[dpdk-dev] [PATCH V16 2/4] eal: add device event monitor framework

[Tan, Jianfeng]

> -----Original Message-----
> From: Guo, Jia
> Sent: Monday, March 26, 2018 7:21 PM
> To: stephen at networkplumber.org; Richardson, Bruce; Yigit, Ferruh;
> Ananyev, Konstantin; gaetan.rivet at 6wind.com; Wu, Jingjing;
> thomas at monjalon.net; motih at mellanox.com; Van Haaren, Harry; Tan,
> Jianfeng
> Cc: jblunck at infradead.org; shreyansh.jain at nxp.com; dev at dpdk.org; Guo,
> Jia; Zhang, Helin
> Subject: [PATCH V16 2/4] eal: add device event monitor framework
> 
> This patch aims to add a general device event monitor mechanism at

mechanism -> framework?
Linux uevent is a mechanism.

> EAL device layer, for device hotplug awareness and actions adopted
> accordingly. It could also expand for all other type of device event
> monitor, but not in this scope at the stage.
> 
> To get started, users firstly register or unregister callbacks through
> the new added APIs. Callbacks can be some device specific, or for all
> devices.
>   -rte_dev_callback_register
>   -rte_dev_callback_unregister
> 

New APIs shall be added into rte_eal_version.map.

And also, we shall update the release note.

> Then application shall call below new added APIs to enable/disable the
> mechanism:
>   - rte_dev_event_monitor_start
>   - rte_dev_event_monitor_stop

Do we really have the use case to keep the callbacks, but stop monitoring? I don't think we really need these two APIs to enable/disable.

Instead, if we have a callback registered, then enable it; if we don't have any callbacks, then it's definitely disabled. 

> 
> Use hotplug case for example, when device hotplug insertion or hotplug
> removal, we will get notified from kernel, then call user's callbacks
> accordingly to handle it, such as detach or attach the device from the
> bus, and could be benifit for futher fail-safe or live-migration.

Typo: "be benifit " -> "benefit"
Typo: " futher" -> "further"

> 
> Signed-off-by: Jeff Guo <jia.guo at intel.com>
> ---
> v16->v15:
> 1.remove some linux related code out of eal common layer
> 2.fix some uneasy readble issue.
> ---
>  lib/librte_eal/bsdapp/eal/Makefile      |   1 +
>  lib/librte_eal/bsdapp/eal/eal_dev.c     |  19 +++++
>  lib/librte_eal/common/eal_common_dev.c  | 145
> ++++++++++++++++++++++++++++++++
>  lib/librte_eal/common/eal_private.h     |  24 ++++++
>  lib/librte_eal/common/include/rte_dev.h |  92 ++++++++++++++++++++
>  lib/librte_eal/linuxapp/eal/Makefile    |   1 +
>  lib/librte_eal/linuxapp/eal/eal_dev.c   |  20 +++++
>  7 files changed, 302 insertions(+)
>  create mode 100644 lib/librte_eal/bsdapp/eal/eal_dev.c
>  create mode 100644 lib/librte_eal/linuxapp/eal/eal_dev.c
> 
> diff --git a/lib/librte_eal/bsdapp/eal/Makefile
> b/lib/librte_eal/bsdapp/eal/Makefile
> index dd455e6..c0921dd 100644
> --- a/lib/librte_eal/bsdapp/eal/Makefile
> +++ b/lib/librte_eal/bsdapp/eal/Makefile
> @@ -33,6 +33,7 @@ SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) +=
> eal_lcore.c
>  SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += eal_timer.c
>  SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += eal_interrupts.c
>  SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += eal_alarm.c
> +SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += eal_dev.c
> 
>  # from common dir
>  SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += eal_common_lcore.c
> diff --git a/lib/librte_eal/bsdapp/eal/eal_dev.c
> b/lib/librte_eal/bsdapp/eal/eal_dev.c
> new file mode 100644
> index 0000000..ad606b3
> --- /dev/null
> +++ b/lib/librte_eal/bsdapp/eal/eal_dev.c
> @@ -0,0 +1,19 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2018 Intel Corporation
> + */
> +
> +#include <rte_log.h>
> +
> +int __rte_experimental
> +rte_dev_event_monitor_start(void)
> +{
> +	RTE_LOG(ERR, EAL, "Device event is not supported for FreeBSD\n");
> +	return -1;
> +}
> +
> +int __rte_experimental
> +rte_dev_event_monitor_stop(void)
> +{
> +	RTE_LOG(ERR, EAL, "Device event is not supported for FreeBSD\n");
> +	return -1;
> +}
> diff --git a/lib/librte_eal/common/eal_common_dev.c
> b/lib/librte_eal/common/eal_common_dev.c
> index cd07144..3a1bbb6 100644
> --- a/lib/librte_eal/common/eal_common_dev.c
> +++ b/lib/librte_eal/common/eal_common_dev.c
> @@ -14,9 +14,34 @@
>  #include <rte_devargs.h>
>  #include <rte_debug.h>
>  #include <rte_log.h>
> +#include <rte_spinlock.h>
> +#include <rte_malloc.h>
> 
>  #include "eal_private.h"
> 
> +/* spinlock for device callbacks */
> +static rte_spinlock_t dev_event_lock = RTE_SPINLOCK_INITIALIZER;
> +
> +/**
> + * The device event callback description.
> + *
> + * It contains callback address to be registered by user application,
> + * the pointer to the parameters for callback, and the device name.
> + */
> +struct dev_event_callback {
> +	TAILQ_ENTRY(dev_event_callback) next; /**< Callbacks list */
> +	rte_dev_event_cb_fn cb_fn;                /**< Callback address */
> +	void *cb_arg;                           /**< Callback parameter */
> +	char *dev_name;	 /**< Callback devcie name, NULL is for all
> device */

