123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767/*
 * Copyright 2014 Advanced Micro Devices, Inc.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 *
 */

#include <linux/ratelimit.h>
#include <linux/printk.h>
#include <linux/slab.h>
#include <linux/list.h>
#include <linux/types.h>
#include <linux/bitops.h>
#include <linux/sched.h>
#include "kfd_priv.h"
#include "kfd_device_queue_manager.h"
#include "kfd_mqd_manager.h"
#include "cik_regs.h"
#include "kfd_kernel_queue.h"

/* Size of the per-pipe EOP queue */
#define CIK_HPD_EOP_BYTES_LOG2 11
#define CIK_HPD_EOP_BYTES (1U << CIK_HPD_EOP_BYTES_LOG2)

static int set_pasid_vmid_mapping(struct device_queue_manager *dqm,
					unsigned int pasid, unsigned int vmid);

static int create_compute_queue_nocpsch(struct device_queue_manager *dqm,
					struct queue *q,
					struct qcm_process_device *qpd);

static int execute_queues_cpsch(struct device_queue_manager *dqm,
				enum kfd_unmap_queues_filter filter,
				uint32_t filter_param);
static int unmap_queues_cpsch(struct device_queue_manager *dqm,
				enum kfd_unmap_queues_filter filter,
				uint32_t filter_param);

static int map_queues_cpsch(struct device_queue_manager *dqm);

static int create_sdma_queue_nocpsch(struct device_queue_manager *dqm,
					struct queue *q,
					struct qcm_process_device *qpd);

static void deallocate_sdma_queue(struct device_queue_manager *dqm,
				unsigned int sdma_queue_id);

static inline
enum KFD_MQD_TYPE get_mqd_type_from_queue_type(enum kfd_queue_type type)
{
	if (type == KFD_QUEUE_TYPE_SDMA)
		return KFD_MQD_TYPE_SDMA;
	return KFD_MQD_TYPE_CP;
}

static bool is_pipe_enabled(struct device_queue_manager *dqm, int mec, int pipe)
{
	int i;
	int pipe_offset = mec * dqm->dev->shared_resources.num_pipe_per_mec
		+ pipe * dqm->dev->shared_resources.num_queue_per_pipe;

	/* queue is available for KFD usage if bit is 1 */
	for (i = 0; i <  dqm->dev->shared_resources.num_queue_per_pipe; ++i)
		if (test_bit(pipe_offset + i,
			      dqm->dev->shared_resources.queue_bitmap))
			return true;
	return false;
}

unsigned int get_queues_num(struct device_queue_manager *dqm)
{
	return bitmap_weight(dqm->dev->shared_resources.queue_bitmap,
				KGD_MAX_QUEUES);
}

unsigned int get_queues_per_pipe(struct device_queue_manager *dqm)
{
	return dqm->dev->shared_resources.num_queue_per_pipe;
}

unsigned int get_pipes_per_mec(struct device_queue_manager *dqm)
{
	return dqm->dev->shared_resources.num_pipe_per_mec;
}

void program_sh_mem_settings(struct device_queue_manager *dqm,
					struct qcm_process_device *qpd)
{
	return dqm->dev->kfd2kgd->program_sh_mem_settings(
						dqm->dev->kgd, qpd->vmid,
						qpd->sh_mem_config,
						qpd->sh_mem_ape1_base,
						qpd->sh_mem_ape1_limit,
						qpd->sh_mem_bases);
}

static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q)
{
	struct kfd_dev *dev = qpd->dqm->dev;

	if (!KFD_IS_SOC15(dev->device_info->asic_family)) {
		/* On pre-SOC15 chips we need to use the queue ID to
		 * preserve the user mode ABI.
		 */
		q->doorbell_id = q->properties.queue_id;
	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
		/* For SDMA queues on SOC15, use static doorbell
		 * assignments based on the engine and queue.
		 */
		q->doorbell_id = dev->shared_resources.sdma_doorbell
			[q->properties.sdma_engine_id]
			[q->properties.sdma_queue_id];
	} else {
		/* For CP queues on SOC15 reserve a free doorbell ID */
		unsigned int found;

		found = find_first_zero_bit(qpd->doorbell_bitmap,
					    KFD_MAX_NUM_OF_QUEUES_PER_PROCESS);
		if (found >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS) {
			pr_debug("No doorbells available");
			return -EBUSY;
		}
		set_bit(found, qpd->doorbell_bitmap);
		q->doorbell_id = found;
	}

	q->properties.doorbell_off =
		kfd_doorbell_id_to_offset(dev, q->process,
					  q->doorbell_id);

	return 0;
}

static void deallocate_doorbell(struct qcm_process_device *qpd,
				struct queue *q)
{
	unsigned int old;
	struct kfd_dev *dev = qpd->dqm->dev;

	if (!KFD_IS_SOC15(dev->device_info->asic_family) ||
	    q->properties.type == KFD_QUEUE_TYPE_SDMA)
		return;

	old = test_and_clear_bit(q->doorbell_id, qpd->doorbell_bitmap);
	WARN_ON(!old);
}

static int allocate_vmid(struct device_queue_manager *dqm,
			struct qcm_process_device *qpd,
			struct queue *q)
{
	int bit, allocated_vmid;

	if (dqm->vmid_bitmap == 0)
		return -ENOMEM;

	bit = ffs(dqm->vmid_bitmap) - 1;
	dqm->vmid_bitmap &= ~(1 << bit);

	allocated_vmid = bit + dqm->dev->vm_info.first_vmid_kfd;
	pr_debug("vmid allocation %d\n", allocated_vmid);
	qpd->vmid = allocated_vmid;
	q->properties.vmid = allocated_vmid;

	set_pasid_vmid_mapping(dqm, q->process->pasid, q->properties.vmid);
	program_sh_mem_settings(dqm, qpd);

	/* qpd->page_table_base is set earlier when register_process()
	 * is called, i.e. when the first queue is created.
	 */
	dqm->dev->kfd2kgd->set_vm_context_page_table_base(dqm->dev->kgd,
			qpd->vmid,
			qpd->page_table_base);
	/* invalidate the VM context after pasid and vmid mapping is set up */
	kfd_flush_tlb(qpd_to_pdd(qpd));

	return 0;
}

static int flush_texture_cache_nocpsch(struct kfd_dev *kdev,
				struct qcm_process_device *qpd)
{
	const struct packet_manager_funcs *pmf = qpd->dqm->packets.pmf;
	int ret;

	if (!qpd->ib_kaddr)
		return -ENOMEM;

	ret = pmf->release_mem(qpd->ib_base, (uint32_t *)qpd->ib_kaddr);
	if (ret)
		return ret;

	return kdev->kfd2kgd->submit_ib(kdev->kgd, KGD_ENGINE_MEC1, qpd->vmid,
				qpd->ib_base, (uint32_t *)qpd->ib_kaddr,
				pmf->release_mem_size / sizeof(uint32_t));
}

static void deallocate_vmid(struct device_queue_manager *dqm,
				struct qcm_process_device *qpd,
				struct queue *q)
{
	int bit = qpd->vmid - dqm->dev->vm_info.first_vmid_kfd;

