2019-02-16 05:39:11 +07:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Copyright 2016-2019 HabanaLabs, Ltd.
|
|
|
|
* All Rights Reserved.
|
|
|
|
*/
|
|
|
|
|
2019-04-10 19:18:46 +07:00
|
|
|
#define pr_fmt(fmt) "habanalabs: " fmt
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
#include "habanalabs.h"
|
|
|
|
|
|
|
|
#include <linux/pci.h>
|
2019-02-16 05:39:20 +07:00
|
|
|
#include <linux/sched/signal.h>
|
2019-02-16 05:39:19 +07:00
|
|
|
#include <linux/hwmon.h>
|
2019-03-24 15:15:44 +07:00
|
|
|
#include <uapi/misc/habanalabs.h>
|
2019-02-16 05:39:11 +07:00
|
|
|
|
2019-03-13 18:36:28 +07:00
|
|
|
#define HL_PLDM_PENDING_RESET_PER_SEC (HL_PENDING_RESET_PER_SEC * 10)
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
bool hl_device_disabled_or_in_reset(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
if ((hdev->disabled) || (atomic_read(&hdev->in_reset)))
|
|
|
|
return true;
|
|
|
|
else
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2019-03-24 15:15:44 +07:00
|
|
|
enum hl_device_status hl_device_status(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
enum hl_device_status status;
|
|
|
|
|
|
|
|
if (hdev->disabled)
|
|
|
|
status = HL_DEVICE_STATUS_MALFUNCTION;
|
|
|
|
else if (atomic_read(&hdev->in_reset))
|
|
|
|
status = HL_DEVICE_STATUS_IN_RESET;
|
|
|
|
else
|
|
|
|
status = HL_DEVICE_STATUS_OPERATIONAL;
|
|
|
|
|
|
|
|
return status;
|
|
|
|
};
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
static void hpriv_release(struct kref *ref)
|
|
|
|
{
|
|
|
|
struct hl_fpriv *hpriv;
|
|
|
|
struct hl_device *hdev;
|
|
|
|
|
|
|
|
hpriv = container_of(ref, struct hl_fpriv, refcount);
|
|
|
|
|
|
|
|
hdev = hpriv->hdev;
|
|
|
|
|
|
|
|
put_pid(hpriv->taskpid);
|
|
|
|
|
2019-02-16 05:39:24 +07:00
|
|
|
hl_debugfs_remove_file(hpriv);
|
|
|
|
|
2019-02-16 05:39:21 +07:00
|
|
|
mutex_destroy(&hpriv->restore_phase_mutex);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
list_del(&hpriv->dev_node);
|
2019-07-30 15:49:36 +07:00
|
|
|
hdev->compute_ctx = NULL;
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
|
|
|
kfree(hpriv);
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void hl_hpriv_get(struct hl_fpriv *hpriv)
|
|
|
|
{
|
|
|
|
kref_get(&hpriv->refcount);
|
|
|
|
}
|
|
|
|
|
|
|
|
void hl_hpriv_put(struct hl_fpriv *hpriv)
|
|
|
|
{
|
|
|
|
kref_put(&hpriv->refcount, hpriv_release);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_device_release - release function for habanalabs device
|
|
|
|
*
|
|
|
|
* @inode: pointer to inode structure
|
|
|
|
* @filp: pointer to file structure
|
|
|
|
*
|
|
|
|
* Called when process closes an habanalabs device
|
|
|
|
*/
|
|
|
|
static int hl_device_release(struct inode *inode, struct file *filp)
|
|
|
|
{
|
|
|
|
struct hl_fpriv *hpriv = filp->private_data;
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
hl_cb_mgr_fini(hpriv->hdev, &hpriv->cb_mgr);
|
2019-02-16 05:39:14 +07:00
|
|
|
hl_ctx_mgr_fini(hpriv->hdev, &hpriv->ctx_mgr);
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
filp->private_data = NULL;
|
|
|
|
|
|
|
|
hl_hpriv_put(hpriv);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2019-07-30 13:10:50 +07:00
|
|
|
static int hl_device_release_ctrl(struct inode *inode, struct file *filp)
|
|
|
|
{
|
|
|
|
struct hl_fpriv *hpriv = filp->private_data;
|
|
|
|
struct hl_device *hdev;
|
|
|
|
|
|
|
|
filp->private_data = NULL;
|
|
|
|
|
|
|
|
hdev = hpriv->hdev;
|
|
|
|
|
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
list_del(&hpriv->dev_node);
|
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
|
|
|
kfree(hpriv);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
/*
|
|
|
|
* hl_mmap - mmap function for habanalabs device
|
|
|
|
*
|
|
|
|
* @*filp: pointer to file structure
|
|
|
|
* @*vma: pointer to vm_area_struct of the process
|
|
|
|
*
|
|
|
|
* Called when process does an mmap on habanalabs device. Call the device's mmap
|
|
|
|
* function at the end of the common code.
|
|
|
|
*/
|
|
|
|
static int hl_mmap(struct file *filp, struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
struct hl_fpriv *hpriv = filp->private_data;
|
|
|
|
|
|
|
|
if ((vma->vm_pgoff & HL_MMAP_CB_MASK) == HL_MMAP_CB_MASK) {
|
|
|
|
vma->vm_pgoff ^= HL_MMAP_CB_MASK;
|
|
|
|
return hl_cb_mmap(hpriv, vma);
|
|
|
|
}
|
|
|
|
|
2019-02-27 17:15:16 +07:00
|
|
|
return -EINVAL;
|
2019-02-16 05:39:15 +07:00
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
static const struct file_operations hl_ops = {
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.open = hl_device_open,
|
2019-02-16 05:39:15 +07:00
|
|
|
.release = hl_device_release,
|
|
|
|
.mmap = hl_mmap,
|
|
|
|
.unlocked_ioctl = hl_ioctl,
|
|
|
|
.compat_ioctl = hl_ioctl
|
2019-02-16 05:39:11 +07:00
|
|
|
};
|
|
|
|
|
2019-07-30 13:10:50 +07:00
|
|
|
static const struct file_operations hl_ctrl_ops = {
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.open = hl_device_open_ctrl,
|
|
|
|
.release = hl_device_release_ctrl,
|
|
|
|
.unlocked_ioctl = hl_ioctl_control,
|
|
|
|
.compat_ioctl = hl_ioctl_control
|
|
|
|
};
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
static void device_release_func(struct device *dev)
|
|
|
|
{
|
|
|
|
kfree(dev);
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
/*
|
2019-08-08 19:25:52 +07:00
|
|
|
* device_init_cdev - Initialize cdev and device for habanalabs device
|
2019-02-16 05:39:11 +07:00
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
* @hclass: pointer to the class object of the device
|
|
|
|
* @minor: minor number of the specific device
|
2019-07-30 13:10:02 +07:00
|
|
|
* @fpos: file operations to install for this device
|
|
|
|
* @name: name of the device as it will appear in the filesystem
|
2019-08-08 19:25:52 +07:00
|
|
|
* @cdev: pointer to the char device object that will be initialized
|
|
|
|
* @dev: pointer to the device object that will be initialized
|
2019-02-16 05:39:11 +07:00
|
|
|
*
|
2019-08-08 19:25:52 +07:00
|
|
|
* Initialize a cdev and a Linux device for habanalabs's device.
