Changes:
$ sudo yum install -y git "kernel-devel-uname-r == $(uname -r)"
$ sudo yum groupinstall -y "Development Tools"
$ sudo modprobe dm-mod
$ sudo ./ddi-setup.sh -d /dev/vdb1 create /data
| /* | |
| * Copyright (C) 2005-2007 Red Hat GmbH | |
| * | |
| * A target that delays reads and/or writes and can send | |
| * them to different devices. | |
| * | |
| * This file is released under the GPL. | |
| * | |
| * | |
| * ddi - Disk Delay Injection | |
| * A DM driver which injects delay in block device I/O, allowing dynamic parameter control. | |
| * Based on dm-delay from the kernel upstream. | |
| */ | |
| #include <linux/module.h> | |
| #include <linux/init.h> | |
| #include <linux/blkdev.h> | |
| #include <linux/bio.h> | |
| #include <linux/slab.h> | |
| #include <linux/device-mapper.h> | |
| #define DM_MSG_PREFIX "ddi" | |
| struct delay_class { | |
| struct dm_dev *dev; | |
| sector_t start; | |
| unsigned delay; | |
| unsigned ops; | |
| }; | |
| struct delay_c { | |
| struct timer_list delay_timer; | |
| struct mutex timer_lock; | |
| struct workqueue_struct *kdelayd_wq; | |
| struct work_struct flush_expired_bios; | |
| struct list_head delayed_bios; | |
| atomic_t may_delay; | |
| struct delay_class read; | |
| struct delay_class write; | |
| struct delay_class flush; | |
| struct kobject *kobj; | |
| struct kobj_attribute read_delay_attr; | |
| struct kobj_attribute write_delay_attr; | |
| int argc; | |
| }; | |
| struct dm_delay_info { | |
| struct delay_c *context; | |
| struct delay_class *class; | |
| struct list_head list; | |
| unsigned long expires; | |
| }; | |
| static DEFINE_MUTEX(delayed_bios_lock); | |
| /* Sysfs implementation for dynamic parameter control.*/ | |
| static struct kobject *ddi_kobj; | |
| static ssize_t show_delay(unsigned delay, char *buf) | |
| { | |
| /* The buffer allocation size is PAGE_SIZE(=4k typically) so it is safe to print an int | |
| * without limiting length. | |
| * ref: https://www.kernel.org/doc/Documentation/filesystems/sysfs.txt | |
| */ | |
| return sprintf(buf, "%u\n", delay); | |
| } | |
| static void queue_timeout(struct delay_c *dc, unsigned long expires); | |
| static ssize_t store_delay(struct delay_c *dc, unsigned *delay, const char *buf, size_t count) | |
| { | |
| unsigned new_delay; | |
| unsigned long expires; | |
| /* buf is null-terminated. */ | |
| sscanf(buf, "%d", &new_delay); | |
| if (new_delay < 0) { | |
| printk(KERN_WARNING "Not setting an invalid delay: %d\n", new_delay); | |
| return count; | |
| } | |
| printk(KERN_DEBUG "Updating delay %u => %u\n", *delay, new_delay); | |
| *delay = new_delay; | |
| smp_wmb(); | |
| /* Update timer to cancel possibly existing too long timeout. */ | |
| expires = jiffies + msecs_to_jiffies(new_delay); | |
| queue_timeout(dc, expires); | |
| return count; | |
| } | |
| static ssize_t read_delay_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) | |
| { | |
| struct delay_c *dc = container_of(attr, struct delay_c, read_delay_attr); | |
| return show_delay(dc->read.delay, buf); | |
| } | |
| static ssize_t read_delay_store(struct kobject *kobj, struct kobj_attribute *attr, | |
| const char *buf, size_t count) | |
| { | |
| struct delay_c *dc = container_of(attr, struct delay_c, read_delay_attr); | |
| return store_delay(dc, &dc->read.delay, buf, count); | |
| } | |
| static ssize_t write_delay_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) | |
| { | |
| struct delay_c *dc = container_of(attr, struct delay_c, write_delay_attr); | |
