[PATCH 09/18] drm/amdkfd: CRIU add queues support
Felix Kuehling
felix.kuehling at amd.com
Mon Aug 23 18:29:24 UTC 2021
kfd_chardev.c contains the ioctl API, but not the whole implementation
of everything. I think it would make sense to move the criu_dump_queue*
functions into kfd_process_queue_manager.c.
Regards,
Felix
Am 2021-08-19 um 9:37 a.m. schrieb David Yat Sin:
> Add support to existing CRIU ioctl's to save number of queues and queue
> properties for each queue during checkpoint and re-create queues on
> restore.
>
> Signed-off-by: David Yat Sin <david.yatsin at amd.com>
> ---
> drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 380 ++++++++++++++++++++++-
> drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 22 +-
> 2 files changed, 400 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> index 24e5c53261f5..6f1c9fb8d46c 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
> @@ -1965,6 +1965,213 @@ static int criu_dump_bos(struct kfd_process *p, struct kfd_ioctl_criu_dumper_arg
> return ret;
> }
>
> +static void get_queue_data_sizes(struct kfd_process_device *pdd,
> + struct queue *q,
> + uint32_t *cu_mask_size)
> +{
> + *cu_mask_size = sizeof(uint32_t) * (q->properties.cu_mask_count / 32);
> +}
> +
> +int get_process_queue_info(struct kfd_process *p, uint32_t *num_queues, uint32_t *q_data_sizes)
> +{
> + u32 data_sizes = 0;
> + u32 q_index = 0;
> + struct queue *q;
> + int i;
> +
> + /* Run over all PDDs of the process */
> + for (i = 0; i < p->n_pdds; i++) {
> + struct kfd_process_device *pdd = p->pdds[i];
> +
> + list_for_each_entry(q, &pdd->qpd.queues_list, list) {
> + if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
> + q->properties.type == KFD_QUEUE_TYPE_SDMA ||
> + q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
> + u32 cu_mask_size;
> +
> + get_queue_data_sizes(pdd, q, &cu_mask_size);
> +
> + data_sizes += cu_mask_size;
> + q_index++;
> + } else {
> + pr_err("Unsupported queue type (%d)\n", q->properties.type);
> + return -EOPNOTSUPP;
> + }
> + }
> + }
> + *num_queues = q_index;
> + *q_data_sizes = data_sizes;
> +
> + return 0;
> +}
> +
> +static void criu_dump_queue(struct kfd_process_device *pdd,
> + struct queue *q,
> + struct kfd_criu_queue_bucket *q_bucket,
> + void *private_data)
> +{
> + struct kfd_criu_queue_priv_data *q_data = (struct kfd_criu_queue_priv_data *) private_data;
> + uint8_t *cu_mask;
> +
> + cu_mask = (void *)(q_data + 1);
> +
> + q_bucket->gpu_id = pdd->dev->id;
> + q_data->type = q->properties.type;
> + q_data->format = q->properties.format;
> + q_data->q_id = q->properties.queue_id;
> + q_data->q_address = q->properties.queue_address;
> + q_data->q_size = q->properties.queue_size;
> + q_data->priority = q->properties.priority;
> + q_data->q_percent = q->properties.queue_percent;
> + q_data->read_ptr_addr = (uint64_t)q->properties.read_ptr;
> + q_data->write_ptr_addr = (uint64_t)q->properties.write_ptr;
> + q_data->doorbell_id = q->doorbell_id;
> +
> + q_data->sdma_id = q->sdma_id;
> +
> + q_data->eop_ring_buffer_address =
> + q->properties.eop_ring_buffer_address;
> +
> + q_data->eop_ring_buffer_size = q->properties.eop_ring_buffer_size;
> +
> + q_data->ctx_save_restore_area_address =
> + q->properties.ctx_save_restore_area_address;
> +
> + q_data->ctx_save_restore_area_size =
> + q->properties.ctx_save_restore_area_size;
> +
> + if (q_data->cu_mask_size)
> + memcpy(cu_mask, q->properties.cu_mask, q_data->cu_mask_size);
> +
> + pr_debug("Dumping Queue: gpu_id:%x queue_id:%u\n", q_bucket->gpu_id, q_data->q_id);
> +}
> +
> +static int criu_dump_queues_device(struct kfd_process_device *pdd,
> + unsigned int *q_index,