|
2019-02-16 05:39:11 +07:00
|
|
|
*/
|
2019-08-08 19:25:52 +07:00
|
|
|
static int device_init_cdev(struct hl_device *hdev, struct class *hclass,
|
2019-07-30 13:10:02 +07:00
|
|
|
int minor, const struct file_operations *fops,
|
|
|
|
char *name, struct cdev *cdev,
|
|
|
|
struct device **dev)
|
2019-02-16 05:39:11 +07:00
|
|
|
{
|
2019-07-30 13:10:02 +07:00
|
|
|
cdev_init(cdev, fops);
|
|
|
|
cdev->owner = THIS_MODULE;
|
2019-08-08 19:25:52 +07:00
|
|
|
|
|
|
|
*dev = kzalloc(sizeof(**dev), GFP_KERNEL);
|
|
|
|
if (!*dev)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
device_initialize(*dev);
|
|
|
|
(*dev)->devt = MKDEV(hdev->major, minor);
|
|
|
|
(*dev)->class = hclass;
|
|
|
|
(*dev)->release = device_release_func;
|
|
|
|
dev_set_drvdata(*dev, hdev);
|
|
|
|
dev_set_name(*dev, "%s", name);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int device_cdev_sysfs_add(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
rc = cdev_device_add(&hdev->cdev, hdev->dev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"failed to add a char device to the system\n");
|
|
|
|
return rc;
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
rc = cdev_device_add(&hdev->cdev_ctrl, hdev->dev_ctrl);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"failed to add a control char device to the system\n");
|
|
|
|
goto delete_cdev_device;
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
/* hl_sysfs_init() must be done after adding the device to the system */
|
|
|
|
rc = hl_sysfs_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "failed to initialize sysfs\n");
|
|
|
|
goto delete_ctrl_cdev_device;
|
|
|
|
}
|
|
|
|
|
|
|
|
hdev->cdev_sysfs_created = true;
|
2019-02-16 05:39:11 +07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
delete_ctrl_cdev_device:
|
|
|
|
cdev_device_del(&hdev->cdev_ctrl, hdev->dev_ctrl);
|
|
|
|
delete_cdev_device:
|
|
|
|
cdev_device_del(&hdev->cdev, hdev->dev);
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void device_cdev_sysfs_del(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
/* device_release() won't be called so must free devices explicitly */
|
|
|
|
if (!hdev->cdev_sysfs_created) {
|
|
|
|
kfree(hdev->dev_ctrl);
|
|
|
|
kfree(hdev->dev);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
hl_sysfs_fini(hdev);
|
|
|
|
cdev_device_del(&hdev->cdev_ctrl, hdev->dev_ctrl);
|
|
|
|
cdev_device_del(&hdev->cdev, hdev->dev);
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* device_early_init - do some early initialization for the habanalabs device
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
* Install the relevant function pointers and call the early_init function,
|
|
|
|
* if such a function exists
|
|
|
|
*/
|
|
|
|
static int device_early_init(struct hl_device *hdev)
|
|
|
|
{
|
2019-02-16 05:39:13 +07:00
|
|
|
int rc;
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
switch (hdev->asic_type) {
|
|
|
|
case ASIC_GOYA:
|
2019-02-16 05:39:13 +07:00
|
|
|
goya_set_asic_funcs(hdev);
|
2019-02-16 05:39:11 +07:00
|
|
|
strlcpy(hdev->asic_name, "GOYA", sizeof(hdev->asic_name));
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
dev_err(hdev->dev, "Unrecognized ASIC type %d\n",
|
|
|
|
hdev->asic_type);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
rc = hdev->asic_funcs->early_init(hdev);
|
|
|
|
if (rc)
|
|
|
|
return rc;
|
|
|
|
|
2019-02-16 05:39:14 +07:00
|
|
|
rc = hl_asid_init(hdev);
|
|
|
|
if (rc)
|
|
|
|
goto early_fini;
|
|
|
|
|
2019-02-16 05:39:17 +07:00
|
|
|
hdev->cq_wq = alloc_workqueue("hl-free-jobs", WQ_UNBOUND, 0);
|
|
|
|
if (hdev->cq_wq == NULL) {
|
|
|
|
dev_err(hdev->dev, "Failed to allocate CQ workqueue\n");
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto asid_fini;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:18 +07:00
|
|
|
hdev->eq_wq = alloc_workqueue("hl-events", WQ_UNBOUND, 0);
|
|
|
|
if (hdev->eq_wq == NULL) {
|
|
|
|
dev_err(hdev->dev, "Failed to allocate EQ workqueue\n");
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto free_cq_wq;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
hdev->hl_chip_info = kzalloc(sizeof(struct hwmon_chip_info),
|
|
|
|
GFP_KERNEL);
|
|
|
|
if (!hdev->hl_chip_info) {
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto free_eq_wq;
|
|
|
|
}
|
|
|
|
|
2019-08-28 21:32:04 +07:00
|
|
|
hdev->idle_busy_ts_arr = kmalloc_array(HL_IDLE_BUSY_TS_ARR_SIZE,
|
|
|
|
sizeof(struct hl_device_idle_busy_ts),
|
|
|
|
(GFP_KERNEL | __GFP_ZERO));
|
|
|
|
if (!hdev->idle_busy_ts_arr) {
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto free_chip_info;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
hl_cb_mgr_init(&hdev->kernel_cb_mgr);
|
|
|
|
|
2019-02-16 05:39:17 +07:00
|
|
|
mutex_init(&hdev->send_cpu_message_lock);
|
2019-05-04 21:36:06 +07:00
|
|
|
mutex_init(&hdev->debug_lock);
|
2019-05-13 16:13:39 +07:00
|
|
|
mutex_init(&hdev->mmu_cache_lock);
|
2019-02-16 05:39:21 +07:00
|
|
|
INIT_LIST_HEAD(&hdev->hw_queues_mirror_list);
|
|
|
|
spin_lock_init(&hdev->hw_queues_mirror_lock);
|
2019-07-30 15:56:09 +07:00
|
|
|
INIT_LIST_HEAD(&hdev->fpriv_list);
|
|
|
|
mutex_init(&hdev->fpriv_list_lock);
|
2019-02-16 05:39:20 +07:00
|
|
|
atomic_set(&hdev->in_reset, 0);
|
2019-02-16 05:39:14 +07:00
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
return 0;
|
2019-02-16 05:39:14 +07:00
|
|
|
|
2019-08-28 21:32:04 +07:00
|
|
|
free_chip_info:
|
|
|
|
kfree(hdev->hl_chip_info);
|
2019-02-16 05:39:19 +07:00
|
|
|
free_eq_wq:
|
|
|
|
destroy_workqueue(hdev->eq_wq);
|
2019-02-16 05:39:18 +07:00
|
|
|
free_cq_wq:
|
|
|
|
destroy_workqueue(hdev->cq_wq);
|
2019-02-16 05:39:17 +07:00
|
|
|
asid_fini:
|
|
|
|
hl_asid_fini(hdev);
|
2019-02-16 05:39:14 +07:00
|
|
|
early_fini:
|
|
|
|
if (hdev->asic_funcs->early_fini)
|
|
|
|
hdev->asic_funcs->early_fini(hdev);
|
|
|
|
|
|
|
|
return rc;
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* device_early_fini - finalize all that was done in device_early_init
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
static void device_early_fini(struct hl_device *hdev)
|
|
|
|
{
|
2019-05-13 16:13:39 +07:00
|
|
|
mutex_destroy(&hdev->mmu_cache_lock);
|
2019-05-04 21:36:06 +07:00
|
|
|
mutex_destroy(&hdev->debug_lock);
|
2019-02-16 05:39:17 +07:00
|
|
|
mutex_destroy(&hdev->send_cpu_message_lock);
|
2019-02-16 05:39:13 +07:00
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_destroy(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
