Home
last modified time | relevance | path

Searched full:queues (Results 1 – 25 of 1644) sorted by relevance

12345678910>>...66

/linux-5.10/Documentation/networking/device_drivers/ethernet/huawei/
Dhinic.rst55 Asynchronous Event Queues(AEQs) - The event queues for receiving messages from
69 Completion Event Queues(CEQs) - The completion Event Queues that describe IO
72 Work Queues(WQ) - Contain the memory and operations for use by CMD queues and
77 Command Queues(CMDQ) - The queues for sending commands for IO management and is
82 Queue Pairs(QPs) - The HW Receive and Send queues for Receiving and Transmitting
104 Tx Queues - Logical Tx Queues that use the HW Send Queues for transmit.
108 Rx Queues - Logical Rx Queues that use the HW Receive Queues for receive.
112 hinic_dev - de/constructs the Logical Tx and Rx Queues.
/linux-5.10/net/sched/
Dsch_multiq.c25 struct Qdisc **queues; member
54 return q->queues[0]; in multiq_classify()
56 return q->queues[band]; in multiq_classify()
105 qdisc = q->queues[q->curband]; in multiq_dequeue()
137 qdisc = q->queues[curband]; in multiq_peek()
154 qdisc_reset(q->queues[band]); in multiq_reset()
167 qdisc_put(q->queues[band]); in multiq_destroy()
169 kfree(q->queues); in multiq_destroy()
197 if (q->queues[i] != &noop_qdisc) { in multiq_tune()
198 struct Qdisc *child = q->queues[i]; in multiq_tune()
[all …]
Dsch_prio.c26 struct Qdisc *queues[TCQ_PRIO_BANDS]; member
57 return q->queues[q->prio2band[band & TC_PRIO_MAX]]; in prio_classify()
63 return q->queues[q->prio2band[0]]; in prio_classify()
65 return q->queues[band]; in prio_classify()
103 struct Qdisc *qdisc = q->queues[prio]; in prio_peek()
117 struct Qdisc *qdisc = q->queues[prio]; in prio_dequeue()
137 qdisc_reset(q->queues[prio]); in prio_reset()
175 qdisc_put(q->queues[prio]); in prio_destroy()
182 struct Qdisc *queues[TCQ_PRIO_BANDS]; in prio_tune() local
200 queues[i] = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, in prio_tune()
[all …]
/linux-5.10/Documentation/ABI/testing/
Dsysfs-class-net-queues1 What: /sys/class/<iface>/queues/rx-<queue>/rps_cpus
11 What: /sys/class/<iface>/queues/rx-<queue>/rps_flow_cnt
19 What: /sys/class/<iface>/queues/tx-<queue>/tx_timeout
27 What: /sys/class/<iface>/queues/tx-<queue>/tx_maxrate
35 What: /sys/class/<iface>/queues/tx-<queue>/xps_cpus
45 What: /sys/class/<iface>/queues/tx-<queue>/xps_rxqs
56 What: /sys/class/<iface>/queues/tx-<queue>/byte_queue_limits/hold_time
65 What: /sys/class/<iface>/queues/tx-<queue>/byte_queue_limits/inflight
73 What: /sys/class/<iface>/queues/tx-<queue>/byte_queue_limits/limit
82 What: /sys/class/<iface>/queues/tx-<queue>/byte_queue_limits/limit_max
[all …]
/linux-5.10/Documentation/devicetree/bindings/soc/ti/
Dkeystone-navigator-qmss.txt9 management of the packet queues. Packets are queued/de-queued by writing or
32 -- managed-queues : the actual queues managed by each queue manager
33 instance, specified as <"base queue #" "# of queues">.
51 - qpend : pool of qpend(interruptible) queues
52 - general-purpose : pool of general queues, primarily used
53 as free descriptor queues or the
54 transmit DMA queues.
55 - accumulator : pool of queues on PDSP accumulator channel
57 -- qrange : number of queues to use per queue range, specified as
58 <"base queue #" "# of queues">.
[all …]
/linux-5.10/Documentation/block/
Dblk-mq.rst37 spawns multiple queues with individual entry points local to the CPU, removing
49 blk-mq has two group of queues: software staging queues and hardware dispatch
50 queues. When the request arrives at the block layer, it will try the shortest
56 Then, after the requests are processed by software queues, they will be placed
62 Software staging queues
65 The block IO subsystem adds requests in the software staging queues
71 the number of queues is defined by a per-CPU or per-node basis.
