. - #. **Crypto Device Driver** See :doc:`../cryptodevs/octeontx2` for CPT crypto device driver information. diff --git a/doc/guides/rawdevs/index.rst b/doc/guides/rawdevs/index.rst index 7fbae40ea9..228d4a7743 100644 --- a/doc/guides/rawdevs/index.rst +++ b/doc/guides/rawdevs/index.rst @@ -17,5 +17,4 @@ application through rawdev API. ifpga ioat ntb -octeontx2_dma octeontx2_ep diff --git a/doc/guides/rawdevs/octeontx2_dma.rst b/doc/guides/rawdevs/octeontx2_dma.rst deleted file mode 100644 index 6887da5278..00 --- a/doc/guides/rawdevs/octeontx2_dma.rst +++ /dev/null @@ -1,103 +0,0 @@ -.. SPDX-License-Identifier: BSD-3-Clause -Copyright(c) 2019 Marvell International Ltd. - -OCTEON TX2 DMA Driver -= - -OCTEON TX2 has an internal DMA unit which can be used by applications to initiate -DMA transaction internally, from/to host when OCTEON TX2 operates in PCIe End -Point mode. The DMA PF function supports 8 VFs corresponding to 8 DMA queues. -Each DMA queue was exposed as a VF function when SRIOV enabled. - -Features - - -This DMA PMD supports below 3 modes of memory transfers - -#. Internal - OCTEON TX2 DRAM to DRAM without core intervention - -#. Inbound - Host DRAM to OCTEON TX2 DRAM without host/OCTEON TX2 cores involvement - -#. Outbound - OCTEON TX2 DRAM to Host DRAM without host/OCTEON TX2 cores involvement - -Prerequisites and Compilation procedure - - See :doc:`../platform/octeontx2` for setup information. - - -Enabling logs -- - -For enabling logs, use the following EAL parameter: - -.. code-block:: console - - ./your_dma_application --log-level=pmd.raw.octeontx2.dpi, - -Using ``pmd.raw.octeontx2.dpi`` as log matching criteria, all Event PMD logs -can be enabled which are lower than logging ``level``. - -Initialization --- - -The number of DMA VFs (queues) enabled can be controlled by setting sysfs -entry, `sriov_numvfs` for the corresponding PF driver. - -.. code-block:: console - - echo > /sys/bus/pci/drivers/octeontx2-dpi/\:05\:00.0/sriov_numvfs - -Once the required VFs are enabled, to be accessible from DPDK, VFs need to be -bound to vfio-pci driver. - -Device Setup -- - -The OCTEON TX2 DPI DMA HW devices will need to be bound to a -user-space IO driver for use. The script ``dpdk-devbind.py`` script -included with DPDK can be used to view the state of the devices and to bind -them to a suitabl
[dpdk-dev] [PATCH v4 2/2] drivers/raw: remove octeontx2-ep driver
Removing the rawdev based octeontx2-ep driver as the dependent
common/octeontx2 will soon be going away. Moreover this driver is no
longer required as the net/octeontx_ep driver is sufficient.
Signed-off-by: Radha Mohan Chintakuntla
---
Changes from v3:
Fixed patch application failure due to conflict on main branch.
Changes from v2:
Fixed DPDK CI reported issues for more documentation failure.
Changes from v1:
Fixed compilation issues in documentation
MAINTAINERS | 6 -
doc/guides/rawdevs/index.rst | 1 -
doc/guides/rawdevs/octeontx2_ep.rst | 82 ---
drivers/raw/meson.build | 1 -
drivers/raw/octeontx2_ep/meson.build | 11 -
drivers/raw/octeontx2_ep/otx2_ep_enqdeq.c | 846 --
drivers/raw/octeontx2_ep/otx2_ep_enqdeq.h | 52 --
drivers/raw/octeontx2_ep/otx2_ep_rawdev.c | 362 -
drivers/raw/octeontx2_ep/otx2_ep_rawdev.h | 499 -
drivers/raw/octeontx2_ep/otx2_ep_test.c | 172 -
drivers/raw/octeontx2_ep/otx2_ep_vf.c | 475
drivers/raw/octeontx2_ep/otx2_ep_vf.h | 10 -
drivers/raw/octeontx2_ep/version.map | 3 -
13 files changed, 2520 deletions(-)
delete mode 100644 doc/guides/rawdevs/octeontx2_ep.rst
delete mode 100644 drivers/raw/octeontx2_ep/meson.build
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_enqdeq.c
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_enqdeq.h
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_rawdev.c
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_rawdev.h
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_test.c
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_vf.c
delete mode 100644 drivers/raw/octeontx2_ep/otx2_ep_vf.h
delete mode 100644 drivers/raw/octeontx2_ep/version.map
diff --git a/MAINTAINERS b/MAINTAINERS
index 9f9aa37c68..1d6a408f49 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1321,12 +1321,6 @@ M: Tomasz Duszynski
F: doc/guides/rawdevs/cnxk_bphy.rst
F: drivers/raw/cnxk_bphy/
-Marvell OCTEON TX2 EP
-M: Radha Mohan Chintakuntla
-M: Veerasenareddy Burru
-F: drivers/raw/octeontx2_ep/
-F: doc/guides/rawdevs/octeontx2_ep.rst
-
NTB
M: Xiaoyun Li
M: Jingjing Wu
diff --git a/doc/guides/rawdevs/index.rst b/doc/guides/rawdevs/index.rst
index 228d4a7743..b6cf917443 100644
--- a/doc/guides/rawdevs/index.rst
+++ b/doc/guides/rawdevs/index.rst
@@ -17,4 +17,3 @@ application through rawdev API.
ifpga
ioat
ntb
-octeontx2_ep
diff --git a/doc/guides/rawdevs/octeontx2_ep.rst
b/doc/guides/rawdevs/octeontx2_ep.rst
deleted file mode 100644
index fb9d346ccf..00
--- a/doc/guides/rawdevs/octeontx2_ep.rst
+++ /dev/null
@@ -1,82 +0,0 @@
-.. SPDX-License-Identifier: BSD-3-Clause
-Copyright(c) 2019 Marvell International Ltd.
-
-Marvell OCTEON TX2 End Point Rawdev Driver
-==
-
-OCTEON TX2 has an internal SDP unit which provides End Point mode of operation
-by exposing its IOQs to Host, IOQs are used for packet I/O between Host and
-OCTEON TX2. Each OCTEON TX2 SDP PF supports a max of 128 VFs and Each VF is
-associated with a set of IOQ pairs.
-
-Features
-
-
-This OCTEON TX2 End Point mode PMD supports
-
-#. Packet Input - Host to OCTEON TX2 with direct data instruction mode.
-
-#. Packet Output - OCTEON TX2 to Host with info pointer mode.
-
-
-Initialization
---
-
-The number of SDP VFs enabled, can be controlled by setting sysfs
-entry `sriov_numvfs` for the corresponding PF driver.
-
-.. code-block:: console
-
- echo > /sys/bus/pci/drivers/octeontx2-ep/\:04\:00.0/sriov_numvfs
-
-Once the required VFs are enabled, to be accessible from DPDK, VFs need to be
-bound to vfio-pci driver.
-
-Device Setup
-
-
-The OCTEON TX2 SDP End Point VF devices will need to be bound to a
-user-space IO driver for use. The script ``dpdk-devbind.py`` script
-included with DPDK can be used to view the state of the devices and to bind
-them to a suitable DPDK-supported kernel driver. When querying the status
-of the devices, they will appear under the category of "Misc (rawdev)
-devices", i.e. the command ``dpdk-devbind.py --status-dev misc`` can be
-used to see the state of those devices alone.
-
-Device Configuration
-
-
-Configuring SDP EP rawdev device is done using the ``rte_rawdev_configure()``
-API, which takes the mempool as parameter. PMD uses this pool to send/receive
-packets to/from the HW.
-
-The following code shows how the device is configured
-
-.. code-block:: c
-
- struct sdp_rawdev_info config = {0};
- struct rte_rawdev_info rdev_info = {.dev_private = &config};
- config.enqdeq_mpool = (void *)rte_mempool_create(...);
-
- rte_rawdev_configure(dev_id, (rte_rawdev_obj_t)&rdev_info,
-sizeof(config));
-
-Performing Data Transfer
-
-
-To perform data transfer using SDP VF EP rawdev devices use standard
-``rte_rawdev_enqueue_buffe
Re: [dpdk-dev] [PATCH] test/service_cores: fix wrong comment
> -Original Message-
> From: Joyce Kong
> Sent: Thursday, August 19, 2021 7:11 AM
> To: jerin.ja...@caviumnetworks.com; Van Haaren, Harry
> ; honnappa.nagaraha...@arm.com;
> ruifeng.w...@arm.com
> Cc: dev@dpdk.org; n...@arm.com; sta...@dpdk.org
> Subject: [PATCH] test/service_cores: fix wrong comment
>
> Change the inaccurate comment of 'set pass flag'
> to 'clear pass flag' as the '*pass_test = 0' code
> actually implements clearing.
>
> Fixes: f038a81e1c56 ("service: add unit tests")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Joyce Kong
> Reviewed-by: Ruifeng Wang
Acked-by: Harry van Haaren
Re: [dpdk-dev] [PATCH] net/bnxt: fix to reset Rx next consumer index
On Mon, Aug 9, 2021 at 11:11 PM Somnath Kotur
wrote:
> In bnxt_init_one_rx_ring(), reset this variable internal to the driver
> ring to 0, so that there is no mismatch with actual value in HW on
> traffic resumption.
>
> Fixes: 03c8f2fe111c ("net/bnxt: detect bad opaque in Rx completion")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Somnath Kotur
> Reviewed-by: Kalesh AP
> Reviewed-by: Ajit Khaparde
>
Patch applied to dpdk-next-net-brcm for-next-net branch.
---
> drivers/net/bnxt/bnxt_rxr.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/net/bnxt/bnxt_rxr.c b/drivers/net/bnxt/bnxt_rxr.c
> index aea71703d1..73fbdd17d1 100644
> --- a/drivers/net/bnxt/bnxt_rxr.c
> +++ b/drivers/net/bnxt/bnxt_rxr.c
> @@ -1379,6 +1379,9 @@ int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq)
> }
> PMD_DRV_LOG(DEBUG, "TPA alloc Done!\n");
>
> + /* Explicitly reset this driver internal tracker on a ring init */
> + rxr->rx_next_cons = 0;
> +
> return 0;
> }
>
> --
> 2.28.0.497.g54e85e7
>
>
Re: [dpdk-dev] [PATCH] net/bnxt: update ring group after ring stop start
On Tue, Aug 17, 2021 at 10:39 PM Ajit Khaparde
wrote:
> A Rx ring stop start sequence may result in the FW returning
> a different set of Rx ring and AGG ring IDs. If the ring group
> is not updated with the new IDs, the HW sees the host driver using
> incorrect BD types for the Rx ring and AGG ring. This can cause
> the chip to go into a bad state or encounter RE_flush issue
> or leak MBUFs in the HW.
>
> Fix this by issuing a bnxt_hwrm_ring_grp_free() and an
> bnxt_hwrm_ring_grp_alloc() to refresh the ring group information.
>
> Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Ajit Khaparde
> Reviewed-by: Somnath Kotur
>
Patch applied to dpdk-next-net-brcm for-next-net branch.
> ---
> drivers/net/bnxt/bnxt_hwrm.c | 3 +++
> drivers/net/bnxt/bnxt_ring.c | 6 ++
> 2 files changed, 9 insertions(+)
>
> diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c
> index 73deb41f81..eb9de45cb9 100644
> --- a/drivers/net/bnxt/bnxt_hwrm.c
> +++ b/drivers/net/bnxt/bnxt_hwrm.c
> @@ -2717,6 +2717,9 @@ void bnxt_free_hwrm_rx_ring(struct bnxt *bp, int
> queue_index)
> struct bnxt_ring *ring = rxr->rx_ring_struct;
> struct bnxt_cp_ring_info *cpr = rxq->cp_ring;
>
> + if (BNXT_HAS_RING_GRPS(bp))
> + bnxt_hwrm_ring_grp_free(bp, queue_index);
> +
> bnxt_hwrm_ring_free(bp, ring,
> HWRM_RING_FREE_INPUT_RING_TYPE_RX,
> cpr->cp_ring_struct->fw_ring_id);
> diff --git a/drivers/net/bnxt/bnxt_ring.c b/drivers/net/bnxt/bnxt_ring.c
> index b05c470766..957b175f1b 100644
> --- a/drivers/net/bnxt/bnxt_ring.c
> +++ b/drivers/net/bnxt/bnxt_ring.c
> @@ -631,6 +631,12 @@ int bnxt_alloc_hwrm_rx_ring(struct bnxt *bp, int
> queue_index)
> if (rc)
> goto err_out;
>
> + if (BNXT_HAS_RING_GRPS(bp)) {
> + rc = bnxt_hwrm_ring_grp_alloc(bp, queue_index);
> + if (rc)
> + goto err_out;
> + }
> +
> if (rxq->rx_started) {
> if (bnxt_init_one_rx_ring(rxq)) {
> PMD_DRV_LOG(ERR, "bnxt_init_one_rx_ring
> failed!\n");
> --
> 2.21.1 (Apple Git-122.3)
>
>
[dpdk-dev] [PATCH] examples/ptpclient: enable Rx queue configuration
This patch adds support to enable per-queue Rx offloads so that
it can convey the configuration to PMD.
Signed-off-by: Simei Su
---
examples/ptpclient/ptpclient.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/examples/ptpclient/ptpclient.c b/examples/ptpclient/ptpclient.c
index 4f32ade..d083a2e 100644
--- a/examples/ptpclient/ptpclient.c
+++ b/examples/ptpclient/ptpclient.c
@@ -217,8 +217,13 @@ port_init(uint16_t port, struct rte_mempool *mbuf_pool)
/* Allocate and set up 1 RX queue per Ethernet port. */
for (q = 0; q < rx_rings; q++) {
+ struct rte_eth_rxconf *rxconf;
+
+ rxconf = &dev_info.default_rxconf;
+ rxconf->offloads = port_conf.rxmode.offloads;
+
retval = rte_eth_rx_queue_setup(port, q, nb_rxd,
- rte_eth_dev_socket_id(port), NULL, mbuf_pool);
+ rte_eth_dev_socket_id(port), rxconf, mbuf_pool);
if (retval < 0)
return retval;
--
2.9.5
Re: [dpdk-dev] [PATCH v13 09/10] eal: add EAL argument for setting thread priority
On Wed, Aug 18, 2021 at 02:28:33PM -0700, Stephen Hemminger wrote:
> On Tue, 3 Aug 2021 12:01:30 -0700 Narcisa Ana Maria Vasile
> wrote:
>
> > +static int +eal_parse_thread_priority(const char *arg) +{ +
> > struct internal_config *internal_conf = +
> > eal_get_internal_configuration(); + enum rte_thread_priority priority;
> > + + if (!strncmp("normal", arg, sizeof("normal"))) +
> > priority = RTE_THREAD_PRIORITY_NORMAL; +else if
> > (!strncmp("realtime", arg, sizeof("realtime"))) + priority =
> > RTE_THREAD_PRIORITY_REALTIME_CRITICAL; +else + return -1;
> > + + internal_conf->thread_priority = priority; +return 0; +} +
>
> In my experience using real time priority is dangerous and risks
> starvation and deadlock. The problem is that DPDK applications are
> typically 100% CPU poll mode with no system calls; but the kernel has a
> number of worker threads that can be required on those CPUs.
>
> The typical failure is a disk completion interrupt happens on a CPU that
> is being used by DPDK lcore thread. With RT priority, the kernel thread
> to process that I/O completion never runs because the RT user thread has
> higher priority than the kernel I/O completion thread.
>
> It maybe possible to workaround this with lots of hand crafting through
> sysfs to reassign threads and irq's. Also, later kernels with full RT
> might also help.
>
> Before putting this in as part of DPDK in EAL, a full set of testing and
> documentation of how to configure these kind of applications and systems
> is needed.
>
I would tend to agree caution here, based on my experience of having locked
up a number of systems in the past when testing running DPDK apps with RT
priority!
Re: [dpdk-dev] Windows community call: MoM 2021-08-04
On Wed, Aug 18, 2021 at 04:40:02PM -0700, William Tu wrote:
> On Tue, Aug 17, 2021 at 7:14 AM Bruce Richardson
> wrote:
> >
> > On Fri, Aug 06, 2021 at 04:17:32PM -0700, William Tu wrote:
> > > On Thu, Aug 5, 2021 at 12:15 PM Dmitry Kozlyuk
> > > wrote:
> > > >
> > > snip
> > >
> > > > # Porting OvS build system to meson (William Tu)
> > > >
> > > > Status: OvS compiles with some features disabled, with a lot of
> > > > warnings.
> > > > Issues:
> > > >
> > > > * vhost-user is Linux-specific.
> > > > [Omar] Microsoft is working on functional equivalent.
> > > > * rte_version* not exported.
> > > > AI William to send patches.
> > > > * rte_open_logstream() implementation relies on Linux-specific
> > > > fopencookie().
> > > > We need a more generic facility to redirect logs.
> > > > AI William and DmitryK to discuss.
> > > > * meson not finding DPDK with pkg-config, maybe meson bug.
> > >
> > > To give more details about this for people who are interested.
> > > At OVS side, we tried to link the DPDK library, by doing below at
> > > meson.build file
> > > libdpdk = dependency('libdpdk', method: 'pkg-config')" , or give
> > > it a specific path
> > > libdpdk = cc.find_library('dpdk', dirs: ['C:\\temp\\dpdk\\lib'])
> > >
> > > However, it doesn't work, with error below
> > > Run-time dependency libdpdk found: NO (tried pkgconfig)
> > > meson.build:45:4: ERROR: Dependency "libdpdk" not found, tried
> > > pkgconfig
> > >
> >
> > Can you share the meson log file snippet where it tries the pkg-config
> > call? Does PKG_CONFIG_PATH have to be set to a special value to get the .pc
> > files found?
>
> Hi Bruce,
> Thanks!
> The log below:
> ---
> Target machine cpu: x86_64
> Pkg-config binary for MachineChoice.HOST is not cached.
> Pkg-config binary missing from cross or native file, or env var undefined.
> Trying a default Pkg-config fallback at pkg-config
> Found pkg-config:
> C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0
> .28-1\bin\pkg-config.EXE (0.28)
> Determining dependency 'libdpdk' with pkg-config executable
> 'C:\\ProgramData\\chocola
> tey\\lib\\pkgconfiglite\\tools\\pkg-config-lite-0.28-1\\bin\\pkg-config.EXE'
> env[PKG_CONFIG_PATH]:
> Called
> `C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0.28-1\bin\
> pkg-config.EXE --modversion libdpdk` -> 1
>
> Run-time dependency libdpdk found: NO (tried pkgconfig)
> ---
> However, I copy the command and run it on powershell
> PS C:\k8s-antrea-dpdk-win\ovs>
> C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-
> config-lite-0.28-1\bin\pkg-config.EXE --modversion libdpdk
> 21.08.0-rc1
>
> Yes, for example:
> $env:PKG_CONFIG_PATH = 'C:\temp\DPDK\lib\pkgconfig'
>
Looking at the meson log, it appears that the PKG_CONFIG_PATH is empty for
the meson call to pkg-config "env[PKG_CONFIG_PATH]:", which may explain why
DPDK is not found. I suspect meson only uses the PKG_CONFIG_PATH from the
environment on first run, but beyond that you can configure a
PKG_CONFIG_PATH using a meson configuration option "-Dpkg_config_path". Can
you try explicitly setting that in your build and see if it fixes it for
you?
/Bruce
[dpdk-dev] [PATCH] stack: reload head when pop fails (generic)
The previous fix 18effad9cfa7 ("stack: reload head when pop fails") only
changed C11 implementation, not generic implementation.
List head must be loaded right before continue (when failed to find the
new head). Without this, one thread might keep trying and failing to pop
items without ever loading the new correct head.
Fixes: 3340202f5954 ("stack: add lock-free implementation")
Cc: gage.e...@intel.com
Cc: sta...@dpdk.org
Signed-off-by: Julien Meunier
---
lib/stack/rte_stack_lf_generic.h | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/lib/stack/rte_stack_lf_generic.h b/lib/stack/rte_stack_lf_generic.h
index 4850a05ee7..7fa29cedb2 100644
--- a/lib/stack/rte_stack_lf_generic.h
+++ b/lib/stack/rte_stack_lf_generic.h
@@ -128,8 +128,10 @@ __rte_stack_lf_pop_elems(struct rte_stack_lf_list *list,
/* If NULL was encountered, the list was modified while
* traversing it. Retry.