hl_cb_mgr_fini(hdev, &hdev->kernel_cb_mgr);
|
|
|
|
|
2019-08-28 21:32:04 +07:00
|
|
|
kfree(hdev->idle_busy_ts_arr);
|
2019-02-16 05:39:19 +07:00
|
|
|
kfree(hdev->hl_chip_info);
|
|
|
|
|
2019-02-16 05:39:18 +07:00
|
|
|
destroy_workqueue(hdev->eq_wq);
|
2019-02-16 05:39:17 +07:00
|
|
|
destroy_workqueue(hdev->cq_wq);
|
|
|
|
|
2019-02-16 05:39:14 +07:00
|
|
|
hl_asid_fini(hdev);
|
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
if (hdev->asic_funcs->early_fini)
|
|
|
|
hdev->asic_funcs->early_fini(hdev);
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
static void set_freq_to_low_job(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct hl_device *hdev = container_of(work, struct hl_device,
|
|
|
|
work_freq.work);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
|
|
|
|
if (!hdev->compute_ctx)
|
2019-02-16 05:39:19 +07:00
|
|
|
hl_device_set_frequency(hdev, PLL_LOW);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
schedule_delayed_work(&hdev->work_freq,
|
|
|
|
usecs_to_jiffies(HL_PLL_LOW_JOB_FREQ_USEC));
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
static void hl_device_heartbeat(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct hl_device *hdev = container_of(work, struct hl_device,
|
|
|
|
work_heartbeat.work);
|
|
|
|
|
|
|
|
if (hl_device_disabled_or_in_reset(hdev))
|
|
|
|
goto reschedule;
|
|
|
|
|
|
|
|
if (!hdev->asic_funcs->send_heartbeat(hdev))
|
|
|
|
goto reschedule;
|
|
|
|
|
|
|
|
dev_err(hdev->dev, "Device heartbeat failed!\n");
|
|
|
|
hl_device_reset(hdev, true, false);
|
|
|
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
reschedule:
|
|
|
|
schedule_delayed_work(&hdev->work_heartbeat,
|
|
|
|
usecs_to_jiffies(HL_HEARTBEAT_PER_USEC));
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
/*
|
|
|
|
* device_late_init - do late stuff initialization for the habanalabs device
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
* Do stuff that either needs the device H/W queues to be active or needs
|
|
|
|
* to happen after all the rest of the initialization is finished
|
|
|
|
*/
|
|
|
|
static int device_late_init(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
2019-05-29 18:24:51 +07:00
|
|
|
if (hdev->asic_funcs->late_init) {
|
|
|
|
rc = hdev->asic_funcs->late_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"failed late initialization for the H/W\n");
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
hdev->high_pll = hdev->asic_prop.high_pll;
|
|
|
|
|
|
|
|
/* force setting to low frequency */
|
2019-07-30 15:56:09 +07:00
|
|
|
hdev->curr_pll_profile = PLL_LOW;
|
2019-02-16 05:39:19 +07:00
|
|
|
|
|
|
|
if (hdev->pm_mng_profile == PM_AUTO)
|
|
|
|
hdev->asic_funcs->set_pll_profile(hdev, PLL_LOW);
|
|
|
|
else
|
|
|
|
hdev->asic_funcs->set_pll_profile(hdev, PLL_LAST);
|
|
|
|
|
2019-05-29 18:24:51 +07:00
|
|
|
INIT_DELAYED_WORK(&hdev->work_freq, set_freq_to_low_job);
|
2019-02-16 05:39:19 +07:00
|
|
|
schedule_delayed_work(&hdev->work_freq,
|
2019-05-29 18:24:51 +07:00
|
|
|
usecs_to_jiffies(HL_PLL_LOW_JOB_FREQ_USEC));
|
2019-02-16 05:39:19 +07:00
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
if (hdev->heartbeat) {
|
|
|
|
INIT_DELAYED_WORK(&hdev->work_heartbeat, hl_device_heartbeat);
|
|
|
|
schedule_delayed_work(&hdev->work_heartbeat,
|
|
|
|
usecs_to_jiffies(HL_HEARTBEAT_PER_USEC));
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
hdev->late_init_done = true;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* device_late_fini - finalize all that was done in device_late_init
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
static void device_late_fini(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
if (!hdev->late_init_done)
|
|
|
|
return;
|
|
|
|
|
|
|
|
cancel_delayed_work_sync(&hdev->work_freq);
|
2019-02-16 05:39:20 +07:00
|
|
|
if (hdev->heartbeat)
|
|
|
|
cancel_delayed_work_sync(&hdev->work_heartbeat);
|
2019-02-16 05:39:19 +07:00
|
|
|
|
|
|
|
if (hdev->asic_funcs->late_fini)
|
|
|
|
hdev->asic_funcs->late_fini(hdev);
|
|
|
|
|
|
|
|
hdev->late_init_done = false;
|
|
|
|
}
|
|
|
|
|
2019-08-28 21:32:04 +07:00
|
|
|
uint32_t hl_device_utilization(struct hl_device *hdev, uint32_t period_ms)
|
|
|
|
{
|
|
|
|
struct hl_device_idle_busy_ts *ts;
|
|
|
|
ktime_t zero_ktime, curr = ktime_get();
|
|
|
|
u32 overlap_cnt = 0, last_index = hdev->idle_busy_ts_idx;
|
|
|
|
s64 period_us, last_start_us, last_end_us, last_busy_time_us,
|
|
|
|
total_busy_time_us = 0, total_busy_time_ms;
|
|
|
|
|
|
|
|
zero_ktime = ktime_set(0, 0);
|
|
|
|
period_us = period_ms * USEC_PER_MSEC;
|
|
|
|
ts = &hdev->idle_busy_ts_arr[last_index];
|
|
|
|
|
|
|
|
/* check case that device is currently in idle */
|
|
|
|
if (!ktime_compare(ts->busy_to_idle_ts, zero_ktime) &&
|
|
|
|
!ktime_compare(ts->idle_to_busy_ts, zero_ktime)) {
|
|
|
|
|
|
|
|
last_index--;
|
|
|
|
/* Handle case idle_busy_ts_idx was 0 */
|
|
|
|
if (last_index > HL_IDLE_BUSY_TS_ARR_SIZE)
|
|
|
|
last_index = HL_IDLE_BUSY_TS_ARR_SIZE - 1;
|
|
|
|
|
|
|
|
ts = &hdev->idle_busy_ts_arr[last_index];
|
|
|
|
}
|
|
|
|
|
|
|
|
while (overlap_cnt < HL_IDLE_BUSY_TS_ARR_SIZE) {
|
|
|
|
/* Check if we are in last sample case. i.e. if the sample
|
|
|
|
* begun before the sampling period. This could be a real
|
|
|
|
* sample or 0 so need to handle both cases
|
|
|
|
*/
|
|
|
|
last_start_us = ktime_to_us(
|
|
|
|
ktime_sub(curr, ts->idle_to_busy_ts));
|
|
|
|
|
|
|
|
if (last_start_us > period_us) {
|
|
|
|
|
|
|
|
/* First check two cases:
|
|
|
|
* 1. If the device is currently busy
|
|
|
|
* 2. If the device was idle during the whole sampling
|
|
|
|
* period
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (!ktime_compare(ts->busy_to_idle_ts, zero_ktime)) {
|
|
|
|
/* Check if the device is currently busy */
|
|
|
|
if (ktime_compare(ts->idle_to_busy_ts,
|
|
|
|
zero_ktime))
|
|
|
|
return 100;
|
|
|
|
|
|
|
|
/* We either didn't have any activity or we
|
|
|
|
* reached an entry which is 0. Either way,
|
|
|
|
* exit and return what was accumulated so far
|
|
|
|
*/
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If sample has finished, check it is relevant */
|
|
|
|
last_end_us = ktime_to_us(
|
|
|
|
ktime_sub(curr, ts->busy_to_idle_ts));
|
|
|
|
|
|
|
|
if (last_end_us > period_us)
|
|
|
|
break;
|
|
|
|
|
|
|
|
/* It is relevant so add it but with adjustment */
|
|
|
|
last_busy_time_us = ktime_to_us(
|
|
|
|
ktime_sub(ts->busy_to_idle_ts,
|
|
|
|
ts->idle_to_busy_ts));
|
|
|
|