> + unsigned int max_num_queues,
> + struct kfd_criu_queue_bucket *q_buckets,
> + uint8_t *user_priv_data,
> + uint64_t *queues_priv_data_offset)
> +{
> + struct queue *q;
> + uint8_t *q_private_data = NULL; /* Local buffer to store individual queue private data */
> + unsigned int q_private_data_size = 0;
> + int ret = 0;
> +
> + list_for_each_entry(q, &pdd->qpd.queues_list, list) {
> + struct kfd_criu_queue_bucket q_bucket;
> + struct kfd_criu_queue_priv_data *q_data;
> + uint64_t q_data_size;
> + uint32_t cu_mask_size;
> +
> + if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE &&
> + q->properties.type != KFD_QUEUE_TYPE_SDMA &&
> + q->properties.type != KFD_QUEUE_TYPE_SDMA_XGMI) {
> +
> + pr_err("Unsupported queue type (%d)\n", q->properties.type);
> + return -EOPNOTSUPP;
> + }
> +
> + memset(&q_bucket, 0, sizeof(q_bucket));
> +
> + get_queue_data_sizes(pdd, q, &cu_mask_size);
> +
> + q_data_size = sizeof(*q_data) + cu_mask_size;
> +
> + /* Increase local buffer space if needed */
> + if (q_private_data_size < q_data_size) {
> + kfree(q_private_data);
> +
> + q_private_data = kzalloc(q_data_size, GFP_KERNEL);
> + if (!q_private_data) {
> + ret = -ENOMEM;
> + break;
> + }
> + q_private_data_size = q_data_size;
> + }
> +
> + q_data = (struct kfd_criu_queue_priv_data *)q_private_data;
> +
> + q_data->cu_mask_size = cu_mask_size;
> +
> + criu_dump_queue(pdd, q, &q_bucket, q_data);
> +
> + q_bucket.priv_data_offset = *queues_priv_data_offset;
> + q_bucket.priv_data_size = q_data_size;
> +
> + ret = copy_to_user((void __user *) (user_priv_data + q_bucket.priv_data_offset),
> + q_private_data, q_bucket.priv_data_size);
> + if (ret) {
> + ret = -EFAULT;
> + break;
> + }
> + *queues_priv_data_offset += q_data_size;
> +
> + ret = copy_to_user((void __user *)&q_buckets[*q_index],
> + &q_bucket, sizeof(q_bucket));
> + if (ret) {
> + pr_err("Failed to copy queue information to user\n");
> + ret = -EFAULT;
> + break;
> + }
> + *q_index = *q_index + 1;
> + }
> +
> + kfree(q_private_data);
> +
> + return ret;
> +}
> +
> +static int criu_dump_queues(struct kfd_process *p, struct kfd_ioctl_criu_dumper_args *args)
> +{
> + struct kfd_criu_queue_bucket *queue_buckets;
> + uint32_t num_queues, queue_extra_data_sizes;
> + uint64_t queues_priv_data_offset = 0;
> + int ret = 0, pdd_index, q_index = 0;
> + void *private_data; /* Pointer to first private data in userspace */
> +
> + ret = get_process_queue_info(p, &num_queues, &queue_extra_data_sizes);
> + if (ret)
> + return ret;
> +
> + if (args->num_objects != num_queues) {
> + pr_err("Mismatch with number of queues (current:%d user:%lld)\n",
> + num_queues, args->num_objects);
> + return -EINVAL;
> + }
> +
> + if (args->objects_size != queue_extra_data_sizes +
> + (num_queues * (sizeof(*queue_buckets) +
> + sizeof(struct kfd_criu_queue_priv_data)))) {
> + pr_err("Invalid objects size for queues\n");
> + return -EINVAL;
> + }
> +
> + /* Queue private data size for each queue can vary in size as it also includes cu_mask, mqd
> + * and ctl_stack. First queue private data starts after all queue_buckets
> + */
> +
> + queue_buckets = (struct kfd_criu_queue_bucket *)args->objects;
> + private_data = (void *)(queue_buckets + args->num_objects);
> +
> + for (pdd_index = 0; pdd_index < p->n_pdds; pdd_index++) {
> + struct kfd_process_device *pdd = p->pdds[pdd_index];
> +
> + /* criu_dump_queues_device will copy data to user */
> + ret = criu_dump_queues_device(pdd,
> + &q_index,
> + args->num_objects,
> + queue_buckets,
> + private_data,
> + &queues_priv_data_offset);
> +
> + if (ret)
> + break;
> + }
> +
> + return ret;
> +}
> +
> static int kfd_ioctl_criu_dumper(struct file *filep,