*/
- if (i != num)
+ if (i != num) {
+ old_head = list->head;
continue;
+ }
new_head.top = tmp;
new_head.cnt = old_head.cnt + 1;
--
2.17.1
[dpdk-dev] [PATCH] cryptodev: add telemetry callbacks
The cryptodev library now registers commands with telemetry, and
implements the corresponding callback functions. These commands
allow a list of cryptodevs and stats for a cryptodev to be
queried.
An example usage can be seen below:
Connecting to /var/run/dpdk/rte/dpdk_telemetry.v2
{"version": "DPDK 21.11.0-rc0", "pid": 1135019, "max_output_len": 16384}
--> /
{"/": ["/", "/cryptodev/list", "/cryptodev/stats", ...]}
--> /cryptodev/list
{"/cryptodev/list": {":1a:01.0_qat_sym": 0, ":1a:01.0_qat_asym": \
1}}
--> /cryptodev/stats,0
{"/cryptodev/stats": {"enqueued_count": 0, "dequeued_count": 0, \
"enqueue_err_count": 0, "dequeue_err_count": 0}}
Signed-off-by: Rebecca Troy
---
lib/cryptodev/rte_cryptodev.c | 62 +++
1 file changed, 62 insertions(+)
diff --git a/lib/cryptodev/rte_cryptodev.c b/lib/cryptodev/rte_cryptodev.c
index 447aa9d519..1e3ab633cc 100644
--- a/lib/cryptodev/rte_cryptodev.c
+++ b/lib/cryptodev/rte_cryptodev.c
@@ -36,6 +36,7 @@
#include
#include
#include
+#include
#include "rte_crypto.h"
#include "rte_cryptodev.h"
@@ -2427,3 +2428,64 @@ rte_cryptodev_allocate_driver(struct cryptodev_driver
*crypto_drv,
return nb_drivers++;
}
+
+static int
+cryptodev_handle_dev_list(const char *cmd __rte_unused,
+ const char *params __rte_unused,
+ struct rte_tel_data *d)
+{
+ int dev_id;
+
+ if (rte_cryptodev_count() < 1)
+ return -1;
+
+ rte_tel_data_start_dict(d);
+ for (dev_id = 0; dev_id < RTE_CRYPTO_MAX_DEVS; dev_id++)
+ if (rte_cryptodev_pmd_is_valid_dev(dev_id))
+ rte_tel_data_add_dict_int(d,
+ rte_cryptodev_name_get(dev_id), dev_id);
+
+ return 0;
+}
+
+#define ADD_DICT_STAT(stats, s) rte_tel_data_add_dict_u64(d, #s, stats.s)
+
+static int
+cryptodev_handle_dev_stats(const char *cmd __rte_unused,
+ const char *params,
+ struct rte_tel_data *d)
+{
+ struct rte_cryptodev_stats cryptodev_stats;
+ int dev_id, ret;
+ char *end_param;
+
+ if (params == NULL || strlen(params) == 0 || !isdigit(*params))
+ return -1;
+
+ dev_id = strtoul(params, &end_param, 0);
+ if (*end_param != '\0')
+ CDEV_LOG_ERR("Extra parameters passed to cryptodev telemetry
command, ignoring");
+ if (!rte_cryptodev_pmd_is_valid_dev(dev_id))
+ return -1;
+
+ ret = rte_cryptodev_stats_get(dev_id, &cryptodev_stats);
+ if (ret < 0)
+ return -1;
+
+ rte_tel_data_start_dict(d);
+ ADD_DICT_STAT(cryptodev_stats, enqueued_count);
+ ADD_DICT_STAT(cryptodev_stats, dequeued_count);
+ ADD_DICT_STAT(cryptodev_stats, enqueue_err_count);
+ ADD_DICT_STAT(cryptodev_stats, dequeue_err_count);
+
+ return 0;
+}
+
+RTE_INIT(cryptodev_init_telemetry)
+{
+ rte_telemetry_register_cmd("/cryptodev/list", cryptodev_handle_dev_list,
+ "Returns list of available cryptodev names and IDs.");
+ rte_telemetry_register_cmd("/cryptodev/stats",
+ cryptodev_handle_dev_stats,
+ "Returns the stats for a cryptodev. Parameters: int
dev_id");
+}
--
2.25.1
Re: [dpdk-dev] 20.11.3 patches review and test
Hi, > -Original Message- > From: luca.bocca...@gmail.com > Sent: Monday, August 9, 2021 11:44 AM > To: sta...@dpdk.org > Cc: dev@dpdk.org; Abhishek Marathe ; > Akhil Goyal ; Ali Alnubani ; > benjamin.wal...@intel.com; David Christensen ; > hariprasad.govindhara...@intel.com; Hemant Agrawal > ; Ian Stokes ; Jerin > Jacob ; John McNamara ; > Ju-Hyoung Lee ; Kevin Traynor > ; Luca Boccassi ; Pei Zhang > ; pingx...@intel.com; qian.q...@intel.com; Raslan > Darawsheh ; NBU-Contact-Thomas Monjalon > ; yuan.p...@intel.com; zhaoyan.c...@intel.com > Subject: 20.11.3 patches review and test > > Hi all, > > Here is a list of patches targeted for stable release 20.11.3. > > The planned date for the final release is August 23rd. > > Please help with testing and validation of your use cases and report any > issues/results with reply-all to this mail. For the final release the fixes > and > reported validations will be added to the release notes. > > A release candidate tarball can be found at: > > https://dpdk.org/browse/dpdk-stable/tag/?id=v20.11.3-rc1 > > These patches are located at branch 20.11 of dpdk-stable repo: > https://dpdk.org/browse/dpdk-stable/ > > Thanks. > > Luca Boccassi > > --- The following covers the functional tests that we ran on Mellanox hardware for this release: - Basic functionality: Send and receive multiple types of traffic. - testpmd xstats counter test. - testpmd timestamp test. - Changing/checking link status through testpmd. - RTE flow tests: Items: - eth - gre - gre_key gtp - icmp - icmp6 - ipv4 - ipv6 - ipv6_ext - ipv6_frag_ext - meta - mpls - nvgre - tcp - udp - vlan - vxlan - vxlan-gpe Actions: - age (shared and non-shared) - count - dec_tcp_ack - dec_tcp_seq - dec_ttl - drop - flag - inc_tcp_ack - inc_tcp_seq - jump - mark - meter - of_pop_vlan - of_push_vlan - of_set_vlan_pcp - of_set_vlan_vid - queue - raw_decap - raw_encap - rss (shared and non-shared) - set_ipv4_dscp - set_ipv4_dst - set_ipv4_src - set_ipv6_dscp - set_ipv6_dst - set_ipv6_src - set_mac_dst - set_mac_src - set_meta - set_tag - set_tp_dst - set_tp_src - set_ttl - vxlan_decap - vxlan_encap - Some RSS tests. - VLAN filtering, stripping and insertion tests. - Checksum and TSO tests. - ptype tests. - link_status_interrupt example application tests. - l3fwd-power example application tests. - Multi-process example applications tests. - Hardware LRO tests. Functional tests ran on: - NIC: ConnectX-4 Lx / OS: Ubuntu 20.04 LTS / Driver: MLNX_OFED_LINUX-5.4-1.0.3.0 / Firmware: 14.31.1014 - NIC: ConnectX-4 Lx / OS: Ubuntu 20.04 LTS / kernel: 5.14.0-rc6 / Driver: rdma-core v36.0 / Firmware: 14.31.1014 - NIC: ConnectX-5 / OS: Ubuntu 20.04 LTS / Driver: MLNX_OFED_LINUX-5.4-1.0.3.0 / Firmware: 16.31.1014 - NIC: ConnectX-5 / OS: Ubuntu 20.04 LTS / kernel: 5.14.0-rc6 / Driver: v36.0 / Firmware: 16.31.1014 Compilation tests with multiple configurations in the following OS/driver combinations are also passing: - Ubuntu 20.04.2 with MLNX_OFED_LINUX-5.4-1.0.3.0. - Ubuntu 20.04.2 with rdma-core master (64d1ae5). - Ubuntu 20.04.2 with rdma-core v28.0. - Ubuntu 18.04.5 with rdma-core v17.1. - Ubuntu 18.04.5 with rdma-core master (5b0f5b2) (i386). - Ubuntu 16.04.7 with rdma-core v22.7. - Fedora 34 with rdma-core v36.0. - Fedora 36 (Rawhide) with rdma-core v36.0 (only with gcc). - CentOS 7 7.9.2009 with rdma-core master (64d1ae5). - CentOS 7 7.9.2009 with MLNX_OFED_LINUX-5.4-1.0.3.0. - CentOS 8 8.3.2011 with rdma-core master (64d1ae5). - OpenSUSE Leap 15.3 with rdma-core v31.0. We don't see any critical issues blocking this release. Regards, Ali
[dpdk-dev] [PATCH] net/ice: fix deadlock in ice_flow_query()
It will cause deadlock when use switch-default, so fix it Signed-off-by: Yu Wenjun --- drivers/net/ice/ice_generic_flow.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/net/ice/ice_generic_flow.c b/drivers/net/ice/ice_generic_flow.c index 66b5743abf..c2fa75f165 100644 --- a/drivers/net/ice/ice_generic_flow.c +++ b/drivers/net/ice/ice_generic_flow.c @@ -2518,15 +2518,16 @@ ice_flow_query(struct rte_eth_dev *dev, ret = flow->engine->query_count(ad, flow, count, error); break; default: - return rte_flow_error_set(error, ENOTSUP, + ret = rte_flow_error_set(error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION, actions, "action not supported"); + goto out; } } +out: rte_spinlock_unlock(&pf->flow_ops_lock); - return ret; } -- 2.32.0.windows.1
[dpdk-dev] [PATCH] app/testpmd: configurable number of flows in flowgen
Make number of flows in flowgen configurable by setting parameter
--flowgen-flows=N.
Signed-off-by: Zhihong Wang
---
Depends-on: series-18277 ("app/testpmd: flowgen fixes and improvements")
app/test-pmd/flowgen.c| 22 ++
app/test-pmd/parameters.c | 10 ++
app/test-pmd/testpmd.c| 1 +
app/test-pmd/testpmd.h| 1 +
doc/guides/testpmd_app_ug/run_app.rst | 5 +
5 files changed, 31 insertions(+), 8 deletions(-)
diff --git a/app/test-pmd/flowgen.c b/app/test-pmd/flowgen.c
index 9348618d0f..9910a4dc53 100644
--- a/app/test-pmd/flowgen.c
+++ b/app/test-pmd/flowgen.c
@@ -40,8 +40,6 @@
#include "testpmd.h"
-/* hardcoded configuration (for now) */
-static unsigned cfg_n_flows= 1024;
static uint32_t cfg_ip_src = RTE_IPV4(10, 254, 0, 0);
static uint32_t cfg_ip_dst = RTE_IPV4(10, 253, 0, 0);
static uint16_t cfg_udp_src= 1000;
@@ -76,6 +74,7 @@ pkt_burst_flow_gen(struct fwd_stream *fs)
uint64_t ol_flags = 0;
uint16_t nb_rx;
uint16_t nb_tx;
+ uint16_t nb_dropped;
uint16_t nb_pkt;
uint16_t nb_clones = nb_pkt_flowgen_clones;
uint16_t i;
@@ -165,7 +164,7 @@ pkt_burst_flow_gen(struct fwd_stream *fs)
}
pkts_burst[nb_pkt] = pkt;
- if (++next_flow >= (int)cfg_n_flows)
+ if (++next_flow >= nb_flows_flowgen)
next_flow = 0;
}
@@ -184,13 +183,14 @@ pkt_burst_flow_gen(struct fwd_stream *fs)
fs->tx_packets += nb_tx;
inc_tx_burst_stats(fs, nb_tx);
- if (unlikely(nb_tx < nb_pkt)) {
+ nb_dropped = nb_pkt - nb_tx;
+ if (unlikely(nb_dropped > 0)) {
/* Back out the flow counter. */
- next_flow -= (nb_pkt - nb_tx);
+ next_flow -= nb_dropped;
while (next_flow < 0)
- next_flow += cfg_n_flows;
+ next_flow += nb_flows_flowgen;
- fs->fwd_dropped += nb_pkt - nb_tx;
+ fs->fwd_dropped += nb_dropped;
do {
rte_pktmbuf_free(pkts_burst[nb_tx]);
} while (++nb_tx < nb_pkt);
@@ -201,9 +201,15 @@ pkt_burst_flow_gen(struct fwd_stream *fs)
get_end_cycles(fs, start_tsc);
}
+static void
+flowgen_begin(portid_t pi)
+{
+ printf("nb flows from port %u: %d\n", pi, nb_flows_flowgen);
+}
+
struct fwd_engine flow_gen_engine = {
.fwd_mode_name = "flowgen",
- .port_fwd_begin = NULL,
+ .port_fwd_begin = flowgen_begin,
.port_fwd_end = NULL,
.packet_fwd = pkt_burst_flow_gen,
};
diff --git a/app/test-pmd/parameters.c b/app/test-pmd/parameters.c
index 7c13210f04..825275e683 100644
--- a/app/test-pmd/parameters.c
+++ b/app/test-pmd/parameters.c
@@ -143,6 +143,7 @@ usage(char* progname)
"N.\n");
printf(" --burst=N: set the number of packets per burst to N.\n");
printf(" --flowgen-clones=N: set the number of single packet clones to
send in flowgen mode. Should be less than burst value.\n");
+ printf(" --flowgen-flows=N: set the number of flows in flowgen mode to
N (1 <= N <= 2147483647).\n");
printf(" --mbcache=N: set the cache of mbuf memory pool to N.\n");
printf(" --rxpt=N: set prefetch threshold register of RX rings to
N.\n");
printf(" --rxht=N: set the host threshold register of RX rings to
N.\n");
@@ -586,6 +587,7 @@ launch_args_parse(int argc, char** argv)
{ "hairpin-mode", 1, 0, 0 },
{ "burst", 1, 0, 0 },
{ "flowgen-clones", 1, 0, 0 },
+ { "flowgen-flows", 1, 0, 0 },
{ "mbcache",1, 0, 0 },
{ "txpt", 1, 0, 0 },
{ "txht", 1, 0, 0 },
@@ -1122,6 +1124,14 @@ launch_args_parse(int argc, char** argv)
rte_exit(EXIT_FAILURE,
"clones must be >= 0 and <=
current burst\n");
}
+ if (!strcmp(lgopts[opt_idx].name, "flowgen-flows")) {
+ n = atoi(optarg);
+ if (n > 0)
+ nb_flows_flowgen = (int) n;
+ else
+ rte_exit(EXIT_FAILURE,
+"flows must be >= 1\n");
+ }
if (!strcmp(lgopts[opt_idx].name, "mbcache")) {
n = atoi(optarg);
if ((n >= 0) &&
diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 6cbe9ba3c8..9061cbf637 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testp
Re: [dpdk-dev] Windows community call: MoM 2021-08-04
[...] > > Target machine cpu: x86_64 > > Pkg-config binary for MachineChoice.HOST is not cached. > > Pkg-config binary missing from cross or native file, or env var undefined. > > Trying a default Pkg-config fallback at pkg-config > > Found pkg-config: > > C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0 > > .28-1\bin\pkg-config.EXE (0.28) > > Determining dependency 'libdpdk' with pkg-config executable > > 'C:\\ProgramData\\chocola > > tey\\lib\\pkgconfiglite\\tools\\pkg-config-lite-0.28-1\\bin\\pkg-config.EXE' > > env[PKG_CONFIG_PATH]: > > Called > > `C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0.28-1\bin\ > > pkg-config.EXE --modversion libdpdk` -> 1 > > > > Run-time dependency libdpdk found: NO (tried pkgconfig) > > --- > > However, I copy the command and run it on powershell > > PS C:\k8s-antrea-dpdk-win\ovs> > > C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg- > > config-lite-0.28-1\bin\pkg-config.EXE --modversion libdpdk > > 21.08.0-rc1 > > > > Yes, for example: > > $env:PKG_CONFIG_PATH = 'C:\temp\DPDK\lib\pkgconfig' > > > > Looking at the meson log, it appears that the PKG_CONFIG_PATH is empty for > the meson call to pkg-config "env[PKG_CONFIG_PATH]:", which may explain why > DPDK is not found. I suspect meson only uses the PKG_CONFIG_PATH from the > environment on first run, but beyond that you can configure a > PKG_CONFIG_PATH using a meson configuration option "-Dpkg_config_path". Can > you try explicitly setting that in your build and see if it fixes it for > you? Hi Bruce, It works, thanks a lot! I did PS C:\k8s-antrea-dpdk-win\ovs> meson --reconfigure build -Dpkg_config_path=C:\temp\dpdk\lib\pkgconfig\ Log below, if you're interested The Meson build system Version: 0.59.1 ... Pkg-config binary for MachineChoice.HOST is not cached. Pkg-config binary missing from cross or native file, or env var undefined. Trying a default Pkg-config fallback at pkg-config Found pkg-config: C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0.28-1\bin\pkg-config.EXE (0.28) Determining dependency 'libdpdk' with pkg-config executable 'C:\\ProgramData\\chocolatey\\lib\\pkgconfiglit e\\tools\\pkg-config-lite-0.28-1\\bin\\pkg-config.EXE' env[PKG_CONFIG_PATH]: C:/temp/dpdk/lib/pkgconfig Called `C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0.28-1\bin\pkg-config.EXE --modversion libdpdk` -> 0 21.08.0-rc1 env[PKG_CONFIG_PATH]: C:/temp/dpdk/lib/pkgconfig Called `C:\ProgramData\chocolatey\lib\pkgconfiglite\tools\pkg-config-lite-0.28-1\bin\pkg-config.EXE --cflag s libdpdk` -> 0 -include rte_config.h -march=native -IC:/temp/dpdk/include env[PKG_CONFIG_ALLOW_SYSTEM_LIBS]: 1 env[PKG_CONFIG_PATH]: C:/temp/dpdk/lib/pkgconfig --- William
Re: [dpdk-dev] 20.11.3 patches review and test
On Thu, 2021-08-19 at 12:10 +, Ali Alnubani wrote: > Hi, > > > -Original Message- > > From: luca.bocca...@gmail.com > > Sent: Monday, August 9, 2021 11:44 AM > > To: sta...@dpdk.org > > Cc: dev@dpdk.org; Abhishek Marathe ; > > Akhil Goyal ; Ali Alnubani ; > > benjamin.wal...@intel.com; David Christensen ; > > hariprasad.govindhara...@intel.com; Hemant Agrawal > > ; Ian Stokes ; Jerin > > Jacob ; John McNamara ; > > Ju-Hyoung Lee ; Kevin Traynor > > ; Luca Boccassi ; Pei Zhang > > ; pingx...@intel.com; qian.q...@intel.com; Raslan > > Darawsheh ; NBU-Contact-Thomas Monjalon > > ; yuan.p...@intel.com; zhaoyan.c...@intel.com > > Subject: 20.11.3 patches review and test > > > > Hi all, > > > > Here is a list of patches targeted for stable release 20.11.3. > > > > The planned date for the final release is August 23rd. > > > > Please help with testing and validation of your use cases and report any > > issues/results with reply-all to this mail. For the final release the fixes > > and > > reported validations will be added to the release notes. > > > > A release candidate tarball can be found at: > > > > https://dpdk.org/browse/dpdk-stable/tag/?id=v20.11.3-rc1 > > > > These patches are located at branch 20.11 of dpdk-stable repo: > > https://dpdk.org/browse/dpdk-stable/ > > > > Thanks. > > > > Luca Boccassi > > > > --- > > The following covers the functional tests that we ran on Mellanox hardware > for this release: > - Basic functionality: > Send and receive multiple types of traffic. > - testpmd xstats counter test. > - testpmd timestamp test. > - Changing/checking link status through testpmd. > - RTE flow tests: > Items: > - eth > - gre > - gre_key > gtp > - icmp > - icmp6 > - ipv4 > - ipv6 > - ipv6_ext > - ipv6_frag_ext > - meta > - mpls > - nvgre > - tcp > - udp > - vlan > - vxlan > - vxlan-gpe > Actions: > - age (shared and non-shared) > - count > - dec_tcp_ack > - dec_tcp_seq > - dec_ttl > - drop > - flag > - inc_tcp_ack > - inc_tcp_seq > - jump > - mark > - meter > - of_pop_vlan > - of_push_vlan > - of_set_vlan_pcp > - of_set_vlan_vid > - queue > - raw_decap > - raw_encap > - rss (shared and non-shared) > - set_ipv4_dscp > - set_ipv4_dst > - set_ipv4_src > - set_ipv6_dscp > - set_ipv6_dst > - set_ipv6_src > - set_mac_dst > - set_mac_src > - set_meta > - set_tag > - set_tp_dst > - set_tp_src > - set_ttl > - vxlan_decap > - vxlan_encap > > - Some RSS tests. > - VLAN filtering, stripping and insertion tests. > - Checksum and TSO tests. > - ptype tests. > - link_status_interrupt example application tests. > - l3fwd-power example application tests. > - Multi-process example applications tests. > - Hardware LRO tests. > > Functional tests ran on: > - NIC: ConnectX-4 Lx / OS: Ubuntu 20.04 LTS / Driver: > MLNX_OFED_LINUX-5.4-1.0.3.0 / Firmware: 14.31.1014 > - NIC: ConnectX-4 Lx / OS: Ubuntu 20.04 LTS / kernel: 5.14.0-rc6 / Driver: > rdma-core v36.0 / Firmware: 14.31.1014 > - NIC: ConnectX-5 / OS: Ubuntu 20.04 LTS / Driver: > MLNX_OFED_LINUX-5.4-1.0.3.0 / Firmware: 16.31.1014 > - NIC: ConnectX-5 / OS: Ubuntu 20.04 LTS / kernel: 5.14.0-rc6 / Driver: v36.0 > / Firmware: 16.31.1014 > > Compilation tests with multiple configurations in the following OS/driver > combinations are also passing: > - Ubuntu 20.04.2 with MLNX_OFED_LINUX-5.4-1.0.3.0. > - Ubuntu 20.04.2 with rdma-core master (64d1ae5). > - Ubuntu 20.04.2 with rdma-core v28.0. > - Ubuntu 18.04.5 with rdma-core v17.1. > - Ubuntu 18.04.5 with rdma-core master (5b0f5b2) (i386). > - Ubuntu 16.04.7 with rdma-core v22.7. > - Fedora 34 with rdma-core v36.0. > - Fedora 36 (Rawhide) with rdma-core v36.0 (only with gcc). > - CentOS 7 7.9.2009 with rdma-core master (64d1ae5). > - CentOS 7 7.9.2009 with MLNX_OFED_LINUX-5.4-1.0.3.0. > - CentOS 8 8.3.2011 with rdma-core master (64d1ae5). > - OpenSUSE Leap 15.3 with rdma-core v31.0. > > We don't see any critical issues blocking this release. > > Regards, > Ali Thank you! -- Kind regards, Luca Boccassi
Re: [dpdk-dev] [PATCH v15 1/6] dmadev: introduce DMA device library public APIs
On Fri, Aug 13, 2021 at 05:09:29PM +0800, Chengwen Feng wrote:
> The 'dmadevice' is a generic type of DMA device.