	/* On GFX v7, CP doesn't flush TC at dequeue */
	if (q->device->device_info->asic_family == CHIP_HAWAII)
		if (flush_texture_cache_nocpsch(q->device, qpd))
			pr_err("Failed to flush TC\n");

	kfd_flush_tlb(qpd_to_pdd(qpd));

	/* Release the vmid mapping */
	set_pasid_vmid_mapping(dqm, 0, qpd->vmid);

	dqm->vmid_bitmap |= (1 << bit);
	qpd->vmid = 0;
	q->properties.vmid = 0;
}

static int create_queue_nocpsch(struct device_queue_manager *dqm,
				struct queue *q,
				struct qcm_process_device *qpd)
{
	int retval;

	print_queue(q);

	mutex_lock(&dqm->lock);

	if (dqm->total_queue_count >= max_num_of_queues_per_device) {
		pr_warn("Can't create new usermode queue because %d queues were already created\n",
				dqm->total_queue_count);
		retval = -EPERM;
		goto out_unlock;
	}

	if (list_empty(&qpd->queues_list)) {
		retval = allocate_vmid(dqm, qpd, q);
		if (retval)
			goto out_unlock;
	}
	q->properties.vmid = qpd->vmid;
	/*
	 * Eviction state logic: we only mark active queues as evicted
	 * to avoid the overhead of restoring inactive queues later
	 */
	if (qpd->evicted)
		q->properties.is_evicted = (q->properties.queue_size > 0 &&
					    q->properties.queue_percent > 0 &&
					    q->properties.queue_address != 0);

	q->properties.tba_addr = qpd->tba_addr;
	q->properties.tma_addr = qpd->tma_addr;

	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE)
		retval = create_compute_queue_nocpsch(dqm, q, qpd);
	else if (q->properties.type == KFD_QUEUE_TYPE_SDMA)
		retval = create_sdma_queue_nocpsch(dqm, q, qpd);
	else
		retval = -EINVAL;

	if (retval) {
		if (list_empty(&qpd->queues_list))
			deallocate_vmid(dqm, qpd, q);
		goto out_unlock;
	}

	list_add(&q->list, &qpd->queues_list);
	qpd->queue_count++;
	if (q->properties.is_active)
		dqm->queue_count++;

	if (q->properties.type == KFD_QUEUE_TYPE_SDMA)
		dqm->sdma_queue_count++;

	/*
	 * Unconditionally increment this counter, regardless of the queue's
	 * type or whether the queue is active.
	 */
	dqm->total_queue_count++;
	pr_debug("Total of %d queues are accountable so far\n",
			dqm->total_queue_count);

out_unlock:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int allocate_hqd(struct device_queue_manager *dqm, struct queue *q)
{
	bool set;
	int pipe, bit, i;

	set = false;

	for (pipe = dqm->next_pipe_to_allocate, i = 0;
			i < get_pipes_per_mec(dqm);
			pipe = ((pipe + 1) % get_pipes_per_mec(dqm)), ++i) {

		if (!is_pipe_enabled(dqm, 0, pipe))
			continue;

		if (dqm->allocated_queues[pipe] != 0) {
			bit = ffs(dqm->allocated_queues[pipe]) - 1;
			dqm->allocated_queues[pipe] &= ~(1 << bit);
			q->pipe = pipe;
			q->queue = bit;
			set = true;
			break;
		}
	}

	if (!set)
		return -EBUSY;

	pr_debug("hqd slot - pipe %d, queue %d\n", q->pipe, q->queue);
	/* horizontal hqd allocation */
	dqm->next_pipe_to_allocate = (pipe + 1) % get_pipes_per_mec(dqm);

	return 0;
}

static inline void deallocate_hqd(struct device_queue_manager *dqm,
				struct queue *q)
{
	dqm->allocated_queues[q->pipe] |= (1 << q->queue);
}

static int create_compute_queue_nocpsch(struct device_queue_manager *dqm,
					struct queue *q,
					struct qcm_process_device *qpd)
{
	int retval;
	struct mqd_manager *mqd;

	mqd = dqm->ops.get_mqd_manager(dqm, KFD_MQD_TYPE_COMPUTE);
	if (!mqd)
		return -ENOMEM;

	retval = allocate_hqd(dqm, q);
	if (retval)
		return retval;

	retval = allocate_doorbell(qpd, q);
	if (retval)
		goto out_deallocate_hqd;

	retval = mqd->init_mqd(mqd, &q->mqd, &q->mqd_mem_obj,
				&q->gart_mqd_addr, &q->properties);
	if (retval)
		goto out_deallocate_doorbell;

	pr_debug("Loading mqd to hqd on pipe %d, queue %d\n",
			q->pipe, q->queue);

	dqm->dev->kfd2kgd->set_scratch_backing_va(
			dqm->dev->kgd, qpd->sh_hidden_private_base, qpd->vmid);

	if (!q->properties.is_active)
		return 0;

	retval = mqd->load_mqd(mqd, q->mqd, q->pipe, q->queue, &q->properties,
			       q->process->mm);
	if (retval)
		goto out_uninit_mqd;

	return 0;

out_uninit_mqd:
	mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
out_deallocate_doorbell:
	deallocate_doorbell(qpd, q);
out_deallocate_hqd:
	deallocate_hqd(dqm, q);

	return retval;
}

/* Access to DQM has to be locked before calling destroy_queue_nocpsch_locked
 * to avoid asynchronized access
 */
static int destroy_queue_nocpsch_locked(struct device_queue_manager *dqm,
				struct qcm_process_device *qpd,
				struct queue *q)
{
	int retval;
	struct mqd_manager *mqd;

	mqd = dqm->ops.get_mqd_manager(dqm,
		get_mqd_type_from_queue_type(q->properties.type));
	if (!mqd)
		return -ENOMEM;

