Merge branch 'amdkfd-next-3.19' of git://people.freedesktop.org/~gabbayo/linux into drm-next

- Fixes for sparse warnings
- Memory leak fixes
- Fix for deadlock between amdkfd and iommu

* 'amdkfd-next-3.19' of git://people.freedesktop.org/~gabbayo/linux:
  amdkfd: delete some dead code
  amdkfd: Fix memory leak of mqds on dqm fini
  amdkfd: fix an error handling bug in pqm_create_queue()
  amdkfd: fix some error handling in ioctl
  amdkfd: Remove DRM_AMDGPU dependency from Kconfig
  amdkfd: explicitely include io.h in kfd_doorbell.c
  amdkfd: Clear ctx cb before suspend
  amdkfd: Instead of using get function, use container_of
  amdkfd: use schedule() in sync_with_hw
  amdkfd: Fix memory leak on process deregistration
  amdkfd: add __iomem attribute to doorbell_ptr
  amdkfd: fence_wait_timeout() can be static
  amdkfd: is_occupied() can be static
  amdkfd: Fix sparse warnings in kfd_flat_memory.c
  amdkfd: pqm_get_kernel_queue() can be static
  amdkfd: test_kq() can be static
  amdkfd: Fix sparse warnings in kfd_topology.c
  amdkfd: Fix sparse warnings in kfd_chardev.c

drivers/gpu/drm/amd/amdkfd/Kconfig

@@ -4,6 +4,6 @@
 
 config HSA_AMD
 	tristate "HSA kernel driver for AMD GPU devices"
-	depends on (DRM_RADEON || DRM_AMDGPU) && AMD_IOMMU_V2 && X86_64
+	depends on DRM_RADEON && AMD_IOMMU_V2 && X86_64
 	help
 	  Enable this if you want to use HSA features on AMD GPU devices.

drivers/gpu/drm/amd/amdkfd/kfd_chardev.c

@@ -149,7 +149,9 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties,
 	}
 
 	if ((args->ring_base_address) &&
-		(!access_ok(VERIFY_WRITE, args->ring_base_address, sizeof(uint64_t)))) {
+		(!access_ok(VERIFY_WRITE,
+			(const void __user *) args->ring_base_address,
+			sizeof(uint64_t)))) {
 		pr_err("kfd: can't access ring base address\n");
 		return -EFAULT;
 	}
@@ -159,12 +161,16 @@ static int set_queue_properties_from_user(struct queue_properties *q_properties,
 		return -EINVAL;
 	}
 
-	if (!access_ok(VERIFY_WRITE, args->read_pointer_address, sizeof(uint32_t))) {
+	if (!access_ok(VERIFY_WRITE,
+			(const void __user *) args->read_pointer_address,
+			sizeof(uint32_t))) {
 		pr_err("kfd: can't access read pointer\n");
 		return -EFAULT;
 	}
 
-	if (!access_ok(VERIFY_WRITE, args->write_pointer_address, sizeof(uint32_t))) {
+	if (!access_ok(VERIFY_WRITE,
+			(const void __user *) args->write_pointer_address,
+			sizeof(uint32_t))) {
 		pr_err("kfd: can't access write pointer\n");
 		return -EFAULT;
 	}
@@ -236,7 +242,7 @@ static long kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
 	mutex_lock(&p->mutex);
 
 	pdd = kfd_bind_process_to_device(dev, p);
-	if (IS_ERR(pdd) < 0) {
+	if (IS_ERR(pdd)) {
 		err = PTR_ERR(pdd);
 		goto err_bind_process;
 	}
@@ -325,7 +331,9 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p,
 	}
 
 	if ((args.ring_base_address) &&
-		(!access_ok(VERIFY_WRITE, args.ring_base_address, sizeof(uint64_t)))) {
+		(!access_ok(VERIFY_WRITE,
+			(const void __user *) args.ring_base_address,
+			sizeof(uint64_t)))) {
 		pr_err("kfd: can't access ring base address\n");
 		return -EFAULT;
 	}
@@ -381,7 +389,7 @@ static long kfd_ioctl_set_memory_policy(struct file *filep,
 	mutex_lock(&p->mutex);
 
 	pdd = kfd_bind_process_to_device(dev, p);
-	if (IS_ERR(pdd) < 0) {
+	if (IS_ERR(pdd)) {
 		err = PTR_ERR(pdd);
 		goto out;
 	}

drivers/gpu/drm/amd/amdkfd/kfd_device.c

@@ -267,6 +267,7 @@ void kgd2kfd_suspend(struct kfd_dev *kfd)
 
 	if (kfd->init_complete) {
 		kfd->dqm->stop(kfd->dqm);
+		amd_iommu_set_invalidate_ctx_cb(kfd->pdev, NULL);
 		amd_iommu_free_device(kfd->pdev);
 	}
 }

drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c

@@ -67,26 +67,21 @@ static inline unsigned int get_pipes_num_cpsch(void)
 	return PIPE_PER_ME_CP_SCHEDULING;
 }
 