>
> This patch introduce the 'dmadevice' public APIs which expose generic
> operations that can enable configuration and I/O with the DMA devices.
>
> Signed-off-by: Chengwen Feng
> Acked-by: Bruce Richardson
> Acked-by: Morten Brørup
> Acked-by: Jerin Jacob
> ---
one minor comment for clarification
> +/**
> + * rte_dmadev_stats - running statistics.
> + */
> +struct rte_dmadev_stats {
> + uint64_t submitted_count;
> + /**< Count of operations which were submitted to hardware. */
> + uint64_t completed_fail_count;
> + /**< Count of operations which failed to complete. */
> + uint64_t completed_count;
> + /**< Count of operations which successfully complete. */
> +};
The name of the last variable and the comment on it seem mismatched. The
name implies that it's all completed ops, i.e. to get successful only you
do "stats.completed_count - stats.completed_fail_count", while the comment
says that it's successful only. Therefore I suggest:
* We rename the last two vars to "completed_fail" and "completed_success"
for clarity OR
* We redefine "completed_count" to be the full completed count of both
success and failure.
I have a slight preference for the latter option, but either can work.
/Bruce
PS: We probably don't need "count" on any of these values, based on two
options above suggest structs as:
struct rte_dmadev_stats {
uint64_t submitted;
uint64_t failed;
uint64_t successful;
};
OR:
struct rte_dmadev_stats {
uint64_t submitted;
uint64_t completed;
uint64_t errors;
}
Re: [dpdk-dev] [PATCH v2 2/2] examples/flow_classify: add an ACL table for tcp
Hi Sowmini,
> -Original Message-
> From: Sowmini Varadhan
> Sent: Wednesday, August 18, 2021 4:01 PM
> To: sowmin...@gmail.com; Iremonger, Bernard
> ; dev@dpdk.org;
> sovar...@linux.microsoft.com
> Cc: tho...@monjalon.net
> Subject: [PATCH v2 2/2] examples/flow_classify: add an ACL table for tcp
>
> v2 updates: typo fixes for checkpatch, bernard.iremonger comments
The above line should not be added to the commit message.
~/dpdk/devtools/check-git-log.sh -1
Wrong tag:
v2 fixes: typo fixes, get_tcp_flags returns -EINVAL
The check-git-log.sh script is in your dpdk checkout, best to run this script
before sending patches
>
> The struct rte_flow_classifier can have up to
> RTE_FLOW_CLASSIFY_TABLE_MAX
> (32) classifier tables, but the existing flow_classify examples only adds a
> single table for the L4 5-tuple.
>
> When dealing with tcp flows, we frequently want to add ACLs and filters to
> filter based on the state of the TCP connection, e.g., by looking at the tcp
> flags field.
>
> So we enhance flow_classify to add an additional acl table for tcp 5-tuples.
> If
> the input-file-parser detects a filter for a tcp flow with a non-wildcard
> argument for tcp_flags, the IP4_TCP_5TUPLE table is used by flow_classify.
>
> Signed-off-by: Sowmini Varadhan
> ---
If you want comment on the patch updates (optional), it can be added here after
the --- line
> examples/flow_classify/flow_classify.c | 41 +++---
> lib/flow_classify/rte_flow_classify.c | 87 +
> lib/flow_classify/rte_flow_classify.h | 19 +
> lib/flow_classify/rte_flow_classify_parse.c | 8 +-
> 4 files changed, 142 insertions(+), 13 deletions(-)
This patch contains changes to the flow_classify example and the flow_classify
library.
It needs to be split in two, a patch for the flow_classify example and a patch
for the flow classify library.
>
> diff --git a/examples/flow_classify/flow_classify.c
> b/examples/flow_classify/flow_classify.c
> index ff4c8bc2f5..8f6aacae03 100644
> --- a/examples/flow_classify/flow_classify.c
> +++ b/examples/flow_classify/flow_classify.c
> @@ -723,11 +723,6 @@ add_classify_rule(struct rte_eth_ntuple_filter
> *ntuple_filter,
> return ret;
> }
>
> - /* XXX but this only adds table_type of
> - * RTE_FLOW_CLASSIFY_TABLE_ACL_IP4_5TUPLE
> - * i.e., it only ever does allocate_acl_ipv4_5tuple_rule()
> - * so the tcp_flags is ignored!
> - */
> rule = rte_flow_classify_table_entry_add(
> cls_app->cls, &attr, pattern_ipv4_5tuple,
> actions, &key_found, &error);
> @@ -856,7 +851,8 @@ main(int argc, char *argv[])
> int ret;
> int socket_id;
> struct rte_table_acl_params table_acl_params;
> - struct rte_flow_classify_table_params cls_table_params;
> + struct rte_table_acl_params table_acl_tcp_params;
> + struct rte_flow_classify_table_params cls_table_params[2];
> struct flow_classifier *cls_app;
> struct rte_flow_classifier_params cls_params;
> uint32_t size;
> @@ -923,21 +919,42 @@ main(int argc, char *argv[])
> memcpy(table_acl_params.field_format, ipv4_defs,
> sizeof(ipv4_defs));
>
> /* initialise table create params */
> - cls_table_params.ops = &rte_table_acl_ops;
> - cls_table_params.arg_create = &table_acl_params;
> - cls_table_params.type =
> RTE_FLOW_CLASSIFY_TABLE_ACL_IP4_5TUPLE;
> + cls_table_params[0].ops = &rte_table_acl_ops;
> + cls_table_params[0].arg_create = &table_acl_params;
> + cls_table_params[0].type =
> RTE_FLOW_CLASSIFY_TABLE_ACL_IP4_5TUPLE;
> +
> + /* initialise ACL table params */
> + table_acl_tcp_params.name = "table_acl_ipv4_tcp_5tuple";
> + table_acl_tcp_params.n_rules = FLOW_CLASSIFY_MAX_RULE_NUM;
> + table_acl_tcp_params.n_rule_fields = RTE_DIM(ipv4_defs);
> + memcpy(table_acl_tcp_params.field_format, ipv4_defs,
> +sizeof(ipv4_defs));
> +
> + /* initialise table create params */
> + cls_table_params[1].ops = &rte_table_acl_ops;
> + cls_table_params[1].arg_create = &table_acl_tcp_params;
> + cls_table_params[1].type =
> RTE_FLOW_CLASSIFY_TABLE_ACL_IP4_TCP_5TUPLE;
>
> - ret = rte_flow_classify_table_create(cls_app->cls,
> &cls_table_params);
> + ret = rte_flow_classify_table_create(cls_app->cls,
> + &cls_table_params[0]);
> if (ret) {
> rte_flow_classifier_free(cls_app->cls);
> rte_free(cls_app);
> rte_exit(EXIT_FAILURE, "Failed to create classifier table\n");
> }
> + ret = rte_flow_classify_table_create(cls_app->cls,
> + &cls_table_params[1]);
> + if (ret) {
> + rte_flow_classifier_free(cls_app->cls);
> + rte_free(cls_app);
> + rte_exit(EXIT_FAILURE,
> + "Failed to create classifier table\n");
> + }
>
Re: [dpdk-dev] [PATCH v2 2/2] examples/flow_classify: add an ACL table for tcp
On (08/19/21 15:06), Iremonger, Bernard wrote: > > v2 updates: typo fixes for checkpatch, bernard.iremonger comments > > The above line should not be added to the commit message. > ~/dpdk/devtools/check-git-log.sh -1 > Wrong tag: > v2 fixes: typo fixes, get_tcp_flags returns -EINVAL > > The check-git-log.sh script is in your dpdk checkout, best to run this > script before sending patches apologies, I was not aware of this, I will fix and send out v3. I did run check-git-log.sh but before editing the vX updates. thanks for your patience. --Sowmini
Re: [dpdk-dev] [PATCH v2 1/2] examples/flow_classify: hooks for filters on tcp flags
HI Sowmini,
> -Original Message-
> From: Sowmini Varadhan
> Sent: Wednesday, August 18, 2021 4:01 PM
> To: sowmin...@gmail.com; Iremonger, Bernard
> ; dev@dpdk.org;
> sovar...@linux.microsoft.com
> Cc: tho...@monjalon.net
> Subject: [PATCH v2 1/2] examples/flow_classify: hooks for filters on tcp flags
>
> v2 fixes: typo fixes, get_tcp_flags returns -EINVAL
The above line should not be added to the commit message.
~/dpdk/devtools/check-git-log.sh -1
Wrong tag:
v2 fixes: typo fixes, get_tcp_flags returns -EINVAL
The check-git-log.sh script is in your dpdk checkout, best to run this script
before sending patches
>
> The rte_eth_ntuple_filter allows tcp_flags which can check for things like
> #define RTE_TCP_CWR_FLAG 0x80 /**< Congestion Window Reduced */
> #define RTE_TCP_ECE_FLAG 0x40 /**< ECN-Echo */
> #define RTE_TCP_URG_FLAG 0x20 /**< Urgent Pointer field significant */
> #define RTE_TCP_ACK_FLAG 0x10 /**< Acknowledgment field significant
> */
> #define RTE_TCP_PSH_FLAG 0x08 /**< Push Function */
> #define RTE_TCP_RST_FLAG 0x04 /**< Reset the connection */
> #define RTE_TCP_SYN_FLAG 0x02 /**< Synchronize sequence numbers */
> #define RTE_TCP_FIN_FLAG 0x01 /**< No more data from sender */ but
> there are no existing examples that demonstrate how to use this feature.
>
> This patch extends the existing classification support to allow an optional
> flags in the input file. The flags string can be any concatenation of
> characters
> from {C, E, U, A, P, R, S, F} and "*" indicates "don't care". These flags are
> set
> in the ntuple_filter and are used to construct the tcp_spec and tcp_mask
> sent to the driver
>
> The rte_acl_field_def is updated to use the (u8) tcp flag as lookup key.
> Note that, as per
> https://doc.dpdk.org/guides/prog_guide/packet_classif_access_ctrl.html
> this field MUST be allocated for 4 bytes, thus it has sizeof(uint32_t).
>
> However, also note the XXX in this commit: additional updates are needed to
> the rte_flow_classify_table_entry_add() so that it does not ignore any key
> fields other than the 5-tuple.
>
> Signed-off-by: Sowmini Varadhan
> ---
If you want comment on the patch updates (optional), it can be added here after
the --- line
> examples/flow_classify/flow_classify.c | 87 --
> examples/flow_classify/ipv4_rules_file.txt | 22 +++---
> 2 files changed, 91 insertions(+), 18 deletions(-)
>
> diff --git a/examples/flow_classify/flow_classify.c
> b/examples/flow_classify/flow_classify.c
> index db71f5aa04..ff4c8bc2f5 100644
> --- a/examples/flow_classify/flow_classify.c
> +++ b/examples/flow_classify/flow_classify.c
> @@ -51,6 +51,7 @@ enum {
> CB_FLD_DST_PORT_MASK,
> CB_FLD_PROTO,
> CB_FLD_PRIORITY,
> + CB_FLD_TCP_FLAGS,
> CB_FLD_NUM,
> };
>
> @@ -86,6 +87,7 @@ enum {
> DST_FIELD_IPV4,
> SRCP_FIELD_IPV4,
> DSTP_FIELD_IPV4,
> + TCP_FLAGS_FIELD,
I think it would be better to rename TCP_FLAGS_FIELD to TCP_FLAGS_FIELD_IPV4
inline with the other values.
> NUM_FIELDS_IPV4
> };
>
> @@ -93,7 +95,8 @@ enum {
> PROTO_INPUT_IPV4,
> SRC_INPUT_IPV4,
> DST_INPUT_IPV4,
> - SRCP_DESTP_INPUT_IPV4
> + SRCP_DESTP_INPUT_IPV4,
> + TCP_FLAGS_INDEX,
I think it would be better to rename TCP_FLAGS_INDEX to TCP_FLAGS_INPUT_IPV4
inline with the other values.
> };
>
> static struct rte_acl_field_def ipv4_defs[NUM_FIELDS_IPV4] = { @@ -150,6
> +153,17 @@ static struct rte_acl_field_def ipv4_defs[NUM_FIELDS_IPV4] = {
> sizeof(struct rte_ipv4_hdr) +
> offsetof(struct rte_tcp_hdr, dst_port),
> },
> + /* next field must be 4 bytes, even though flags is only 1 byte */
> + {
> + /* rte_flags */
> + .type = RTE_ACL_FIELD_TYPE_BITMASK,
> + .size = sizeof(uint32_t),
> + .field_index = TCP_FLAGS_FIELD,
> + .input_index = TCP_FLAGS_INDEX,
> + .offset = sizeof(struct rte_ether_hdr) +
> + sizeof(struct rte_ipv4_hdr) +
> + offsetof(struct rte_tcp_hdr, tcp_flags),
> + },
> };
> /* >8 End of creation of ACL table. */
>
> @@ -285,12 +299,14 @@ lcore_main(struct flow_classifier *cls_app)
> int ret;
> int i = 0;
>
> - ret = rte_flow_classify_table_entry_delete(cls_app->cls,
> - rules[7]);
> - if (ret)
> - printf("table_entry_delete failed [7] %d\n\n", ret);
> - else
> - printf("table_entry_delete succeeded [7]\n\n");
> + if (rules[7]) {
> + ret = rte_flow_classify_table_entry_delete(cls_app->cls,
> + rules[7]);
> + if (ret)
> + printf("table_entry_delete failed [7] %d\n\n", ret);
> + else
> + printf("table_entry_delete succeeded [7]\n\n");
> + }
>
> /*
>* Check
Re: [dpdk-dev] [PATCH v2 01/15] ethdev: introduce shared Rx queue
> -Original Message-
> From: Jerin Jacob
> Sent: Thursday, August 19, 2021 1:27 PM
> To: Xueming(Steven) Li
> Cc: dpdk-dev ; Ferruh Yigit ;
> NBU-Contact-Thomas Monjalon ;
> Andrew Rybchenko
> Subject: Re: [PATCH v2 01/15] ethdev: introduce shared Rx queue
>
> On Wed, Aug 18, 2021 at 4:44 PM Xueming(Steven) Li
> wrote:
> >
> >
> >
> > > -Original Message-
> > > From: Jerin Jacob
> > > Sent: Tuesday, August 17, 2021 11:12 PM
> > > To: Xueming(Steven) Li
> > > Cc: dpdk-dev ; Ferruh Yigit ;
> > > NBU-Contact-Thomas Monjalon ; Andrew Rybchenko
> > >
> > > Subject: Re: [PATCH v2 01/15] ethdev: introduce shared Rx queue
> > >
> > > On Tue, Aug 17, 2021 at 5:01 PM Xueming(Steven) Li
> > > wrote:
> > > >
> > > >
> > > >
> > > > > -Original Message-
> > > > > From: Jerin Jacob
> > > > > Sent: Tuesday, August 17, 2021 5:33 PM
> > > > > To: Xueming(Steven) Li
> > > > > Cc: dpdk-dev ; Ferruh Yigit
> > > > > ; NBU-Contact-Thomas Monjalon
> > > > > ; Andrew Rybchenko
> > > > >
> > > > > Subject: Re: [PATCH v2 01/15] ethdev: introduce shared Rx queue
> > > > >
> > > > > On Wed, Aug 11, 2021 at 7:34 PM Xueming Li
> > > > > wrote:
> > > > > >
> > > > > > In current DPDK framework, each RX queue is pre-loaded with
> > > > > > mbufs for incoming packets. When number of representors scale
> > > > > > out in a switch domain, the memory consumption became
> > > > > > significant. Most important, polling all ports leads to high
> > > > > > cache miss, high latency and low throughput.
> > > > > >
> > > > > > This patch introduces shared RX queue. Ports with same
> > > > > > configuration in a switch domain could share RX queue set by
> > > > > > specifying sharing group.
> > > > > > Polling any queue using same shared RX queue receives packets
> > > > > > from all member ports. Source port is identified by mbuf->port.
> > > > > >
> > > > > > Port queue number in a shared group should be identical. Queue
> > > > > > index is
> > > > > > 1:1 mapped in shared group.
> > > > > >
> > > > > > Share RX queue must be polled on single thread or core.
> > > > > >
> > > > > > Multiple groups is supported by group ID.
> > > > > >
> > > > > > Signed-off-by: Xueming Li
> > > > > > Cc: Jerin Jacob
> > > > > > ---
> > > > > > Rx queue object could be used as shared Rx queue object, it's
> > > > > > important to clear all queue control callback api that using queue
> > > > > > object:
> > > > > > https://mails.dpdk.org/archives/dev/2021-July/215574.html
> > > > >
> > > > > > #undef RTE_RX_OFFLOAD_BIT2STR diff --git
> > > > > > a/lib/ethdev/rte_ethdev.h b/lib/ethdev/rte_ethdev.h index
> > > > > > d2b27c351f..a578c9db9d 100644
> > > > > > --- a/lib/ethdev/rte_ethdev.h
> > > > > > +++ b/lib/ethdev/rte_ethdev.h
> > > > > > @@ -1047,6 +1047,7 @@ struct rte_eth_rxconf {
> > > > > > uint8_t rx_drop_en; /**< Drop packets if no descriptors are
> > > > > > available. */
> > > > > > uint8_t rx_deferred_start; /**< Do not start queue with
> > > > > > rte_eth_dev_start(). */
> > > > > > uint16_t rx_nseg; /**< Number of descriptions in rx_seg
> > > > > > array.
> > > > > > */
> > > > > > + uint32_t shared_group; /**< Shared port group index in
> > > > > > + switch domain. */
> > > > >
> > > > > Not to able to see anyone setting/creating this group ID test
> > > > > application.