Typo: " devcie" -> "device" 

> +	uint32_t active;                        /**< Callback is executing */
> +};
> +
> +/** @internal Structure to keep track of registered callbacks */
> +TAILQ_HEAD(dev_event_cb_list, dev_event_callback);
> +
> +/* The device event callback list for all registered callbacks. */
> +static struct dev_event_cb_list dev_event_cbs;
> +
>  static int cmp_detached_dev_name(const struct rte_device *dev,
>  	const void *_name)
>  {
> @@ -207,3 +232,123 @@ rte_eal_hotplug_remove(const char *busname,
> const char *devname)
>  	rte_eal_devargs_remove(busname, devname);
>  	return ret;
>  }
> +
> +static struct dev_event_callback * __rte_experimental

We don't have to flag an internal function as " __rte_experimental ".

> +dev_event_cb_find(const char *device_name, rte_dev_event_cb_fn cb_fn,
> +				void *cb_arg)
> +{
> +	struct dev_event_callback *event_cb = NULL;
> +
> +	TAILQ_FOREACH(event_cb, &(dev_event_cbs), next) {
> +		if (event_cb->cb_fn == cb_fn && event_cb->cb_arg == cb_arg) {
> +			if (device_name == NULL && event_cb->dev_name == NULL)
> +				break;
> +			if (device_name == NULL || event_cb->dev_name == NULL)
> +				continue;
> +			if (!strcmp(event_cb->dev_name, device_name))
> +				break;
> +		}
> +	}
> +	return event_cb;
> +}
> +
> +int __rte_experimental
> +rte_dev_callback_register(const char *device_name, rte_dev_event_cb_fn cb_fn,
> +				void *cb_arg)

"rte_dev_event_callback_register" sounds more reasonable?

> +{
> +	struct dev_event_callback *event_cb = NULL;

We don't need to initialize it to NULL.

> +
> +	if (!cb_fn)
> +		return -EINVAL;
> +
> +	rte_spinlock_lock(&dev_event_lock);
> +
> +	if (TAILQ_EMPTY(&(dev_event_cbs)))
> +		TAILQ_INIT(&(dev_event_cbs));
> +
> +	event_cb = dev_event_cb_find(device_name, cb_fn, cb_arg);
> +
> +	/* create a new callback. */
> +	if (event_cb == NULL) {
> +		event_cb = malloc(sizeof(struct dev_event_callback));
> +		if (event_cb != NULL) {
> +			event_cb->cb_fn = cb_fn;
> +			event_cb->cb_arg = cb_arg;
> +			event_cb->dev_name = strdup(device_name);
> +			if (event_cb->dev_name == NULL)
> +				return -EINVAL;
> +			TAILQ_INSERT_TAIL(&(dev_event_cbs), event_cb, next);
> +		} else {
> +			RTE_LOG(ERR, EAL,
> +				"Failed to allocate memory for device event callback");

Miss the unlock here.