93 requests from different queues, otherwise there would be cache trashing and a
99 queue (a.k.a. run the hardware queue), the software queues mapped to that
102 Hardware dispatch queues
[all …]
/linux-5.10/Documentation/networking/
Dscaling.rst27 Contemporary NICs support multiple receive and transmit descriptor queues
29 queues to distribute processing among CPUs. The NIC distributes packets by
47 Some advanced NICs allow steering packets to queues based on
57 module parameter for specifying the number of hardware queues to
60 for each CPU if the device supports enough queues, or otherwise at least
66 default mapping is to distribute the queues evenly in the table, but the
69 indirection table could be done to give different queues different
80 of queues to IRQs can be determined from /proc/interrupts. By default,
95 is to allocate as many queues as there are CPUs in the system (or the
97 is likely the one with the smallest number of receive queues where no
[all …]
Dmultiqueue.rst18 the subqueue memory, as well as netdev configuration of where the queues
21 The base driver will also need to manage the queues as it does the global
33 A new round-robin qdisc, sch_multiq also supports multiple hardware queues. The
35 bands and queues based on the value in skb->queue_mapping. Use this field in
42 On qdisc load, the number of bands is based on the number of queues on the
56 The qdisc will allocate the number of bands to equal the number of queues that
58 queues, the band mapping would look like::
/linux-5.10/drivers/staging/wfx/
Dqueue.c234 struct wfx_queue *queues[IEEE80211_NUM_ACS * ARRAY_SIZE(wdev->vif)]; in wfx_tx_queues_get_skb() local
240 // sort the queues in wfx_tx_queues_get_skb()
244 WARN_ON(num_queues >= ARRAY_SIZE(queues)); in wfx_tx_queues_get_skb()
245 queues[num_queues] = &wvif->tx_queue[i]; in wfx_tx_queues_get_skb()
247 if (wfx_tx_queue_get_weight(queues[j]) < in wfx_tx_queues_get_skb()
248 wfx_tx_queue_get_weight(queues[j - 1])) in wfx_tx_queues_get_skb()
249 swap(queues[j - 1], queues[j]); in wfx_tx_queues_get_skb()
259 skb = skb_dequeue(&queues[i]->cab); in wfx_tx_queues_get_skb()
267 WARN_ON(queues[i] != in wfx_tx_queues_get_skb()
269 atomic_inc(&queues[i]->pending_frames); in wfx_tx_queues_get_skb()
[all …]
/linux-5.10/Documentation/arm/keystone/
Dknav-qmss.rst15 management of the packet queues. Packets are queued/de-queued by writing or
24 knav qmss driver provides a set of APIs to drivers to open/close qmss queues,
25 allocate descriptor pools, map the descriptors, push/pop to queues etc. For
31 Accumulator QMSS queues using PDSP firmware
34 queue or multiple contiguous queues. drivers/soc/ti/knav_qmss_acc.c is the
37 1 or 32 queues per channel. More description on the firmware is available in
56 Use of accumulated queues requires the firmware image to be present in the
57 file system. The driver doesn't acc queues to the supported queue range if
/linux-5.10/drivers/scsi/snic/
Dvnic_resource.h27 RES_TYPE_WQ, /* Work queues */
28 RES_TYPE_RQ, /* Receive queues */
29 RES_TYPE_CQ, /* Completion queues */
45 RES_TYPE_MQ_WQ, /* MQ Work queues */
46 RES_TYPE_MQ_RQ, /* MQ Receive queues */
47 RES_TYPE_MQ_CQ, /* MQ Completion queues */
/linux-5.10/drivers/scsi/fnic/
Dvnic_resource.h27 RES_TYPE_WQ, /* Work queues */
28 RES_TYPE_RQ, /* Receive queues */
29 RES_TYPE_CQ, /* Completion queues */
45 RES_TYPE_MQ_WQ, /* MQ Work queues */
46 RES_TYPE_MQ_RQ, /* MQ Receive queues */
47 RES_TYPE_MQ_CQ, /* MQ Completion queues */
/linux-5.10/drivers/scsi/aacraid/
Dcomminit.c237 * Fill in addresses of the Comm Area Headers and Queues in aac_alloc_comm()
373 struct aac_entry * queues; in aac_comm_init() local
375 struct aac_queue_block * comm = dev->queues; in aac_comm_init()
394 queues = (struct aac_entry *)(((ulong)headers) + hdrsize); in aac_comm_init()
397 comm->queue[HostNormCmdQueue].base = queues; in aac_comm_init()
399 queues += HOST_NORM_CMD_ENTRIES; in aac_comm_init()
403 comm->queue[HostHighCmdQueue].base = queues; in aac_comm_init()
406 queues += HOST_HIGH_CMD_ENTRIES; in aac_comm_init()
410 comm->queue[AdapNormCmdQueue].base = queues; in aac_comm_init()