total_busy_time_us += last_busy_time_us -
|
|
|
|
(last_start_us - period_us);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Check if the sample is finished or still open */
|
|
|
|
if (ktime_compare(ts->busy_to_idle_ts, zero_ktime))
|
|
|
|
last_busy_time_us = ktime_to_us(
|
|
|
|
ktime_sub(ts->busy_to_idle_ts,
|
|
|
|
ts->idle_to_busy_ts));
|
|
|
|
else
|
|
|
|
last_busy_time_us = ktime_to_us(
|
|
|
|
ktime_sub(curr, ts->idle_to_busy_ts));
|
|
|
|
|
|
|
|
total_busy_time_us += last_busy_time_us;
|
|
|
|
|
|
|
|
last_index--;
|
|
|
|
/* Handle case idle_busy_ts_idx was 0 */
|
|
|
|
if (last_index > HL_IDLE_BUSY_TS_ARR_SIZE)
|
|
|
|
last_index = HL_IDLE_BUSY_TS_ARR_SIZE - 1;
|
|
|
|
|
|
|
|
ts = &hdev->idle_busy_ts_arr[last_index];
|
|
|
|
|
|
|
|
overlap_cnt++;
|
|
|
|
}
|
|
|
|
|
|
|
|
total_busy_time_ms = DIV_ROUND_UP_ULL(total_busy_time_us,
|
|
|
|
USEC_PER_MSEC);
|
|
|
|
|
|
|
|
return DIV_ROUND_UP_ULL(total_busy_time_ms * 100, period_ms);
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
/*
|
|
|
|
* hl_device_set_frequency - set the frequency of the device
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
* @freq: the new frequency value
|
|
|
|
*
|
2019-07-30 15:56:09 +07:00
|
|
|
* Change the frequency if needed. This function has no protection against
|
|
|
|
* concurrency, therefore it is assumed that the calling function has protected
|
|
|
|
* itself against the case of calling this function from multiple threads with
|
|
|
|
* different values
|
|
|
|
*
|
|
|
|
* Returns 0 if no change was done, otherwise returns 1
|
2019-02-16 05:39:19 +07:00
|
|
|
*/
|
|
|
|
int hl_device_set_frequency(struct hl_device *hdev, enum hl_pll_frequency freq)
|
|
|
|
{
|
2019-07-30 15:56:09 +07:00
|
|
|
if ((hdev->pm_mng_profile == PM_MANUAL) ||
|
|
|
|
(hdev->curr_pll_profile == freq))
|
2019-02-16 05:39:19 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
dev_dbg(hdev->dev, "Changing device frequency to %s\n",
|
|
|
|
freq == PLL_HIGH ? "high" : "low");
|
|
|
|
|
|
|
|
hdev->asic_funcs->set_pll_profile(hdev, freq);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
hdev->curr_pll_profile = freq;
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2019-05-04 21:36:06 +07:00
|
|
|
int hl_device_set_debug_mode(struct hl_device *hdev, bool enable)
|
|
|
|
{
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
mutex_lock(&hdev->debug_lock);
|
|
|
|
|
|
|
|
if (!enable) {
|
|
|
|
if (!hdev->in_debug) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to disable debug mode because device was not in debug mode\n");
|
|
|
|
rc = -EFAULT;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
hdev->asic_funcs->halt_coresight(hdev);
|
|
|
|
hdev->in_debug = 0;
|
|
|
|
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (hdev->in_debug) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to enable debug mode because device is already in debug mode\n");
|
|
|
|
rc = -EFAULT;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
hdev->in_debug = 1;
|
|
|
|
|
|
|
|
out:
|
|
|
|
mutex_unlock(&hdev->debug_lock);
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
/*
|
|
|
|
* hl_device_suspend - initiate device suspend
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
* Puts the hw in the suspend state (all asics).
|
|
|
|
* Returns 0 for success or an error on failure.
|
|
|
|
* Called at driver suspend.
|
|
|
|
*/
|
|
|
|
int hl_device_suspend(struct hl_device *hdev)
|
|
|
|
{
|
2019-02-16 05:39:13 +07:00
|
|
|
int rc;
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
pci_save_state(hdev->pdev);
|
|
|
|
|
2019-03-04 03:29:20 +07:00
|
|
|
/* Block future CS/VM/JOB completion operations */
|
|
|
|
rc = atomic_cmpxchg(&hdev->in_reset, 0, 1);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Can't suspend while in reset\n");
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* This blocks all other stuff that is not blocked by in_reset */
|
|
|
|
hdev->disabled = true;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Flush anyone that is inside the critical section of enqueue
|
|
|
|
* jobs to the H/W
|
|
|
|
*/
|
|
|
|
hdev->asic_funcs->hw_queues_lock(hdev);
|
|
|
|
hdev->asic_funcs->hw_queues_unlock(hdev);
|
|
|
|
|
|
|
|
/* Flush processes that are sending message to CPU */
|
|
|
|
mutex_lock(&hdev->send_cpu_message_lock);
|
|
|
|
mutex_unlock(&hdev->send_cpu_message_lock);
|
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
rc = hdev->asic_funcs->suspend(hdev);
|
|
|
|
if (rc)
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to disable PCI access of device CPU\n");
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
/* Shut down the device */
|
|
|
|
pci_disable_device(hdev->pdev);
|
|
|
|
pci_set_power_state(hdev->pdev, PCI_D3hot);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_device_resume - initiate device resume
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
* Bring the hw back to operating state (all asics).
|
|
|
|
* Returns 0 for success or an error on failure.
|
|
|
|
* Called at driver resume.
|
|
|
|
*/
|
|
|
|
int hl_device_resume(struct hl_device *hdev)
|
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
pci_set_power_state(hdev->pdev, PCI_D0);
|
|
|
|
pci_restore_state(hdev->pdev);
|
2019-03-04 03:29:20 +07:00
|
|
|
rc = pci_enable_device_mem(hdev->pdev);
|
2019-02-16 05:39:11 +07:00
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to enable PCI device in resume\n");
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
2019-03-04 03:29:20 +07:00
|
|
|
pci_set_master(hdev->pdev);
|
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
rc = hdev->asic_funcs->resume(hdev);
|
|
|
|
if (rc) {
|
2019-03-04 03:29:20 +07:00
|
|
|
dev_err(hdev->dev, "Failed to resume device after suspend\n");
|
|
|
|
goto disable_device;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
hdev->disabled = false;
|
|
|
|
atomic_set(&hdev->in_reset, 0);
|
|
|
|
|
|
|
|
rc = hl_device_reset(hdev, true, false);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Failed to reset device during resume\n");
|
|
|
|
goto disable_device;
|
2019-02-16 05:39:13 +07:00
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
return 0;
|
2019-03-04 03:29:20 +07:00
|
|
|
|
|
|
|
disable_device:
|
|
|
|
pci_clear_master(hdev->pdev);
|
|
|
|
pci_disable_device(hdev->pdev);
|
|
|
|
|
|
|
|
return rc;
|
2019-02-16 05:39:11 +07:00
|
|
|
}
|
|
|
|
|
2019-04-06 17:23:54 +07:00
|
|
|
static void device_kill_open_processes(struct hl_device *hdev)