> > > > > How this group is created?
> > > >
> > > > Nice catch, the initial testpmd version only support one default
> > > > group(0).
> > > > All ports that supports shared-rxq assigned in same group.
> > > >
> > > > We should be able to change "--rxq-shared" to "--rxq-shared-group"
> > > > to support group other than default.
> > > >
> > > > To support more groups simultaneously, need to consider testpmd
> > > > forwarding stream core assignment, all streams in same group need to
> > > > stay on same core.
> > > > It's possible to specify how many ports to increase group number,
> > > > but user must schedule stream affinity carefully - error prone.
> > > >
> > > > On the other hand, one group should be sufficient for most
> > > > customer, the doubt is whether it valuable to support multiple groups
> > > > test.
> > >
> > > Ack. One group is enough in testpmd.
> > >
> > > My question was more about who and how this group is created, Should
> > > n't we need API to create shared_group? If we do the following, at least,
> > > I can think, how it can be implemented in SW or other HW.
> > >
> > > - Create aggregation queue group
> > > - Attach multiple Rx queues to the aggregation queue group
> > > - Pull the packets from the queue group(which internally fetch from
> > > the Rx queues _attached_)
> > >
> > > Does the above kind of sequence, break your representor use case?
> >
> > Seems more like a set of EAL wrapper. Current API tries to minimize the
> > application efforts to adapt shared-rxq.
> > - step 1, not sure how important it is to create group with API, in
> > rte_flow, group is created
Re: [dpdk-dev] [PATCH v2 2/2] examples/flow_classify: add an ACL table for tcp
Hi Sowmini, Looking closer at this patchset, I am not sure that a second ACL table is needed. The existing ACL table handles UDP, TCP and SCP, however it is not processing the TCP flags. I think it just needs to be modified to process the TCP flags. Could you take another look to see if the above proposed solution will work for you. Regards, Bernard. > -Original Message- > From: Sowmini Varadhan > Sent: Thursday, August 19, 2021 4:21 PM > To: Iremonger, Bernard > Cc: sowmin...@gmail.com; dev@dpdk.org; tho...@monjalon.net > Subject: Re: [PATCH v2 2/2] examples/flow_classify: add an ACL table for tcp > > On (08/19/21 15:06), Iremonger, Bernard wrote: > > > v2 updates: typo fixes for checkpatch, bernard.iremonger comments > > > > The above line should not be added to the commit message. > > ~/dpdk/devtools/check-git-log.sh -1 > > Wrong tag: > > v2 fixes: typo fixes, get_tcp_flags returns -EINVAL > > > > The check-git-log.sh script is in your dpdk checkout, best to run this > > script > before sending patches > > apologies, I was not aware of this, I will fix and send out v3. > I did run check-git-log.sh but before editing the vX updates. > > thanks for your patience. > > --Sowmini
Re: [dpdk-dev] [PATCH v2 2/2] examples/flow_classify: add an ACL table for tcp
On (08/19/21 16:21), Iremonger, Bernard wrote: > > Looking closer at this patchset, I am not sure that a second ACL table is > needed. > The existing ACL table handles UDP, TCP and SCP, however it is not processing > the TCP flags. > I think it just needs to be modified to process the TCP flags. > Could you take another look to see if the above proposed solution will work > for you. I'm not sure it would. As I pointed out in the original rfc at https://inbox.dpdk.org/dev/cover.1578936382.git.sowmini.varad...@microsoft.com/ we need to add a key for the 8 bit flags, but the multibit trie lookup moves in steps of 4 bytes. However, it has admittedly been a while since I tinkereed with this, and I can give it a shot and get back. Thanks --Sowmini
[dpdk-dev] [PATCH v2 0/6] bbdev update related to CRC usage
v2: typo missed in documentation. Capturing in this serie minor change to bbdev to support a more comprehensive set of options for CRC used for 4G and 5G (exposed as optional capabilities). This also includes clarifications in documention and minor update to log print. Nicolas Chautru (6): bbdev: add capability for CRC16 check baseband/turbo_sw: add support for CRC16 bbdev: add capability for 4G CB CRC DROP baseband/acc100: add support for 4G CRC drop doc: clarification of usage of HARQ in bbdev doc bbdev: reduce warning level for one scenario app/test-bbdev/test_bbdev_vector.c | 4 +++ doc/guides/bbdevs/acc100.rst | 1 + doc/guides/prog_guide/bbdev.rst | 26 doc/guides/rel_notes/release_21_11.rst | 8 + drivers/baseband/acc100/rte_acc100_pmd.c | 12 ++-- drivers/baseband/turbo_sw/bbdev_turbo_software.c | 17 +++ lib/bbdev/rte_bbdev.c| 7 +++-- lib/bbdev/rte_bbdev_op.h | 39 +--- 8 files changed, 91 insertions(+), 23 deletions(-) -- 1.8.3.1
[dpdk-dev] [PATCH v2 2/6] baseband/turbo_sw: add support for CRC16
This is to support the case for operation
where CRC16 is to be appended or checked.
Signed-off-by: Nicolas Chautru
---
doc/guides/rel_notes/release_21_11.rst | 3 +++
drivers/baseband/turbo_sw/bbdev_turbo_software.c | 17 +
2 files changed, 20 insertions(+)
diff --git a/doc/guides/rel_notes/release_21_11.rst
b/doc/guides/rel_notes/release_21_11.rst
index 69dd518..8ca59b7 100644
--- a/doc/guides/rel_notes/release_21_11.rst
+++ b/doc/guides/rel_notes/release_21_11.rst
@@ -55,6 +55,9 @@ New Features
Also, make sure to start the actual text at the margin.
===
+* **Updated the turbo_sw bbdev PMD.**
+
+ Added support for more comprehensive CRC options.
Removed Items
-
diff --git a/drivers/baseband/turbo_sw/bbdev_turbo_software.c
b/drivers/baseband/turbo_sw/bbdev_turbo_software.c
index 77e9a2e..e570044 100644
--- a/drivers/baseband/turbo_sw/bbdev_turbo_software.c
+++ b/drivers/baseband/turbo_sw/bbdev_turbo_software.c
@@ -199,6 +199,7 @@ struct turbo_sw_queue {
.cap.ldpc_enc = {
.capability_flags =
RTE_BBDEV_LDPC_RATE_MATCH |
+ RTE_BBDEV_LDPC_CRC_16_ATTACH |
RTE_BBDEV_LDPC_CRC_24A_ATTACH |
RTE_BBDEV_LDPC_CRC_24B_ATTACH,
.num_buffers_src =
@@ -211,6 +212,7 @@ struct turbo_sw_queue {
.type = RTE_BBDEV_OP_LDPC_DEC,
.cap.ldpc_dec = {
.capability_flags =
+ RTE_BBDEV_LDPC_CRC_TYPE_16_CHECK |
RTE_BBDEV_LDPC_CRC_TYPE_24B_CHECK |
RTE_BBDEV_LDPC_CRC_TYPE_24A_CHECK |
RTE_BBDEV_LDPC_CRC_TYPE_24B_DROP |
@@ -880,6 +882,12 @@ struct turbo_sw_queue {
crc_req.len = in_length_in_bits - 24;
crc_resp.data = q->enc_in;
bblib_lte_crc24b_gen(&crc_req, &crc_resp);
+ } else if (enc->op_flags & RTE_BBDEV_LDPC_CRC_16_ATTACH) {
+ rte_memcpy(q->enc_in, in, in_length_in_bytes - 2);
+ crc_req.data = in;
+ crc_req.len = in_length_in_bits - 16;
+ crc_resp.data = q->enc_in;
+ bblib_lte_crc16_gen(&crc_req, &crc_resp);
} else
rte_memcpy(q->enc_in, in, in_length_in_bytes);
@@ -1492,6 +1500,15 @@ struct turbo_sw_queue {
if (!crc_resp.check_passed)
op->status |= 1 << RTE_BBDEV_CRC_ERROR;
}
+ if (check_bit(dec->op_flags, RTE_BBDEV_LDPC_CRC_TYPE_16_CHECK)) {
+ crc_req.data = adapter_input;
+ crc_req.len = K - dec->n_filler - 16;
+ crc_resp.check_passed = false;
+ crc_resp.data = adapter_input;
+ bblib_lte_crc16_check(&crc_req, &crc_resp);
+ if (!crc_resp.check_passed)
+ op->status |= 1 << RTE_BBDEV_CRC_ERROR;
+ }
#ifdef RTE_BBDEV_OFFLOAD_COST
q_stats->acc_offload_cycles += rte_rdtsc_precise() - start_time;
--
1.8.3.1
[dpdk-dev] [PATCH v2 1/6] bbdev: add capability for CRC16 check
Adding a missing operation when CRC16
is being used for TB CRC check.
Signed-off-by: Nicolas Chautru
---
app/test-bbdev/test_bbdev_vector.c | 2 ++
doc/guides/prog_guide/bbdev.rst| 3 +++
doc/guides/rel_notes/release_21_11.rst | 1 +
lib/bbdev/rte_bbdev_op.h | 34 ++
4 files changed, 24 insertions(+), 16 deletions(-)
diff --git a/app/test-bbdev/test_bbdev_vector.c
b/app/test-bbdev/test_bbdev_vector.c
index 614dbd1..8d796b1 100644
--- a/app/test-bbdev/test_bbdev_vector.c
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -167,6 +167,8 @@
*op_flag_value = RTE_BBDEV_LDPC_CRC_TYPE_24B_CHECK;
else if (!strcmp(token, "RTE_BBDEV_LDPC_CRC_TYPE_24B_DROP"))
*op_flag_value = RTE_BBDEV_LDPC_CRC_TYPE_24B_DROP;
+ else if (!strcmp(token, "RTE_BBDEV_LDPC_CRC_TYPE_16_CHECK"))
+ *op_flag_value = RTE_BBDEV_LDPC_CRC_TYPE_16_CHECK;
else if (!strcmp(token, "RTE_BBDEV_LDPC_DEINTERLEAVER_BYPASS"))
*op_flag_value = RTE_BBDEV_LDPC_DEINTERLEAVER_BYPASS;
else if (!strcmp(token, "RTE_BBDEV_LDPC_HQ_COMBINE_IN_ENABLE"))
diff --git a/doc/guides/prog_guide/bbdev.rst b/doc/guides/prog_guide/bbdev.rst
index 9619280..8bd7cba 100644
--- a/doc/guides/prog_guide/bbdev.rst
+++ b/doc/guides/prog_guide/bbdev.rst
@@ -891,6 +891,9 @@ given below.
|RTE_BBDEV_LDPC_CRC_TYPE_24B_DROP|
| Set to drop the last CRC bits decoding output |
++
+|RTE_BBDEV_LDPC_CRC_TYPE_16_CHECK|
+| Set for code block CRC-16 checking |
+++
|RTE_BBDEV_LDPC_DEINTERLEAVER_BYPASS |
| Set for bit-level de-interleaver bypass on input stream|
++
diff --git a/doc/guides/rel_notes/release_21_11.rst
b/doc/guides/rel_notes/release_21_11.rst
index d707a55..69dd518 100644
--- a/doc/guides/rel_notes/release_21_11.rst
+++ b/doc/guides/rel_notes/release_21_11.rst
@@ -84,6 +84,7 @@ API Changes
Also, make sure to start the actual text at the margin.
===
+* bbdev: Added capability related to more comprehensive CRC options.
ABI Changes
---
diff --git a/lib/bbdev/rte_bbdev_op.h b/lib/bbdev/rte_bbdev_op.h
index f946842..7c44ddd 100644
--- a/lib/bbdev/rte_bbdev_op.h
+++ b/lib/bbdev/rte_bbdev_op.h
@@ -142,51 +142,53 @@ enum rte_bbdev_op_ldpcdec_flag_bitmasks {
RTE_BBDEV_LDPC_CRC_TYPE_24B_CHECK = (1ULL << 1),
/** Set to drop the last CRC bits decoding output */
RTE_BBDEV_LDPC_CRC_TYPE_24B_DROP = (1ULL << 2),
+ /** Set for transport block CRC-16 checking */
+ RTE_BBDEV_LDPC_CRC_TYPE_16_CHECK = (1ULL << 3),
/** Set for bit-level de-interleaver bypass on Rx stream. */
- RTE_BBDEV_LDPC_DEINTERLEAVER_BYPASS = (1ULL << 3),
+ RTE_BBDEV_LDPC_DEINTERLEAVER_BYPASS = (1ULL << 4),
/** Set for HARQ combined input stream enable. */
- RTE_BBDEV_LDPC_HQ_COMBINE_IN_ENABLE = (1ULL << 4),
+ RTE_BBDEV_LDPC_HQ_COMBINE_IN_ENABLE = (1ULL << 5),
/** Set for HARQ combined output stream enable. */
- RTE_BBDEV_LDPC_HQ_COMBINE_OUT_ENABLE = (1ULL << 5),
+ RTE_BBDEV_LDPC_HQ_COMBINE_OUT_ENABLE = (1ULL << 6),
/** Set for LDPC decoder bypass.
* RTE_BBDEV_LDPC_HQ_COMBINE_OUT_ENABLE must be set.
*/
- RTE_BBDEV_LDPC_DECODE_BYPASS = (1ULL << 6),
+ RTE_BBDEV_LDPC_DECODE_BYPASS = (1ULL << 7),
/** Set for soft-output stream enable */
- RTE_BBDEV_LDPC_SOFT_OUT_ENABLE = (1ULL << 7),
+ RTE_BBDEV_LDPC_SOFT_OUT_ENABLE = (1ULL << 8),
/** Set for Rate-Matching bypass on soft-out stream. */
- RTE_BBDEV_LDPC_SOFT_OUT_RM_BYPASS = (1ULL << 8),
+ RTE_BBDEV_LDPC_SOFT_OUT_RM_BYPASS = (1ULL << 9),
/** Set for bit-level de-interleaver bypass on soft-output stream. */
- RTE_BBDEV_LDPC_SOFT_OUT_DEINTERLEAVER_BYPASS = (1ULL << 9),
+ RTE_BBDEV_LDPC_SOFT_OUT_DEINTERLEAVER_BYPASS = (1ULL << 10),
/** Set for iteration stopping on successful decode condition
* i.e. a successful syndrome check.
*/
- RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE = (1ULL << 10),
+ RTE_BBDEV_LDPC_ITERATION_STOP_ENABLE = (1ULL << 11),
/** Set if a device supports decoder dequeue interrupts. */
- RTE_BBDEV_LDPC_DEC_INTERRUPTS = (1ULL << 11),
+ RTE_BBDEV_LDPC_DEC_INTERRUPTS = (1ULL << 12),
/** Set if a device supports scatter-gather functionality. */
- RTE_BBDEV_LDPC_DEC_SCATTER_GATHER = (1ULL << 12),
+ RTE_BBDEV_LDPC_DEC_SCATTER_GATHER = (1ULL << 13),
/** Set if a device supports input/output HARQ compression. */
- RTE_B
[dpdk-dev] [PATCH v2 3/6] bbdev: add capability for 4G CB CRC DROP
Adding option to drop CRC24B to align with existing
feature for 5G
Signed-off-by: Nicolas Chautru
---
app/test-bbdev/test_bbdev_vector.c | 2 ++
lib/bbdev/rte_bbdev_op.h | 5 -
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/app/test-bbdev/test_bbdev_vector.c
b/app/test-bbdev/test_bbdev_vector.c
index 8d796b1..f020836 100644
--- a/app/test-bbdev/test_bbdev_vector.c
+++ b/app/test-bbdev/test_bbdev_vector.c
@@ -149,6 +149,8 @@
*op_flag_value = RTE_BBDEV_TURBO_DEC_SCATTER_GATHER;
else if (!strcmp(token, "RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP"))
*op_flag_value = RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP;
+ else if (!strcmp(token, "RTE_BBDEV_TURBO_DEC_CRC_24B_DROP"))
+ *op_flag_value = RTE_BBDEV_TURBO_DEC_CRC_24B_DROP;
else {
printf("The given value is not a turbo decoder flag\n");
return -1;
diff --git a/lib/bbdev/rte_bbdev_op.h b/lib/bbdev/rte_bbdev_op.h
index 7c44ddd..5512859 100644
--- a/lib/bbdev/rte_bbdev_op.h
+++ b/lib/bbdev/rte_bbdev_op.h
@@ -114,7 +114,10 @@ enum rte_bbdev_op_td_flag_bitmasks {
/** Set to keep CRC24B bits appended while decoding. Only usable when
* decoding Transport Block mode.
*/
- RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP = (1ULL << 16)
+ RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP = (1ULL << 16),
+ /** Set to drop CRC24B bits not to be appended while decoding.
+*/
+ RTE_BBDEV_TURBO_DEC_CRC_24B_DROP = (1ULL << 17)
};
--
1.8.3.1
[dpdk-dev] [PATCH v2 4/6] baseband/acc100: add support for 4G CRC drop
This implements in PMD the option to drop the CB CRC
after 4G decoding to help transport block concatenation.
Signed-off-by: Nicolas Chautru
---
doc/guides/bbdevs/acc100.rst | 1 +
doc/guides/rel_notes/release_21_11.rst | 4
drivers/baseband/acc100/rte_acc100_pmd.c | 12 +---
3 files changed, 14 insertions(+), 3 deletions(-)
diff --git a/doc/guides/bbdevs/acc100.rst b/doc/guides/bbdevs/acc100.rst
index ff0fa4b..9fff6ab 100644
--- a/doc/guides/bbdevs/acc100.rst
+++ b/doc/guides/bbdevs/acc100.rst
@@ -58,6 +58,7 @@ ACC100 5G/4G FEC PMD supports the following BBDEV
capabilities:
- ``RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN`` : set if negative LLR encoder i/p
is supported
- ``RTE_BBDEV_TURBO_POS_LLR_1_BIT_IN`` : set if positive LLR encoder i/p
is supported
- ``RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP`` : keep CRC24B bits appended
while decoding
+ - ``RTE_BBDEV_TURBO_DEC_CRC_24B_DROP`` : option to drop the code block CRC
after decoding
- ``RTE_BBDEV_TURBO_EARLY_TERMINATION`` : set early termination feature
- ``RTE_BBDEV_TURBO_DEC_SCATTER_GATHER`` : supports scatter-gather for
input/output data
- ``RTE_BBDEV_TURBO_HALF_ITERATION_EVEN`` : set half iteration granularity
diff --git a/doc/guides/rel_notes/release_21_11.rst
b/doc/guides/rel_notes/release_21_11.rst
index 8ca59b7..f7843bc 100644
--- a/doc/guides/rel_notes/release_21_11.rst
+++ b/doc/guides/rel_notes/release_21_11.rst
@@ -59,6 +59,10 @@ New Features
Added support for more comprehensive CRC options.
+* **Updated the ACC100 bbdev PMD.**
+
+ Added support for more comprehensive CRC options.
+
Removed Items
-
diff --git a/drivers/baseband/acc100/rte_acc100_pmd.c
b/drivers/baseband/acc100/rte_acc100_pmd.c
index 68ba523..891be81 100644
--- a/drivers/baseband/acc100/rte_acc100_pmd.c
+++ b/drivers/baseband/acc100/rte_acc100_pmd.c
@@ -980,6 +980,7 @@
RTE_BBDEV_TURBO_NEG_LLR_1_BIT_IN |
RTE_BBDEV_TURBO_MAP_DEC |
RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP |
+ RTE_BBDEV_TURBO_DEC_CRC_24B_DROP |
RTE_BBDEV_TURBO_DEC_SCATTER_GATHER,
.max_llr_modulus = INT8_MAX,
.num_buffers_src =
@@ -1708,8 +1709,12 @@
}
if ((op->turbo_dec.code_block_mode == RTE_BBDEV_TRANSPORT_BLOCK)
- && !check_bit(op->turbo_dec.op_flags,
- RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP))
+ && !check_bit(op->turbo_dec.op_flags,
+ RTE_BBDEV_TURBO_DEC_TB_CRC_24B_KEEP))
+ crc24_overlap = 24;
+ if ((op->turbo_dec.code_block_mode == RTE_BBDEV_CODE_BLOCK)
+ && check_bit(op->turbo_dec.op_flags,
+ RTE_BBDEV_TURBO_DEC_CRC_24B_DROP))
crc24_overlap = 24;
/* Calculates circular buffer size.