> struct kfd_process *p, void *data)
> {
> @@ -2000,6 +2207,8 @@ static int kfd_ioctl_criu_dumper(struct file *filep,
> ret = criu_dump_bos(p, args);
> break;
> case KFD_CRIU_OBJECT_TYPE_QUEUE:
> + ret = criu_dump_queues(p, args);
> + break;
> case KFD_CRIU_OBJECT_TYPE_EVENT:
> case KFD_CRIU_OBJECT_TYPE_DEVICE:
> case KFD_CRIU_OBJECT_TYPE_SVM_RANGE:
> @@ -2274,6 +2483,163 @@ static int criu_restore_bos(struct kfd_process *p, struct kfd_ioctl_criu_restore
> return ret;
> }
>
> +static int set_queue_properties_from_criu(struct queue_properties *qp,
> + struct kfd_criu_queue_bucket *q_bucket,
> + struct kfd_criu_queue_priv_data *q_data,
> + void *cu_mask)
> +{
> + qp->is_interop = false;
> + qp->is_gws = q_data->is_gws;
> + qp->queue_percent = q_data->q_percent;
> + qp->priority = q_data->priority;
> + qp->queue_address = q_data->q_address;
> + qp->queue_size = q_data->q_size;
> + qp->read_ptr = (uint32_t *) q_data->read_ptr_addr;
> + qp->write_ptr = (uint32_t *) q_data->write_ptr_addr;
> + qp->eop_ring_buffer_address = q_data->eop_ring_buffer_address;
> + qp->eop_ring_buffer_size = q_data->eop_ring_buffer_size;
> + qp->ctx_save_restore_area_address = q_data->ctx_save_restore_area_address;
> + qp->ctx_save_restore_area_size = q_data->ctx_save_restore_area_size;
> + qp->ctl_stack_size = q_data->ctl_stack_size;
> + qp->type = q_data->type;
> + qp->format = q_data->format;
> +
> + if (q_data->cu_mask_size) {
> + qp->cu_mask = kzalloc(q_data->cu_mask_size, GFP_KERNEL);
> + if (!qp->cu_mask)
> + return -ENOMEM;
> +
> + /* CU mask is stored after q_data */
> + memcpy(qp->cu_mask, cu_mask, q_data->cu_mask_size);
> + qp->cu_mask_count = (q_data->cu_mask_size / sizeof(uint32_t)) * 32;
> + }
> +
> + return 0;
> +}
> +
> +static int criu_restore_queue(struct kfd_process *p,
> + struct kfd_dev *dev,
> + struct kfd_process_device *pdd,
> + struct kfd_criu_queue_bucket *q_bucket,
> + void *private_data)
> +{
> + struct kfd_criu_queue_priv_data *q_data = (struct kfd_criu_queue_priv_data *) private_data;
> + uint8_t *cu_mask, *mqd, *ctl_stack;
> + struct queue_properties qp;
> + unsigned int queue_id;
> + int ret = 0;
> +
> + pr_debug("Restoring Queue: gpu_id:%x queue_id:%u\n", q_bucket->gpu_id, q_data->q_id);
> +
> + /* data stored in this order: cu_mask, mqd, ctl_stack */
> + cu_mask = (void *)(q_data + 1);
> + mqd = cu_mask + q_data->cu_mask_size;
> + ctl_stack = mqd + q_data->mqd_size;
> +
> + memset(&qp, 0, sizeof(qp));
> + ret = set_queue_properties_from_criu(&qp, q_bucket, q_data, cu_mask);
> + if (ret)
> + goto err_create_queue;
> +
> + print_queue_properties(&qp);
> +
> + ret = pqm_create_queue(&p->pqm, dev, NULL, &qp, &queue_id, NULL);
> + if (ret) {
> + pr_err("Failed to create new queue err:%d\n", ret);
> + ret = -EINVAL;
> + goto err_create_queue;
> + }
> +
> + pr_debug("Queue id %d was restored successfully\n", queue_id);
> +
> + return 0;
> +err_create_queue:
> + kfree(qp.cu_mask);
> +
> + return ret;
> +}
> +
> +static int criu_restore_queues(struct kfd_process *p,
> + struct kfd_ioctl_criu_restorer_args *args)
> +{
> + int ret = 0, i;
> + struct kfd_criu_queue_bucket *user_buckets;
> + uint8_t *all_private_data; /* Pointer to first private data in userspace */
> + uint8_t *q_private_data = NULL; /* Local buffer for individual queue private data */
> + unsigned int q_private_data_size = 0;
> +
> + user_buckets = (struct kfd_criu_queue_bucket *)args->objects;
> + all_private_data = (void *)(user_buckets + args->num_objects);
> +
> + /*
> + * This process will not have any queues at this point, but we are
> + * setting all the dqm's for this process to evicted state.