> +			return -ENOMEM;
> +		}
> +	}
> +
> +	rte_spinlock_unlock(&dev_event_lock);
> +	return (event_cb == NULL) ? -EEXIST : 0;
> +}
> +
> +int __rte_experimental
> +rte_dev_callback_unregister(const char *device_name,
> rte_dev_event_cb_fn cb_fn,
> +				void *cb_arg)
> +{
> +	int ret = -1;
> +	struct dev_event_callback *event_cb = NULL;
> +
> +	if (!cb_fn)
> +		return -EINVAL;
> +
> +	rte_spinlock_lock(&dev_event_lock);
> +
> +	event_cb = dev_event_cb_find(device_name, cb_fn, cb_arg);
> +
> +	/*
> +	 * if this callback is not executing right now,
> +	 * then remove it.
> +	 */

This note is not in right place.

> +	if (event_cb != NULL) {
> +		if (event_cb->active == 0) {
> +			TAILQ_REMOVE(&(dev_event_cbs), event_cb, next);
> +			rte_free(event_cb);
> +			ret = 0;
> +		}
> +		ret = -EBUSY;

Miss "else" for busy cb.

> +	}

Miss "else" for a cb which is not existed. And print an error log here.

> +
> +	rte_spinlock_unlock(&dev_event_lock);
> +	return ret;
> +}
> +
> +int __rte_experimental
> +dev_callback_process(char *device_name, enum rte_dev_event_type event,
> +				void *cb_arg)

>From interrupt thread, there is no cb_arg.

> +{
> +	struct dev_event_callback *cb_lst;
> +	int rc = 0;
> +
> +	rte_spinlock_lock(&dev_event_lock);
> +
> +	if (device_name == NULL)
> +		return -EINVAL;

Put such check out of lock. Or it's very easy to miss the unlock which is happening now.

> +
> +	TAILQ_FOREACH(cb_lst, &dev_event_cbs, next) {
> +		if (!cb_lst->dev_name)
> +			break;
> +		else if (!strcmp(cb_lst->dev_name, device_name))
> +			break;
> +	}

We invoke only one callback. But for this device, we could have many callback to call.

> +	if (cb_lst) {
> +		cb_lst->active = 1;
> +		if (cb_arg)
> +			cb_lst->cb_arg = cb_arg;

What's the reason of overwriting this cb_arg?

> +		rte_spinlock_unlock(&dev_event_lock);
> +		rc = cb_lst->cb_fn(device_name, event,
> +				cb_lst->cb_arg);
> +		rte_spinlock_lock(&dev_event_lock);
> +		cb_lst->active = 0;
> +	}
> +
> +	rte_spinlock_unlock(&dev_event_lock);
> +	return rc;

What's the reason of returning the ret of a callback? I don't think we need to return anything here.

> +}
> diff --git a/lib/librte_eal/common/eal_private.h
> b/lib/librte_eal/common/eal_private.h
> index 0b28770..d55cd68 100644
> --- a/lib/librte_eal/common/eal_private.h
> +++ b/lib/librte_eal/common/eal_private.h
> @@ -9,6 +9,8 @@
>  #include <stdint.h>
>  #include <stdio.h>
> 
> +#include <rte_dev.h>
> +
>  /**
>   * Initialize the memzone subsystem (private to eal).
>   *
> @@ -205,4 +207,26 @@ struct rte_bus
> *rte_bus_find_by_device_name(const char *str);
> 
>  int rte_mp_channel_init(void);
> 
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice

It's not an API. We don't need this.

> + *
> + * internal Executes all the user application registered callbacks for
> + * the specific device. It is for DPDK internal user only. User
> + * application should not call it directly.
> + *
> + * @param device_name
> + *  The device name.
> + * @param event
> + *  the device event type
> + * @param cb_arg
> + *  callback parameter.
> + *
> + * @return
> + *  - On success, return zero.
> + *  - On failure, a negative value.
> + */
> +int __rte_experimental

It's not an API, so don't add "__rte_experimental" flag.

> +dev_callback_process(char *device_name, enum rte_dev_event_type
> event,
> +				void *cb_arg);

As mentioned above, we don't have cb_arg from interrupt thread.

>  #endif /* _EAL_PRIVATE_H_ */
> diff --git a/lib/librte_eal/common/include/rte_dev.h
> b/lib/librte_eal/common/include/rte_dev.h
> index b688f1e..8867de6 100644
> --- a/lib/librte_eal/common/include/rte_dev.h
> +++ b/lib/librte_eal/common/include/rte_dev.h
> @@ -24,6 +24,26 @@ extern "C" {
>  #include <rte_compat.h>
>  #include <rte_log.h>
> 
> +/**
> + * The device event type.
> + */
> +enum rte_dev_event_type {
> +	RTE_DEV_EVENT_UNKNOWN,	/**< unknown event type */

Again, why do we report an "unknown" event to applications?