| return show_delay(dc->write.delay, buf); | |
| } | |
| static ssize_t write_delay_store(struct kobject *kobj, struct kobj_attribute *attr, | |
| const char *buf, size_t count) | |
| { | |
| struct delay_c *dc = container_of(attr, struct delay_c, write_delay_attr); | |
| if (!dc->write.dev) { | |
| printk(KERN_WARNING "Write device is not configured\n"); | |
| return count; | |
| } | |
| return store_delay(dc, &dc->write.delay, buf, count); | |
| } | |
| static int init_dev_kobject(struct delay_c *dc) | |
| { | |
| int ret = 0; | |
| static struct attribute *attrs[3]; | |
| static struct attribute_group attr_group = { .attrs = attrs }; | |
| attrs[0] = &dc->read_delay_attr.attr; | |
| attrs[1] = &dc->write_delay_attr.attr; | |
| attrs[2] = NULL; | |
| dc->kobj = kobject_create_and_add(dc->read.dev->name, ddi_kobj); | |
| if (!dc->kobj) | |
| return -ENOMEM; | |
| dc->read_delay_attr = (struct kobj_attribute)__ATTR(read_delay, 0664, read_delay_show, read_delay_store); | |
| dc->write_delay_attr = (struct kobj_attribute)__ATTR(write_delay, 0664, write_delay_show, write_delay_store); | |
| ret = sysfs_create_group(dc->kobj, &attr_group); | |
| if (ret) | |
| kobject_put(dc->kobj); | |
| return ret; | |
| } | |
| static void destroy_dev_kobject(struct delay_c *dc) | |
| { | |
| kobject_put(dc->kobj); | |
| } | |
| /* Device Mapper implementation. */ | |
| static void handle_delayed_timer(struct timer_list *t) | |
| { | |
| struct delay_c *dc = from_timer(dc, t, delay_timer); | |
| queue_work(dc->kdelayd_wq, &dc->flush_expired_bios); | |
| } | |
| static void queue_timeout(struct delay_c *dc, unsigned long expires) | |
| { | |
| mutex_lock(&dc->timer_lock); | |
| if (!timer_pending(&dc->delay_timer) || expires < dc->delay_timer.expires) | |
| mod_timer(&dc->delay_timer, expires); | |
| mutex_unlock(&dc->timer_lock); | |
| } | |
| static void flush_bios(struct bio *bio) | |
| { | |
| struct bio *n; | |
| while (bio) { | |
| n = bio->bi_next; | |
| bio->bi_next = NULL; | |
| submit_bio_noacct(bio); | |
| bio = n; | |
| } | |
| } | |
| static struct bio *flush_delayed_bios(struct delay_c *dc, int flush_all) | |
| { | |
| struct dm_delay_info *delayed, *next; | |
| unsigned long next_expires = 0; | |
| unsigned long start_timer = 0; | |
| struct bio_list flush_bios = { }; | |
| mutex_lock(&delayed_bios_lock); | |
| list_for_each_entry_safe(delayed, next, &dc->delayed_bios, list) { | |
| if (flush_all || time_after_eq(jiffies, delayed->expires)) { | |
| struct bio *bio = dm_bio_from_per_bio_data(delayed, | |
| sizeof(struct dm_delay_info)); | |
| list_del(&delayed->list); | |
| bio_list_add(&flush_bios, bio); | |
| delayed->class->ops--; | |
| continue; | |
| } | |
| if (!start_timer) { | |
| start_timer = 1; | |
| next_expires = delayed->expires; | |
| } else | |
| next_expires = min(next_expires, delayed->expires); | |
| } | |
| mutex_unlock(&delayed_bios_lock); | |
| if (start_timer) | |
| queue_timeout(dc, next_expires); | |
| return bio_list_get(&flush_bios); | |
| } | |
| static void flush_expired_bios(struct work_struct *work) | |
| { | |
| struct delay_c *dc; | |
| dc = container_of(work, struct delay_c, flush_expired_bios); | |
| flush_bios(flush_delayed_bios(dc, 0)); | |
| } | |
| static void delay_dtr(struct dm_target *ti) | |
| { | |
| struct delay_c *dc = ti->private; | |
| destroy_dev_kobject(dc); | |
| if (dc->kdelayd_wq) | |
| destroy_workqueue(dc->kdelayd_wq); | |
| if (dc->read.dev) | |
| dm_put_device(ti, dc->read.dev); | |
| if (dc->write.dev) | |
