下面來看各個(gè)子系統(tǒng)對(duì)cgroup的支持，第一篇先研究blkio子系統(tǒng)

blkio子系統(tǒng)支持三種類型的QoS控制：

blkio.weight， blkio.weight_device：這些是基于設(shè)備權(quán)重值的控制方式

blkio.throttle.read_bps_device，blkio.throttle.write_bps_device：這些是基于帶寬的控制方式

blkio.throttle.read_iops_device，blkio.throttle.write_iops_device：這些是基于iops的控制方式

其中基于權(quán)重的控制方式，必須依賴于CFQ調(diào)度器，而基于throttle的控制方式則只需要在通用塊層實(shí)現(xiàn)就可以了

1) 基于blkio的cgroup_subsys的定義如下：

struct cgroup_subsys blkio_subsys = {

.name = "blkio",

.create = blkiocg_create,

.can_attach_task = blkiocg_can_attach_task,

.attach_task = blkiocg_attach_task,

.destroy = blkiocg_destroy,

.populate = blkiocg_populate,

#ifdef CONFIG_BLK_CGROUP

/* note: blkio_subsys_id is otherwise defined in blk-cgroup.h */

.subsys_id = blkio_subsys_id,

#endif

.use_id = 1,

.module = THIS_MODULE,

};

blkiocg_create(struct cgroup_subsys *subsys, struct cgroup *cgroup)：初始化一個(gè)blkio_cgroup結(jié)構(gòu)，并初始化blkio_cgroup->policy_list, blkio_cgroup->blkg_list

blkiocg_destroy(struct cgroup_subsys *subsys, struct cgroup *cgroup)：略過

blkiocg_populate(struct cgroup_subsys *subsys, struct cgroup *cgroup)：初始化好blkio_files里所有的blkio_policy_node對(duì)應(yīng)的cgroup文件系統(tǒng)的文件

blkiocg_can_attach_task(struct cgroup *cgrp, struct task_struct *tsk)：

/*

* We cannot support shared io contexts, as we have no mean to support

* two tasks with the same ioc in two different groups without major rework

* of the main cic data structures. ?For now we allow a task to change

* its cgroup only if it's the only owner of its ioc.

*/

2) 基于blkio的policy的數(shù)據(jù)結(jié)構(gòu)定義如下：

struct blkio_policy_node {

struct list_head node;

dev_t dev;

/* This node belongs to max bw policy or porportional weight policy */

enum blkio_policy_id plid;

/* cgroup file to which this rule belongs to */

int fileid;

union {

unsigned int weight;

/*

* Rate read/write in terms of byptes per second

* Whether this rate represents read or write is determined

* by file type "fileid".

*/

u64 bps;

unsigned int iops;

} val;

};

struct blkio_policy_ops {

blkio_unlink_group_fn *blkio_unlink_group_fn;

blkio_update_group_weight_fn *blkio_update_group_weight_fn;

blkio_update_group_read_bps_fn *blkio_update_group_read_bps_fn;

blkio_update_group_write_bps_fn *blkio_update_group_write_bps_fn;

blkio_update_group_read_iops_fn *blkio_update_group_read_iops_fn;

blkio_update_group_write_iops_fn *blkio_update_group_write_iops_fn;

};

enum blkio_policy_id {

BLKIO_POLICY_PROP = 0,/* Proportional Bandwidth division */

BLKIO_POLICY_THROTL,/* Throttling */

};

struct blkio_policy_type {

struct list_head list;

struct blkio_policy_ops ops;

enum blkio_policy_id plid;

};

blkio_policy_node，基本上可以看做一個(gè)cgroup文件系統(tǒng)下的一個(gè)配置文件對(duì)應(yīng)一個(gè)blkio_policy_node，一個(gè)cgroup目錄的所有的policy_node都會(huì)被鏈在一個(gè)blkio_cgroup->policy_list的鏈表中

blkio_policy_type根據(jù)不同的blkio_policy_id有不同的blkio_policy_ops，blkio_policy_register在cfq_init，throtl_init時(shí)被調(diào)用，這兩個(gè)初始化函數(shù)分別對(duì)應(yīng)基于權(quán)重的控制和基于閥值的控制，目前有兩個(gè)全局的blkio_policy_type的變量：

static struct blkio_policy_type blkio_policy_cfq = {

.ops = {

.blkio_unlink_group_fn =cfq_unlink_blkio_group,

.blkio_update_group_weight_fn =cfq_update_blkio_group_weight,

},

.plid = BLKIO_POLICY_PROP,

};

以及

static struct blkio_policy_type blkio_policy_throtl = {

.ops = {

.blkio_unlink_group_fn = throtl_unlink_blkio_group,

.blkio_update_group_read_bps_fn =

throtl_update_blkio_group_read_bps,

.blkio_update_group_write_bps_fn =

throtl_update_blkio_group_write_bps,

.blkio_update_group_read_iops_fn =

throtl_update_blkio_group_read_iops,

.blkio_update_group_write_iops_fn =

throtl_update_blkio_group_write_iops,

},

.plid = BLKIO_POLICY_THROTL,

};