	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE) {
		deallocate_hqd(dqm, q);
	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
		dqm->sdma_queue_count--;
		deallocate_sdma_queue(dqm, q->sdma_id);
	} else {
		pr_debug("q->properties.type %d is invalid\n",
				q->properties.type);
		return -EINVAL;
	}
	dqm->total_queue_count--;

	deallocate_doorbell(qpd, q);

	retval = mqd->destroy_mqd(mqd, q->mqd,
				KFD_PREEMPT_TYPE_WAVEFRONT_RESET,
				KFD_UNMAP_LATENCY_MS,
				q->pipe, q->queue);
	if (retval == -ETIME)
		qpd->reset_wavefronts = true;

	mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);

	list_del(&q->list);
	if (list_empty(&qpd->queues_list)) {
		if (qpd->reset_wavefronts) {
			pr_warn("Resetting wave fronts (nocpsch) on dev %p\n",
					dqm->dev);
			/* dbgdev_wave_reset_wavefronts has to be called before
			 * deallocate_vmid(), i.e. when vmid is still in use.
			 */
			dbgdev_wave_reset_wavefronts(dqm->dev,
					qpd->pqm->process);
			qpd->reset_wavefronts = false;
		}

		deallocate_vmid(dqm, qpd, q);
	}
	qpd->queue_count--;
	if (q->properties.is_active)
		dqm->queue_count--;

	return retval;
}

static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
				struct qcm_process_device *qpd,
				struct queue *q)
{
	int retval;

	mutex_lock(&dqm->lock);
	retval = destroy_queue_nocpsch_locked(dqm, qpd, q);
	mutex_unlock(&dqm->lock);

	return retval;
}

static int update_queue(struct device_queue_manager *dqm, struct queue *q)
{
	int retval;
	struct mqd_manager *mqd;
	struct kfd_process_device *pdd;
	bool prev_active = false;

	mutex_lock(&dqm->lock);
	pdd = kfd_get_process_device_data(q->device, q->process);
	if (!pdd) {
		retval = -ENODEV;
		goto out_unlock;
	}
	mqd = dqm->ops.get_mqd_manager(dqm,
			get_mqd_type_from_queue_type(q->properties.type));
	if (!mqd) {
		retval = -ENOMEM;
		goto out_unlock;
	}
	/*
	 * Eviction state logic: we only mark active queues as evicted
	 * to avoid the overhead of restoring inactive queues later
	 */
	if (pdd->qpd.evicted)
		q->properties.is_evicted = (q->properties.queue_size > 0 &&
					    q->properties.queue_percent > 0 &&
					    q->properties.queue_address != 0);

	/* Save previous activity state for counters */
	prev_active = q->properties.is_active;

	/* Make sure the queue is unmapped before updating the MQD */
	if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS) {
		retval = unmap_queues_cpsch(dqm,
				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
		if (retval) {
			pr_err("unmap queue failed\n");
			goto out_unlock;
		}
	} else if (prev_active &&
		   (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
		    q->properties.type == KFD_QUEUE_TYPE_SDMA)) {
		retval = mqd->destroy_mqd(mqd, q->mqd,
				KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN,
				KFD_UNMAP_LATENCY_MS, q->pipe, q->queue);
		if (retval) {
			pr_err("destroy mqd failed\n");
			goto out_unlock;
		}
	}

	retval = mqd->update_mqd(mqd, q->mqd, &q->properties);

	/*
	 * check active state vs. the previous state and modify
	 * counter accordingly. map_queues_cpsch uses the
	 * dqm->queue_count to determine whether a new runlist must be
	 * uploaded.
	 */
	if (q->properties.is_active && !prev_active)
		dqm->queue_count++;
	else if (!q->properties.is_active && prev_active)
		dqm->queue_count--;

	if (dqm->sched_policy != KFD_SCHED_POLICY_NO_HWS)
		retval = map_queues_cpsch(dqm);
	else if (q->properties.is_active &&
		 (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
		  q->properties.type == KFD_QUEUE_TYPE_SDMA))
		retval = mqd->load_mqd(mqd, q->mqd, q->pipe, q->queue,
				       &q->properties, q->process->mm);

out_unlock:
	mutex_unlock(&dqm->lock);
	return retval;
}

static struct mqd_manager *get_mqd_manager(
		struct device_queue_manager *dqm, enum KFD_MQD_TYPE type)
{
	struct mqd_manager *mqd;

	if (WARN_ON(type >= KFD_MQD_TYPE_MAX))
		return NULL;

	pr_debug("mqd type %d\n", type);

	mqd = dqm->mqds[type];
	if (!mqd) {
		mqd = mqd_manager_init(type, dqm->dev);
		if (!mqd)
			pr_err("mqd manager is NULL");
		dqm->mqds[type] = mqd;
	}

	return mqd;
}

static int evict_process_queues_nocpsch(struct device_queue_manager *dqm,
					struct qcm_process_device *qpd)
{
	struct queue *q;
	struct mqd_manager *mqd;
	struct kfd_process_device *pdd;
	int retval = 0;

	mutex_lock(&dqm->lock);
	if (qpd->evicted++ > 0) /* already evicted, do nothing */
		goto out;

	pdd = qpd_to_pdd(qpd);
	pr_info_ratelimited("Evicting PASID %u queues\n",
			    pdd->process->pasid);

	/* unactivate all active queues on the qpd */
	list_for_each_entry(q, &qpd->queues_list, list) {
		if (!q->properties.is_active)
			continue;
		mqd = dqm->ops.get_mqd_manager(dqm,
			get_mqd_type_from_queue_type(q->properties.type));
		if (!mqd) { /* should not be here */
			pr_err("Cannot evict queue, mqd mgr is NULL\n");
			retval = -ENOMEM;
			goto out;
		}
		q->properties.is_evicted = true;
		q->properties.is_active = false;
		retval = mqd->destroy_mqd(mqd, q->mqd,
				KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN,
				KFD_UNMAP_LATENCY_MS, q->pipe, q->queue);
		if (retval)
			goto out;
		dqm->queue_count--;
	}

out:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int evict_process_queues_cpsch(struct device_queue_manager *dqm,
				      struct qcm_process_device *qpd)
{
	struct queue *q;
	struct kfd_process_device *pdd;
	int retval = 0;

	mutex_lock(&dqm->lock);
	if (qpd->evicted++ > 0) /* already evicted, do nothing */
		goto out;

	pdd = qpd_to_pdd(qpd);
	pr_info_ratelimited("Evicting PASID %u queues\n",
			    pdd->process->pasid);

	/* unactivate all active queues on the qpd */
	list_for_each_entry(q, &qpd->queues_list, list) {
		if (!q->properties.is_active)
			continue;
		q->properties.is_evicted = true;
		q->properties.is_active = false;
		dqm->queue_count--;
	}
	retval = execute_queues_cpsch(dqm,
				qpd->is_debug ?
				KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES :
				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);

out:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int restore_process_queues_nocpsch(struct device_queue_manager *dqm,
					  struct qcm_process_device *qpd)
{
	struct queue *q;
	struct mqd_manager *mqd;
	struct kfd_process_device *pdd;
	uint32_t pd_base;
	int retval = 0;

	pdd = qpd_to_pdd(qpd);
	/* Retrieve PD base */
	pd_base = dqm->dev->kfd2kgd->get_process_page_dir(pdd->vm);

	mutex_lock(&dqm->lock);
	if (WARN_ON_ONCE(!qpd->evicted)) /* already restored, do nothing */
		goto out;
	if (qpd->evicted > 1) { /* ref count still > 0, decrement & quit */
		qpd->evicted--;
		goto out;
	}

	pr_info_ratelimited("Restoring PASID %u queues\n",
			    pdd->process->pasid);