|
2019-02-16 05:39:20 +07:00
|
|
|
{
|
2019-03-13 18:36:28 +07:00
|
|
|
u16 pending_total, pending_cnt;
|
2019-07-30 15:56:09 +07:00
|
|
|
struct hl_fpriv *hpriv;
|
2019-02-16 05:39:20 +07:00
|
|
|
struct task_struct *task = NULL;
|
|
|
|
|
2019-03-13 18:36:28 +07:00
|
|
|
if (hdev->pldm)
|
|
|
|
pending_total = HL_PLDM_PENDING_RESET_PER_SEC;
|
|
|
|
else
|
|
|
|
pending_total = HL_PENDING_RESET_PER_SEC;
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
/* Giving time for user to close FD, and for processes that are inside
|
|
|
|
* hl_device_open to finish
|
|
|
|
*/
|
|
|
|
if (!list_empty(&hdev->fpriv_list))
|
2019-02-16 05:39:20 +07:00
|
|
|
ssleep(1);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
|
|
|
|
/* This section must be protected because we are dereferencing
|
|
|
|
* pointers that are freed if the process exits
|
|
|
|
*/
|
|
|
|
list_for_each_entry(hpriv, &hdev->fpriv_list, dev_node) {
|
|
|
|
task = get_pid_task(hpriv->taskpid, PIDTYPE_PID);
|
2019-02-16 05:39:20 +07:00
|
|
|
if (task) {
|
2019-07-30 13:10:50 +07:00
|
|
|
dev_info(hdev->dev, "Killing user process pid=%d\n",
|
|
|
|
task_pid_nr(task));
|
2019-02-16 05:39:20 +07:00
|
|
|
send_sig(SIGKILL, task, 1);
|
2019-07-30 15:56:09 +07:00
|
|
|
usleep_range(1000, 10000);
|
2019-02-16 05:39:20 +07:00
|
|
|
|
|
|
|
put_task_struct(task);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-04-06 17:23:54 +07:00
|
|
|
/* We killed the open users, but because the driver cleans up after the
|
|
|
|
* user contexts are closed (e.g. mmu mappings), we need to wait again
|
|
|
|
* to make sure the cleaning phase is finished before continuing with
|
|
|
|
* the reset
|
|
|
|
*/
|
|
|
|
|
2019-03-13 18:36:28 +07:00
|
|
|
pending_cnt = pending_total;
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
while ((!list_empty(&hdev->fpriv_list)) && (pending_cnt)) {
|
|
|
|
dev_info(hdev->dev,
|
|
|
|
"Waiting for all unmap operations to finish before hard reset\n");
|
2019-03-13 18:36:28 +07:00
|
|
|
|
|
|
|
pending_cnt--;
|
|
|
|
|
|
|
|
ssleep(1);
|
|
|
|
}
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
if (!list_empty(&hdev->fpriv_list))
|
2019-03-13 18:36:28 +07:00
|
|
|
dev_crit(hdev->dev,
|
|
|
|
"Going to hard reset with open user contexts\n");
|
2019-04-06 17:23:54 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static void device_hard_reset_pending(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct hl_device_reset_work *device_reset_work =
|
|
|
|
container_of(work, struct hl_device_reset_work, reset_work);
|
|
|
|
struct hl_device *hdev = device_reset_work->hdev;
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
hl_device_reset(hdev, true, true);
|
|
|
|
|
|
|
|
kfree(device_reset_work);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_device_reset - reset the device
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
* @hard_reset: should we do hard reset to all engines or just reset the
|
|
|
|
* compute/dma engines
|
|
|
|
*
|
|
|
|
* Block future CS and wait for pending CS to be enqueued
|
|
|
|
* Call ASIC H/W fini
|
|
|
|
* Flush all completions
|
|
|
|
* Re-initialize all internal data structures
|
|
|
|
* Call ASIC H/W init, late_init
|
|
|
|
* Test queues
|
|
|
|
* Enable device
|
|
|
|
*
|
|
|
|
* Returns 0 for success or an error on failure.
|
|
|
|
*/
|
|
|
|
int hl_device_reset(struct hl_device *hdev, bool hard_reset,
|
|
|
|
bool from_hard_reset_thread)
|
|
|
|
{
|
|
|
|
int i, rc;
|
|
|
|
|
|
|
|
if (!hdev->init_done) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Can't reset before initialization is done\n");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Prevent concurrency in this function - only one reset should be
|
|
|
|
* done at any given time. Only need to perform this if we didn't
|
|
|
|
* get from the dedicated hard reset thread
|
|
|
|
*/
|
|
|
|
if (!from_hard_reset_thread) {
|
|
|
|
/* Block future CS/VM/JOB completion operations */
|
|
|
|
rc = atomic_cmpxchg(&hdev->in_reset, 0, 1);
|
|
|
|
if (rc)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* This also blocks future CS/VM/JOB completion operations */
|
|
|
|
hdev->disabled = true;
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
/* Flush anyone that is inside the critical section of enqueue
|
2019-02-16 05:39:20 +07:00
|
|
|
* jobs to the H/W
|
|
|
|
*/
|
|
|
|
hdev->asic_funcs->hw_queues_lock(hdev);
|
|
|
|
hdev->asic_funcs->hw_queues_unlock(hdev);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
/* Flush anyone that is inside device open */
|
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
dev_err(hdev->dev, "Going to RESET device!\n");
|
|
|
|
}
|
|
|
|
|
|
|
|
again:
|
|
|
|
if ((hard_reset) && (!from_hard_reset_thread)) {
|
|
|
|
struct hl_device_reset_work *device_reset_work;
|
|
|
|
|
2019-04-06 19:41:35 +07:00
|
|
|
hdev->hard_reset_pending = true;
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
device_reset_work = kzalloc(sizeof(*device_reset_work),
|
|
|
|
GFP_ATOMIC);
|
|
|
|
if (!device_reset_work) {
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Because the reset function can't run from interrupt or
|
|
|
|
* from heartbeat work, we need to call the reset function
|
|
|
|
* from a dedicated work
|
|
|
|
*/
|
|
|
|
INIT_WORK(&device_reset_work->reset_work,
|
2019-04-06 17:23:54 +07:00
|
|
|
device_hard_reset_pending);
|
2019-02-16 05:39:20 +07:00
|
|
|
device_reset_work->hdev = hdev;
|
|
|
|
schedule_work(&device_reset_work->reset_work);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (hard_reset) {
|
|
|
|
device_late_fini(hdev);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Now that the heartbeat thread is closed, flush processes
|
|
|
|
* which are sending messages to CPU
|
|
|
|
*/
|
|
|
|
mutex_lock(&hdev->send_cpu_message_lock);
|
|
|
|
mutex_unlock(&hdev->send_cpu_message_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Halt the engines and disable interrupts so we won't get any more
|
|
|
|
* completions from H/W and we won't have any accesses from the
|
|
|
|
* H/W to the host machine
|
|
|
|
*/
|
|
|
|
hdev->asic_funcs->halt_engines(hdev, hard_reset);
|
|
|
|
|
2019-02-16 05:39:21 +07:00
|
|
|
/* Go over all the queues, release all CS and their jobs */
|
|
|
|
hl_cs_rollback_all(hdev);
|
|
|
|
|
2019-11-17 22:41:57 +07:00
|
|
|
if (hard_reset) {
|
|
|
|
/* Kill processes here after CS rollback. This is because the
|
|
|
|