@@ -1744,7 +1749,8 @@
next_triplet = acc100_dma_fill_blk_type_out(
desc, h_output, *h_out_offset,
- k >> 3, next_triplet, ACC100_DMA_BLKID_OUT_HARD);
+ (k - crc24_overlap) >> 3, next_triplet,
+ ACC100_DMA_BLKID_OUT_HARD);
if (unlikely(next_triplet < 0)) {
rte_bbdev_log(ERR,
"Mismatch between data to process and mbuf data
length in bbdev_op: %p",
--
1.8.3.1
[dpdk-dev] [PATCH v2 6/6] bbdev: reduce warning level for one scenario
Queue setup may genuinely fail when adding incremental queues
for a given priority level. In that case application would
attempt to configure a queue at a different priority level.
Not an actual error.
Signed-off-by: Nicolas Chautru
---
lib/bbdev/rte_bbdev.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/lib/bbdev/rte_bbdev.c b/lib/bbdev/rte_bbdev.c
index fc37236..defddcf 100644
--- a/lib/bbdev/rte_bbdev.c
+++ b/lib/bbdev/rte_bbdev.c
@@ -528,9 +528,10 @@ struct rte_bbdev *
ret = dev->dev_ops->queue_setup(dev, queue_id, (conf != NULL) ?
conf : &dev_info.default_queue_conf);
if (ret < 0) {
- rte_bbdev_log(ERR,
- "Device %u queue %u setup failed", dev_id,
- queue_id);
+ /* This may happen when trying different priority levels */
+ rte_bbdev_log(INFO,
+ "Device %u queue %u setup failed",
+ dev_id, queue_id);
return ret;
}
--
1.8.3.1
[dpdk-dev] [PATCH v2 5/6] doc: clarification of usage of HARQ in bbdev doc
New paragraph detailing typical VRAN usecase and mapping to bbdev API usage. Signed-off-by: Nicolas Chautru --- doc/guides/prog_guide/bbdev.rst | 23 +++ 1 file changed, 23 insertions(+) diff --git a/doc/guides/prog_guide/bbdev.rst b/doc/guides/prog_guide/bbdev.rst index 8bd7cba..f39b62f 100644 --- a/doc/guides/prog_guide/bbdev.rst +++ b/doc/guides/prog_guide/bbdev.rst @@ -1054,6 +1054,29 @@ capability RTE_BBDEV_LDPC_INTERNAL_HARQ_MEMORY_OUT_ENABLE is set then the HARQ is stored in memory internal to the device and not visible to BBDEV. +.. note:: +More explicitly for a typical usage of HARQ retransmission in a VRAN +application using a HW PMD, there will be 2 cases. + +For 1st transmission, only the HARQ output is enabled : + +- the harq_combined_output.offset is provided to a given address. ie. typically an integer index * 32K, where the index is tracked by the application based on code block index for a given UE and HARQ process. + +- the related operation flag would notably include RTE_BBDEV_LDPC_HQ_COMBINE_OUT_ENABLE and RTE_BBDEV_LDPC_HARQ_6BIT_COMPRESSION. + +- note that not explicit flush or reset of the memory is required. + +For 2nd transmission, an input is also required to benefit from HARQ combination gain: + +- the changes mentioned above are the same (note that rvIndex may be adjusted). + +- the operation flag would additionally include the LDPC_HQ_COMBINE_IN_ENABLE flag. + +- the harq_combined_input.offset must set to the address of the related code block (ie. same as the harq_combine_output index above for the same code block, HARQ process, UE). + +- the harq_combined_input.length must be set to the length which was provided back in the related harq_combined_output.length when it has processed and dequeued (previous HARQ iteration). + + The output mbuf data structures are expected to be allocated by the application with enough room for the output data. -- 1.8.3.1
Re: [dpdk-dev] [PATCH v13 09/10] eal: add EAL argument for setting thread priority
On Thu, Aug 19, 2021 at 10:06:06AM +0100, Bruce Richardson wrote:
> On Wed, Aug 18, 2021 at 02:28:33PM -0700, Stephen Hemminger wrote:
> > On Tue, 3 Aug 2021 12:01:30 -0700 Narcisa Ana Maria Vasile
> > wrote:
> >
> > > +static int +eal_parse_thread_priority(const char *arg) +{ +
> > > struct internal_config *internal_conf = +
> > > eal_get_internal_configuration(); + enum rte_thread_priority
> > > priority;
> > > + + if (!strncmp("normal", arg, sizeof("normal"))) +
> > > priority = RTE_THREAD_PRIORITY_NORMAL; + else if
> > > (!strncmp("realtime", arg, sizeof("realtime"))) + priority =
> > > RTE_THREAD_PRIORITY_REALTIME_CRITICAL; + else + return -1;
> > > + + internal_conf->thread_priority = priority; +return 0; +} +
> >
> > In my experience using real time priority is dangerous and risks
> > starvation and deadlock. The problem is that DPDK applications are
> > typically 100% CPU poll mode with no system calls; but the kernel has a
> > number of worker threads that can be required on those CPUs.
> >
> > The typical failure is a disk completion interrupt happens on a CPU that
> > is being used by DPDK lcore thread. With RT priority, the kernel thread
> > to process that I/O completion never runs because the RT user thread has
> > higher priority than the kernel I/O completion thread.
> >
> > It maybe possible to workaround this with lots of hand crafting through
> > sysfs to reassign threads and irq's. Also, later kernels with full RT
> > might also help.
> >
> > Before putting this in as part of DPDK in EAL, a full set of testing and
> > documentation of how to configure these kind of applications and systems
> > is needed.
> >
> I would tend to agree caution here, based on my experience of having locked
> up a number of systems in the past when testing running DPDK apps with RT
> priority!
Thank you for the comments! I've added this option since it was requested by
multiple users. I understand RT priority causes issues on Linux platforms.
On Windows we want to be able to use REALTIME priority in certain scenarios.
Would it be acceptable to replace this option with a "HIGH_PRIORITY" one
and keep it realtime on Windows and choose a higher (but non-realtime) option
on Linux?
However, there are 2 issues here:
* We will have different behaviors between the 2 platforms.
* Not sure if I can set a normal but higher priority on Linux. SCHED_OTHER
only allows
one value and Linux "nice" values don't help. If anyone knows of a way to
accomplish
this on Linux, please do advise.
Alternatively, we can have this option for Windows only.
In the meantime, I've removed this patch from this patchset in v14 as the
cmdline option is not
being enabled yet, as DmitryK noted.
[dpdk-dev] [PATCH v14 0/9] eal: Add EAL API for threading
From: Narcisa Vasile
EAL thread API
**Problem Statement**
DPDK currently uses the pthread interface to create and manage threads.
Windows does not support the POSIX thread programming model,
so it currently
relies on a header file that hides the Windows calls under
pthread matched interfaces. Given that EAL should isolate the environment
specifics from the applications and libraries and mediate
all the communication with the operating systems, a new EAL interface
is needed for thread management.
**Goals**
* Introduce a generic EAL API for threading support that will remove
the current Windows pthread.h shim.
* Replace references to pthread_* across the DPDK codebase with the new
RTE_THREAD_* API.
* Allow users to choose between using the RTE_THREAD_* API or a
3rd party thread library through a configuration option.
**Design plan**
New API main files:
* rte_thread.h (librte_eal/include)
* rte_thread.c (librte_eal/windows)
* rte_thread.c (librte_eal/common)
**A schematic example of the design**
--
lib/librte_eal/include/rte_thread.h
int rte_thread_create();
lib/librte_eal/common/rte_thread.c
int rte_thread_create()
{
return pthread_create();
}
lib/librte_eal/windows/rte_thread.c
int rte_thread_create()
{
return CreateThread();
}
-
**Thread attributes**
When or after a thread is created, specific characteristics of the thread
can be adjusted. Given that the thread characteristics that are of interest
for DPDK applications are affinity and priority, the following structure
that represents thread attributes has been defined:
typedef struct
{
enum rte_thread_priority priority;
rte_cpuset_t cpuset;
} rte_thread_attr_t;
The *rte_thread_create()* function can optionally receive
an rte_thread_attr_t
object that will cause the thread to be created with the
affinity and priority
described by the attributes object. If no rte_thread_attr_t is passed
(parameter is NULL), the default affinity and priority are used.
An rte_thread_attr_t object can also be set to the default values
by calling *rte_thread_attr_init()*.
*Priority* is represented through an enum that currently advertises
two values for priority:
- RTE_THREAD_PRIORITY_NORMAL
- RTE_THREAD_PRIORITY_REALTIME_CRITICAL
The enum can be extended to allow for multiple priority levels.
rte_thread_set_priority - sets the priority of a thread
rte_thread_attr_set_priority - updates an rte_thread_attr_t object
with a new value for priority
*Affinity* is described by the already known “rte_cpuset_t” type.
rte_thread_attr_set/get_affinity - sets/gets the affinity field in a
rte_thread_attr_t object
rte_thread_set/get_affinity – sets/gets the affinity of a thread
**Errors**
A translation function that maps Windows error codes to errno-style
error codes is provided.
**Future work**
The long term plan is for EAL to provide full threading support:
* Add support for conditional variables
* Add support for pthread_mutex_trylock
* Additional functionality offered by pthread_*
(such as pthread_setname_np, etc.)
v14:
- Remove patch "eal: add EAL argument for setting thread priority"
This will be added later when enabling the new threading API.
- Remove priority enum value "_UNDEFINED". NORMAL is used
as the default.
- Fix issue with thread return value.
v13:
- Fix syntax error in unit tests
v12:
- Fix freebsd warning about initializer in unit tests
v11:
- Add unit tests for thread API
- Rebase
v10:
- Remove patch no. 10. It will be broken down in subpatches
and sent as a different patchset that depends on this one.
This is done due to the ABI breaks that would be caused by patch 10.
- Replace unix/rte_thread.c with common/rte_thread.c
- Remove initializations that may prevent compiler from issuing useful
warnings.
- Remove rte_thread_types.h and rte_windows_thread_types.h
- Remove unneeded priority macros (EAL_THREAD_PRIORITY*)
- Remove functions that retrieves thread handle from process handle
- Remove rte_thread_cancel() until same behavior is obtained on
all platforms.
- Fix rte_thread_detach() function description,
return value and remove empty line.
- Reimplement mutex functions. Add compatible representation for mutex
identifier. Add macro to replace static mutex initialization instances.
- Fix commit messages (lines too long, remove unicode symbols)
v9:
- Sign patches
v8:
- Rebase
- Add rte_thread_detach() API
- Set default priority, when user did not specify a value
v7:
Based on DmitryK's review:
- Change thread id representation
- Change mutex id representation
- Implement static mutex inititalizer for Windows
- Change barrier identifier representation
- Improve commit messages
- Add missing doxygen comments
- Split error translation function
- Improve name for affinity function
- Remo
[dpdk-dev] [PATCH v14 1/9] eal: add basic threading functions
From: Narcisa Vasile
Use a portable, type-safe representation for the thread identifier.
Add functions for comparing thread ids and obtaining the thread id
for the current thread.
Signed-off-by: Narcisa Vasile
---
lib/eal/common/meson.build| 1 +
lib/eal/{unix => common}/rte_thread.c | 57 ---
lib/eal/include/rte_thread.h | 48 +-
lib/eal/unix/meson.build | 1 -
lib/eal/version.map | 3 ++
lib/eal/windows/rte_thread.c | 17
6 files changed, 95 insertions(+), 32 deletions(-)
rename lib/eal/{unix => common}/rte_thread.c (66%)
diff --git a/lib/eal/common/meson.build b/lib/eal/common/meson.build
index edfca9..eda250247b 100644
--- a/lib/eal/common/meson.build
+++ b/lib/eal/common/meson.build
@@ -80,6 +80,7 @@ sources += files(
'rte_random.c',
'rte_reciprocal.c',
'rte_service.c',
+'rte_thread.c',
'rte_version.c',
)
diff --git a/lib/eal/unix/rte_thread.c b/lib/eal/common/rte_thread.c
similarity index 66%
rename from lib/eal/unix/rte_thread.c
rename to lib/eal/common/rte_thread.c
index c72d619ec1..92a7451b0a 100644
--- a/lib/eal/unix/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -1,5 +1,6 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright 2021 Mellanox Technologies, Ltd
+ * Copyright(c) 2021 Microsoft Corporation
*/
#include
@@ -16,25 +17,41 @@ struct eal_tls_key {
pthread_key_t thread_index;
};
+rte_thread_t
+rte_thread_self(void)
+{
+ rte_thread_t thread_id;
+
+ thread_id.opaque_id = (uintptr_t)pthread_self();
+
+ return thread_id;
+}
+
+int
+rte_thread_equal(rte_thread_t t1, rte_thread_t t2)
+{
+ return pthread_equal((pthread_t)t1.opaque_id, (pthread_t)t2.opaque_id);
+}
+
int
rte_thread_key_create(rte_thread_key *key, void (*destructor)(void *))
{
int err;
+ rte_thread_key k;
- *key = malloc(sizeof(**key));
- if ((*key) == NULL) {
+ k = malloc(sizeof(*k));
+ if (k == NULL) {
RTE_LOG(DEBUG, EAL, "Cannot allocate TLS key.\n");
- rte_errno = ENOMEM;
- return -1;
+ return EINVAL;
}
- err = pthread_key_create(&((*key)->thread_index), destructor);
- if (err) {
+ err = pthread_key_create(&(k->thread_index), destructor);
+ if (err != 0) {
RTE_LOG(DEBUG, EAL, "pthread_key_create failed: %s\n",
strerror(err));
- free(*key);
- rte_errno = ENOEXEC;
- return -1;
+ free(k);
+ return err;
}
+ *key = k;
return 0;
}
@@ -43,18 +60,16 @@ rte_thread_key_delete(rte_thread_key key)
{
int err;
- if (!key) {
+ if (key == NULL) {
RTE_LOG(DEBUG, EAL, "Invalid TLS key.\n");
- rte_errno = EINVAL;
- return -1;
+ return EINVAL;
}
err = pthread_key_delete(key->thread_index);
- if (err) {
+ if (err != 0) {
RTE_LOG(DEBUG, EAL, "pthread_key_delete failed: %s\n",
strerror(err));
free(key);
- rte_errno = ENOEXEC;
- return -1;
+ return err;
}
free(key);
return 0;
@@ -65,17 +80,15 @@ rte_thread_value_set(rte_thread_key key, const void *value)
{
int err;
- if (!key) {
+ if (key == NULL) {
RTE_LOG(DEBUG, EAL, "Invalid TLS key.\n");
- rte_errno = EINVAL;
- return -1;
+ return EINVAL;
}
err = pthread_setspecific(key->thread_index, value);
- if (err) {
+ if (err != 0) {
RTE_LOG(DEBUG, EAL, "pthread_setspecific failed: %s\n",
strerror(err));
- rte_errno = ENOEXEC;
- return -1;
+ return err;
}
return 0;
}
@@ -83,7 +96,7 @@ rte_thread_value_set(rte_thread_key key, const void *value)
void *
rte_thread_value_get(rte_thread_key key)
{
- if (!key) {
+ if (key == NULL) {
RTE_LOG(DEBUG, EAL, "Invalid TLS key.\n");
rte_errno = EINVAL;
return NULL;
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 8be8ed8f36..748f64d230 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -1,6 +1,8 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(c) 2021 Mellanox Technologies, Ltd
+ * Copyright(c) 2021 Microsoft Corporation
*/
+#include
#include
#include
@@ -20,11 +22,45 @@
extern "C" {
#endif
+#include
+
+/**
+ * Thread id descriptor.
+ */
+typedef struct rte_thread_tag {
+ uintptr_t opaque_id; /**< thread identifier */
+} rte_thread_t;
+
/**
* TLS key type, an opaque pointer.
*/
typedef struct eal_tls_key *
[dpdk-dev] [PATCH v14 2/9] eal: add thread attributes
From: Narcisa Vasile
Implement thread attributes for:
* thread affinity
* thread priority
Implement functions for managing thread attributes.
Priority is represented through an enum that allows for two levels:
- RTE_THREAD_PRIORITY_NORMAL
- RTE_THREAD_PRIORITY_REALTIME_CRITICAL
Affinity is described by the rte_cpuset_t type.
An rte_thread_attr_t object can be set to the default values
by calling rte_thread_attr_init().
Signed-off-by: Narcisa Vasile
---
lib/eal/common/rte_thread.c | 46 ++
lib/eal/include/rte_thread.h | 91
lib/eal/version.map | 4 ++
lib/eal/windows/rte_thread.c | 44 +
4 files changed, 185 insertions(+)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index 92a7451b0a..e1a4d7eae4 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -9,6 +9,7 @@
#include
#include
+#include
#include
#include
#include
@@ -33,6 +34,51 @@ rte_thread_equal(rte_thread_t t1, rte_thread_t t2)
return pthread_equal((pthread_t)t1.opaque_id, (pthread_t)t2.opaque_id);
}
+int
+rte_thread_attr_init(rte_thread_attr_t *attr)
+{
+ RTE_VERIFY(attr != NULL);
+
+ CPU_ZERO(&attr->cpuset);
+ attr->priority = RTE_THREAD_PRIORITY_NORMAL;
+
+ return 0;
+}
+
+int
+rte_thread_attr_set_affinity(rte_thread_attr_t *thread_attr,
+rte_cpuset_t *cpuset)
+{
+ RTE_VERIFY(thread_attr != NULL);
+ RTE_VERIFY(cpuset != NULL);
+
+ thread_attr->cpuset = *cpuset;
+
+ return 0;
+}
+
+int
+rte_thread_attr_get_affinity(rte_thread_attr_t *thread_attr,
+rte_cpuset_t *cpuset)
+{
+ RTE_VERIFY(thread_attr != NULL);
+ RTE_VERIFY(cpuset != NULL);
+
+ *cpuset = thread_attr->cpuset;
+
+ return 0;
+}
+
+int
+rte_thread_attr_set_priority(rte_thread_attr_t *thread_attr,
+enum rte_thread_priority priority)
+{
+ RTE_VERIFY(thread_attr != NULL);
+
+ thread_attr->priority = priority;
+ return 0;
+}
+
int
rte_thread_key_create(rte_thread_key *key, void (*destructor)(void *))
{
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 748f64d230..4ac36957ce 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -31,6 +31,28 @@ typedef struct rte_thread_tag {
uintptr_t opaque_id; /**< thread identifier */
} rte_thread_t;
+/**
+ * Thread priority values.
+ */
+enum rte_thread_priority {
+ RTE_THREAD_PRIORITY_NORMAL= 0,
+ /**< normal thread priority, the default */
+ RTE_THREAD_PRIORITY_REALTIME_CRITICAL = 1,
+ /**< highest thread priority allowed */
+};
+
+#ifdef RTE_HAS_CPUSET
+
+/**
+ * Representation for thread attributes.
+ */
+typedef struct {
+ enum rte_thread_priority priority; /**< thread priority */
+ rte_cpuset_t cpuset; /**< thread affinity */
+} rte_thread_attr_t;
+
+#endif /* RTE_HAS_CPUSET */
+
/**
* TLS key type, an opaque pointer.
*/
@@ -63,6 +85,75 @@ int rte_thread_equal(rte_thread_t t1, rte_thread_t t2);
#ifdef RTE_HAS_CPUSET
+/**
+ * Initialize the attributes of a thread.
+ * These attributes can be passed to the rte_thread_create() function
+ * that will create a new thread and set its attributes according to attr.
+ *
+ * @param attr
+ * Thread attributes to initialize.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_attr_init(rte_thread_attr_t *attr);
+
+/**
+ * Set the CPU affinity value in the thread attributes pointed to
+ * by 'thread_attr'.
+ *
+ * @param thread_attr
+ * Points to the thread attributes in which affinity will be updated.
+ *
+ * @param cpuset
+ * Points to the value of the affinity to be set.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_attr_set_affinity(rte_thread_attr_t *thread_attr,
+ rte_cpuset_t *cpuset);
+
+/**
+ * Get the value of CPU affinity that is set in the thread attributes pointed
+ * to by 'thread_attr'.
+ *
+ * @param thread_attr
+ * Points to the thread attributes from which affinity will be retrieved.
+ *
+ * @param cpuset
+ * Pointer to the memory that will store the affinity.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_attr_get_affinity(rte_thread_attr_t *thread_attr,
+ rte_cpuset_t *cpuset);
+
+/**
+ * Set the thread priority value in the thread attributes pointed to
+ * by 'thread_attr'.
+ *
+ * @param thread_attr
+ * Points to the thread attributes in which priority will be updated.
+ *
+ * @param priority
+ * Points to the value of the priority to be set.