	/* Update PD Base in QPD */
	qpd->page_table_base = pd_base;
	pr_debug("Updated PD address to 0x%08x\n", pd_base);

	if (!list_empty(&qpd->queues_list)) {
		dqm->dev->kfd2kgd->set_vm_context_page_table_base(
				dqm->dev->kgd,
				qpd->vmid,
				qpd->page_table_base);
		kfd_flush_tlb(pdd);
	}

	/* activate all active queues on the qpd */
	list_for_each_entry(q, &qpd->queues_list, list) {
		if (!q->properties.is_evicted)
			continue;
		mqd = dqm->ops.get_mqd_manager(dqm,
			get_mqd_type_from_queue_type(q->properties.type));
		if (!mqd) { /* should not be here */
			pr_err("Cannot restore queue, mqd mgr is NULL\n");
			retval = -ENOMEM;
			goto out;
		}
		q->properties.is_evicted = false;
		q->properties.is_active = true;
		retval = mqd->load_mqd(mqd, q->mqd, q->pipe,
				       q->queue, &q->properties,
				       q->process->mm);
		if (retval)
			goto out;
		dqm->queue_count++;
	}
	qpd->evicted = 0;
out:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int restore_process_queues_cpsch(struct device_queue_manager *dqm,
					struct qcm_process_device *qpd)
{
	struct queue *q;
	struct kfd_process_device *pdd;
	uint32_t pd_base;
	int retval = 0;

	pdd = qpd_to_pdd(qpd);
	/* Retrieve PD base */
	pd_base = dqm->dev->kfd2kgd->get_process_page_dir(pdd->vm);

	mutex_lock(&dqm->lock);
	if (WARN_ON_ONCE(!qpd->evicted)) /* already restored, do nothing */
		goto out;
	if (qpd->evicted > 1) { /* ref count still > 0, decrement & quit */
		qpd->evicted--;
		goto out;
	}

	pr_info_ratelimited("Restoring PASID %u queues\n",
			    pdd->process->pasid);

	/* Update PD Base in QPD */
	qpd->page_table_base = pd_base;
	pr_debug("Updated PD address to 0x%08x\n", pd_base);

	/* activate all active queues on the qpd */
	list_for_each_entry(q, &qpd->queues_list, list) {
		if (!q->properties.is_evicted)
			continue;
		q->properties.is_evicted = false;
		q->properties.is_active = true;
		dqm->queue_count++;
	}
	retval = execute_queues_cpsch(dqm,
				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
	if (!retval)
		qpd->evicted = 0;
out:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int register_process(struct device_queue_manager *dqm,
					struct qcm_process_device *qpd)
{
	struct device_process_node *n;
	struct kfd_process_device *pdd;
	uint32_t pd_base;
	int retval;

	n = kzalloc(sizeof(*n), GFP_KERNEL);
	if (!n)
		return -ENOMEM;

	n->qpd = qpd;

	pdd = qpd_to_pdd(qpd);
	/* Retrieve PD base */
	pd_base = dqm->dev->kfd2kgd->get_process_page_dir(pdd->vm);

	mutex_lock(&dqm->lock);
	list_add(&n->list, &dqm->queues);

	/* Update PD Base in QPD */
	qpd->page_table_base = pd_base;

	retval = dqm->asic_ops.update_qpd(dqm, qpd);

	dqm->processes_count++;

	mutex_unlock(&dqm->lock);

	return retval;
}

static int unregister_process(struct device_queue_manager *dqm,
					struct qcm_process_device *qpd)
{
	int retval;
	struct device_process_node *cur, *next;

	pr_debug("qpd->queues_list is %s\n",
			list_empty(&qpd->queues_list) ? "empty" : "not empty");

	retval = 0;
	mutex_lock(&dqm->lock);

	list_for_each_entry_safe(cur, next, &dqm->queues, list) {
		if (qpd == cur->qpd) {
			list_del(&cur->list);
			kfree(cur);
			dqm->processes_count--;
			goto out;
		}
	}
	/* qpd not found in dqm list */
	retval = 1;
out:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int
set_pasid_vmid_mapping(struct device_queue_manager *dqm, unsigned int pasid,
			unsigned int vmid)
{
	uint32_t pasid_mapping;

	pasid_mapping = (pasid == 0) ? 0 :
		(uint32_t)pasid |
		ATC_VMID_PASID_MAPPING_VALID;

	return dqm->dev->kfd2kgd->set_pasid_vmid_mapping(
						dqm->dev->kgd, pasid_mapping,
						vmid);
}

static void init_interrupts(struct device_queue_manager *dqm)
{
	unsigned int i;

	for (i = 0 ; i < get_pipes_per_mec(dqm) ; i++)
		if (is_pipe_enabled(dqm, 0, i))
			dqm->dev->kfd2kgd->init_interrupts(dqm->dev->kgd, i);
}

static int initialize_nocpsch(struct device_queue_manager *dqm)
{
	int pipe, queue;

	pr_debug("num of pipes: %d\n", get_pipes_per_mec(dqm));

	dqm->allocated_queues = kcalloc(get_pipes_per_mec(dqm),
					sizeof(unsigned int), GFP_KERNEL);
	if (!dqm->allocated_queues)
		return -ENOMEM;

	mutex_init(&dqm->lock);
	INIT_LIST_HEAD(&dqm->queues);
	dqm->queue_count = dqm->next_pipe_to_allocate = 0;
	dqm->sdma_queue_count = 0;

	for (pipe = 0; pipe < get_pipes_per_mec(dqm); pipe++) {
		int pipe_offset = pipe * get_queues_per_pipe(dqm);

		for (queue = 0; queue < get_queues_per_pipe(dqm); queue++)
			if (test_bit(pipe_offset + queue,
				     dqm->dev->shared_resources.queue_bitmap))
				dqm->allocated_queues[pipe] |= 1 << queue;
	}

	dqm->vmid_bitmap = (1 << dqm->dev->vm_info.vmid_num_kfd) - 1;
	dqm->sdma_bitmap = (1 << CIK_SDMA_QUEUES) - 1;

	return 0;
}

static void uninitialize(struct device_queue_manager *dqm)
{
	int i;