-static unsigned int get_sh_mem_bases_nybble_64(struct kfd_process *process,
-						struct kfd_dev *dev)
+static inline unsigned int
+get_sh_mem_bases_nybble_64(struct kfd_process_device *pdd)
 {
-	struct kfd_process_device *pdd;
 	uint32_t nybble;
 
-	pdd = kfd_get_process_device_data(dev, process, 1);
 	nybble = (pdd->lds_base >> 60) & 0x0E;
 
 	return nybble;
 
 }
 
-static unsigned int get_sh_mem_bases_32(struct kfd_process *process,
-					struct kfd_dev *dev)
+static inline unsigned int get_sh_mem_bases_32(struct kfd_process_device *pdd)
 {
-	struct kfd_process_device *pdd;
 	unsigned int shared_base;
 
-	pdd = kfd_get_process_device_data(dev, process, 1);
 	shared_base = (pdd->lds_base >> 16) & 0xFF;
 
 	return shared_base;
@@ -96,10 +91,13 @@ static uint32_t compute_sh_mem_bases_64bit(unsigned int top_address_nybble);
 static void init_process_memory(struct device_queue_manager *dqm,
 				struct qcm_process_device *qpd)
 {
+	struct kfd_process_device *pdd;
 	unsigned int temp;
 
 	BUG_ON(!dqm || !qpd);
 
+	pdd = qpd_to_pdd(qpd);
+
 	/* check if sh_mem_config register already configured */
 	if (qpd->sh_mem_config == 0) {
 		qpd->sh_mem_config =
@@ -111,11 +109,11 @@ static void init_process_memory(struct device_queue_manager *dqm,
 	}
 
 	if (qpd->pqm->process->is_32bit_user_mode) {
-		temp = get_sh_mem_bases_32(qpd->pqm->process, dqm->dev);
+		temp = get_sh_mem_bases_32(pdd);
 		qpd->sh_mem_bases = SHARED_BASE(temp);
 		qpd->sh_mem_config |= PTR32;
 	} else {
-		temp = get_sh_mem_bases_nybble_64(qpd->pqm->process, dqm->dev);
+		temp = get_sh_mem_bases_nybble_64(pdd);
 		qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
 	}
 
@@ -409,6 +407,7 @@ static int unregister_process_nocpsch(struct device_queue_manager *dqm,
 	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;
 		}
@@ -576,11 +575,15 @@ static int initialize_nocpsch(struct device_queue_manager *dqm)
 
 static void uninitialize_nocpsch(struct device_queue_manager *dqm)
 {
+	int i;
+
 	BUG_ON(!dqm);
 
 	BUG_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);
 	kfd2kgd->free_mem(dqm->dev->kgd,
 			(struct kgd_mem *) dqm->pipeline_mem);
@@ -706,8 +709,7 @@ static int stop_cpsch(struct device_queue_manager *dqm)
 	destroy_queues_cpsch(dqm, true);
 
 	list_for_each_entry(node, &dqm->queues, list) {
-		pdd = kfd_get_process_device_data(dqm->dev,
-						node->qpd->pqm->process, 1);
+		pdd = qpd_to_pdd(node->qpd);
 		pdd->bound = false;
 	}
 	kfd2kgd->free_mem(dqm->dev->kgd,
@@ -789,7 +791,8 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
 	return retval;
 }
 
-int fence_wait_timeout(unsigned int *fence_addr, unsigned int fence_value,
+static int fence_wait_timeout(unsigned int *fence_addr,
+				unsigned int fence_value,
 				unsigned long timeout)
 {
 	BUG_ON(!fence_addr);

drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c

@@ -23,6 +23,7 @@
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/slab.h>
+#include <linux/io.h>
 