* process can't really exit until all its CSs are done, which
|
|
|
|
* is what we do in cs rollback
|
|
|
|
*/
|
2019-07-22 21:37:22 +07:00
|
|
|
device_kill_open_processes(hdev);
|
|
|
|
|
2019-11-17 22:41:57 +07:00
|
|
|
/* Flush the Event queue workers to make sure no other thread is
|
|
|
|
* reading or writing to registers during the reset
|
|
|
|
*/
|
|
|
|
flush_workqueue(hdev->eq_wq);
|
|
|
|
}
|
|
|
|
|
2019-03-17 14:12:29 +07:00
|
|
|
/* Release kernel context */
|
|
|
|
if ((hard_reset) && (hl_ctx_put(hdev->kernel_ctx) == 1))
|
2019-02-16 05:39:20 +07:00
|
|
|
hdev->kernel_ctx = NULL;
|
|
|
|
|
|
|
|
/* Reset the H/W. It will be in idle state after this returns */
|
|
|
|
hdev->asic_funcs->hw_fini(hdev, hard_reset);
|
|
|
|
|
2019-02-16 05:39:22 +07:00
|
|
|
if (hard_reset) {
|
|
|
|
hl_vm_fini(hdev);
|
2019-05-29 18:43:04 +07:00
|
|
|
hl_mmu_fini(hdev);
|
2019-02-16 05:39:20 +07:00
|
|
|
hl_eq_reset(hdev, &hdev->event_queue);
|
2019-02-16 05:39:22 +07:00
|
|
|
}
|
2019-02-16 05:39:20 +07:00
|
|
|
|
|
|
|
/* Re-initialize PI,CI to 0 in all queues (hw queue, cq) */
|
|
|
|
hl_hw_queue_reset(hdev, hard_reset);
|
|
|
|
for (i = 0 ; i < hdev->asic_prop.completion_queues_count ; i++)
|
|
|
|
hl_cq_reset(hdev, &hdev->completion_queue[i]);
|
|
|
|
|
2019-08-28 21:32:04 +07:00
|
|
|
hdev->idle_busy_ts_idx = 0;
|
|
|
|
hdev->idle_busy_ts_arr[0].busy_to_idle_ts = ktime_set(0, 0);
|
|
|
|
hdev->idle_busy_ts_arr[0].idle_to_busy_ts = ktime_set(0, 0);
|
|
|
|
|
|
|
|
if (hdev->cs_active_cnt)
|
|
|
|
dev_crit(hdev->dev, "CS active cnt %d is not 0 during reset\n",
|
|
|
|
hdev->cs_active_cnt);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-04-26 00:15:42 +07:00
|
|
|
/* Make sure the context switch phase will run again */
|
2019-07-30 15:49:36 +07:00
|
|
|
if (hdev->compute_ctx) {
|
|
|
|
atomic_set(&hdev->compute_ctx->thread_ctx_switch_token, 1);
|
|
|
|
hdev->compute_ctx->thread_ctx_switch_wait_token = 0;
|
2019-02-16 05:39:21 +07:00
|
|
|
}
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
/* Finished tear-down, starting to re-initialize */
|
|
|
|
|
|
|
|
if (hard_reset) {
|
2019-02-28 15:46:12 +07:00
|
|
|
hdev->device_cpu_disabled = false;
|
2019-04-06 19:41:35 +07:00
|
|
|
hdev->hard_reset_pending = false;
|
2019-02-28 15:46:12 +07:00
|
|
|
|
2019-03-17 14:12:29 +07:00
|
|
|
if (hdev->kernel_ctx) {
|
|
|
|
dev_crit(hdev->dev,
|
|
|
|
"kernel ctx was alive during hard reset, something is terribly wrong\n");
|
|
|
|
rc = -EBUSY;
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
2019-05-29 18:43:04 +07:00
|
|
|
rc = hl_mmu_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to initialize MMU S/W after hard reset\n");
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
/* Allocate the kernel context */
|
|
|
|
hdev->kernel_ctx = kzalloc(sizeof(*hdev->kernel_ctx),
|
|
|
|
GFP_KERNEL);
|
|
|
|
if (!hdev->kernel_ctx) {
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
2019-07-30 15:49:36 +07:00
|
|
|
hdev->compute_ctx = NULL;
|
2019-02-16 05:39:20 +07:00
|
|
|
|
|
|
|
rc = hl_ctx_init(hdev, hdev->kernel_ctx, true);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"failed to init kernel ctx in hard reset\n");
|
|
|
|
kfree(hdev->kernel_ctx);
|
|
|
|
hdev->kernel_ctx = NULL;
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
rc = hdev->asic_funcs->hw_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"failed to initialize the H/W after reset\n");
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
|
|
|
hdev->disabled = false;
|
|
|
|
|
|
|
|
/* Check that the communication with the device is working */
|
|
|
|
rc = hdev->asic_funcs->test_queues(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to detect if device is alive after reset\n");
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (hard_reset) {
|
|
|
|
rc = device_late_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed late init after hard reset\n");
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:22 +07:00
|
|
|
rc = hl_vm_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to init memory module after hard reset\n");
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
hl_set_max_power(hdev, hdev->max_power);
|
|
|
|
} else {
|
|
|
|
rc = hdev->asic_funcs->soft_reset_late_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed late init after soft reset\n");
|
|
|
|
goto out_err;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
atomic_set(&hdev->in_reset, 0);
|
|
|
|
|
|
|
|
if (hard_reset)
|
|
|
|
hdev->hard_reset_cnt++;
|
|
|
|
else
|
|
|
|
hdev->soft_reset_cnt++;
|
|
|
|
|
2019-08-08 20:48:55 +07:00
|
|
|
dev_warn(hdev->dev, "Successfully finished resetting the device\n");
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
out_err:
|
|
|
|
hdev->disabled = true;
|
|
|
|
|
|
|
|
if (hard_reset) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to reset! Device is NOT usable\n");
|
|
|
|
hdev->hard_reset_cnt++;
|
|
|
|
} else {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to do soft-reset, trying hard reset\n");
|
|
|
|
hdev->soft_reset_cnt++;
|
|
|
|
hard_reset = true;
|
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
|
|
|
|
atomic_set(&hdev->in_reset, 0);
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
/*
|
|
|
|
* hl_device_init - main initialization function for habanalabs device
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
* Allocate an id for the device, do early initialization and then call the
|
|
|
|
* ASIC specific initialization functions. Finally, create the cdev and the
|
|
|
|
* Linux device to expose it to the user
|
|
|
|
*/
|
|
|
|
int hl_device_init(struct hl_device *hdev, struct class *hclass)
|
|
|
|
{
|
2019-02-16 05:39:17 +07:00
|
|
|
int i, rc, cq_ready_cnt;
|
2019-07-30 13:10:02 +07:00
|
|
|
char *name;
|
2019-08-08 19:25:52 +07:00
|
|
|
bool add_cdev_sysfs_on_err = false;
|
2019-07-30 13:10:02 +07:00
|
|
|
|
2019-07-30 13:10:50 +07:00
|
|
|
name = kasprintf(GFP_KERNEL, "hl%d", hdev->id / 2);
|
2019-08-08 19:25:52 +07:00
|
|
|
if (!name) {
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
2019-02-16 05:39:11 +07:00
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
/* Initialize cdev and device structures */
|
|
|
|
rc = device_init_cdev(hdev, hclass, hdev->id, &hl_ops, name,
|
2019-07-30 13:10:02 +07:00
|
|
|
&hdev->cdev, &hdev->dev);
|
|
|
|
|
|
|
|
kfree(name);
|
2019-02-16 05:39:11 +07:00
|
|
|
|
|
|
|
if (rc)
|
|
|
|
goto out_disabled;
|
|
|
|
|
2019-07-30 13:10:50 +07:00
|
|
|
