Re: [PATCHv2 10/10] zram: add dynamic device add/remove functionality

[Minchan Kim]

On Thu, Apr 16, 2015 at 08:55:56PM +0900, Sergey Senozhatsky wrote:
> We currently don't support on-demand device creation. The one and only way
> to have N zram devices is to specify num_devices module parameter (default
> value 1). That means that if, for some reason, at some point, user wants
> to have N + 1 devies he/she must umount all the existing devices, unload
> the module, load the module passing num_devices equals to N + 1. And do
> this again, if needed.
> 
> This patch introduces zram control sysfs class, which has two sysfs
> attrs:
> - zram_add      -- add a new zram device
> - zram_remove   -- remove a specific (device_id) zram device
> 
> zram_add sysfs attr is read-only and has only automatic device id assignment
> mode (as requested by Minchan Kim). read operation performed on this attr
> creates a new zram device and returns back its device_id or error status.
> 
> Usage example:
> 	# add a new specific zram device
> 	cat /sys/class/zram-control/zram_add
> 	2
> 
> 	# remove a specific zram device
> 	echo 4 > /sys/class/zram-control/zram_remove
> 
> Returning zram_add() error code back to user (-ENOMEM in this case)
> 
> 	cat /sys/class/zram-control/zram_add
> 	cat: /sys/class/zram-control/zram_add: Cannot allocate memory
> 
> NOTE, there might be users who already depend on the fact that at
> least zram0 device gets always created by zram_init(). Preserve this
> behavior.
> 
> [Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>: fix comment layout]
> Reported-by: Minchan Kim <minchan@xxxxxxxxxx>
> Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>
> ---
>  Documentation/ABI/testing/sysfs-class-zram |  24 ++++++
>  Documentation/blockdev/zram.txt            |  23 +++++-
>  drivers/block/zram/zram_drv.c              | 124 ++++++++++++++++++++++++++++-
>  3 files changed, 166 insertions(+), 5 deletions(-)
>  create mode 100644 Documentation/ABI/testing/sysfs-class-zram
> 
> diff --git a/Documentation/ABI/testing/sysfs-class-zram b/Documentation/ABI/testing/sysfs-class-zram
> new file mode 100644
> index 0000000..6f62ef5
> --- /dev/null
> +++ b/Documentation/ABI/testing/sysfs-class-zram
> @@ -0,0 +1,24 @@
> +What:		/sys/class/zram-control/
> +Date:		August 2015
> +KernelVersion:	4.1
> +Contact:	Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>
> +Description:
> +		The zram-control/ class sub-directory belongs to zram
> +		device class
> +
> +What:		/sys/class/zram-control/zram_add
> +Date:		August 2015
> +KernelVersion:	4.1
> +Contact:	Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>
> +Description:
> +		RO attribute. Read operation will cause zram to add a new
> +		device and return its device id back to user (so one can
> +		use /dev/zram<id>), or error code.
> +
> +What:		/sys/class/zram-control/zram_add
> +Date:		August 2015
> +KernelVersion:	4.1
> +Contact:	Sergey Senozhatsky <sergey.senozhatsky@xxxxxxxxx>
> +Description:
> +		Remove a specific /dev/zramX device, where X is a device_id
> +		provided by user
> diff --git a/Documentation/blockdev/zram.txt b/Documentation/blockdev/zram.txt
> index 2ccc741..44b1a77 100644
> --- a/Documentation/blockdev/zram.txt
> +++ b/Documentation/blockdev/zram.txt
> @@ -99,7 +99,24 @@ size of the disk when not in use so a huge zram is wasteful.
>  	mkfs.ext4 /dev/zram1
>  	mount /dev/zram1 /tmp
>  
> -7) Stats:
> +7) Add/remove zram devices
> +
> +zram provides a control interface, which enables dynamic (on-demand) device
> +addition and removal.
> +
> +In order to add a new /dev/zramX device, perform read operation on zram_add
> +attribute. This will return either new device's device id (meaning that you
> +can use /dev/zram<id>) or error code.
> +
> +Example:
> +	cat /sys/class/zram-control/zram_add

Why do we put zram-contol there rather than /sys/block/zram
> +	1
> +
> +To remove the existing /dev/zramX device (where X is a device id)
> +execute
> +	echo X > /sys/class/zram-control/zram_remove
> +
> +8) Stats:
>  Per-device statistics are exported as various nodes under /sys/block/zram<id>/
>  
>  A brief description of exported device attritbutes. For more details please
> @@ -178,11 +195,11 @@ line of text and contains the following stats separated by whitespace:
>  	num_migrated
>  
>  
> -8) Deactivate:
> +9) Deactivate:
>  	swapoff /dev/zram0
>  	umount /dev/zram1
>  
> -9) Reset:
> +10) Reset:
>  	Write any positive value to 'reset' sysfs node
>  	echo 1 > /sys/block/zram0/reset
>  	echo 1 > /sys/block/zram1/reset
> diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> index 2c2e7cc..848222a 100644
> --- a/drivers/block/zram/zram_drv.c
> +++ b/drivers/block/zram/zram_drv.c
> @@ -33,10 +33,14 @@
>  #include <linux/vmalloc.h>
>  #include <linux/err.h>
>  #include <linux/idr.h>
> +#include <linux/sysfs.h>
>  
>  #include "zram_drv.h"
>  
>  static DEFINE_IDR(zram_index_idr);
> +/* idr index must be protected */
> +static DEFINE_MUTEX(zram_index_mutex);
> +
>  static int zram_major;
>  static const char *default_compressor = "lzo";
>  
> @@ -1168,8 +1172,15 @@ static int zram_add(int device_id)

Why do zram_add need device_id?
We decided to remove option user pass device_id.