 /*
  * This extension supports a kernel level doorbells management for

drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c

@@ -276,21 +276,22 @@
  */
 
 #define MAKE_GPUVM_APP_BASE(gpu_num) \
-	(((uint64_t)(gpu_num) << 61) + 0x1000000000000)
+	(((uint64_t)(gpu_num) << 61) + 0x1000000000000L)
 
 #define MAKE_GPUVM_APP_LIMIT(base) \
-	(((uint64_t)(base) & 0xFFFFFF0000000000) | 0xFFFFFFFFFF)
+	(((uint64_t)(base) & \
+		0xFFFFFF0000000000UL) | 0xFFFFFFFFFFL)
 
 #define MAKE_SCRATCH_APP_BASE(gpu_num) \
-	(((uint64_t)(gpu_num) << 61) + 0x100000000)
+	(((uint64_t)(gpu_num) << 61) + 0x100000000L)
 
 #define MAKE_SCRATCH_APP_LIMIT(base) \
-	(((uint64_t)base & 0xFFFFFFFF00000000) | 0xFFFFFFFF)
+	(((uint64_t)base & 0xFFFFFFFF00000000UL) | 0xFFFFFFFF)
 
 #define MAKE_LDS_APP_BASE(gpu_num) \
 	(((uint64_t)(gpu_num) << 61) + 0x0)
 #define MAKE_LDS_APP_LIMIT(base) \
-	(((uint64_t)(base) & 0xFFFFFFFF00000000) | 0xFFFFFFFF)
+	(((uint64_t)(base) & 0xFFFFFFFF00000000UL) | 0xFFFFFFFF)
 
 int kfd_init_apertures(struct kfd_process *process)
 {

drivers/gpu/drm/amd/amdkfd/kfd_kernel_queue.c

@@ -25,6 +25,7 @@
 #include <linux/mutex.h>
 #include <linux/slab.h>
 #include <linux/printk.h>
+#include <linux/sched.h>
 #include "kfd_kernel_queue.h"
 #include "kfd_priv.h"
 #include "kfd_device_queue_manager.h"
@@ -66,8 +67,7 @@ static bool initialize(struct kernel_queue *kq, struct kfd_dev *dev,
 	if (kq->mqd == NULL)
 		return false;
 
-	prop.doorbell_ptr =
-		(uint32_t *)kfd_get_kernel_doorbell(dev, &prop.doorbell_off);
+	prop.doorbell_ptr = kfd_get_kernel_doorbell(dev, &prop.doorbell_off);
 
 	if (prop.doorbell_ptr == NULL)
 		goto err_get_kernel_doorbell;
@@ -172,7 +172,7 @@ static bool initialize(struct kernel_queue *kq, struct kfd_dev *dev,
 	kfd2kgd->free_mem(dev->kgd, (struct kgd_mem *) kq->pq);
 err_pq_allocate_vidmem:
 	pr_err("kfd: error init pq\n");
-	kfd_release_kernel_doorbell(dev, (u32 *)prop.doorbell_ptr);
+	kfd_release_kernel_doorbell(dev, prop.doorbell_ptr);
 err_get_kernel_doorbell:
 	pr_err("kfd: error init doorbell");
 	return false;
@@ -195,7 +195,7 @@ static void uninitialize(struct kernel_queue *kq)
 	kfd2kgd->free_mem(kq->dev->kgd, (struct kgd_mem *) kq->wptr_mem);
 	kfd2kgd->free_mem(kq->dev->kgd, (struct kgd_mem *) kq->pq);
 	kfd_release_kernel_doorbell(kq->dev,
-				(u32 *)kq->queue->properties.doorbell_ptr);
+					kq->queue->properties.doorbell_ptr);
 	uninit_queue(kq->queue);
 }
 
@@ -255,7 +255,7 @@ static void submit_packet(struct kernel_queue *kq)
 #endif
 
 	*kq->wptr_kernel = kq->pending_wptr;
-	write_kernel_doorbell((u32 *)kq->queue->properties.doorbell_ptr,
+	write_kernel_doorbell(kq->queue->properties.doorbell_ptr,
 				kq->pending_wptr);
 }
 
@@ -275,7 +275,7 @@ static int sync_with_hw(struct kernel_queue *kq, unsigned long timeout_ms)
 				*kq->wptr_kernel, *kq->rptr_kernel);
 			return -ETIME;
 		}
-		cpu_relax();
+		schedule();
 	}
 
 	return 0;
@@ -321,7 +321,7 @@ void kernel_queue_uninit(struct kernel_queue *kq)
 	kfree(kq);
 }
 