> + */
> + kfd_process_evict_queues(p);
> +
> + for (i = 0; i < args->num_objects; i++) {
> + struct kfd_process_device *pdd;
> + struct kfd_dev *dev;
> + struct kfd_criu_queue_bucket q_bucket;
> +
> + ret = copy_from_user(&q_bucket, (void __user *)&user_buckets[i],
> + sizeof(struct kfd_criu_queue_bucket));
> +
> + if (ret) {
> + ret = -EFAULT;
> + goto exit;
> + }
> +
> + /* Increase local buffer space if needed */
> + if (q_bucket.priv_data_size > q_private_data_size) {
> + kfree(q_private_data);
> +
> + q_private_data = kmalloc(q_bucket.priv_data_size, GFP_KERNEL);
> + if (!q_private_data) {
> + ret = -ENOMEM;
> + goto exit;
> + }
> + q_private_data_size = q_bucket.priv_data_size;
> + }
> +
> + ret = copy_from_user(q_private_data,
> + (void __user *) (all_private_data + q_bucket.priv_data_offset),
> + q_bucket.priv_data_size);
> + if (ret) {
> + ret = -EFAULT;
> + goto exit;
> + }
> +
> + dev = kfd_device_by_id(q_bucket.gpu_id);
> + if (!dev) {
> + pr_err("Could not get kfd_dev from gpu_id = 0x%x\n",
> + q_bucket.gpu_id);
> +
> + ret = -EINVAL;
> + goto exit;
> + }
> +
> + pdd = kfd_get_process_device_data(dev, p);
> + if (!pdd) {
> + pr_err("Failed to get pdd\n");
> + ret = -EFAULT;
> + return ret;
> + }
> +
> + ret = criu_restore_queue(p, dev, pdd, &q_bucket, q_private_data);
> + if (ret) {
> + pr_err("Failed to restore queue (%d)\n", ret);
> + goto exit;
> + }
> +
> + }
> +
> +exit:
> + kfree(q_private_data);
> +
> + return ret;
> +}
> +
> static int kfd_ioctl_criu_restorer(struct file *filep,
> struct kfd_process *p, void *data)
> {
> @@ -2293,6 +2659,8 @@ static int kfd_ioctl_criu_restorer(struct file *filep,
> ret = criu_restore_bos(p, args);
> break;
> case KFD_CRIU_OBJECT_TYPE_QUEUE:
> + ret = criu_restore_queues(p, args);
> + break;
> case KFD_CRIU_OBJECT_TYPE_EVENT:
> case KFD_CRIU_OBJECT_TYPE_DEVICE:
> case KFD_CRIU_OBJECT_TYPE_SVM_RANGE:
> @@ -2368,6 +2736,7 @@ static int kfd_ioctl_criu_process_info(struct file *filep,
> struct kfd_process *p, void *data)
> {
> struct kfd_ioctl_criu_process_info_args *args = data;
> + uint32_t queues_extra_data_size;
> int ret = 0;
>
> pr_debug("Inside %s\n", __func__);
> @@ -2387,7 +2756,16 @@ static int kfd_ioctl_criu_process_info(struct file *filep,
> args->total_bos = get_process_num_bos(p);
> args->bos_priv_data_size = args->total_bos * sizeof(struct kfd_criu_bo_priv_data);
>
> - dev_dbg(kfd_device, "Num of bos:%llu\n", args->total_bos);
> + ret = get_process_queue_info(p, &args->total_queues, &queues_extra_data_size);
> + if (ret)
> + goto err_unlock;
> +
> + args->queues_priv_data_size = queues_extra_data_size +
> + (args->total_queues * sizeof(struct kfd_criu_queue_priv_data));
> +
> + dev_dbg(kfd_device, "Num of bos:%llu queues:%u\n",
> + args->total_bos,
> + args->total_queues);
> err_unlock:
> mutex_unlock(&p->mutex);
> return ret;
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
> index 0b8165729cde..4b4808b191f2 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
> @@ -1044,7 +1044,27 @@ struct kfd_criu_svm_range_priv_data {
> };
>
> struct kfd_criu_queue_priv_data {
> - uint64_t reserved;
> + uint64_t q_address;
> + uint64_t q_size;
> + uint64_t read_ptr_addr;
> + uint64_t write_ptr_addr;
> + uint64_t doorbell_off;
> + uint64_t eop_ring_buffer_address;
> + uint64_t ctx_save_restore_area_address;
> + uint32_t gpu_id;
> + uint32_t type;
> + uint32_t format;
> + uint32_t q_id;
> + uint32_t priority;
> + uint32_t q_percent;
> + uint32_t doorbell_id;
> + uint32_t is_gws;
> + uint32_t sdma_id;
> + uint32_t eop_ring_buffer_size;
> + uint32_t ctx_save_restore_area_size;
> + uint32_t ctl_stack_size;
> + uint32_t cu_mask_size;
> + uint32_t mqd_size;
> };
>
> struct kfd_criu_event_priv_data {
More information about the amd-gfx
mailing list