3) 基于blkio的cgroup文件系統(tǒng)的數(shù)據(jù)結(jié)構(gòu)如下：

struct cftype blkio_files[] = {

{

.name = "weight_device",

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,

BLKIO_PROP_weight_device),

.read_seq_string = blkiocg_file_read,

.write_string = blkiocg_file_write,

.max_write_len = 256,

},

{

.name = "weight",

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,

BLKIO_PROP_weight),

.read_u64 = blkiocg_file_read_u64,

.write_u64 = blkiocg_file_write_u64,

},

{

.name = "throttle.read_bps_device",

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,

BLKIO_THROTL_read_bps_device),

.read_seq_string = blkiocg_file_read,

.write_string = blkiocg_file_write,

.max_write_len = 256,

},

{

.name = "throttle.write_bps_device",

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,

BLKIO_THROTL_write_bps_device),

.read_seq_string = blkiocg_file_read,

.write_string = blkiocg_file_write,

.max_write_len = 256,

},

{

.name = "throttle.read_iops_device",

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,

BLKIO_THROTL_read_iops_device),

.read_seq_string = blkiocg_file_read,

.write_string = blkiocg_file_write,

.max_write_len = 256,

},

{

.name = "throttle.write_iops_device",

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_THROTL,

BLKIO_THROTL_write_iops_device),

.read_seq_string = blkiocg_file_read,

.write_string = blkiocg_file_write,

.max_write_len = 256,

},

基本上調(diào)用的都是blkiocg_file_read，blkiocg_file_write

blkio_files中的struct cftype有個(gè)private成員變量，通過BLKIOFILE_PRIVATE宏來賦值，e.g.

.private = BLKIOFILE_PRIVATE(BLKIO_POLICY_PROP,?BLKIO_PROP_weight_device)

之后可以通過BLKIOFILE_POLICY獲取其policy類型：BLKIO_POLICY_THROTL或者BLKIO_POLICY_PROP，通過BLKIOFILE_ATTR獲取其文件名，所有的配置文件都有如下定義：

/* cgroup files owned by proportional weight policy */

enum blkcg_file_name_prop {

BLKIO_PROP_weight = 1,

BLKIO_PROP_weight_device,

BLKIO_PROP_io_service_bytes,

BLKIO_PROP_io_serviced,

BLKIO_PROP_time,

BLKIO_PROP_sectors,

BLKIO_PROP_unaccounted_time,

BLKIO_PROP_io_service_time,

BLKIO_PROP_io_wait_time,

BLKIO_PROP_io_merged,

BLKIO_PROP_io_queued,

BLKIO_PROP_avg_queue_size,

BLKIO_PROP_group_wait_time,

BLKIO_PROP_idle_time,

BLKIO_PROP_empty_time,

BLKIO_PROP_dequeue,

};

/* cgroup files owned by throttle policy */

enum blkcg_file_name_throtl {

BLKIO_THROTL_read_bps_device,

BLKIO_THROTL_write_bps_device,

BLKIO_THROTL_read_iops_device,

BLKIO_THROTL_write_iops_device,

BLKIO_THROTL_io_service_bytes,

BLKIO_THROTL_io_serviced,

};

static int blkiocg_file_read(struct cgroup *cgrp, struct cftype *cft,?struct seq_file *m)：通過cftype得到POLICY_ID, POLICY_FILE_NAME，通過struct cgroup得到struct blkio_cgroup，然后調(diào)用blkio_read_policy_node_files，按照一定格式存到一個(gè)seq_file里面，可以參考blkio_print_policy_node函數(shù)

static int blkiocg_file_write(struct cgroup *cgrp, struct cftype *cft,?const char *buffer)：先調(diào)用blkio_policy_parse_and_set生成一個(gè)新的blkio_policy_node，下面的步驟就是先刪了已有的policy node，再把新的policy node插入到blkio_cgroup->policy_list里面，最后調(diào)用blkio_update_policy_node_blkg，該函數(shù)對(duì)blkio_cgroup下面的所有blkio_group，都調(diào)用blkio_update_blkg_policy，該函數(shù)會(huì)根據(jù)blkio_policy_node的plid, fileid，調(diào)用不同的 blkio_update_xxxxx函數(shù)，以weight為例，最終調(diào)用到blkio_update_group_weight，后者又調(diào)用cfq_update_blkio_group_weight(這是跟CFQ緊耦合的一個(gè)函數(shù)，這里不做介紹了)

4) 幾個(gè)關(guān)鍵的數(shù)據(jù)結(jié)構(gòu)blkio_cgroup和blkio_group

struct blkio_cgroup {

struct cgroup_subsys_state css;

unsigned int weight;

spinlock_t lock;

struct hlist_head blkg_list;

struct list_head policy_list; /* list of blkio_policy_node */

};

struct blkio_group {

/* An rcu protected unique identifier for the group */

void *key;

struct hlist_node blkcg_node;

unsigned short blkcg_id;

/* Store cgroup path */

char path[128];

/* The device MKDEV(major, minor), this group has been created for */

dev_t dev;

/* policy which owns this blk group */

enum blkio_policy_id plid;

/* Need to serialize the stats in the case of reset/update */

spinlock_t stats_lock;

struct blkio_group_stats stats;

/* Per cpu stats pointer */

struct blkio_group_stats_cpu __percpu *stats_cpu;

};

blkio_cgroup代表了一個(gè)cgroup，但是這個(gè)cgroup里的進(jìn)程有可能會(huì)讀寫多個(gè)塊設(shè)備，所有通過一個(gè)cfq_data或者throtl_data的結(jié)構(gòu)作為紅黑樹的key，把多個(gè)blkio_group關(guān)聯(lián)到一個(gè)blkio_cgroup中。每個(gè)cfq_data或者throtl_data(根據(jù)policy的不同)實(shí)際上代表了一個(gè)塊設(shè)備

《新程序員》：云原生和全面數(shù)字化實(shí)踐50位技術(shù)專家共同創(chuàng)作，文字、視頻、音頻交互閱讀

總結(jié)

以上是生活随笔為你收集整理的linux qos 实现机制,linux的qos机制 - cgroup篇 (4)的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。