-void test_kq(struct kfd_dev *dev)
+static __attribute__((unused)) void test_kq(struct kfd_dev *dev)
 {
 	struct kernel_queue *kq;
 	uint32_t *buffer, i;

drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager.c

@@ -179,7 +179,7 @@ static int destroy_mqd(struct mqd_manager *mm, void *mqd,
 					pipe_id, queue_id);
 }
 
-bool is_occupied(struct mqd_manager *mm, void *mqd,
+static bool is_occupied(struct mqd_manager *mm, void *mqd,
 			uint64_t queue_address,	uint32_t pipe_id,
 			uint32_t queue_id)
 {

drivers/gpu/drm/amd/amdkfd/kfd_priv.h

@@ -279,7 +279,7 @@ struct queue_properties {
 	uint32_t queue_percent;
 	uint32_t *read_ptr;
 	uint32_t *write_ptr;
-	uint32_t *doorbell_ptr;
+	uint32_t __iomem *doorbell_ptr;
 	uint32_t doorbell_off;
 	bool is_interop;
 	bool is_active;
@@ -414,6 +414,8 @@ struct kfd_process_device {
 	bool bound;
 };
 
+#define qpd_to_pdd(x) container_of(x, struct kfd_process_device, qpd)
+
 /* Process data */
 struct kfd_process {
 	/*

drivers/gpu/drm/amd/amdkfd/kfd_process.c

@@ -348,11 +348,6 @@ struct kfd_process_device *kfd_bind_process_to_device(struct kfd_dev *dev,
 	if (err < 0)
 		return ERR_PTR(err);
 
-	if (err < 0) {
-		amd_iommu_unbind_pasid(dev->pdev, p->pasid);
-		return ERR_PTR(err);
-	}
-
 	pdd->bound = true;
 
 	return pdd;

drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c

@@ -208,7 +208,7 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	case KFD_QUEUE_TYPE_DIQ:
 		kq = kernel_queue_init(dev, KFD_QUEUE_TYPE_DIQ);
 		if (kq == NULL) {
-			kernel_queue_uninit(kq);
+			retval = -ENOMEM;
 			goto err_create_queue;
 		}
 		kq->queue->properties.queue_id = *qid;
@@ -325,7 +325,8 @@ int pqm_update_queue(struct process_queue_manager *pqm, unsigned int qid,
 	return 0;
 }
 
-struct kernel_queue *pqm_get_kernel_queue(struct process_queue_manager *pqm,
+static __attribute__((unused)) struct kernel_queue *pqm_get_kernel_queue(
+					struct process_queue_manager *pqm,
 					unsigned int qid)
 {
 	struct process_queue_node *pqn;

drivers/gpu/drm/amd/amdkfd/kfd_topology.c

@@ -96,7 +96,7 @@ static int kfd_topology_get_crat_acpi(void *crat_image, size_t *size)
 		return -EINVAL;
 	}
 
-	if (*size >= crat_table->length && crat_image != 0)
+	if (*size >= crat_table->length && crat_image != NULL)
 		memcpy(crat_image, crat_table, crat_table->length);
 
 	*size = crat_table->length;
@@ -183,7 +183,7 @@ static int kfd_parse_subtype_mem(struct crat_subtype_memory *mem)
 	list_for_each_entry(dev, &topology_device_list, list) {
 		if (mem->promixity_domain == i) {
 			props = kfd_alloc_struct(props);
-			if (props == 0)
+			if (props == NULL)
 				return -ENOMEM;
 
 			if (dev->node_props.cpu_cores_count == 0)
@@ -231,7 +231,7 @@ static int kfd_parse_subtype_cache(struct crat_subtype_cache *cache)
 		if (id == dev->node_props.cpu_core_id_base ||
 		    id == dev->node_props.simd_id_base) {
 			props = kfd_alloc_struct(props);
-			if (props == 0)
+			if (props == NULL)
 				return -ENOMEM;
 
 			props->processor_id_low = id;
@@ -282,7 +282,7 @@ static int kfd_parse_subtype_iolink(struct crat_subtype_iolink *iolink)
 	list_for_each_entry(dev, &topology_device_list, list) {
 		if (id_from == i) {
 			props = kfd_alloc_struct(props);
-			if (props == 0)
+			if (props == NULL)
 				return -ENOMEM;
 
 			props->node_from = id_from;
@@ -415,9 +415,9 @@ static struct kfd_topology_device *kfd_create_topology_device(void)
 	struct kfd_topology_device *dev;
 
 	dev = kfd_alloc_struct(dev);
-	if (dev == 0) {
+	if (dev == NULL) {
 		pr_err("No memory to allocate a topology device");
-		return 0;
+		return NULL;
 	}
 