>  	if (!zram)
>  		return -ENOMEM;
>  
> -	ret = idr_alloc(&zram_index_idr, zram, device_id,
> -			device_id + 1, GFP_KERNEL);
> +	if (device_id < 0) {
> +		/* generate new device_id */
> +		ret = idr_alloc(&zram_index_idr, zram, 0, 0, GFP_KERNEL);
> +		device_id = ret;
> +	} else {
> +		/* use provided device_id */
> +		ret = idr_alloc(&zram_index_idr, zram, device_id,
> +				device_id + 1, GFP_KERNEL);
> +	}
>  	if (ret < 0)
>  		goto out_free_dev;
>  
> @@ -1278,6 +1289,105 @@ static void zram_remove(struct zram *zram)
>  	kfree(zram);
>  }
>  
> +/*
> + * Lookup if there is any device pointer that match the given device_id.
> + * return device pointer if so, or ERR_PTR() otherwise.
> + */
> +static struct zram *zram_lookup(int dev_id)
> +{
> +	struct zram *zram;
> +
> +	zram = idr_find(&zram_index_idr, dev_id);
> +	if (zram)
> +		return zram;
> +	return ERR_PTR(-ENODEV);

Just return NULL which is more simple and readable.

> +}
> +
> +/* zram module control sysfs attributes */
> +static ssize_t zram_add_show(struct class *class,
> +			struct class_attribute *attr,
> +			char *buf)
> +{
> +	int ret;
> +
> +	mutex_lock(&zram_index_mutex);
> +	/* pick up available device_id */
> +	ret = zram_add(-1);
> +	mutex_unlock(&zram_index_mutex);
> +
> +	if (ret < 0)
> +		return ret;
> +	return scnprintf(buf, PAGE_SIZE, "%d\n", ret);
> +}
> +
> +static ssize_t zram_remove_store(struct class *class,
> +			struct class_attribute *attr,
> +			const char *buf,
> +			size_t count)
> +{
> +	struct zram *zram;
> +	int ret, err, dev_id;
> +
> +	mutex_lock(&zram_index_mutex);
> +
> +	/* dev_id is gendisk->first_minor, which is `int' */
> +	ret = kstrtoint(buf, 10, &dev_id);
> +	if (ret || dev_id < 0) {
> +		ret = -EINVAL;
> +		goto out;
> +	}
> +
> +	zram = zram_lookup(dev_id);
> +	if (IS_ERR(zram)) {
> +		ret = PTR_ERR(zram);
> +		goto out;
> +	}
> +
> +	/*
> +	 * First, make ->disksize device attr RO, closing
> +	 * zram_remove() vs disksize_store() race window

Why don't you use zram->init_lock to protect the race?

> +	 */
> +	ret = sysfs_chmod_file(&disk_to_dev(zram->disk)->kobj,
> +			&dev_attr_disksize.attr, S_IRUGO);
> +	if (ret)
> +		goto out;
> +
> +	ret = zram_reset_device(zram);
> +	if (ret == 0) {
> +		/* ->disksize is RO and there are no ->bd_openers */
> +		zram_remove(zram);
> +		goto out;
> +	}
> +
> +	/*
> +	 * If there are still device bd_openers, try to make ->disksize
> +	 * RW again and return. even if we fail to make ->disksize RW,
> +	 * user still has RW ->reset attr. so it's possible to destroy
> +	 * that device.
> +	 */
> +	err = sysfs_chmod_file(&disk_to_dev(zram->disk)->kobj,
> +			&dev_attr_disksize.attr,
> +			S_IWUSR | S_IRUGO);
> +	if (err)
> +		ret = err;
> +
> +out:
> +	mutex_unlock(&zram_index_mutex);
> +	return ret ? ret : count;
> +}
> +
> +static struct class_attribute zram_control_class_attrs[] = {
> +	__ATTR_RO(zram_add),
> +	__ATTR_WO(zram_remove),
> +	__ATTR_NULL,
> +};
> +
> +static struct class zram_control_class = {
> +	.name		= "zram-control",
> +	.owner		= THIS_MODULE,
> +	.class_attrs	= zram_control_class_attrs,
> +};
> +
>  static int zram_exit_cb(int id, void *ptr, void *data)
>  {
>  	zram_remove(ptr);
> @@ -1286,6 +1396,7 @@ static int zram_exit_cb(int id, void *ptr, void *data)
>  
>  static void destroy_devices(void)
>  {
> +	class_unregister(&zram_control_class);
>  	idr_for_each(&zram_index_idr, &zram_exit_cb, NULL);
>  	idr_destroy(&zram_index_idr);
>  	unregister_blkdev(zram_major, "zram");
> @@ -1295,14 +1406,23 @@ static int __init zram_init(void)
>  {
>  	int ret, dev_id;
>  
> +	ret = class_register(&zram_control_class);
> +	if (ret) {
> +		pr_warn("Unable to register zram-control class\n");
> +		return ret;
> +	}
> +
>  	zram_major = register_blkdev(0, "zram");
>  	if (zram_major <= 0) {
>  		pr_warn("Unable to get major number\n");
> +		class_unregister(&zram_control_class);
>  		return -EBUSY;
>  	}
>  
>  	for (dev_id = 0; dev_id < num_devices; dev_id++) {
> +		mutex_lock(&zram_index_mutex);
>  		ret = zram_add(dev_id);
> +		mutex_unlock(&zram_index_mutex);
>  		if (ret < 0)
>  			goto out_error;
>  	}
> -- 
> 2.4.0.rc2
> 

-- 
Kind regards,
Minchan Kim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/