	WARN_ON(dqm->queue_count > 0 || dqm->processes_count > 0);

	kfree(dqm->allocated_queues);
	for (i = 0 ; i < KFD_MQD_TYPE_MAX ; i++)
		kfree(dqm->mqds[i]);
	mutex_destroy(&dqm->lock);
	kfd_gtt_sa_free(dqm->dev, dqm->pipeline_mem);
}

static int start_nocpsch(struct device_queue_manager *dqm)
{
	init_interrupts(dqm);
	return pm_init(&dqm->packets, dqm);
}

static int stop_nocpsch(struct device_queue_manager *dqm)
{
	pm_uninit(&dqm->packets);
	return 0;
}

static int allocate_sdma_queue(struct device_queue_manager *dqm,
				unsigned int *sdma_queue_id)
{
	int bit;

	if (dqm->sdma_bitmap == 0)
		return -ENOMEM;

	bit = ffs(dqm->sdma_bitmap) - 1;
	dqm->sdma_bitmap &= ~(1 << bit);
	*sdma_queue_id = bit;

	return 0;
}

static void deallocate_sdma_queue(struct device_queue_manager *dqm,
				unsigned int sdma_queue_id)
{
	if (sdma_queue_id >= CIK_SDMA_QUEUES)
		return;
	dqm->sdma_bitmap |= (1 << sdma_queue_id);
}

static int create_sdma_queue_nocpsch(struct device_queue_manager *dqm,
					struct queue *q,
					struct qcm_process_device *qpd)
{
	struct mqd_manager *mqd;
	int retval;

	mqd = dqm->ops.get_mqd_manager(dqm, KFD_MQD_TYPE_SDMA);
	if (!mqd)
		return -ENOMEM;

	retval = allocate_sdma_queue(dqm, &q->sdma_id);
	if (retval)
		return retval;

	q->properties.sdma_queue_id = q->sdma_id / CIK_SDMA_QUEUES_PER_ENGINE;
	q->properties.sdma_engine_id = q->sdma_id % CIK_SDMA_QUEUES_PER_ENGINE;

	retval = allocate_doorbell(qpd, q);
	if (retval)
		goto out_deallocate_sdma_queue;

	pr_debug("SDMA id is:    %d\n", q->sdma_id);
	pr_debug("SDMA queue id: %d\n", q->properties.sdma_queue_id);
	pr_debug("SDMA engine id: %d\n", q->properties.sdma_engine_id);

	dqm->asic_ops.init_sdma_vm(dqm, q, qpd);
	retval = mqd->init_mqd(mqd, &q->mqd, &q->mqd_mem_obj,
				&q->gart_mqd_addr, &q->properties);
	if (retval)
		goto out_deallocate_doorbell;

	retval = mqd->load_mqd(mqd, q->mqd, 0, 0, &q->properties, NULL);
	if (retval)
		goto out_uninit_mqd;

	return 0;

out_uninit_mqd:
	mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
out_deallocate_doorbell:
	deallocate_doorbell(qpd, q);
out_deallocate_sdma_queue:
	deallocate_sdma_queue(dqm, q->sdma_id);

	return retval;
}

/*
 * Device Queue Manager implementation for cp scheduler
 */

static int set_sched_resources(struct device_queue_manager *dqm)
{
	int i, mec;
	struct scheduling_resources res;

	res.vmid_mask = dqm->dev->shared_resources.compute_vmid_bitmap;

	res.queue_mask = 0;
	for (i = 0; i < KGD_MAX_QUEUES; ++i) {
		mec = (i / dqm->dev->shared_resources.num_queue_per_pipe)
			/ dqm->dev->shared_resources.num_pipe_per_mec;

		if (!test_bit(i, dqm->dev->shared_resources.queue_bitmap))
			continue;

		/* only acquire queues from the first MEC */
		if (mec > 0)
			continue;

		/* This situation may be hit in the future if a new HW
		 * generation exposes more than 64 queues. If so, the
		 * definition of res.queue_mask needs updating
		 */
		if (WARN_ON(i >= (sizeof(res.queue_mask)*8))) {
			pr_err("Invalid queue enabled by amdgpu: %d\n", i);
			break;
		}

		res.queue_mask |= (1ull << i);
	}
	res.gws_mask = res.oac_mask = res.gds_heap_base =
						res.gds_heap_size = 0;

	pr_debug("Scheduling resources:\n"
			"vmid mask: 0x%8X\n"
			"queue mask: 0x%8llX\n",
			res.vmid_mask, res.queue_mask);

	return pm_send_set_resources(&dqm->packets, &res);
}

static int initialize_cpsch(struct device_queue_manager *dqm)
{
	pr_debug("num of pipes: %d\n", get_pipes_per_mec(dqm));

	mutex_init(&dqm->lock);
	INIT_LIST_HEAD(&dqm->queues);
	dqm->queue_count = dqm->processes_count = 0;
	dqm->sdma_queue_count = 0;
	dqm->active_runlist = false;
	dqm->sdma_bitmap = (1 << CIK_SDMA_QUEUES) - 1;

	return 0;
}

static int start_cpsch(struct device_queue_manager *dqm)
{
	int retval;

	retval = 0;

	retval = pm_init(&dqm->packets, dqm);
	if (retval)
		goto fail_packet_manager_init;

	retval = set_sched_resources(dqm);
	if (retval)
		goto fail_set_sched_resources;

	pr_debug("Allocating fence memory\n");

	/* allocate fence memory on the gart */
	retval = kfd_gtt_sa_allocate(dqm->dev, sizeof(*dqm->fence_addr),
					&dqm->fence_mem);

	if (retval)
		goto fail_allocate_vidmem;

	dqm->fence_addr = dqm->fence_mem->cpu_ptr;
	dqm->fence_gpu_addr = dqm->fence_mem->gpu_addr;

	init_interrupts(dqm);

	mutex_lock(&dqm->lock);
	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
	mutex_unlock(&dqm->lock);

	return 0;
fail_allocate_vidmem:
fail_set_sched_resources:
	pm_uninit(&dqm->packets);
fail_packet_manager_init:
	return retval;
}

static int stop_cpsch(struct device_queue_manager *dqm)
{
	mutex_lock(&dqm->lock);
	unmap_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0);
	mutex_unlock(&dqm->lock);

	kfd_gtt_sa_free(dqm->dev, dqm->fence_mem);
	pm_uninit(&dqm->packets);

	return 0;
}

static int create_kernel_queue_cpsch(struct device_queue_manager *dqm,
					struct kernel_queue *kq,
					struct qcm_process_device *qpd)
{
	mutex_lock(&dqm->lock);
	if (dqm->total_queue_count >= max_num_of_queues_per_device) {
		pr_warn("Can't create new kernel queue because %d queues were already created\n",
				dqm->total_queue_count);
		mutex_unlock(&dqm->lock);
		return -EPERM;
	}