413 queues += ADAP_NORM_CMD_ENTRIES; in aac_comm_init()
[all …]
/linux-5.10/drivers/nvme/target/
Dloop.c30 struct nvme_loop_queue *queues; member
71 return queue - queue->ctrl->queues; in nvme_loop_queue_idx()
176 struct nvme_loop_queue *queue = &ctrl->queues[0]; in nvme_loop_submit_async_event()
198 iod->queue = &ctrl->queues[queue_idx]; in nvme_loop_init_iod()
218 struct nvme_loop_queue *queue = &ctrl->queues[hctx_idx + 1]; in nvme_loop_init_hctx()
230 struct nvme_loop_queue *queue = &ctrl->queues[0]; in nvme_loop_init_admin_hctx()
254 clear_bit(NVME_LOOP_Q_LIVE, &ctrl->queues[0].flags); in nvme_loop_destroy_admin_queue()
255 nvmet_sq_destroy(&ctrl->queues[0].nvme_sq); in nvme_loop_destroy_admin_queue()
276 kfree(ctrl->queues); in nvme_loop_free_ctrl()
287 clear_bit(NVME_LOOP_Q_LIVE, &ctrl->queues[i].flags); in nvme_loop_destroy_io_queues()
[all …]
/linux-5.10/drivers/net/ethernet/intel/ixgbe/
Dixgbe_lib.c215 /* FCoE uses a linear block of queues so just assigning 1:1 */ in ixgbe_cache_ring_sriov()
236 /* FCoE uses a linear block of queues so just assigning 1:1 */ in ixgbe_cache_ring_sriov()
313 * ixgbe_set_dcb_sriov_queues: Allocate queues for SR-IOV devices w/ DCB
316 * When SR-IOV (Single Root IO Virtualiztion) is enabled, allocate queues
317 * and VM pools where appropriate. Also assign queues based on DCB
339 /* limit VMDq instances on the PF by number of Tx queues */ in ixgbe_set_dcb_sriov_queues()
356 /* queues in the remaining pools are available for FCoE */ in ixgbe_set_dcb_sriov_queues()
394 /* alloc queues for FCoE separately */ in ixgbe_set_dcb_sriov_queues()
398 /* add queues to adapter */ in ixgbe_set_dcb_sriov_queues()
428 /* Map queue offset and counts onto allocated tx queues */ in ixgbe_set_dcb_queues()
[all …]
/linux-5.10/drivers/net/ethernet/cisco/enic/
Dvnic_resource.h34 RES_TYPE_WQ, /* Work queues */
35 RES_TYPE_RQ, /* Receive queues */
36 RES_TYPE_CQ, /* Completion queues */
52 RES_TYPE_MQ_WQ, /* MQ Work queues */
53 RES_TYPE_MQ_RQ, /* MQ Receive queues */
54 RES_TYPE_MQ_CQ, /* MQ Completion queues */
/linux-5.10/Documentation/devicetree/bindings/misc/
Dintel,ixp4xx-ahb-queue-manager.yaml14 The IXP4xx AHB Queue Manager maintains queues as circular buffers in
17 IXP4xx for accelerating queues, especially for networking. Clients pick
18 queues from the queue manager with foo-queue = <&qmgr N> where the
33 - description: Interrupt for queues 0-31
34 - description: Interrupt for queues 32-63
/linux-5.10/tools/perf/util/
Darm-spe.c40 struct auxtrace_queues queues; member
141 queue = &speq->spe->queues.queue_array[speq->queue_nr]; in arm_spe_get_trace()
437 for (i = 0; i < spe->queues.nr_queues; i++) { in arm_spe__setup_queues()
438 ret = arm_spe__setup_queue(spe, &spe->queues.queue_array[i], i); in arm_spe__setup_queues()
448 if (spe->queues.new_data) { in arm_spe__update_queues()
449 spe->queues.new_data = false; in arm_spe__update_queues()
516 queue = &spe->queues.queue_array[queue_nr]; in arm_spe_process_queues()
552 struct auxtrace_queues *queues = &spe->queues; in arm_spe_process_timeless_queues() local
556 for (i = 0; i < queues->nr_queues; i++) { in arm_spe_process_timeless_queues()
557 struct auxtrace_queue *queue = &spe->queues.queue_array[i]; in arm_spe_process_timeless_queues()
[all …]
Dintel-bts.c46 struct auxtrace_queues queues; member
211 for (i = 0; i < bts->queues.nr_queues; i++) { in intel_bts_setup_queues()
212 ret = intel_bts_setup_queue(bts, &bts->queues.queue_array[i], in intel_bts_setup_queues()
222 if (bts->queues.new_data) { in intel_bts_update_queues()
223 bts->queues.new_data = false; in intel_bts_update_queues()
465 queue = &btsq->bts->queues.queue_array[btsq->queue_nr]; in intel_bts_process_queue()
539 struct auxtrace_queues *queues = &bts->queues; in intel_bts_process_tid_exit() local
542 for (i = 0; i < queues->nr_queues; i++) { in intel_bts_process_tid_exit()
543 struct auxtrace_queue *queue = &bts->queues.queue_array[i]; in intel_bts_process_tid_exit()
568 queue = &bts->queues.queue_array[queue_nr]; in intel_bts_process_queues()
[all …]
/linux-5.10/include/linux/avf/
Dvirtchnl.h29 * have a maximum of sixteen queues for all of its VSIs.