 	INIT_LIST_HEAD(&dev->mem_props);
@@ -428,7 +428,7 @@ static struct kfd_topology_device *kfd_create_topology_device(void)
 	sys_props.num_devices++;
 
 	return dev;
-	}
+}
 
 static int kfd_parse_crat_table(void *crat_image)
 {
@@ -752,11 +752,11 @@ static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
 			if (iolink->kobj) {
 				kfd_remove_sysfs_file(iolink->kobj,
 							&iolink->attr);
-				iolink->kobj = 0;
+				iolink->kobj = NULL;
 			}
 		kobject_del(dev->kobj_iolink);
 		kobject_put(dev->kobj_iolink);
-		dev->kobj_iolink = 0;
+		dev->kobj_iolink = NULL;
 	}
 
 	if (dev->kobj_cache) {
@@ -764,22 +764,22 @@ static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
 			if (cache->kobj) {
 				kfd_remove_sysfs_file(cache->kobj,
 							&cache->attr);
-				cache->kobj = 0;
+				cache->kobj = NULL;
 			}
 		kobject_del(dev->kobj_cache);
 		kobject_put(dev->kobj_cache);
-		dev->kobj_cache = 0;
+		dev->kobj_cache = NULL;
 	}
 
 	if (dev->kobj_mem) {
 		list_for_each_entry(mem, &dev->mem_props, list)
 			if (mem->kobj) {
 				kfd_remove_sysfs_file(mem->kobj, &mem->attr);
-				mem->kobj = 0;
+				mem->kobj = NULL;
 			}
 		kobject_del(dev->kobj_mem);
 		kobject_put(dev->kobj_mem);
-		dev->kobj_mem = 0;
+		dev->kobj_mem = NULL;
 	}
 
 	if (dev->kobj_node) {
@@ -788,7 +788,7 @@ static void kfd_remove_sysfs_node_entry(struct kfd_topology_device *dev)
 		sysfs_remove_file(dev->kobj_node, &dev->attr_props);
 		kobject_del(dev->kobj_node);
 		kobject_put(dev->kobj_node);
-		dev->kobj_node = 0;
+		dev->kobj_node = NULL;
 	}
 }
 
@@ -939,7 +939,7 @@ static int kfd_topology_update_sysfs(void)
 	int ret;
 
 	pr_info("Creating topology SYSFS entries\n");
-	if (sys_props.kobj_topology == 0) {
+	if (sys_props.kobj_topology == NULL) {
 		sys_props.kobj_topology =
 				kfd_alloc_struct(sys_props.kobj_topology);
 		if (!sys_props.kobj_topology)
@@ -989,17 +989,17 @@ static void kfd_topology_release_sysfs(void)
 		if (sys_props.kobj_nodes) {
 			kobject_del(sys_props.kobj_nodes);
 			kobject_put(sys_props.kobj_nodes);
-			sys_props.kobj_nodes = 0;
+			sys_props.kobj_nodes = NULL;
 		}
 		kobject_del(sys_props.kobj_topology);
 		kobject_put(sys_props.kobj_topology);
-		sys_props.kobj_topology = 0;
+		sys_props.kobj_topology = NULL;
 	}
 }
 
 int kfd_topology_init(void)
 {
-	void *crat_image = 0;
+	void *crat_image = NULL;
 	size_t image_size = 0;
 	int ret;
 
@@ -1094,12 +1094,12 @@ static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu)
 static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu)
 {
 	struct kfd_topology_device *dev;
-	struct kfd_topology_device *out_dev = 0;
+	struct kfd_topology_device *out_dev = NULL;
 
 	BUG_ON(!gpu);
 
 	list_for_each_entry(dev, &topology_device_list, list)
-		if (dev->gpu == 0 && dev->node_props.simd_count > 0) {
+		if (dev->gpu == NULL && dev->node_props.simd_count > 0) {
 			dev->gpu = gpu;
 			out_dev = dev;
 			break;