| dm_put_device(ti, dc->write.dev); | |
| if (dc->flush.dev) | |
| dm_put_device(ti, dc->flush.dev); | |
| mutex_destroy(&dc->timer_lock); | |
| kfree(dc); | |
| } | |
| static int delay_class_ctr(struct dm_target *ti, struct delay_class *c, char **argv) | |
| { | |
| int ret; | |
| unsigned long long tmpll; | |
| char dummy; | |
| if (sscanf(argv[1], "%llu%c", &tmpll, &dummy) != 1 || tmpll != (sector_t)tmpll) { | |
| ti->error = "Invalid device sector"; | |
| return -EINVAL; | |
| } | |
| c->start = tmpll; | |
| if (sscanf(argv[2], "%u%c", &c->delay, &dummy) != 1) { | |
| ti->error = "Invalid delay"; | |
| return -EINVAL; | |
| } | |
| ret = dm_get_device(ti, argv[0], dm_table_get_mode(ti->table), &c->dev); | |
| if (ret) { | |
| ti->error = "Device lookup failed"; | |
| return ret; | |
| } | |
| return 0; | |
| } | |
| /* | |
| * Mapping parameters: | |
| * <device> <offset> <delay> [<write_device> <write_offset> <write_delay>] | |
| * | |
| * With separate write parameters, the first set is only used for reads. | |
| * Offsets are specified in sectors. | |
| * Delays are specified in milliseconds. | |
| */ | |
| static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) | |
| { | |
| struct delay_c *dc; | |
| int ret; | |
| if (argc != 3 && argc != 6 && argc != 9) { | |
| ti->error = "Requires exactly 3, 6 or 9 arguments"; | |
| return -EINVAL; | |
| } | |
| dc = kzalloc(sizeof(*dc), GFP_KERNEL); | |
| if (!dc) { | |
| ti->error = "Cannot allocate context"; | |
| return -ENOMEM; | |
| } | |
| ti->private = dc; | |
| timer_setup(&dc->delay_timer, handle_delayed_timer, 0); | |
| INIT_WORK(&dc->flush_expired_bios, flush_expired_bios); | |
| INIT_LIST_HEAD(&dc->delayed_bios); | |
| mutex_init(&dc->timer_lock); | |
| atomic_set(&dc->may_delay, 1); | |
| dc->argc = argc; | |
| ret = delay_class_ctr(ti, &dc->read, argv); | |
| if (ret) | |
| goto bad; | |
| if (argc == 3) { | |
| ret = delay_class_ctr(ti, &dc->write, argv); | |
| if (ret) | |
| goto bad; | |
| ret = delay_class_ctr(ti, &dc->flush, argv); | |
| if (ret) | |
| goto bad; | |
| goto out; | |
| } | |
| ret = delay_class_ctr(ti, &dc->write, argv + 3); | |
| if (ret) | |
| goto bad; | |
| if (argc == 6) { | |
| ret = delay_class_ctr(ti, &dc->flush, argv + 3); | |
| if (ret) | |
| goto bad; | |
| goto out; | |
| } | |
| ret = delay_class_ctr(ti, &dc->flush, argv + 6); | |
| if (ret) | |
| goto bad; | |
| out: | |
| dc->kdelayd_wq = alloc_workqueue("kddid", WQ_MEM_RECLAIM, 0); | |
| if (!dc->kdelayd_wq) { | |
| ret = -EINVAL; | |
| DMERR("Couldn't start kdelayd"); | |
| goto bad; | |
| } | |
| ti->num_flush_bios = 1; | |
| ti->num_discard_bios = 1; | |
| ti->per_io_data_size = sizeof(struct dm_delay_info); | |
| ret = init_dev_kobject(dc); | |
| if (ret) { | |
| DMERR("Failed to setup sysfs"); | |
| goto bad_sysfs; | |
| } | |
| return 0; | |
| bad_sysfs: | |
| destroy_workqueue(dc->kdelayd_wq); | |
| bad: | |
| delay_dtr(ti); | |
| return ret; | |
| } | |
| static int delay_bio(struct delay_c *dc, struct delay_class *c, struct bio *bio) | |
| { | |
| struct dm_delay_info *delayed; | |
| unsigned long expires = 0; | |
| if (!c->delay || !atomic_read(&dc->may_delay)) | |
| return DM_MAPIO_REMAPPED; | |
| delayed = dm_per_bio_data(bio, sizeof(struct dm_delay_info)); | |
| delayed->context = dc; | |
| delayed->expires = expires = jiffies + msecs_to_jiffies(c->delay); | |
| mutex_lock(&delayed_bios_lock); | |
| c->ops++; | |
| list_add_tail(&delayed->list, &dc->delayed_bios); | |
| mutex_unlock(&delayed_bios_lock); | |