name = kasprintf(GFP_KERNEL, "hl_controlD%d", hdev->id / 2);
|
|
|
|
if (!name) {
|
|
|
|
rc = -ENOMEM;
|
2019-08-08 19:25:52 +07:00
|
|
|
goto free_dev;
|
2019-07-30 13:10:50 +07:00
|
|
|
}
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
/* Initialize cdev and device structures for control device */
|
|
|
|
rc = device_init_cdev(hdev, hclass, hdev->id_control, &hl_ctrl_ops,
|
2019-07-30 13:10:50 +07:00
|
|
|
name, &hdev->cdev_ctrl, &hdev->dev_ctrl);
|
|
|
|
|
|
|
|
kfree(name);
|
|
|
|
|
|
|
|
if (rc)
|
2019-08-08 19:25:52 +07:00
|
|
|
goto free_dev;
|
2019-07-30 13:10:50 +07:00
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
/* Initialize ASIC function pointers and perform early init */
|
|
|
|
rc = device_early_init(hdev);
|
|
|
|
if (rc)
|
2019-08-08 19:25:52 +07:00
|
|
|
goto free_dev_ctrl;
|
2019-02-16 05:39:11 +07:00
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
/*
|
|
|
|
* Start calling ASIC initialization. First S/W then H/W and finally
|
|
|
|
* late init
|
|
|
|
*/
|
|
|
|
rc = hdev->asic_funcs->sw_init(hdev);
|
|
|
|
if (rc)
|
|
|
|
goto early_fini;
|
|
|
|
|
2019-02-16 05:39:17 +07:00
|
|
|
/*
|
|
|
|
* Initialize the H/W queues. Must be done before hw_init, because
|
|
|
|
* there the addresses of the kernel queue are being written to the
|
|
|
|
* registers of the device
|
|
|
|
*/
|
|
|
|
rc = hl_hw_queues_create(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "failed to initialize kernel queues\n");
|
|
|
|
goto sw_fini;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Initialize the completion queues. Must be done before hw_init,
|
|
|
|
* because there the addresses of the completion queues are being
|
|
|
|
* passed as arguments to request_irq
|
|
|
|
*/
|
|
|
|
hdev->completion_queue =
|
|
|
|
kcalloc(hdev->asic_prop.completion_queues_count,
|
|
|
|
sizeof(*hdev->completion_queue), GFP_KERNEL);
|
|
|
|
|
|
|
|
if (!hdev->completion_queue) {
|
|
|
|
dev_err(hdev->dev, "failed to allocate completion queues\n");
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto hw_queues_destroy;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0, cq_ready_cnt = 0;
|
|
|
|
i < hdev->asic_prop.completion_queues_count;
|
|
|
|
i++, cq_ready_cnt++) {
|
|
|
|
rc = hl_cq_init(hdev, &hdev->completion_queue[i], i);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"failed to initialize completion queue\n");
|
|
|
|
goto cq_fini;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:18 +07:00
|
|
|
/*
|
|
|
|
* Initialize the event queue. Must be done before hw_init,
|
|
|
|
* because there the address of the event queue is being
|
|
|
|
* passed as argument to request_irq
|
|
|
|
*/
|
|
|
|
rc = hl_eq_init(hdev, &hdev->event_queue);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "failed to initialize event queue\n");
|
|
|
|
goto cq_fini;
|
|
|
|
}
|
|
|
|
|
2019-05-29 18:43:04 +07:00
|
|
|
/* MMU S/W must be initialized before kernel context is created */
|
|
|
|
rc = hl_mmu_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Failed to initialize MMU S/W structures\n");
|
|
|
|
goto eq_fini;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:14 +07:00
|
|
|
/* Allocate the kernel context */
|
|
|
|
hdev->kernel_ctx = kzalloc(sizeof(*hdev->kernel_ctx), GFP_KERNEL);
|
|
|
|
if (!hdev->kernel_ctx) {
|
|
|
|
rc = -ENOMEM;
|
2019-05-29 18:43:04 +07:00
|
|
|
goto mmu_fini;
|
2019-02-16 05:39:14 +07:00
|
|
|
}
|
|
|
|
|
2019-07-30 15:49:36 +07:00
|
|
|
hdev->compute_ctx = NULL;
|
2019-02-16 05:39:14 +07:00
|
|
|
|
|
|
|
rc = hl_ctx_init(hdev, hdev->kernel_ctx, true);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "failed to initialize kernel context\n");
|
2019-08-01 20:57:36 +07:00
|
|
|
kfree(hdev->kernel_ctx);
|
|
|
|
goto mmu_fini;
|
2019-02-16 05:39:14 +07:00
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
rc = hl_cb_pool_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "failed to initialize CB pool\n");
|
|
|
|
goto release_ctx;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:24 +07:00
|
|
|
hl_debugfs_add_device(hdev);
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
if (hdev->asic_funcs->get_hw_state(hdev) == HL_DEVICE_HW_STATE_DIRTY) {
|
|
|
|
dev_info(hdev->dev,
|
|
|
|
"H/W state is dirty, must reset before initializing\n");
|
|
|
|
hdev->asic_funcs->hw_fini(hdev, true);
|
|
|
|
}
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
/*
|
|
|
|
* From this point, in case of an error, add char devices and create
|
|
|
|
* sysfs nodes as part of the error flow, to allow debugging.
|
|
|
|
*/
|
|
|
|
add_cdev_sysfs_on_err = true;
|
|
|
|
|
2019-02-16 05:39:16 +07:00
|
|
|
rc = hdev->asic_funcs->hw_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "failed to initialize the H/W\n");
|
|
|
|
rc = 0;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
|
|
|
|
|
|
|
hdev->disabled = false;
|
|
|
|
|
2019-02-16 05:39:17 +07:00
|
|
|
/* Check that the communication with the device is working */
|
|
|
|
rc = hdev->asic_funcs->test_queues(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Failed to detect if device is alive\n");
|
|
|
|
rc = 0;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
rc = device_late_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Failed late initialization\n");
|
|
|
|
rc = 0;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
|
|
|
|
|
|
|
dev_info(hdev->dev, "Found %s device with %lluGB DRAM\n",
|
|
|
|
hdev->asic_name,
|
|
|
|
hdev->asic_prop.dram_size / 1024 / 1024 / 1024);
|
|
|
|
|
2019-02-16 05:39:22 +07:00
|
|
|
rc = hl_vm_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Failed to initialize memory module\n");
|
|
|
|
rc = 0;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
/*
|
2019-08-08 19:25:52 +07:00
|
|
|
* Expose devices and sysfs nodes to user.
|
|
|
|
* From here there is no need to add char devices and create sysfs nodes
|
|
|
|
* in case of an error.
|
|
|
|
*/
|
|
|
|
add_cdev_sysfs_on_err = false;
|
|
|
|
rc = device_cdev_sysfs_add(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"Failed to add char devices and sysfs nodes\n");
|
|
|
|
rc = 0;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_hwmon_init() must be called after device_late_init(), because only
|
2019-02-16 05:39:19 +07:00
|
|
|
* there we get the information from the device about which
|
2019-08-08 19:25:52 +07:00
|
|
|
* hwmon-related sensors the device supports.
|
|
|
|
* Furthermore, it must be done after adding the device to the system.