38 * queues and interrupts. After these operations are complete, the VF
39 * driver may start its queues, optionally add MAC and VLAN filters, and
198 * When reset is complete, PF must ensure that all queues in all VSIs associated
314 * VF sends this message to set parameters for all active TX and RX queues
316 * PF configures queues and returns status.
317 * If the number of queues specified is greater than the number of queues
318 * associated with the VSI, an error is returned and no queues are configured.
321 /* NOTE: vsi_id and queue_id should be identical for both queues. */
338 * VF sends this message to request the PF to allocate additional queues to
[all …]
/linux-5.10/drivers/net/ethernet/chelsio/cxgb3/
Dfirmware_exports.h98 /* FW_TUNNEL_NUM corresponds to the number of supported TUNNEL Queues. These
99 * queues must start at SGE Egress Context FW_TUNNEL_SGEEC_START and must
109 /* FW_CTRL_NUM corresponds to the number of supported CTRL Queues. These queues
119 /* FW_OFLD_NUM corresponds to the number of supported OFFLOAD Queues. These
120 * queues must start at SGE Egress Context FW_OFLD_SGEEC_START.
123 * OFFLOAD Queues, as the host is responsible for providing the correct TID in
/linux-5.10/Documentation/networking/device_drivers/ethernet/freescale/
Ddpaa.rst86 Tx FQs transmission frame queues
143 confirmation frame queues. The driver is then responsible for freeing the
164 strict priority levels. Each traffic class contains NR_CPU TX queues. By
165 default, only one traffic class is enabled and the lowest priority Tx queues
184 Traffic coming on the DPAA Rx queues or on the DPAA Tx confirmation
185 queues is seen by the CPU as ingress traffic on a certain portal.
191 hardware frame queues using a hash on IP v4/v6 source and destination
195 queues are configured to put the received traffic into a pool channel
197 The default frame queues have the HOLDACTIVE option set, ensuring that
204 128 Rx frame queues that are configured to dedicated channels, in a
[all …]
/linux-5.10/drivers/gpu/drm/amd/amdkfd/
Dkfd_device_queue_manager.h51 * @exeute_queues: Dispatches the queues list to the H/W.
59 * @start: Initializes the resources/modules the the device needs for queues
76 * @process_termination: Clears all process queues belongs to that device.
78 * @evict_process_queues: Evict all active queues of a process
80 * @restore_process_queues: Restore all evicted queues queues of a process
163 * This struct is a base class for the kfd queues scheduler in the
167 * concrete device. This class is the only class in the queues scheduler
180 struct list_head queues; member
/linux-5.10/drivers/staging/qlge/
DTODO13 * rename "rx" queues to "completion" queues. Calling tx completion queues "rx
14 queues" is confusing.
24 frames, resets the link, device and driver buffer queues become
/linux-5.10/arch/mips/include/asm/octeon/
Dcvmx-cmd-queue.h30 * Support functions for managing command queues used for
50 * called "cvmx_cmd_queues". Except for the PKO queues, each
52 * contention on spin locks. The PKO queues are stored such that
54 * allows for queues being in separate cache lines when there
55 * are low number of queues per port. With 16 queues per port,
57 * second queues for each port are in another area, etc. This
59 * 16 queues per port using a minimum of cache lines per core.
60 * All queues for a given core will be isolated in the same
65 * queues. The lock uses a "ticket / now serving" model to
93 * queues. Each hardware block has up to 65536 sub identifiers for
[all …]

12345678910>>...66