> +	RTE_DEV_EVENT_ADD,	/**< device being added */
> +	RTE_DEV_EVENT_REMOVE,	/**< device being removed */
> +	RTE_DEV_EVENT_MAX	/**< max value of this enum */
> +};
> +
> +struct rte_dev_event {
> +	enum rte_dev_event_type type;	/**< device event type */
> +	int subsystem;			/**< subsystem id */
> +	char *devname;			/**< device name */

I prefer to remove such note if we can already get the information from the variable name.

> +};
> +
> +typedef int (*rte_dev_event_cb_fn)(char *device_name,
> +					enum rte_dev_event_type event,
> +					void *cb_arg);

"rte_dev_event_callback" sounds better than "rte_dev_event_cb_fn" to me.

> +
>  __attribute__((format(printf, 2, 0)))
>  static inline void
>  rte_pmd_debug_trace(const char *func_name, const char *fmt, ...)
> @@ -267,4 +287,76 @@ __attribute__((used)) = str
>  }
>  #endif
> 
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice
> + *
> + * It registers the callback for the specific device.

"the" -> "a"

Besides, "It registers a callback for the specific device or all devices"

> + * Multiple callbacks cal be registered at the same time.

"cal" -> "can"

Besides, above sentence sounds like this API can register multiple callbacks. Change to:
"Users can call this API multiple times to register multiple callbacks."

> + *
> + * @param device_name
> + *  The device name, that is the param name of the struct rte_device,
> + *  null value means for all devices.
> + * @param cb_fn
> + *  callback address.
> + * @param cb_arg
> + *  address of parameter for callback.
> + *
> + * @return
> + *  - On success, zero.
> + *  - On failure, a negative value.
> + */
> +int __rte_experimental
> +rte_dev_callback_register(const char *device_name,
> rte_dev_event_cb_fn cb_fn,
> +			void *cb_arg);
> +
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice
> + *
> + * It unregisters the callback according to the specified device.
> + *
> + * @param device_name
> + *  The device name, that is the param name of the struct rte_device,
> + *  null value means for all devices.
> + * @param cb_fn
> + *  callback address.
> + * @param cb_arg
> + *  address of parameter for callback.
> + *
> + * @return
> + *  - On success, return the number of callback entities removed.
> + *  - On failure, a negative value.
> + */
> +int __rte_experimental
> +rte_dev_callback_unregister(const char *device_name,
> rte_dev_event_cb_fn cb_fn,
> +					void *cb_arg);
> +
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice
> + *
> + * Start the device event monitoring.
> + *
> + * @param none
> + * @return
> + *   - On success, zero.
> + *   - On failure, a negative value.
> + */
> +int __rte_experimental
> +rte_dev_event_monitor_start(void);
> +
> +/**
> + * @warning
> + * @b EXPERIMENTAL: this API may change without prior notice
> + *
> + * Stop the device event monitoring .
> + *
> + * @param none
> + * @return
> + *   - On success, zero.
> + *   - On failure, a negative value.
> + */
> +int __rte_experimental
> +rte_dev_event_monitor_stop(void);
>  #endif /* _RTE_DEV_H_ */
> diff --git a/lib/librte_eal/linuxapp/eal/Makefile
> b/lib/librte_eal/linuxapp/eal/Makefile
> index 7e5bbe8..8578796 100644
> --- a/lib/librte_eal/linuxapp/eal/Makefile
> +++ b/lib/librte_eal/linuxapp/eal/Makefile
> @@ -41,6 +41,7 @@ SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) +=
> eal_lcore.c
>  SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_timer.c
>  SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_interrupts.c
>  SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_alarm.c
> +SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_dev.c
> 
>  # from common dir
>  SRCS-$(CONFIG_RTE_EXEC_ENV_LINUXAPP) += eal_common_lcore.c
> diff --git a/lib/librte_eal/linuxapp/eal/eal_dev.c
> b/lib/librte_eal/linuxapp/eal/eal_dev.c
> new file mode 100644
> index 0000000..5ab5830
> --- /dev/null
> +++ b/lib/librte_eal/linuxapp/eal/eal_dev.c
> @@ -0,0 +1,20 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2018 Intel Corporation
> + */
> +
> +#include <rte_log.h>
> +#include <rte_dev.h>
> +
> +int __rte_experimental
> +rte_dev_event_monitor_start(void)
> +{
> +	/* TODO: start uevent monitor for linux */
> +	return 0;
> +}
> +
> +int __rte_experimental
> +rte_dev_event_monitor_stop(void)
> +{
> +	/* TODO: stop uevent monitor for linux */
> +	return 0;
> +}
> --
> 2.7.4