	/*
	 * Unconditionally increment this counter, regardless of the queue's
	 * type or whether the queue is active.
	 */
	dqm->total_queue_count++;
	pr_debug("Total of %d queues are accountable so far\n",
			dqm->total_queue_count);

	list_add(&kq->list, &qpd->priv_queue_list);
	dqm->queue_count++;
	qpd->is_debug = true;
	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
	mutex_unlock(&dqm->lock);

	return 0;
}

static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm,
					struct kernel_queue *kq,
					struct qcm_process_device *qpd)
{
	mutex_lock(&dqm->lock);
	list_del(&kq->list);
	dqm->queue_count--;
	qpd->is_debug = false;
	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0);
	/*
	 * Unconditionally decrement this counter, regardless of the queue's
	 * type.
	 */
	dqm->total_queue_count--;
	pr_debug("Total of %d queues are accountable so far\n",
			dqm->total_queue_count);
	mutex_unlock(&dqm->lock);
}

static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
			struct qcm_process_device *qpd)
{
	int retval;
	struct mqd_manager *mqd;

	retval = 0;

	mutex_lock(&dqm->lock);

	if (dqm->total_queue_count >= max_num_of_queues_per_device) {
		pr_warn("Can't create new usermode queue because %d queues were already created\n",
				dqm->total_queue_count);
		retval = -EPERM;
		goto out_unlock;
	}

	if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
		retval = allocate_sdma_queue(dqm, &q->sdma_id);
		if (retval)
			goto out_unlock;
		q->properties.sdma_queue_id =
			q->sdma_id / CIK_SDMA_QUEUES_PER_ENGINE;
		q->properties.sdma_engine_id =
			q->sdma_id % CIK_SDMA_QUEUES_PER_ENGINE;
	}

	retval = allocate_doorbell(qpd, q);
	if (retval)
		goto out_deallocate_sdma_queue;

	mqd = dqm->ops.get_mqd_manager(dqm,
			get_mqd_type_from_queue_type(q->properties.type));

	if (!mqd) {
		retval = -ENOMEM;
		goto out_deallocate_doorbell;
	}
	/*
	 * Eviction state logic: we only mark active queues as evicted
	 * to avoid the overhead of restoring inactive queues later
	 */
	if (qpd->evicted)
		q->properties.is_evicted = (q->properties.queue_size > 0 &&
					    q->properties.queue_percent > 0 &&
					    q->properties.queue_address != 0);

	dqm->asic_ops.init_sdma_vm(dqm, q, qpd);

	q->properties.tba_addr = qpd->tba_addr;
	q->properties.tma_addr = qpd->tma_addr;
	retval = mqd->init_mqd(mqd, &q->mqd, &q->mqd_mem_obj,
				&q->gart_mqd_addr, &q->properties);
	if (retval)
		goto out_deallocate_doorbell;

	list_add(&q->list, &qpd->queues_list);
	qpd->queue_count++;
	if (q->properties.is_active) {
		dqm->queue_count++;
		retval = execute_queues_cpsch(dqm,
				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
	}

	if (q->properties.type == KFD_QUEUE_TYPE_SDMA)
		dqm->sdma_queue_count++;
	/*
	 * Unconditionally increment this counter, regardless of the queue's
	 * type or whether the queue is active.
	 */
	dqm->total_queue_count++;

	pr_debug("Total of %d queues are accountable so far\n",
			dqm->total_queue_count);

	mutex_unlock(&dqm->lock);
	return retval;

out_deallocate_doorbell:
	deallocate_doorbell(qpd, q);
out_deallocate_sdma_queue:
	if (q->properties.type == KFD_QUEUE_TYPE_SDMA)
		deallocate_sdma_queue(dqm, q->sdma_id);
out_unlock:
	mutex_unlock(&dqm->lock);
	return retval;
}

int amdkfd_fence_wait_timeout(unsigned int *fence_addr,
				unsigned int fence_value,
				unsigned int timeout_ms)
{
	unsigned long end_jiffies = msecs_to_jiffies(timeout_ms) + jiffies;

	while (*fence_addr != fence_value) {
		if (time_after(jiffies, end_jiffies)) {
			pr_err("qcm fence wait loop timeout expired\n");
			return -ETIME;
		}
		schedule();
	}

	return 0;
}

static int unmap_sdma_queues(struct device_queue_manager *dqm,
				unsigned int sdma_engine)
{
	return pm_send_unmap_queue(&dqm->packets, KFD_QUEUE_TYPE_SDMA,
			KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, false,
			sdma_engine);
}

/* dqm->lock mutex has to be locked before calling this function */
static int map_queues_cpsch(struct device_queue_manager *dqm)
{
	int retval;

	if (dqm->queue_count <= 0 || dqm->processes_count <= 0)
		return 0;

	if (dqm->active_runlist)
		return 0;

	retval = pm_send_runlist(&dqm->packets, &dqm->queues);
	if (retval) {
		pr_err("failed to execute runlist\n");
		return retval;
	}
	dqm->active_runlist = true;

	return retval;
}

/* dqm->lock mutex has to be locked before calling this function */
static int unmap_queues_cpsch(struct device_queue_manager *dqm,
				enum kfd_unmap_queues_filter filter,
				uint32_t filter_param)
{
	int retval = 0;

	if (!dqm->active_runlist)
		return retval;

	pr_debug("Before destroying queues, sdma queue count is : %u\n",
		dqm->sdma_queue_count);

	if (dqm->sdma_queue_count > 0) {
		unmap_sdma_queues(dqm, 0);
		unmap_sdma_queues(dqm, 1);
	}

	retval = pm_send_unmap_queue(&dqm->packets, KFD_QUEUE_TYPE_COMPUTE,
			filter, filter_param, false, 0);
	if (retval)
		return retval;

	*dqm->fence_addr = KFD_FENCE_INIT;
	pm_send_query_status(&dqm->packets, dqm->fence_gpu_addr,
				KFD_FENCE_COMPLETED);
	/* should be timed out */
	retval = amdkfd_fence_wait_timeout(dqm->fence_addr, KFD_FENCE_COMPLETED,
				QUEUE_PREEMPT_DEFAULT_TIMEOUT_MS);
	if (retval)
		return retval;

	pm_release_ib(&dqm->packets);
	dqm->active_runlist = false;

	return retval;
}

/* dqm->lock mutex has to be locked before calling this function */
static int execute_queues_cpsch(struct device_queue_manager *dqm,
				enum kfd_unmap_queues_filter filter,
				uint32_t filter_param)
{
	int retval;

	retval = unmap_queues_cpsch(dqm, filter, filter_param);
	if (retval) {
		pr_err("The cp might be in an unrecoverable state due to an unsuccessful queues preemption\n");
		return retval;
	}

	return map_queues_cpsch(dqm);
}

static int destroy_queue_cpsch(struct device_queue_manager *dqm,
				struct qcm_process_device *qpd,
				struct queue *q)
{
	int retval;
	struct mqd_manager *mqd;
	bool preempt_all_queues;

	preempt_all_queues = false;

	retval = 0;

	/* remove queue from list to prevent rescheduling after preemption */
	mutex_lock(&dqm->lock);

	if (qpd->is_debug) {
		/*
		 * error, currently we do not allow to destroy a queue
		 * of a currently debugged process
		 */
		retval = -EBUSY;
		goto failed_try_destroy_debugged_queue;