+ *
+ * @return
+ * On success, return 0.
+ * O
[dpdk-dev] [PATCH v14 4/9] eal: implement functions for thread affinity management
From: Narcisa Vasile
Implement functions for getting/setting thread affinity.
Threads can be pinned to specific cores by setting their
affinity attribute.
Signed-off-by: Narcisa Vasile
Signed-off-by: Dmitry Malloy
---
lib/eal/common/rte_thread.c | 16
lib/eal/include/rte_thread.h | 36 +++
lib/eal/version.map | 2 +
lib/eal/windows/eal_lcore.c | 176 +-
lib/eal/windows/eal_windows.h | 10 ++
lib/eal/windows/rte_thread.c | 125 +++-
6 files changed, 319 insertions(+), 46 deletions(-)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index 27ad1c7eb0..73b7b3141c 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -34,6 +34,22 @@ rte_thread_equal(rte_thread_t t1, rte_thread_t t2)
return pthread_equal((pthread_t)t1.opaque_id, (pthread_t)t2.opaque_id);
}
+int
+rte_thread_set_affinity_by_id(rte_thread_t thread_id,
+ const rte_cpuset_t *cpuset)
+{
+ return pthread_setaffinity_np((pthread_t)thread_id.opaque_id,
+ sizeof(*cpuset), cpuset);
+}
+
+int
+rte_thread_get_affinity_by_id(rte_thread_t thread_id,
+ rte_cpuset_t *cpuset)
+{
+ return pthread_getaffinity_np((pthread_t)thread_id.opaque_id,
+ sizeof(*cpuset), cpuset);
+}
+
int
rte_thread_attr_init(rte_thread_attr_t *attr)
{
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 8a20215a94..5b100cafda 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -85,6 +85,42 @@ int rte_thread_equal(rte_thread_t t1, rte_thread_t t2);
#ifdef RTE_HAS_CPUSET
+/**
+ * Set the affinity of thread 'thread_id' to the cpu set
+ * specified by 'cpuset'.
+ *
+ * @param thread_id
+ *Id of the thread for which to set the affinity.
+ *
+ * @param cpuset
+ * Pointer to CPU affinity to set.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_set_affinity_by_id(rte_thread_t thread_id,
+ const rte_cpuset_t *cpuset);
+
+/**
+ * Get the affinity of thread 'thread_id' and store it
+ * in 'cpuset'.
+ *
+ * @param thread_id
+ *Id of the thread for which to get the affinity.
+ *
+ * @param cpuset
+ * Pointer for storing the affinity value.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_get_affinity_by_id(rte_thread_t thread_id,
+ rte_cpuset_t *cpuset);
+
/**
* Initialize the attributes of a thread.
* These attributes can be passed to the rte_thread_create() function
diff --git a/lib/eal/version.map b/lib/eal/version.map
index b487ab5975..faf0361d37 100644
--- a/lib/eal/version.map
+++ b/lib/eal/version.map
@@ -433,6 +433,8 @@ EXPERIMENTAL {
rte_thread_attr_get_affinity;
rte_thread_attr_set_affinity;
rte_thread_attr_set_priority;
+ rte_thread_get_affinity_by_id;
+ rte_thread_set_affinity_by_id;
};
INTERNAL {
diff --git a/lib/eal/windows/eal_lcore.c b/lib/eal/windows/eal_lcore.c
index 476c2d2bdf..295af50698 100644
--- a/lib/eal/windows/eal_lcore.c
+++ b/lib/eal/windows/eal_lcore.c
@@ -2,7 +2,6 @@
* Copyright(c) 2019 Intel Corporation
*/
-#include
#include
#include
@@ -27,13 +26,15 @@ struct socket_map {
};
struct cpu_map {
- unsigned int socket_count;
unsigned int lcore_count;
+ unsigned int socket_count;
+ unsigned int cpu_count;
struct lcore_map lcores[RTE_MAX_LCORE];
struct socket_map sockets[RTE_MAX_NUMA_NODES];
+ GROUP_AFFINITY cpus[CPU_SETSIZE];
};
-static struct cpu_map cpu_map = { 0 };
+static struct cpu_map cpu_map;
/* eal_create_cpu_map() is called before logging is initialized */
static void
@@ -47,13 +48,118 @@ log_early(const char *format, ...)
va_end(va);
}
+static int
+eal_query_group_affinity(void)
+{
+ SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *infos = NULL;
+ unsigned int *cpu_count = &cpu_map.cpu_count;
+ DWORD infos_size = 0;
+ int ret = 0;
+ USHORT group_count;
+ KAFFINITY affinity;
+ USHORT group_no;
+ unsigned int i;
+
+ if (!GetLogicalProcessorInformationEx(RelationGroup, NULL,
+ &infos_size)) {
+ DWORD error = GetLastError();
+ if (error != ERROR_INSUFFICIENT_BUFFER) {
+ log_early("Cannot get group information size, "
+ "error %lu\n", error);
+ rte_errno = EINVAL;
+ ret = -1;
+ goto cleanup;
+ }
+ }
+
+ infos = malloc(infos_size);
+ if (infos == NULL) {
+ log_early("Cannot allocate memory for NUMA node information\n");
+ rte_errno = ENOMEM;
+
[dpdk-dev] [PATCH v14 3/9] eal/windows: translate Windows errors to errno-style errors
From: Narcisa Vasile
Add function to translate Windows error codes to
errno-style error codes. The possible return values are chosen
so that we have as much semantical compatibility between platforms as
possible.
Signed-off-by: Narcisa Vasile
---
lib/eal/common/rte_thread.c | 6 +--
lib/eal/include/rte_thread.h | 5 +-
lib/eal/windows/rte_thread.c | 95 +++-
3 files changed, 76 insertions(+), 30 deletions(-)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index e1a4d7eae4..27ad1c7eb0 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -47,7 +47,7 @@ rte_thread_attr_init(rte_thread_attr_t *attr)
int
rte_thread_attr_set_affinity(rte_thread_attr_t *thread_attr,
-rte_cpuset_t *cpuset)
+ rte_cpuset_t *cpuset)
{
RTE_VERIFY(thread_attr != NULL);
RTE_VERIFY(cpuset != NULL);
@@ -59,7 +59,7 @@ rte_thread_attr_set_affinity(rte_thread_attr_t *thread_attr,
int
rte_thread_attr_get_affinity(rte_thread_attr_t *thread_attr,
-rte_cpuset_t *cpuset)
+ rte_cpuset_t *cpuset)
{
RTE_VERIFY(thread_attr != NULL);
RTE_VERIFY(cpuset != NULL);
@@ -71,7 +71,7 @@ rte_thread_attr_get_affinity(rte_thread_attr_t *thread_attr,
int
rte_thread_attr_set_priority(rte_thread_attr_t *thread_attr,
-enum rte_thread_priority priority)
+ enum rte_thread_priority priority)
{
RTE_VERIFY(thread_attr != NULL);
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 4ac36957ce..8a20215a94 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -233,9 +233,8 @@ int rte_thread_value_set(rte_thread_key key, const void
*value);
*
* @return
* On success, value data pointer (can also be NULL).
- * On failure, NULL and an error number is set in rte_errno.
- * rte_errno can be: EINVAL - Invalid parameter passed.
- * ENOEXEC - Specific OS error.
+ * On failure, NULL and a positive error number is set in rte_errno.
+ *
*/
__rte_experimental
void *rte_thread_value_get(rte_thread_key key);
diff --git a/lib/eal/windows/rte_thread.c b/lib/eal/windows/rte_thread.c
index 01966e7745..c1ecfbd6ae 100644
--- a/lib/eal/windows/rte_thread.c
+++ b/lib/eal/windows/rte_thread.c
@@ -13,6 +13,54 @@ struct eal_tls_key {
DWORD thread_index;
};
+/* Translates the most common error codes related to threads */
+static int
+thread_translate_win32_error(DWORD error)
+{
+ switch (error) {
+ case ERROR_SUCCESS:
+ return 0;
+
+ case ERROR_INVALID_PARAMETER:
+ return EINVAL;
+
+ case ERROR_INVALID_HANDLE:
+ return EFAULT;
+
+ case ERROR_NOT_ENOUGH_MEMORY:
+ /* FALLTHROUGH */
+ case ERROR_NO_SYSTEM_RESOURCES:
+ return ENOMEM;
+
+ case ERROR_PRIVILEGE_NOT_HELD:
+ /* FALLTHROUGH */
+ case ERROR_ACCESS_DENIED:
+ return EACCES;
+
+ case ERROR_ALREADY_EXISTS:
+ return EEXIST;
+
+ case ERROR_POSSIBLE_DEADLOCK:
+ return EDEADLK;
+
+ case ERROR_INVALID_FUNCTION:
+ /* FALLTHROUGH */
+ case ERROR_CALL_NOT_IMPLEMENTED:
+ return ENOSYS;
+ }
+
+ return EINVAL;
+}
+
+static int
+thread_log_last_error(const char *message)
+{
+ DWORD error = GetLastError();
+ RTE_LOG(DEBUG, EAL, "GetLastError()=%lu: %s\n", error, message);
+
+ return thread_translate_win32_error(error);
+}
+
rte_thread_t
rte_thread_self(void)
{
@@ -42,7 +90,7 @@ rte_thread_attr_init(rte_thread_attr_t *attr)
int
rte_thread_attr_set_affinity(rte_thread_attr_t *thread_attr,
-rte_cpuset_t *cpuset)
+ rte_cpuset_t *cpuset)
{
RTE_VERIFY(thread_attr != NULL);
thread_attr->cpuset = *cpuset;
@@ -52,7 +100,7 @@ rte_thread_attr_set_affinity(rte_thread_attr_t *thread_attr,
int
rte_thread_attr_get_affinity(rte_thread_attr_t *thread_attr,
-rte_cpuset_t *cpuset)
+ rte_cpuset_t *cpuset)
{
RTE_VERIFY(thread_attr != NULL);
@@ -63,7 +111,7 @@ rte_thread_attr_get_affinity(rte_thread_attr_t *thread_attr,
int
rte_thread_attr_set_priority(rte_thread_attr_t *thread_attr,
-enum rte_thread_priority priority)
+ enum rte_thread_priority priority)
{
RTE_VERIFY(thread_attr != NULL);
@@ -76,18 +124,18 @@ int
rte_thread_key_create(rte_thread_key *key,
__rte_unused void (*destructor)(void *))
{
+ int ret;
+
*key = malloc(sizeof(**key));
if ((*key) == NULL) {
RTE_LOG(DEBUG, EAL, "Cannot allocate TLS key.\n");
- rte_errno = ENOMEM;
- return -1;
+ return ENOMEM;
}
(*key)->thread_index = TlsAlloc();
[dpdk-dev] [PATCH v14 6/9] eal: add thread lifetime management
From: Narcisa Vasile
Add functions for thread creation, joining, detaching.
The *rte_thread_create()* function can optionally receive
an rte_thread_attr_t object that will cause the thread to be
created with the affinity and priority described by the
attributes object. If no rte_thread_attr_t is passed (parameter is NULL),
the default affinity and priority are used.
On Windows, the function executed by a thread when the thread starts is
represeneted by a function pointer of type DWORD (*func) (void*).
On other platforms, the function pointer is a void* (*func) (void*).
Performing a cast between these two types of function pointers to
uniformize the API on all platforms may result in undefined behavior.
TO fix this issue, a wrapper that respects the signature required by
CreateThread() has been created on Windows.
Signed-off-by: Narcisa Vasile
---
lib/eal/common/rte_thread.c | 103
lib/eal/include/rte_thread.h| 55 +
lib/eal/version.map | 3 +
lib/eal/windows/include/sched.h | 2 +-
lib/eal/windows/rte_thread.c| 134
5 files changed, 296 insertions(+), 1 deletion(-)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index fcebf7097c..39bb35ae23 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -144,6 +144,109 @@ rte_thread_attr_set_priority(rte_thread_attr_t
*thread_attr,
return 0;
}
+int
+rte_thread_create(rte_thread_t *thread_id,
+ const rte_thread_attr_t *thread_attr,
+ rte_thread_func thread_func, void *args)
+{
+ int ret = 0;
+ pthread_attr_t attr;
+ pthread_attr_t *attrp = NULL;
+ struct sched_param param = {
+ .sched_priority = 0,
+ };
+ int policy = SCHED_OTHER;
+
+ if (thread_attr != NULL) {
+ ret = pthread_attr_init(&attr);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "pthread_attr_init failed\n");
+ goto cleanup;
+ }
+
+ attrp = &attr;
+
+ /*
+* Set the inherit scheduler parameter to explicit,
+* otherwise the priority attribute is ignored.
+*/
+ ret = pthread_attr_setinheritsched(attrp,
+ PTHREAD_EXPLICIT_SCHED);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "pthread_attr_setinheritsched
failed\n");
+ goto cleanup;
+ }
+
+ ret = thread_map_priority_to_os_value(thread_attr->priority,
+ ¶m.sched_priority, &policy);
+ if (ret != 0)
+ goto cleanup;
+
+ ret = pthread_attr_setschedpolicy(attrp, policy);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "pthread_attr_setschedpolicy
failed\n");
+ goto cleanup;
+ }
+
+ ret = pthread_attr_setschedparam(attrp, ¶m);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "pthread_attr_setschedparam
failed\n");
+ goto cleanup;
+ }
+
+ if (CPU_COUNT(&thread_attr->cpuset) > 0) {
+ ret = pthread_attr_setaffinity_np(attrp,
+ sizeof(thread_attr->cpuset),
+ &thread_attr->cpuset);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL,
"pthread_attr_setaffinity_np failed\n");
+ goto cleanup;
+ }
+ }
+ }
+
+ ret = pthread_create((pthread_t *)&thread_id->opaque_id, attrp,
+ thread_func, args);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "pthread_create failed\n");
+ goto cleanup;
+ }
+
+cleanup:
+ if (attrp != NULL)
+ pthread_attr_destroy(&attr);
+
+ return ret;
+}
+
+int
+rte_thread_join(rte_thread_t thread_id, unsigned long *value_ptr)
+{
+ int ret = 0;
+ void *res = NULL;
+ void **pres = NULL;
+
+ if (value_ptr != NULL)
+ pres = &res;
+
+ ret = pthread_join((pthread_t)thread_id.opaque_id, pres);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "pthread_join failed\n");
+ return ret;
+ }
+
+ if (value_ptr != NULL && *pres != NULL)
+ *value_ptr = *(unsigned long *)(*pres);
+
+ return 0;
+}
+
+int
+rte_thread_detach(rte_thread_t thread_id)
+{
+ return pthread_detach((pthread_t)thread_id.opaque_id);
+}
+
int
rte_thread_key_create(rte_thread_key *key, void (*destructor)(void *))
{
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index a9cb6ec81c..8f36cf06e0 100644
--- a/lib/eal/include/rte_thread.h
+++ b/li
[dpdk-dev] [PATCH v14 5/9] eal: implement thread priority management functions
From: Narcisa Vasile
Add function for setting the priority for a thread.
Priorities on multiple platforms are similarly determined by
a priority value and a priority class/policy.
On Linux, the following mapping is created:
RTE_THREAD_PRIORITY_NORMAL corresponds to
* policy SCHED_OTHER
* priority value: (sched_get_priority_min(SCHED_OTHER) +
sched_get_priority_max(SCHED_OTHER))/2;
RTE_THREAD_PRIORITY_REALTIME_CRITICAL corresponds to
* policy SCHED_RR
* priority value: sched_get_priority_max(SCHED_RR);
On Windows, the following mapping is created:
RTE_THREAD_PRIORITY_NORMAL corresponds to
* class NORMAL_PRIORITY_CLASS
* priority THREAD_PRIORITY_NORMAL
RTE_THREAD_PRIORITY_REALTIME_CRITICAL corresponds to
* class REALTIME_PRIORITY_CLASS
* priority THREAD_PRIORITY_TIME_CRITICAL
Signed-off-by: Narcisa Vasile
---
lib/eal/common/rte_thread.c | 49 ++
lib/eal/include/rte_thread.h | 17 ++
lib/eal/version.map | 1 +
lib/eal/windows/rte_thread.c | 66
4 files changed, 133 insertions(+)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index 73b7b3141c..fcebf7097c 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -50,6 +50,55 @@ rte_thread_get_affinity_by_id(rte_thread_t thread_id,
sizeof(*cpuset), cpuset);
}
+static int
+thread_map_priority_to_os_value(enum rte_thread_priority eal_pri,
+ int *os_pri, int *pol)
+{
+ /* Clear the output parameters */
+ *os_pri = sched_get_priority_min(SCHED_OTHER) - 1;
+ *pol = -1;
+
+ switch (eal_pri) {
+ case RTE_THREAD_PRIORITY_NORMAL:
+ *pol = SCHED_OTHER;
+
+ /*
+* Choose the middle of the range to represent
+* the priority 'normal'.
+* On Linux, this should be 0, since both
+* sched_get_priority_min/_max return 0 for SCHED_OTHER.
+*/
+ *os_pri = (sched_get_priority_min(SCHED_OTHER) +
+ sched_get_priority_max(SCHED_OTHER))/2;
+ break;
+ case RTE_THREAD_PRIORITY_REALTIME_CRITICAL:
+ *pol = SCHED_RR;
+ *os_pri = sched_get_priority_max(SCHED_RR);
+ break;
+ default:
+ RTE_LOG(DEBUG, EAL, "The requested priority value is
invalid.\n");
+ return EINVAL;
+ }
+ return 0;
+}
+
+int
+rte_thread_set_priority(rte_thread_t thread_id,
+ enum rte_thread_priority priority)
+{
+ int ret;
+ int policy;
+ struct sched_param param;
+
+ ret = thread_map_priority_to_os_value(priority, ¶m.sched_priority,
+ &policy);
+ if (ret != 0)
+ return ret;
+
+ return pthread_setschedparam((pthread_t)thread_id.opaque_id,
+ policy, ¶m);
+}
+
int
rte_thread_attr_init(rte_thread_attr_t *attr)
{
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 5b100cafda..a9cb6ec81c 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -213,6 +213,23 @@ void rte_thread_get_affinity(rte_cpuset_t *cpusetp);
#endif /* RTE_HAS_CPUSET */
+/**
+ * Set the priority of a thread.
+ *
+ * @param thread_id
+ *Id of the thread for which to set priority.
+ *
+ * @param priority
+ * Priority value to be set.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_set_priority(rte_thread_t thread_id,
+ enum rte_thread_priority priority);
+
/**
* Create a TLS data key visible to all threads in the process.
* the created key is later used to get/set a value.
diff --git a/lib/eal/version.map b/lib/eal/version.map
index faf0361d37..790b5112fb 100644
--- a/lib/eal/version.map
+++ b/lib/eal/version.map
@@ -435,6 +435,7 @@ EXPERIMENTAL {
rte_thread_attr_set_priority;
rte_thread_get_affinity_by_id;
rte_thread_set_affinity_by_id;
+ rte_thread_set_priority;
};
INTERNAL {
diff --git a/lib/eal/windows/rte_thread.c b/lib/eal/windows/rte_thread.c
index 0127119f49..fb04718f58 100644
--- a/lib/eal/windows/rte_thread.c
+++ b/lib/eal/windows/rte_thread.c
@@ -200,6 +200,72 @@ rte_thread_get_affinity_by_id(rte_thread_t thread_id,
return ret;
}
+static int
+thread_map_priority_to_os_value(enum rte_thread_priority eal_pri,
+ int *os_pri, int *pri_class)
+{
+ /* Clear the output parameters */
+ *os_pri = -1;
+ *pri_class = -1;
+
+ switch (eal_pri) {
+ case RTE_THREAD_PRIORITY_NORMAL:
+ *pri_class = NORMAL_PRIORITY_CLASS;
+ *os_pri = THREAD_PRIORITY_NORMAL;
+ break;
+ case RTE_THREAD_PRIORITY_REALTIME_CRITICAL:
+ *pri_class = REALTIME_PRIORITY_CLASS;
+ *os_pri = THREAD_PRIORITY
[dpdk-dev] [PATCH v14 9/9] Add unit tests for thread API
From: Narcisa Vasile
As a new API for threading is introduced,
a set of unit tests have been added to test the new interface.