| queue_timeout(dc, expires); | |
| return DM_MAPIO_SUBMITTED; | |
| } | |
| static void delay_presuspend(struct dm_target *ti) | |
| { | |
| struct delay_c *dc = ti->private; | |
| atomic_set(&dc->may_delay, 0); | |
| del_timer_sync(&dc->delay_timer); | |
| flush_bios(flush_delayed_bios(dc, 1)); | |
| } | |
| static void delay_resume(struct dm_target *ti) | |
| { | |
| struct delay_c *dc = ti->private; | |
| atomic_set(&dc->may_delay, 1); | |
| } | |
| static int delay_map(struct dm_target *ti, struct bio *bio) | |
| { | |
| struct delay_c *dc = ti->private; | |
| struct delay_class *c; | |
| struct dm_delay_info *delayed = dm_per_bio_data(bio, sizeof(struct dm_delay_info)); | |
| if (bio_data_dir(bio) == WRITE) { | |
| if (unlikely(bio->bi_opf & REQ_PREFLUSH)) | |
| c = &dc->flush; | |
| else | |
| c = &dc->write; | |
| } else { | |
| c = &dc->read; | |
| } | |
| delayed->class = c; | |
| bio_set_dev(bio, c->dev->bdev); | |
| if (bio_sectors(bio)) | |
| bio->bi_iter.bi_sector = c->start + dm_target_offset(ti, bio->bi_iter.bi_sector); | |
| return delay_bio(dc, c, bio); | |
| } | |
| #define DMEMIT_DELAY_CLASS(c) \ | |
| DMEMIT("%s %llu %u", (c)->dev->name, (unsigned long long)(c)->start, (c)->delay) | |
| static void delay_status(struct dm_target *ti, status_type_t type, | |
| unsigned status_flags, char *result, unsigned maxlen) | |
| { | |
| struct delay_c *dc = ti->private; | |
| int sz = 0; | |
| switch (type) { | |
| case STATUSTYPE_INFO: | |
| DMEMIT("%u %u %u", dc->read.ops, dc->write.ops, dc->flush.ops); | |
| break; | |
| case STATUSTYPE_TABLE: | |
| DMEMIT_DELAY_CLASS(&dc->read); | |
| if (dc->argc >= 6) { | |
| DMEMIT(" "); | |
| DMEMIT_DELAY_CLASS(&dc->write); | |
| } | |
| if (dc->argc >= 9) { | |
| DMEMIT(" "); | |
| DMEMIT_DELAY_CLASS(&dc->flush); | |
| } | |
| break; | |
| case STATUSTYPE_IMA: | |
| *result = '\0'; | |
| break; | |
| } | |
| } | |
| static int delay_iterate_devices(struct dm_target *ti, | |
| iterate_devices_callout_fn fn, void *data) | |
| { | |
| struct delay_c *dc = ti->private; | |
| int ret = 0; | |
| ret = fn(ti, dc->read.dev, dc->read.start, ti->len, data); | |
| if (ret) | |
| goto out; | |
| ret = fn(ti, dc->write.dev, dc->write.start, ti->len, data); | |
| if (ret) | |
| goto out; | |
| ret = fn(ti, dc->flush.dev, dc->flush.start, ti->len, data); | |
| if (ret) | |
| goto out; | |
| out: | |
| return ret; | |
| } | |
| static struct target_type delay_target = { | |
| .name = "ddi", | |
| .version = {1, 2, 1}, | |
| .features = DM_TARGET_PASSES_INTEGRITY, | |
| .module = THIS_MODULE, | |
| .ctr = delay_ctr, | |
| .dtr = delay_dtr, | |
| .map = delay_map, | |
| .presuspend = delay_presuspend, | |
| .resume = delay_resume, | |
| .status = delay_status, | |
| .iterate_devices = delay_iterate_devices, | |
| }; | |
| static int __init dm_delay_init(void) | |
| { | |
| int r; | |
| r = dm_register_target(&delay_target); | |
| if (r < 0) { | |
| DMERR("register failed %d", r); | |
| goto bad_register; | |
| } | |
| ddi_kobj = kobject_create_and_add("ddi", fs_kobj); | |
| if (!ddi_kobj) | |
| return -ENOMEM; | |
| return 0; | |
| bad_register: | |
| return r; | |
| } | |
| static void __exit dm_delay_exit(void) | |
| { | |
| kobject_put(ddi_kobj); | |
| dm_unregister_target(&delay_target); | |
| } | |
| /* Module hooks */ | |
| module_init(dm_delay_init); | |
| module_exit(dm_delay_exit); | |
| MODULE_DESCRIPTION(DM_NAME " delay target"); | |
| MODULE_AUTHOR("Heinz Mauelshagen <[email protected]>"); | |
| MODULE_LICENSE("GPL"); |