	}

	mqd = dqm->ops.get_mqd_manager(dqm,
			get_mqd_type_from_queue_type(q->properties.type));
	if (!mqd) {
		retval = -ENOMEM;
		goto failed;
	}

	deallocate_doorbell(qpd, q);

	if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
		dqm->sdma_queue_count--;
		deallocate_sdma_queue(dqm, q->sdma_id);
	}

	list_del(&q->list);
	qpd->queue_count--;
	if (q->properties.is_active) {
		dqm->queue_count--;
		retval = execute_queues_cpsch(dqm,
				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
		if (retval == -ETIME)
			qpd->reset_wavefronts = true;
	}

	mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);

	/*
	 * Unconditionally decrement this counter, regardless of the queue's
	 * type
	 */
	dqm->total_queue_count--;
	pr_debug("Total of %d queues are accountable so far\n",
			dqm->total_queue_count);

	mutex_unlock(&dqm->lock);

	return retval;

failed:
failed_try_destroy_debugged_queue:

	mutex_unlock(&dqm->lock);
	return retval;
}

/*
 * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
 * stay in user mode.
 */
#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
/* APE1 limit is inclusive and 64K aligned. */
#define APE1_LIMIT_ALIGNMENT 0xFFFF

static bool set_cache_memory_policy(struct device_queue_manager *dqm,
				   struct qcm_process_device *qpd,
				   enum cache_policy default_policy,
				   enum cache_policy alternate_policy,
				   void __user *alternate_aperture_base,
				   uint64_t alternate_aperture_size)
{
	bool retval = true;

	if (!dqm->asic_ops.set_cache_memory_policy)
		return retval;

	mutex_lock(&dqm->lock);

	if (alternate_aperture_size == 0) {
		/* base > limit disables APE1 */
		qpd->sh_mem_ape1_base = 1;
		qpd->sh_mem_ape1_limit = 0;
	} else {
		/*
		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
		 * Verify that the base and size parameters can be
		 * represented in this format and convert them.
		 * Additionally restrict APE1 to user-mode addresses.
		 */

		uint64_t base = (uintptr_t)alternate_aperture_base;
		uint64_t limit = base + alternate_aperture_size - 1;

		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
			retval = false;
			goto out;
		}

		qpd->sh_mem_ape1_base = base >> 16;
		qpd->sh_mem_ape1_limit = limit >> 16;
	}

	retval = dqm->asic_ops.set_cache_memory_policy(
			dqm,
			qpd,
			default_policy,
			alternate_policy,
			alternate_aperture_base,
			alternate_aperture_size);

	if ((dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) && (qpd->vmid != 0))
		program_sh_mem_settings(dqm, qpd);

	pr_debug("sh_mem_config: 0x%x, ape1_base: 0x%x, ape1_limit: 0x%x\n",
		qpd->sh_mem_config, qpd->sh_mem_ape1_base,
		qpd->sh_mem_ape1_limit);

out:
	mutex_unlock(&dqm->lock);
	return retval;
}

static int set_trap_handler(struct device_queue_manager *dqm,
				struct qcm_process_device *qpd,
				uint64_t tba_addr,
				uint64_t tma_addr)
{
	uint64_t *tma;

	if (dqm->dev->cwsr_enabled) {
		/* Jump from CWSR trap handler to user trap */
		tma = (uint64_t *)(qpd->cwsr_kaddr + KFD_CWSR_TMA_OFFSET);
		tma[0] = tba_addr;
		tma[1] = tma_addr;
	} else {
		qpd->tba_addr = tba_addr;
		qpd->tma_addr = tma_addr;
	}

	return 0;
}

static int process_termination_nocpsch(struct device_queue_manager *dqm,
		struct qcm_process_device *qpd)
{
	struct queue *q, *next;
	struct device_process_node *cur, *next_dpn;
	int retval = 0;

	mutex_lock(&dqm->lock);

	/* Clear all user mode queues */
	list_for_each_entry_safe(q, next, &qpd->queues_list, list) {
		int ret;

		ret = destroy_queue_nocpsch_locked(dqm, qpd, q);
		if (ret)
			retval = ret;
	}

	/* Unregister process */
	list_for_each_entry_safe(cur, next_dpn, &dqm->queues, list) {
		if (qpd == cur->qpd) {
			list_del(&cur->list);
			kfree(cur);
			dqm->processes_count--;
			break;
		}
	}

	mutex_unlock(&dqm->lock);
	return retval;
}


static int process_termination_cpsch(struct device_queue_manager *dqm,
		struct qcm_process_device *qpd)
{
	int retval;
	struct queue *q, *next;
	struct kernel_queue *kq, *kq_next;
	struct mqd_manager *mqd;
	struct device_process_node *cur, *next_dpn;
	enum kfd_unmap_queues_filter filter =
		KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES;

	retval = 0;

	mutex_lock(&dqm->lock);

	/* Clean all kernel queues */
	list_for_each_entry_safe(kq, kq_next, &qpd->priv_queue_list, list) {
		list_del(&kq->list);
		dqm->queue_count--;
		qpd->is_debug = false;
		dqm->total_queue_count--;
		filter = KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES;
	}

	/* Clear all user mode queues */
	list_for_each_entry(q, &qpd->queues_list, list) {
		if (q->properties.type == KFD_QUEUE_TYPE_SDMA) {
			dqm->sdma_queue_count--;
			deallocate_sdma_queue(dqm, q->sdma_id);
		}

		if (q->properties.is_active)
			dqm->queue_count--;

		dqm->total_queue_count--;
	}

	/* Unregister process */
	list_for_each_entry_safe(cur, next_dpn, &dqm->queues, list) {
		if (qpd == cur->qpd) {
			list_del(&cur->list);
			kfree(cur);
			dqm->processes_count--;
			break;
		}
	}

	retval = execute_queues_cpsch(dqm, filter, 0);
	if (retval || qpd->reset_wavefronts) {
		pr_warn("Resetting wave fronts (cpsch) on dev %p\n", dqm->dev);
		dbgdev_wave_reset_wavefronts(dqm->dev, qpd->pqm->process);
		qpd->reset_wavefronts = false;
	}