Signed-off-by: Narcisa Vasile
---
app/test/meson.build| 2 +
app/test/test_threads.c | 419
2 files changed, 421 insertions(+)
create mode 100644 app/test/test_threads.c
diff --git a/app/test/meson.build b/app/test/meson.build
index a7611686ad..57e61ce601 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -140,6 +140,7 @@ test_sources = files(
'test_table_tables.c',
'test_tailq.c',
'test_thash.c',
+'test_threads.c',
'test_timer.c',
'test_timer_perf.c',
'test_timer_racecond.c',
@@ -276,6 +277,7 @@ fast_tests = [
['reorder_autotest', true],
['service_autotest', true],
['thash_autotest', true],
+['threads_autotest', true],
['trace_autotest', true],
]
diff --git a/app/test/test_threads.c b/app/test/test_threads.c
new file mode 100644
index 00..beaa303506
--- /dev/null
+++ b/app/test/test_threads.c
@@ -0,0 +1,419 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright (c) 2021 Microsoft.
+ */
+
+#include
+
+#include
+
+#include "test.h"
+
+#define THREADS_COUNT 20
+
+#define TEST_THREADS_LOG(func) \
+ printf("Error at line %d. %s failed!\n", __LINE__, func)
+
+static void *
+thread_loop_self(void *arg)
+{
+ rte_thread_t *id = arg;
+
+ *id = rte_thread_self();
+
+ return NULL;
+}
+
+static int
+test_thread_self(void)
+{
+ rte_thread_t threads_ids[THREADS_COUNT];
+ rte_thread_t self_ids[THREADS_COUNT] = {};
+ size_t i;
+ size_t j;
+ int ret = 0;
+
+ for (i = 0; i < THREADS_COUNT; ++i) {
+ if (rte_thread_create(&threads_ids[i], NULL, thread_loop_self,
+ &self_ids[i]) != 0) {
+ printf("Error, Only %zu threads created.\n", i);
+ break;
+ }
+ }
+
+ for (j = 0; j < i; ++j) {
+ ret = rte_thread_join(threads_ids[j], NULL);
+ if (ret != 0) {
+ TEST_THREADS_LOG("rte_thread_join()");
+ return -1;
+ }
+
+ if (rte_thread_equal(threads_ids[j], self_ids[j]) == 0)
+ ret = -1;
+ }
+
+ return ret;
+}
+
+struct thread_context {
+ rte_thread_barrier *barrier;
+ size_t *thread_count;
+};
+
+static void *
+thread_loop_barrier(void *arg)
+{
+
+ struct thread_context *ctx = arg;
+
+ (void)__atomic_add_fetch(ctx->thread_count, 1, __ATOMIC_RELAXED);
+
+ if (rte_thread_barrier_wait(ctx->barrier) > 0)
+ TEST_THREADS_LOG("rte_thread_barrier_wait()");
+
+ return NULL;
+}
+
+static int
+test_thread_barrier(void)
+{
+ rte_thread_t threads_ids[THREADS_COUNT];
+ struct thread_context ctx[THREADS_COUNT] = {};
+ rte_thread_barrier barrier;
+ size_t count = 0;
+ size_t i;
+ size_t j;
+ int ret = 0;
+
+ ret = rte_thread_barrier_init(&barrier, THREADS_COUNT + 1);
+ if (ret != 0) {
+ TEST_THREADS_LOG("rte_thread_barrier_init()");
+ return -1;
+ }
+
+ for (i = 0; i < THREADS_COUNT; ++i) {
+ ctx[i].thread_count = &count;
+ ctx[i].barrier = &barrier;
+ if (rte_thread_create(&threads_ids[i], NULL,
+ thread_loop_barrier, &ctx[i]) != 0) {
+ printf("Error, Only %zu threads created.\n", i);
+ ret = -1;
+ goto error;
+ }
+ }
+
+ ret = rte_thread_barrier_wait(ctx->barrier);
+ if (ret > 0) {
+ TEST_THREADS_LOG("rte_thread_barrier_wait()");
+ ret = -1;
+ goto error;
+ }
+
+ if (count != i) {
+ ret = -1;
+ printf("Error, expected thread count(%zu) to be equal "
+ "to the number of threads that wait at the
barrier(%zu)\n",
+ count, i);
+ goto error;
+ }
+
+error:
+ for (j = 0; j < i; ++j) {
+ ret = rte_thread_join(threads_ids[j], NULL);
+ if (ret != 0) {
+ TEST_THREADS_LOG("rte_thread_join()");
+ ret = -1;
+ break;
+ }
+ }
+
+ ret = rte_thread_barrier_destroy(&barrier);
+ if (ret != 0) {
+ TEST_THREADS_LOG("rte_thread_barrier_destroy()");
+ ret = -1;
+ }
+
+ return ret;
+}
+
+static size_t val;
+
+static void *
+thread_loop_mutex(void *arg)
+{
+ rte_thread_mutex *mutex = arg;
+
+ rte_thread_mutex_lock(mutex);
+ val++;
+ rte_thread_mutex_unlock(mutex);
+
+ return NULL;
+}
+
+static int
+test_threa
[dpdk-dev] [PATCH v14 8/9] eal: implement functions for thread barrier management
From: Narcisa Vasile
Add functions for barrier init, destroy, wait.
A portable type is used to represent a barrier identifier.
The rte_thread_barrier_wait() function returns the same value
on all platforms.
Signed-off-by: Narcisa Vasile
---
lib/eal/common/rte_thread.c | 61
lib/eal/include/rte_thread.h | 58 ++
lib/eal/version.map | 3 ++
lib/eal/windows/rte_thread.c | 56 +
4 files changed, 178 insertions(+)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index 05c235c215..ad27ff7d86 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -308,6 +308,67 @@ rte_thread_mutex_destroy(rte_thread_mutex *mutex)
return ret;
}
+int
+rte_thread_barrier_init(rte_thread_barrier *barrier, int count)
+{
+ int ret = 0;
+ pthread_barrier_t *pthread_barrier = NULL;
+
+ RTE_VERIFY(barrier != NULL);
+ RTE_VERIFY(count > 0);
+
+ pthread_barrier = calloc(1, sizeof(*pthread_barrier));
+ if (pthread_barrier == NULL) {
+ RTE_LOG(DEBUG, EAL, "Unable to initialize barrier. Insufficient
memory!\n");
+ ret = ENOMEM;
+ goto cleanup;
+ }
+ ret = pthread_barrier_init(pthread_barrier, NULL, count);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "Failed to init barrier, ret = %d\n", ret);
+ goto cleanup;
+ }
+
+ barrier->barrier_id = pthread_barrier;
+ pthread_barrier = NULL;
+
+cleanup:
+ free(pthread_barrier);
+ return ret;
+}
+
+int
+rte_thread_barrier_wait(rte_thread_barrier *barrier)
+{
+ int ret = 0;
+
+ RTE_VERIFY(barrier != NULL);
+ RTE_VERIFY(barrier->barrier_id != NULL);
+
+ ret = pthread_barrier_wait(barrier->barrier_id);
+ if (ret == PTHREAD_BARRIER_SERIAL_THREAD)
+ ret = RTE_THREAD_BARRIER_SERIAL_THREAD;
+
+ return ret;
+}
+
+int
+rte_thread_barrier_destroy(rte_thread_barrier *barrier)
+{
+ int ret = 0;
+
+ RTE_VERIFY(barrier != NULL);
+
+ ret = pthread_barrier_destroy(barrier->barrier_id);
+ if (ret != 0)
+ RTE_LOG(DEBUG, EAL, "Failed to destroy barrier: %d\n", ret);
+
+ free(barrier->barrier_id);
+ barrier->barrier_id = NULL;
+
+ return ret;
+}
+
int
rte_thread_key_create(rte_thread_key *key, void (*destructor)(void *))
{
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 8015a984b8..b2aec28329 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -74,6 +74,18 @@ typedef struct rte_thread_mutex_tag {
void *mutex_id; /**< mutex identifier */
} rte_thread_mutex;
+/**
+ * Returned by rte_thread_barrier_wait() when call is successful.
+ */
+#define RTE_THREAD_BARRIER_SERIAL_THREAD -1
+
+/**
+ * Thread barrier representation.
+ */
+typedef struct rte_thread_barrier_tag {
+ void *barrier_id; /**< barrrier identifier */
+} rte_thread_barrier;
+
/**
* TLS key type, an opaque pointer.
*/
@@ -379,6 +391,52 @@ int rte_thread_mutex_unlock(rte_thread_mutex *mutex);
__rte_experimental
int rte_thread_mutex_destroy(rte_thread_mutex *mutex);
+/**
+ * Initializes a synchronization barrier.
+ *
+ * @param barrier
+ *A pointer that references the newly created 'barrier' object.
+ *
+ * @param count
+ *The number of threads that must enter the barrier before
+ *the threads can continue execution.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_barrier_init(rte_thread_barrier *barrier, int count);
+
+/**
+ * Causes the calling thread to wait at the synchronization barrier 'barrier'.
+ *
+ * @param barrier
+ *The barrier used for synchronizing the threads.
+ *
+ * @return
+ * Return RTE_THREAD_BARRIER_SERIAL_THREAD for the thread synchronized
+ * at the barrier.
+ * Return 0 for all other threads.
+ * Return a positive errno-style error number, in case of failure.
+ */
+__rte_experimental
+int rte_thread_barrier_wait(rte_thread_barrier *barrier);
+
+/**
+ * Releases all resources used by a synchronization barrier
+ * and uninitializes it.
+ *
+ * @param barrier
+ *The barrier to be destroyed.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_barrier_destroy(rte_thread_barrier *barrier);
+
/**
* Create a TLS data key visible to all threads in the process.
* the created key is later used to get/set a value.
diff --git a/lib/eal/version.map b/lib/eal/version.map
index fe6e4553de..541dc13053 100644
--- a/lib/eal/version.map
+++ b/lib/eal/version.map
@@ -443,6 +443,9 @@ EXPERIMENTAL {
rte_thread_mutex_lock;
rte_thread_mutex_unlock;
rte_thread_mutex_destroy;
+ rte_thread_barrier_init;
+
[dpdk-dev] [PATCH v14 7/9] eal: implement functions for mutex management
From: Narcisa Vasile
Add functions for mutex init, destroy, lock, unlock.
Add RTE_STATIC_MUTEX macro to replace static initialization
of mutexes.
Windows does not have a static initializer.
Initialization is only done through InitializeCriticalSection().
The RTE_STATIC_MUTEX calls into the rte_thread_mutex_init()
function that performs the actual mutex initialization.
Signed-off-by: Narcisa Vasile
---
lib/eal/common/rte_thread.c | 61 +++
lib/eal/include/rte_thread.h | 94
lib/eal/version.map | 4 ++
lib/eal/windows/rte_thread.c | 53
4 files changed, 212 insertions(+)
diff --git a/lib/eal/common/rte_thread.c b/lib/eal/common/rte_thread.c
index 39bb35ae23..05c235c215 100644
--- a/lib/eal/common/rte_thread.c
+++ b/lib/eal/common/rte_thread.c
@@ -247,6 +247,67 @@ rte_thread_detach(rte_thread_t thread_id)
return pthread_detach((pthread_t)thread_id.opaque_id);
}
+int
+rte_thread_mutex_init(rte_thread_mutex *mutex)
+{
+ int ret = 0;
+ pthread_mutex_t *m = NULL;
+
+ RTE_VERIFY(mutex != NULL);
+
+ m = calloc(1, sizeof(*m));
+ if (m == NULL) {
+ RTE_LOG(DEBUG, EAL, "Unable to initialize mutex. Insufficient
memory!\n");
+ ret = ENOMEM;
+ goto cleanup;
+ }
+
+ ret = pthread_mutex_init(m, NULL);
+ if (ret != 0) {
+ RTE_LOG(DEBUG, EAL, "Failed to init mutex. ret = %d\n", ret);
+ goto cleanup;
+ }
+
+ mutex->mutex_id = m;
+ m = NULL;
+
+cleanup:
+ free(m);
+ return ret;
+}
+
+int
+rte_thread_mutex_lock(rte_thread_mutex *mutex)
+{
+ RTE_VERIFY(mutex != NULL);
+
+ return pthread_mutex_lock((pthread_mutex_t *)mutex->mutex_id);
+}
+
+int
+rte_thread_mutex_unlock(rte_thread_mutex *mutex)
+{
+ RTE_VERIFY(mutex != NULL);
+
+ return pthread_mutex_unlock((pthread_mutex_t *)mutex->mutex_id);
+}
+
+int
+rte_thread_mutex_destroy(rte_thread_mutex *mutex)
+{
+ int ret = 0;
+ RTE_VERIFY(mutex != NULL);
+
+ ret = pthread_mutex_destroy((pthread_mutex_t *)mutex->mutex_id);
+ if (ret != 0)
+ RTE_LOG(DEBUG, EAL, "Unable to destroy mutex, ret = %d\n", ret);
+
+ free(mutex->mutex_id);
+ mutex->mutex_id = NULL;
+
+ return ret;
+}
+
int
rte_thread_key_create(rte_thread_key *key, void (*destructor)(void *))
{
diff --git a/lib/eal/include/rte_thread.h b/lib/eal/include/rte_thread.h
index 8f36cf06e0..8015a984b8 100644
--- a/lib/eal/include/rte_thread.h
+++ b/lib/eal/include/rte_thread.h
@@ -54,6 +54,26 @@ typedef struct {
#endif /* RTE_HAS_CPUSET */
+#define RTE_DECLARE_MUTEX(private_lock) rte_thread_mutex private_lock
+
+#define RTE_DEFINE_MUTEX(private_lock)\
+RTE_INIT(__rte_ ## private_lock ## _init)\
+{\
+ RTE_VERIFY(rte_thread_mutex_init(&private_lock) == 0);\
+}
+
+#define RTE_STATIC_MUTEX(private_lock)\
+static RTE_DECLARE_MUTEX(private_lock);\
+RTE_DEFINE_MUTEX(private_lock)
+
+
+/**
+ * Thread mutex representation.
+ */
+typedef struct rte_thread_mutex_tag {
+ void *mutex_id; /**< mutex identifier */
+} rte_thread_mutex;
+
/**
* TLS key type, an opaque pointer.
*/
@@ -266,6 +286,28 @@ int rte_thread_join(rte_thread_t thread_id, unsigned long
*value_ptr);
__rte_experimental
int rte_thread_detach(rte_thread_t thread_id);
+/**
+ * Set core affinity of the current thread.
+ * Support both EAL and non-EAL thread and update TLS.
+ *
+ * @param cpusetp
+ * Pointer to CPU affinity to set.
+ *
+ * @return
+ * On success, return 0; otherwise return -1;
+ */
+int rte_thread_set_affinity(rte_cpuset_t *cpusetp);
+
+/**
+ * Get core affinity of the current thread.
+ *
+ * @param cpusetp
+ * Pointer to CPU affinity of current thread.
+ * It presumes input is not NULL, otherwise it causes panic.
+ *
+ */
+void rte_thread_get_affinity(rte_cpuset_t *cpusetp);
+
#endif /* RTE_HAS_CPUSET */
/**
@@ -285,6 +327,58 @@ __rte_experimental
int rte_thread_set_priority(rte_thread_t thread_id,
enum rte_thread_priority priority);
+/**
+ * Initializes a mutex.
+ *
+ * @param mutex
+ *The mutex to be initialized.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_mutex_init(rte_thread_mutex *mutex);
+
+/**
+ * Locks a mutex.
+ *
+ * @param mutex
+ *The mutex to be locked.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_mutex_lock(rte_thread_mutex *mutex);
+
+/**
+ * Unlocks a mutex.
+ *
+ * @param mutex
+ *The mutex to be unlocked.
+ *
+ * @return
+ * On success, return 0.
+ * On failure, return a positive errno-style error number.
+ */
+__rte_experimental
+int rte_thread_mutex_unlock(rte_thread_mutex *mutex);
+
+/**
+ * Releases all resources associated with a
Re: [dpdk-dev] [PATCH v13 09/10] eal: add EAL argument for setting thread priority
On Thu, 19 Aug 2021 14:30:19 -0700
Narcisa Ana Maria Vasile wrote:
> On Thu, Aug 19, 2021 at 10:06:06AM +0100, Bruce Richardson wrote:
> > On Wed, Aug 18, 2021 at 02:28:33PM -0700, Stephen Hemminger wrote:
> > > On Tue, 3 Aug 2021 12:01:30 -0700 Narcisa Ana Maria Vasile
> > > wrote:
> > >
> > > > +static int +eal_parse_thread_priority(const char *arg) +{ +
> > > > struct internal_config *internal_conf = +
> > > > eal_get_internal_configuration(); + enum rte_thread_priority
> > > > priority;
> > > > + + if (!strncmp("normal", arg, sizeof("normal"))) +
> > > > priority = RTE_THREAD_PRIORITY_NORMAL; +else if
> > > > (!strncmp("realtime", arg, sizeof("realtime"))) +
> > > > priority =
> > > > RTE_THREAD_PRIORITY_REALTIME_CRITICAL; +else + return
> > > > -1;
> > > > + + internal_conf->thread_priority = priority; +return 0; +} +
> > >
> > > In my experience using real time priority is dangerous and risks
> > > starvation and deadlock. The problem is that DPDK applications are
> > > typically 100% CPU poll mode with no system calls; but the kernel has a
> > > number of worker threads that can be required on those CPUs.
> > >
> > > The typical failure is a disk completion interrupt happens on a CPU that
> > > is being used by DPDK lcore thread. With RT priority, the kernel thread
> > > to process that I/O completion never runs because the RT user thread has
> > > higher priority than the kernel I/O completion thread.
> > >
> > > It maybe possible to workaround this with lots of hand crafting through
> > > sysfs to reassign threads and irq's. Also, later kernels with full RT
> > > might also help.
> > >
> > > Before putting this in as part of DPDK in EAL, a full set of testing and
> > > documentation of how to configure these kind of applications and systems
> > > is needed.
> > >
> > I would tend to agree caution here, based on my experience of having locked
> > up a number of systems in the past when testing running DPDK apps with RT
> > priority!
>
> Thank you for the comments! I've added this option since it was requested by
> multiple users. I understand RT priority causes issues on Linux platforms.
> On Windows we want to be able to use REALTIME priority in certain scenarios.
>
> Would it be acceptable to replace this option with a "HIGH_PRIORITY" one
> and keep it realtime on Windows and choose a higher (but non-realtime) option
> on Linux?
> However, there are 2 issues here:
> * We will have different behaviors between the 2 platforms.
> * Not sure if I can set a normal but higher priority on Linux. SCHED_OTHER
> only allows
>one value and Linux "nice" values don't help. If anyone knows of a way to
> accomplish
>this on Linux, please do advise.
> Alternatively, we can have this option for Windows only.
>
> In the meantime, I've removed this patch from this patchset in v14 as the
> cmdline option is not
> being enabled yet, as DmitryK noted.
>
>
I think on Linux it should produce a big a*** warning message.
[dpdk-dev] [PATCH] build: propagate Windows system dependencies to pkg-config
Windows EAL depends on some system libraries. They were linked using
add_project_link_arguments('-l'), which prevented meson from adding
them to Libs.private of pkg-config file. As a result, applications using
pkg-config to find DPDK hit link errors, for example:
librte_eal.a(eal_windows_eal_debug.c.obj) : error LNK2019: unresolved
external symbol __imp_SymInitialize referenced in function
rte_dump_stack
Reference required libraries in EAL using ext_deps meson variable.
bus/pci and net/pcap depend on lib/eal and will pull them automatically.
Drop advapi32 dependency, as MinGW locates VirtualAlloc2() dynamically.
Fixes: 2a5d547a4a9b ("eal/windows: implement basic memory management")
Fixes: c91717eb75c8 ("eal/windows: support exit and panic")
Cc: tal...@nvidia.com
Cc: sta...@dpdk.org
Reported-by: William Tu
Signed-off-by: Dmitry Kozlyuk
---
config/meson.build | 14 ++
lib/eal/windows/meson.build | 10 ++
2 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/config/meson.build b/config/meson.build
index e80421003b..3d5e1e33e7 100644
--- a/config/meson.build
+++ b/config/meson.build
@@ -339,7 +339,8 @@ if is_freebsd
endif
if is_windows
-# VirtualAlloc2() is available since Windows 10 / Server 2016.
+# VirtualAlloc2() is available since Windows 10 / Server 2019.
+# It's essential for EAL, so we don't support older versions.
add_project_arguments('-D_WIN32_WINNT=0x0A00', language: 'c')
# Use MinGW-w64 stdio, because DPDK assumes ANSI-compliant formatting.
@@ -351,17 +352,6 @@ if is_windows
if cc.get_id() == 'clang'
add_project_arguments('-D_CRT_SECURE_NO_WARNINGS', language: 'c')
endif
-
-add_project_link_arguments('-lws2_32', language: 'c')
-
-# Contrary to docs, VirtualAlloc2() is exported by mincore.lib
-# in Windows SDK, while MinGW exports it by advapi32.a.