|
2019-02-16 05:39:19 +07:00
|
|
|
*/
|
|
|
|
rc = hl_hwmon_init(hdev);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(hdev->dev, "Failed to initialize hwmon\n");
|
|
|
|
rc = 0;
|
|
|
|
goto out_disabled;
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
dev_notice(hdev->dev,
|
|
|
|
"Successfully added device to habanalabs driver\n");
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
hdev->init_done = true;
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
return 0;
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
release_ctx:
|
|
|
|
if (hl_ctx_put(hdev->kernel_ctx) != 1)
|
|
|
|
dev_err(hdev->dev,
|
|
|
|
"kernel ctx is still alive on initialization failure\n");
|
2019-05-29 18:43:04 +07:00
|
|
|
mmu_fini:
|
|
|
|
hl_mmu_fini(hdev);
|
2019-02-16 05:39:18 +07:00
|
|
|
eq_fini:
|
|
|
|
hl_eq_fini(hdev, &hdev->event_queue);
|
2019-02-16 05:39:17 +07:00
|
|
|
cq_fini:
|
|
|
|
for (i = 0 ; i < cq_ready_cnt ; i++)
|
|
|
|
hl_cq_fini(hdev, &hdev->completion_queue[i]);
|
|
|
|
kfree(hdev->completion_queue);
|
|
|
|
hw_queues_destroy:
|
|
|
|
hl_hw_queues_destroy(hdev);
|
2019-02-16 05:39:14 +07:00
|
|
|
sw_fini:
|
|
|
|
hdev->asic_funcs->sw_fini(hdev);
|
2019-02-16 05:39:13 +07:00
|
|
|
early_fini:
|
|
|
|
device_early_fini(hdev);
|
2019-08-08 19:25:52 +07:00
|
|
|
free_dev_ctrl:
|
|
|
|
kfree(hdev->dev_ctrl);
|
|
|
|
free_dev:
|
|
|
|
kfree(hdev->dev);
|
2019-02-16 05:39:11 +07:00
|
|
|
out_disabled:
|
|
|
|
hdev->disabled = true;
|
2019-08-08 19:25:52 +07:00
|
|
|
if (add_cdev_sysfs_on_err)
|
|
|
|
device_cdev_sysfs_add(hdev);
|
2019-02-16 05:39:11 +07:00
|
|
|
if (hdev->pdev)
|
|
|
|
dev_err(&hdev->pdev->dev,
|
|
|
|
"Failed to initialize hl%d. Device is NOT usable !\n",
|
2019-09-01 20:13:25 +07:00
|
|
|
hdev->id / 2);
|
2019-02-16 05:39:11 +07:00
|
|
|
else
|
|
|
|
pr_err("Failed to initialize hl%d. Device is NOT usable !\n",
|
2019-09-01 20:13:25 +07:00
|
|
|
hdev->id / 2);
|
2019-02-16 05:39:11 +07:00
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_device_fini - main tear-down function for habanalabs device
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
*
|
|
|
|
* Destroy the device, call ASIC fini functions and release the id
|
|
|
|
*/
|
|
|
|
void hl_device_fini(struct hl_device *hdev)
|
|
|
|
{
|
2019-02-16 05:39:20 +07:00
|
|
|
int i, rc;
|
|
|
|
ktime_t timeout;
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
dev_info(hdev->dev, "Removing device\n");
|
|
|
|
|
2019-02-16 05:39:20 +07:00
|
|
|
/*
|
|
|
|
* This function is competing with the reset function, so try to
|
|
|
|
* take the reset atomic and if we are already in middle of reset,
|
|
|
|
* wait until reset function is finished. Reset function is designed
|
|
|
|
* to always finish (could take up to a few seconds in worst case).
|
|
|
|
*/
|
|
|
|
|
|
|
|
timeout = ktime_add_us(ktime_get(),
|
|
|
|
HL_PENDING_RESET_PER_SEC * 1000 * 1000 * 4);
|
|
|
|
rc = atomic_cmpxchg(&hdev->in_reset, 0, 1);
|
|
|
|
while (rc) {
|
|
|
|
usleep_range(50, 200);
|
|
|
|
rc = atomic_cmpxchg(&hdev->in_reset, 0, 1);
|
|
|
|
if (ktime_compare(ktime_get(), timeout) > 0) {
|
|
|
|
WARN(1, "Failed to remove device because reset function did not finish\n");
|
|
|
|
return;
|
|
|
|
}
|
2019-04-02 19:46:02 +07:00
|
|
|
}
|
2019-02-16 05:39:20 +07:00
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
/* Mark device as disabled */
|
|
|
|
hdev->disabled = true;
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
/* Flush anyone that is inside the critical section of enqueue
|
2019-04-06 17:23:54 +07:00
|
|
|
* jobs to the H/W
|
|
|
|
*/
|
|
|
|
hdev->asic_funcs->hw_queues_lock(hdev);
|
|
|
|
hdev->asic_funcs->hw_queues_unlock(hdev);
|
|
|
|
|
2019-07-30 15:56:09 +07:00
|
|
|
/* Flush anyone that is inside device open */
|
|
|
|
mutex_lock(&hdev->fpriv_list_lock);
|
|
|
|
mutex_unlock(&hdev->fpriv_list_lock);
|
|
|
|
|
2019-04-06 19:41:35 +07:00
|
|
|
hdev->hard_reset_pending = true;
|
|
|
|
|
2019-02-16 05:39:19 +07:00
|
|
|
hl_hwmon_fini(hdev);
|
|
|
|
|
|
|
|
device_late_fini(hdev);
|
|
|
|
|
2019-02-16 05:39:24 +07:00
|
|
|
hl_debugfs_remove_device(hdev);
|
|
|
|
|
2019-02-16 05:39:18 +07:00
|
|
|
/*
|
|
|
|
* Halt the engines and disable interrupts so we won't get any more
|
|
|
|
* completions from H/W and we won't have any accesses from the
|
|
|
|
* H/W to the host machine
|
|
|
|
*/
|
|
|
|
hdev->asic_funcs->halt_engines(hdev, true);
|
|
|
|
|
2019-02-16 05:39:21 +07:00
|
|
|
/* Go over all the queues, release all CS and their jobs */
|
|
|
|
hl_cs_rollback_all(hdev);
|
|
|
|
|
2019-07-22 21:37:22 +07:00
|
|
|
/* Kill processes here after CS rollback. This is because the process
|
|
|
|
* can't really exit until all its CSs are done, which is what we
|
|
|
|
* do in cs rollback
|
|
|
|
*/
|
|
|
|
device_kill_open_processes(hdev);
|
|
|
|
|
2019-02-16 05:39:15 +07:00
|
|
|
hl_cb_pool_fini(hdev);
|
|
|
|
|
2019-02-16 05:39:14 +07:00
|
|
|
/* Release kernel context */
|
|
|
|
if ((hdev->kernel_ctx) && (hl_ctx_put(hdev->kernel_ctx) != 1))
|
|
|
|
dev_err(hdev->dev, "kernel ctx is still alive\n");
|
|
|
|
|
2019-02-16 05:39:16 +07:00
|
|
|
/* Reset the H/W. It will be in idle state after this returns */
|
|
|
|
hdev->asic_funcs->hw_fini(hdev, true);
|
|
|
|
|
2019-02-16 05:39:22 +07:00
|
|
|
hl_vm_fini(hdev);
|
|
|
|
|
2019-05-29 18:43:04 +07:00
|
|
|
hl_mmu_fini(hdev);
|
|
|
|
|
2019-02-16 05:39:18 +07:00
|
|
|
hl_eq_fini(hdev, &hdev->event_queue);
|
|
|
|
|
2019-02-16 05:39:17 +07:00
|
|
|
for (i = 0 ; i < hdev->asic_prop.completion_queues_count ; i++)
|
|
|
|
hl_cq_fini(hdev, &hdev->completion_queue[i]);
|
|
|
|
kfree(hdev->completion_queue);
|
|
|
|
|
|
|
|
hl_hw_queues_destroy(hdev);
|
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
/* Call ASIC S/W finalize function */
|
|
|
|
hdev->asic_funcs->sw_fini(hdev);
|
|
|
|
|
2019-02-16 05:39:11 +07:00
|
|
|
device_early_fini(hdev);
|
|
|
|
|
2019-08-08 19:25:52 +07:00
|
|
|
/* Hide devices and sysfs nodes from user */
|
|
|
|
device_cdev_sysfs_del(hdev);
|
2019-02-16 05:39:11 +07:00
|
|
|
|
|
|
|
pr_info("removed device successfully\n");
|
|
|
|
}
|
|
|
|
|
2019-02-16 05:39:13 +07:00
|
|
|
/*
|
|
|
|
* MMIO register access helper functions.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_rreg - Read an MMIO register
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
* @reg: MMIO register offset (in bytes)
|
|
|
|
*
|
|
|
|
* Returns the value of the MMIO register we are asked to read
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
inline u32 hl_rreg(struct hl_device *hdev, u32 reg)
|
|
|
|
{
|
|
|
|
return readl(hdev->rmmio + reg);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* hl_wreg - Write to an MMIO register
|
|
|
|
*
|
|
|
|
* @hdev: pointer to habanalabs device structure
|
|
|
|
* @reg: MMIO register offset (in bytes)
|
|
|
|
* @val: 32-bit value
|
|
|
|
*
|
|
|
|
* Writes the 32-bit value into the MMIO register
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
inline void hl_wreg(struct hl_device *hdev, u32 reg, u32 val)
|
|
|
|
{
|
|
|
|
writel(val, hdev->rmmio + reg);
|
|
|
|
}
|