	/* lastly, free mqd resources */
	list_for_each_entry_safe(q, next, &qpd->queues_list, list) {
		mqd = dqm->ops.get_mqd_manager(dqm,
			get_mqd_type_from_queue_type(q->properties.type));
		if (!mqd) {
			retval = -ENOMEM;
			goto out;
		}
		list_del(&q->list);
		qpd->queue_count--;
		mqd->uninit_mqd(mqd, q->mqd, q->mqd_mem_obj);
	}

out:
	mutex_unlock(&dqm->lock);
	return retval;
}

struct device_queue_manager *device_queue_manager_init(struct kfd_dev *dev)
{
	struct device_queue_manager *dqm;

	pr_debug("Loading device queue manager\n");

	dqm = kzalloc(sizeof(*dqm), GFP_KERNEL);
	if (!dqm)
		return NULL;

	switch (dev->device_info->asic_family) {
	/* HWS is not available on Hawaii. */
	case CHIP_HAWAII:
	/* HWS depends on CWSR for timely dequeue. CWSR is not
	 * available on Tonga.
	 *
	 * FIXME: This argument also applies to Kaveri.
	 */
	case CHIP_TONGA:
		dqm->sched_policy = KFD_SCHED_POLICY_NO_HWS;
		break;
	default:
		dqm->sched_policy = sched_policy;
		break;
	}

	dqm->dev = dev;
	switch (dqm->sched_policy) {
	case KFD_SCHED_POLICY_HWS:
	case KFD_SCHED_POLICY_HWS_NO_OVERSUBSCRIPTION:
		/* initialize dqm for cp scheduling */
		dqm->ops.create_queue = create_queue_cpsch;
		dqm->ops.initialize = initialize_cpsch;
		dqm->ops.start = start_cpsch;
		dqm->ops.stop = stop_cpsch;
		dqm->ops.destroy_queue = destroy_queue_cpsch;
		dqm->ops.update_queue = update_queue;
		dqm->ops.get_mqd_manager = get_mqd_manager;
		dqm->ops.register_process = register_process;
		dqm->ops.unregister_process = unregister_process;
		dqm->ops.uninitialize = uninitialize;
		dqm->ops.create_kernel_queue = create_kernel_queue_cpsch;
		dqm->ops.destroy_kernel_queue = destroy_kernel_queue_cpsch;
		dqm->ops.set_cache_memory_policy = set_cache_memory_policy;
		dqm->ops.set_trap_handler = set_trap_handler;
		dqm->ops.process_termination = process_termination_cpsch;
		dqm->ops.evict_process_queues = evict_process_queues_cpsch;
		dqm->ops.restore_process_queues = restore_process_queues_cpsch;
		break;
	case KFD_SCHED_POLICY_NO_HWS:
		/* initialize dqm for no cp scheduling */
		dqm->ops.start = start_nocpsch;
		dqm->ops.stop = stop_nocpsch;
		dqm->ops.create_queue = create_queue_nocpsch;
		dqm->ops.destroy_queue = destroy_queue_nocpsch;
		dqm->ops.update_queue = update_queue;
		dqm->ops.get_mqd_manager = get_mqd_manager;
		dqm->ops.register_process = register_process;
		dqm->ops.unregister_process = unregister_process;
		dqm->ops.initialize = initialize_nocpsch;
		dqm->ops.uninitialize = uninitialize;
		dqm->ops.set_cache_memory_policy = set_cache_memory_policy;
		dqm->ops.set_trap_handler = set_trap_handler;
		dqm->ops.process_termination = process_termination_nocpsch;
		dqm->ops.evict_process_queues = evict_process_queues_nocpsch;
		dqm->ops.restore_process_queues =
			restore_process_queues_nocpsch;
		break;
	default:
		pr_err("Invalid scheduling policy %d\n", dqm->sched_policy);
		goto out_free;
	}

	switch (dev->device_info->asic_family) {
	case CHIP_CARRIZO:
		device_queue_manager_init_vi(&dqm->asic_ops);
		break;

	case CHIP_KAVERI:
		device_queue_manager_init_cik(&dqm->asic_ops);
		break;

	case CHIP_HAWAII:
		device_queue_manager_init_cik_hawaii(&dqm->asic_ops);
		break;

	case CHIP_TONGA:
	case CHIP_FIJI:
	case CHIP_POLARIS10:
	case CHIP_POLARIS11:
		device_queue_manager_init_vi_tonga(&dqm->asic_ops);
		break;

	case CHIP_VEGA10:
	case CHIP_RAVEN:
		device_queue_manager_init_v9(&dqm->asic_ops);
		break;
	default:
		WARN(1, "Unexpected ASIC family %u",
		     dev->device_info->asic_family);
		goto out_free;
	}

	if (!dqm->ops.initialize(dqm))
		return dqm;

out_free:
	kfree(dqm);
	return NULL;
}

void device_queue_manager_uninit(struct device_queue_manager *dqm)
{
	dqm->ops.uninitialize(dqm);
	kfree(dqm);
}

#if defined(CONFIG_DEBUG_FS)

static void seq_reg_dump(struct seq_file *m,
			 uint32_t (*dump)[2], uint32_t n_regs)
{
	uint32_t i, count;

	for (i = 0, count = 0; i < n_regs; i++) {
		if (count == 0 ||
		    dump[i-1][0] + sizeof(uint32_t) != dump[i][0]) {
			seq_printf(m, "%s    %08x: %08x",
				   i ? "\n" : "",
				   dump[i][0], dump[i][1]);
			count = 7;
		} else {
			seq_printf(m, " %08x", dump[i][1]);
			count--;
		}
	}

	seq_puts(m, "\n");
}

int dqm_debugfs_hqds(struct seq_file *m, void *data)
{
	struct device_queue_manager *dqm = data;
	uint32_t (*dump)[2], n_regs;
	int pipe, queue;
	int r = 0;

	r = dqm->dev->kfd2kgd->hqd_dump(dqm->dev->kgd,
		KFD_CIK_HIQ_PIPE, KFD_CIK_HIQ_QUEUE, &dump, &n_regs);
	if (!r) {
		seq_printf(m, "  HIQ on MEC %d Pipe %d Queue %d\n",
				KFD_CIK_HIQ_PIPE/get_pipes_per_mec(dqm)+1,
				KFD_CIK_HIQ_PIPE%get_pipes_per_mec(dqm),
				KFD_CIK_HIQ_QUEUE);
		seq_reg_dump(m, dump, n_regs);

		kfree(dump);
	}

	for (pipe = 0; pipe < get_pipes_per_mec(dqm); pipe++) {
		int pipe_offset = pipe * get_queues_per_pipe(dqm);

		for (queue = 0; queue < get_queues_per_pipe(dqm); queue++) {
			if (!test_bit(pipe_offset + queue,
				      dqm->dev->shared_resources.queue_bitmap))
				continue;

			r = dqm->dev->kfd2kgd->hqd_dump(
				dqm->dev->kgd, pipe, queue, &dump, &n_regs);
			if (r)
				break;

			seq_printf(m, "  CP Pipe %d, Queue %d\n",
				  pipe, queue);
			seq_reg_dump(m, dump, n_regs);

			kfree(dump);
		}
	}

	for (pipe = 0; pipe < CIK_SDMA_ENGINE_NUM; pipe++) {
		for (queue = 0; queue < CIK_SDMA_QUEUES_PER_ENGINE; queue++) {
			r = dqm->dev->kfd2kgd->hqd_sdma_dump(
				dqm->dev->kgd, pipe, queue, &dump, &n_regs);
			if (r)
				break;

			seq_printf(m, "  SDMA Engine %d, RLC %d\n",
				  pipe, queue);
			seq_reg_dump(m, dump, n_regs);

			kfree(dump);
		}
	}

	return r;
}

#endif