-if is_ms_linker
-add_project_link_arguments('-lmincore', language: 'c')
-endif
-
-add_project_link_arguments('-ladvapi32', '-lsetupapi', language: 'c')
-add_project_link_arguments('-ldbghelp', language: 'c')
endif
if get_option('b_lto')
diff --git a/lib/eal/windows/meson.build b/lib/eal/windows/meson.build
index fc12fefd0d..845e406ca1 100644
--- a/lib/eal/windows/meson.build
+++ b/lib/eal/windows/meson.build
@@ -24,3 +24,13 @@ sources += files(
)
dpdk_conf.set10('RTE_EAL_NUMA_AWARE_HUGEPAGES', true)
+
+ext_deps += [
+cc.find_library('dbghelp'),
+cc.find_library('setupapi'),
+cc.find_library('ws2_32'),
+]
+if is_ms_linker
+# Contrary to docs, VirtualAlloc2() is exported by mincore.lib.
+ext_deps += cc.find_library('mincore')
+endif
--
2.29.3
Re: [dpdk-dev] [PATCH v7] eal: remove sys/queue.h from public headers.
2021-08-18 23:26 (UTC+), William Tu:
[...]
> diff --git a/lib/eal/linux/include/rte_os.h b/lib/eal/linux/include/rte_os.h
> index 1618b4df22..ce5b0aed52 100644
> --- a/lib/eal/linux/include/rte_os.h
> +++ b/lib/eal/linux/include/rte_os.h
> @@ -11,6 +11,21 @@
> */
>
> #include
> +#include
> +
> +/* These macros are compatible with system's sys/queue.h. */
> +#define RTE_TAILQ_HEAD(name, type) TAILQ_HEAD(name, type)
> +#define RTE_TAILQ_ENTRY(type) TAILQ_ENTRY(type)
> +#define RTE_TAILQ_FOREACH(var, head, field) TAILQ_FOREACH(var, head, field)
> +#define RTE_TAILQ_FOREACH_SAFE(var, head, field, tvar) \
> + for ((var) = TAILQ_FIRST((head)); \
> + (var) && ((tvar) = TAILQ_NEXT((var), field), 1); \
> + (var) = (tvar))
I made a comment to v5 and you were going to fix it, maybe it got lost?
Why duplicate this in rte_os.h (documentation lost, BTW) and add #ifdef?
RTE_TAILQ_FOREACH_SAFE is not needed in headers, it can be left
[in rte_tailq.h].
The important part is duplication in rte_os.h for each platform
and the loss of documentation. I see you already removed #ifdef.
> +#define RTE_TAILQ_FIRST(head) TAILQ_FIRST(head)
> +#define RTE_TAILQ_NEXT(elem, field) TAILQ_NEXT(elem, field)
> +#define RTE_STAILQ_HEAD(name, type) STAILQ_HEAD(name, type)
> +#define RTE_STAILQ_ENTRY(type) STAILQ_ENTRY(type)
> +
>
> #ifdef CPU_SETSIZE /* may require _GNU_SOURCE */
> typedef cpu_set_t rte_cpuset_t;
> diff --git a/lib/eal/windows/eal_alarm.c b/lib/eal/windows/eal_alarm.c
> index e5dc54efb8..103c1f909d 100644
> --- a/lib/eal/windows/eal_alarm.c
> +++ b/lib/eal/windows/eal_alarm.c
> @@ -4,6 +4,7 @@
>
> #include
> #include
> +#include
>
> #include
> #include
> diff --git a/lib/eal/windows/include/rte_os.h
> b/lib/eal/windows/include/rte_os.h
> index 66c711d458..0cbe1dbc1e 100644
> --- a/lib/eal/windows/include/rte_os.h
> +++ b/lib/eal/windows/include/rte_os.h
> @@ -18,6 +18,37 @@
> extern "C" {
> #endif
>
> +/* These macros are compatible with bundled sys/queue.h. */
> +#define RTE_TAILQ_HEAD(name, type) \
> +struct name { \
> + struct type *tqh_first; /* first element */ \
> + struct type **tqh_last; /* addr of last next element */ \
> +}
> +#define RTE_TAILQ_ENTRY(type) \
> +struct { \
> + struct type *tqe_next; /* next element */ \
> + struct type **tqe_prev; /* address of previous next element */ \
> +}
> +#define RTE_TAILQ_FOREACH(var, head, field) \
> + for ((var) = RTE_TAILQ_FIRST((head)); \
> + (var); \
> + (var) = RTE_TAILQ_NEXT((var), field))
> +#define RTE_TAILQ_FOREACH_SAFE(var, head, field, tvar) \
> + for ((var) = TAILQ_FIRST((head)); \
> + (var) && ((tvar) = TAILQ_NEXT((var), field), 1); \
> + (var) = (tvar))
> +#define RTE_TAILQ_FIRST(head) ((head)->tqh_first)
> +#define RTE_TAILQ_NEXT(elm, field) ((elm)->field.tqe_next)
> +#define RTE_STAILQ_HEAD(name, type) \
> +struct name { \
> + struct type *stqh_first;/* first element */ \
> + struct type **stqh_last;/* addr of last next element */ \
> +}
> +#define RTE_STAILQ_ENTRY(type) \
> +struct { \
> + struct type *stqe_next; /* next element */ \
> +}
> +
Please drop the inline comments.
They duplicate what's already in sys/queue.h
and we're not going to maintain them.
[dpdk-dev] [PATCH] eal: remove the deprecated whitelist/blacklist and master/slave API's
Remove the compatiability hooks that were added in 20.11 around
master/slave and blacklist/whitelist.
New API's for these were added in 20.11 and the old API was retained
but marked deprecated. Since 21.11 is the next LTS, it is time
to remove the deprecated ones.
Signed-off-by: Stephen Hemminger
---
doc/guides/rel_notes/release_21_11.rst | 7 +++
lib/eal/common/eal_common_options.c| 14 --
lib/eal/common/eal_options.h | 10 --
lib/eal/include/rte_bus.h | 6 --
lib/eal/include/rte_dev.h | 6 --
lib/eal/include/rte_devargs.h | 6 --
lib/eal/include/rte_launch.h | 4
lib/eal/include/rte_lcore.h| 15 ---
8 files changed, 7 insertions(+), 61 deletions(-)
diff --git a/doc/guides/rel_notes/release_21_11.rst
b/doc/guides/rel_notes/release_21_11.rst
index d707a554efaf..e34c5aa74ad0 100644
--- a/doc/guides/rel_notes/release_21_11.rst
+++ b/doc/guides/rel_notes/release_21_11.rst
@@ -68,6 +68,13 @@ Removed Items
Also, make sure to start the actual text at the margin.
===
+* eal: Removed the deprecated function ``rte_get_master_lcore()``
+ and the iterarator macro ``RTE_LCORE_FOREACH_SLAVE``.
+
+* eal: The old api arguments that were deprecated for
+ blacklist/whitelist are removed. Users must use the new
+ block/allow list arguments.
+
API Changes
---
diff --git a/lib/eal/common/eal_common_options.c
b/lib/eal/common/eal_common_options.c
index ff5861b5f3ef..8853833b108a 100644
--- a/lib/eal/common/eal_common_options.c
+++ b/lib/eal/common/eal_common_options.c
@@ -83,7 +83,6 @@ eal_long_options[] = {
{OPT_TRACE_DIR, 1, NULL, OPT_TRACE_DIR_NUM},
{OPT_TRACE_BUF_SIZE,1, NULL, OPT_TRACE_BUF_SIZE_NUM },
{OPT_TRACE_MODE,1, NULL, OPT_TRACE_MODE_NUM },
- {OPT_MASTER_LCORE, 1, NULL, OPT_MASTER_LCORE_NUM },
{OPT_MAIN_LCORE,1, NULL, OPT_MAIN_LCORE_NUM },
{OPT_MBUF_POOL_OPS_NAME, 1, NULL, OPT_MBUF_POOL_OPS_NAME_NUM},
{OPT_NO_HPET, 0, NULL, OPT_NO_HPET_NUM },
@@ -108,10 +107,6 @@ eal_long_options[] = {
{OPT_NO_TELEMETRY, 0, NULL, OPT_NO_TELEMETRY_NUM },
{OPT_FORCE_MAX_SIMD_BITWIDTH, 1, NULL, OPT_FORCE_MAX_SIMD_BITWIDTH_NUM},
- /* legacy options that will be removed in future */
- {OPT_PCI_BLACKLIST, 1, NULL, OPT_PCI_BLACKLIST_NUM},
- {OPT_PCI_WHITELIST, 1, NULL, OPT_PCI_WHITELIST_NUM},
-
{0, 0, NULL, 0}
};
@@ -1521,10 +1516,6 @@ eal_parse_common_option(int opt, const char *optarg,
static int a_used;
switch (opt) {
- case OPT_PCI_BLACKLIST_NUM:
- fprintf(stderr,
- "Option --pci-blacklist is deprecated, use -b, --block
instead\n");
- /* fallthrough */
case 'b':
if (a_used)
goto ba_conflict;
@@ -1698,11 +1689,6 @@ eal_parse_common_option(int opt, const char *optarg,
conf->process_type = eal_parse_proc_type(optarg);
break;
- case OPT_MASTER_LCORE_NUM:
- fprintf(stderr,
- "Option --" OPT_MASTER_LCORE
- " is deprecated use " OPT_MAIN_LCORE "\n");
- /* fallthrough */
case OPT_MAIN_LCORE_NUM:
if (eal_parse_main_lcore(optarg) < 0) {
RTE_LOG(ERR, EAL, "invalid parameter for --"
diff --git a/lib/eal/common/eal_options.h b/lib/eal/common/eal_options.h
index 7b348e707fb4..8e4f7202a2d5 100644
--- a/lib/eal/common/eal_options.h
+++ b/lib/eal/common/eal_options.h
@@ -18,10 +18,6 @@ enum {
#define OPT_DEV_BLOCK "block"
OPT_DEV_BLOCK_NUM = 'b',
- /* legacy option that will be removed in future */
-#define OPT_PCI_WHITELIST "pci-whitelist"
- OPT_PCI_WHITELIST_NUM = 'w',
-
/* first long only option value must be >= 256, so that we won't
* conflict with short options */
OPT_LONG_MIN_NUM = 256,
@@ -49,8 +45,6 @@ enum {
OPT_TRACE_MODE_NUM,
#define OPT_MAIN_LCORE"main-lcore"
OPT_MAIN_LCORE_NUM,
-#define OPT_MASTER_LCORE "master-lcore"
- OPT_MASTER_LCORE_NUM,
#define OPT_MBUF_POOL_OPS_NAME "mbuf-pool-ops-name"
OPT_MBUF_POOL_OPS_NAME_NUM,
#define OPT_PROC_TYPE "proc-type"
@@ -94,10 +88,6 @@ enum {
#define OPT_FORCE_MAX_SIMD_BITWIDTH "force-max-simd-bitwidth"
OPT_FORCE_MAX_SIMD_BITWIDTH_NUM,
- /* legacy option that will be removed in future */
-#define OPT_PCI_BLACKLIST "pci-blacklist"
- OPT_PCI_BLACKLIST_NUM,
-
OPT_LONG_MAX_NUM
};
diff --git a/lib/eal/include/rte_bus.h b/lib/eal/include/rte_bus.h
index 80b154fb982c..8c4a352bfc8e 100644
--- a/lib/eal/include/rte_bus.h
+++
Re: [dpdk-dev] [PATCH v4 1/2] drivers/raw: remove octeontx2-dma driver
On Thu, Aug 19, 2021 at 1:27 PM Radha Mohan Chintakuntla wrote: > > Removing the rawdev based octeontx2-dma driver as the dependent > common/octeontx2 will be soon be going away. Also a new DMA driver will > be coming in this place once the rte_dmadev library is in. > > Signed-off-by: Radha Mohan Chintakuntla > --- > Changes from v3: > Fixed patch application failure on main due to conflict. > > Changes from v2: > Fixed DPDK CI reported issues for more documentation failure. > > Changes from v1: > Fixed compilation issues in documentation > > MAINTAINERS | 6 - > doc/guides/platform/octeontx2.rst | 3 - > doc/guides/rawdevs/index.rst| 1 - > doc/guides/rawdevs/octeontx2_dma.rst| 103 - > drivers/raw/meson.build | 1 - > drivers/raw/octeontx2_dma/meson.build | 18 - > drivers/raw/octeontx2_dma/otx2_dpi_msg.c| 105 - > drivers/raw/octeontx2_dma/otx2_dpi_rawdev.c | 441 > drivers/raw/octeontx2_dma/otx2_dpi_rawdev.h | 197 - > drivers/raw/octeontx2_dma/otx2_dpi_test.c | 218 -- > drivers/raw/octeontx2_dma/version.map | 3 - > 11 files changed, 1096 deletions(-) > delete mode 100644 doc/guides/rawdevs/octeontx2_dma.rst > delete mode 100644 drivers/raw/octeontx2_dma/meson.build > delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_msg.c > delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_rawdev.c > delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_rawdev.h > delete mode 100644 drivers/raw/octeontx2_dma/otx2_dpi_test.c > delete mode 100644 drivers/raw/octeontx2_dma/version.map > Hi Thomas, The CI is reporting unrelated failures which I think can be ignored. https://lab.dpdk.org/results/dashboard/patchsets/18263/ regards, Radha Mohan
[dpdk-dev] [PATCH] net/ice: fix queue config in DCF
When DCF configures rx_queues, it may cause the pointer of
rx_queues to go out of bounds.
This patch expands the scope of the judgment condition to
fix this issue.
Fixes: 4b0d391f0eab ("net/ice: add queue config in DCF")
Cc: sta...@dpdk.org
Signed-off-by: Jie Wang
---
drivers/net/ice/ice_dcf.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ice/ice_dcf.c b/drivers/net/ice/ice_dcf.c
index 4c2e0c7216..45820a0a82 100644
--- a/drivers/net/ice/ice_dcf.c
+++ b/drivers/net/ice/ice_dcf.c
@@ -880,13 +880,13 @@ ice_dcf_configure_queues(struct ice_dcf_hw *hw)
vc_qp->txq.ring_len = txq[i]->nb_tx_desc;
vc_qp->txq.dma_ring_addr = txq[i]->tx_ring_dma;
}
- vc_qp->rxq.vsi_id = hw->vsi_res->vsi_id;
- vc_qp->rxq.queue_id = i;
- vc_qp->rxq.max_pkt_size = rxq[i]->max_pkt_len;
if (i >= hw->eth_dev->data->nb_rx_queues)
continue;
+ vc_qp->rxq.vsi_id = hw->vsi_res->vsi_id;
+ vc_qp->rxq.queue_id = i;
+ vc_qp->rxq.max_pkt_size = rxq[i]->max_pkt_len;
vc_qp->rxq.ring_len = rxq[i]->nb_rx_desc;
vc_qp->rxq.dma_ring_addr = rxq[i]->rx_ring_dma;
vc_qp->rxq.databuffer_size = rxq[i]->rx_buf_len;
--
2.25.1
[dpdk-dev] [PATCH] vhost: remove copy threshold for async vhost
Copy threshold is introduced in async vhost data path to select
the appropriate copy engine to do copies for higher efficiency.
However, it may cause packets out-of-order, and it also causes
data path performance unpredictable.
Therefore, this patch removes copy threshold support in async vhost
data path.
Signed-off-by: Jiayu Hu
Signed-off-by: Cheng Jiang
---
doc/guides/prog_guide/vhost_lib.rst | 7 -
examples/vhost/main.c | 22 +-
lib/vhost/rte_vhost_async.h | 22 +-
lib/vhost/vhost.c | 6 +-
lib/vhost/vhost.h | 1 -
lib/vhost/virtio_net.c | 439 +---
6 files changed, 116 insertions(+), 381 deletions(-)
diff --git a/doc/guides/prog_guide/vhost_lib.rst
b/doc/guides/prog_guide/vhost_lib.rst
index 8874033..171e009 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -235,13 +235,6 @@ The following is an overview of some key Vhost API
functions:
Currently, only ``RTE_VHOST_ASYNC_INORDER`` capable device is
supported by vhost.
- * ``async_threshold``
-
-The copy length (in bytes) below which CPU copy will be used even if
-applications call async vhost APIs to enqueue/dequeue data.
-
-Typical value is 256~1024 depending on the async device capability.
-
Applications must provide following ``ops`` callbacks for vhost lib to
work with the async copy devices:
diff --git a/examples/vhost/main.c b/examples/vhost/main.c
index bc3d71c..a4a8214 100644
--- a/examples/vhost/main.c
+++ b/examples/vhost/main.c
@@ -891,17 +891,11 @@ drain_vhost(struct vhost_dev *vdev)
if (builtin_net_driver) {
ret = vs_enqueue_pkts(vdev, VIRTIO_RXQ, m, nr_xmit);
} else if (async_vhost_driver) {
- uint32_t cpu_cpl_nr = 0;
uint16_t enqueue_fail = 0;
- struct rte_mbuf *m_cpu_cpl[nr_xmit];
complete_async_pkts(vdev);
- ret = rte_vhost_submit_enqueue_burst(vdev->vid, VIRTIO_RXQ,
- m, nr_xmit, m_cpu_cpl, &cpu_cpl_nr);
- __atomic_add_fetch(&vdev->pkts_inflight, ret - cpu_cpl_nr,
__ATOMIC_SEQ_CST);
-
- if (cpu_cpl_nr)
- free_pkts(m_cpu_cpl, cpu_cpl_nr);
+ ret = rte_vhost_submit_enqueue_burst(vdev->vid, VIRTIO_RXQ, m,
nr_xmit);
+ __atomic_add_fetch(&vdev->pkts_inflight, ret, __ATOMIC_SEQ_CST);
enqueue_fail = nr_xmit - ret;
if (enqueue_fail)
@@ -1222,19 +1216,12 @@ drain_eth_rx(struct vhost_dev *vdev)
enqueue_count = vs_enqueue_pkts(vdev, VIRTIO_RXQ,
pkts, rx_count);
} else if (async_vhost_driver) {
- uint32_t cpu_cpl_nr = 0;
uint16_t enqueue_fail = 0;
- struct rte_mbuf *m_cpu_cpl[MAX_PKT_BURST];
complete_async_pkts(vdev);
enqueue_count = rte_vhost_submit_enqueue_burst(vdev->vid,
- VIRTIO_RXQ, pkts, rx_count,
- m_cpu_cpl, &cpu_cpl_nr);
- __atomic_add_fetch(&vdev->pkts_inflight, enqueue_count -
cpu_cpl_nr,
- __ATOMIC_SEQ_CST);
-
- if (cpu_cpl_nr)
- free_pkts(m_cpu_cpl, cpu_cpl_nr);
+ VIRTIO_RXQ, pkts, rx_count);
+ __atomic_add_fetch(&vdev->pkts_inflight, enqueue_count,
__ATOMIC_SEQ_CST);
enqueue_fail = rx_count - enqueue_count;
if (enqueue_fail)
@@ -1495,7 +1482,6 @@ new_device(int vid)
ioat_check_completed_copies_cb;
config.features = RTE_VHOST_ASYNC_INORDER;
- config.async_threshold = 256;
return rte_vhost_async_channel_register(vid, VIRTIO_RXQ,
config, &channel_ops);
diff --git a/lib/vhost/rte_vhost_async.h b/lib/vhost/rte_vhost_async.h
index b25ff44..ad71555 100644
--- a/lib/vhost/rte_vhost_async.h
+++ b/lib/vhost/rte_vhost_async.h
@@ -103,7 +103,6 @@ enum {
* async channel configuration
*/
struct rte_vhost_async_config {
- uint32_t async_threshold;
uint32_t features;
uint32_t rsvd[2];
};
@@ -182,13 +181,9 @@ int rte_vhost_async_channel_unregister_thread_unsafe(int
vid,
uint16_t queue_id);
/**
- * This function submits enqueue data to async engine. Successfully
- * enqueued packets can be transfer completed or being occupied by DMA
- * engines, when this API returns. Transfer completed packets are returned
- * in comp_pkts, so users need to guarantee its size is greater than or
- * equal to the size of pkts; for packets that are successfully enqueued
- * but not transfer completed, users should poll transfer status by
- * rte_
