mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Marco Felsch <m.felsch@pengutronix.de>
To: Ahmad Fatoum <a.fatoum@pengutronix.de>
Cc: barebox@lists.infradead.org
Subject: Re: [PATCH 12/18] misc: storage-by-uuid: prepare for generalizing to different aliases
Date: Mon, 14 Apr 2025 12:31:21 +0200	[thread overview]
Message-ID: <20250414103121.tak6wi5ze7k3qy5g@pengutronix.de> (raw)
In-Reply-To: <20250414063152.2736649-13-a.fatoum@pengutronix.de>

On 25-04-14, Ahmad Fatoum wrote:
> Besides the probe function and the check_exist function, there is
> nothing UUID specific about the code. As preparation for making it
> reusable for arbitrary cdev aliases, rename it to storage-by-alias
> instead as I couldn't figure out a proper backronym for sbu.
> 
> Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
> ---
>  drivers/misc/storage-by-uuid.c | 204 ++++++++++++++++++---------------
>  1 file changed, 114 insertions(+), 90 deletions(-)
> 
> diff --git a/drivers/misc/storage-by-uuid.c b/drivers/misc/storage-by-uuid.c
> index 7bb987dffbec..2795ff93c07c 100644
> --- a/drivers/misc/storage-by-uuid.c
> +++ b/drivers/misc/storage-by-uuid.c
> @@ -1,4 +1,13 @@
>  // SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * cdev aliases for existing block devices. In future, this should
> + * likely be replaced by a more generic device mapper support.
> + *
> + * Currently supported targets:
> + * - barebox,storage-by-uuid
> + *   Useful for referencing existing EFI disks and their partition
> + *   from device tree by offset
> + */
>  #include <common.h>
>  #include <init.h>
>  #include <io.h>
> @@ -7,200 +16,215 @@
>  #include <envfs.h>
>  #include <fs.h>
>  
> -static LIST_HEAD(sbu_list);
> +static LIST_HEAD(sba_list);
>  
> -struct sbu {
> -	char *uuid;
> +struct sba {
> +	char *alias;
>  	struct device *dev;
>  	struct cdev *rcdev;
>  	struct cdev cdev;
>  	struct list_head list;
>  };
>  
> -void storage_by_uuid_check_exist(struct cdev *cdev);
> -
> -static ssize_t sbu_read(struct cdev *cdev, void *buf, size_t count, loff_t offset, ulong flags)
> +static ssize_t sba_read(struct cdev *cdev, void *buf, size_t count, loff_t offset, ulong flags)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_read(sbu->rcdev, buf, count, offset, flags);
> +	return cdev_read(sba->rcdev, buf, count, offset, flags);
>  }
>  
> -static ssize_t sbu_write(struct cdev *cdev, const void *buf, size_t count, loff_t offset, ulong flags)
> +static ssize_t sba_write(struct cdev *cdev, const void *buf, size_t count, loff_t offset, ulong flags)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_write(sbu->rcdev, buf, count, offset, flags);
> +	return cdev_write(sba->rcdev, buf, count, offset, flags);
>  }
>  
> -static int sbu_ioctl(struct cdev *cdev, unsigned int request, void *buf)
> +static int sba_ioctl(struct cdev *cdev, unsigned int request, void *buf)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_ioctl(sbu->rcdev, request, buf);
> +	return cdev_ioctl(sba->rcdev, request, buf);
>  }
>  
> -static int sbu_open(struct cdev *cdev, unsigned long flags)
> +static int sba_open(struct cdev *cdev, unsigned long flags)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_open(sbu->rcdev, flags);
> +	return cdev_open(sba->rcdev, flags);
>  }
>  
> -static int sbu_close(struct cdev *cdev)
> +static int sba_close(struct cdev *cdev)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	cdev_close(sbu->rcdev);
> +	cdev_close(sba->rcdev);
>  
>  	return 0;
>  }
>  
> -static int sbu_flush(struct cdev *cdev)
> +static int sba_flush(struct cdev *cdev)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_flush(sbu->rcdev);
> +	return cdev_flush(sba->rcdev);
>  }
>  
> -static int sbu_erase(struct cdev *cdev, loff_t count, loff_t offset)
> +static int sba_erase(struct cdev *cdev, loff_t count, loff_t offset)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_erase(sbu->rcdev, count, offset);
> +	return cdev_erase(sba->rcdev, count, offset);
>  }
>  
> -static int sbu_protect(struct cdev *cdev, size_t count, loff_t offset, int prot)
> +static int sba_protect(struct cdev *cdev, size_t count, loff_t offset, int prot)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_protect(sbu->rcdev, count, offset, prot);
> +	return cdev_protect(sba->rcdev, count, offset, prot);
>  }
>  
> -static int sbu_discard_range(struct cdev *cdev, loff_t count, loff_t offset)
> +static int sba_discard_range(struct cdev *cdev, loff_t count, loff_t offset)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_discard_range(sbu->rcdev, count, offset);
> +	return cdev_discard_range(sba->rcdev, count, offset);
>  }
>  
> -static int sbu_memmap(struct cdev *cdev, void **map, int flags)
> +static int sba_memmap(struct cdev *cdev, void **map, int flags)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_memmap(sbu->rcdev, map, flags);
> +	return cdev_memmap(sba->rcdev, map, flags);
>  }
>  
> -static int sbu_truncate(struct cdev *cdev, size_t size)
> +static int sba_truncate(struct cdev *cdev, size_t size)
>  {
> -	struct sbu *sbu = cdev->priv;
> +	struct sba *sba = cdev->priv;
>  
> -	return cdev_truncate(sbu->rcdev, size);
> +	return cdev_truncate(sba->rcdev, size);
>  }
>  
> -static struct cdev_operations sbu_ops = {
> -	.read = sbu_read,
> -	.write = sbu_write,
> -	.ioctl = sbu_ioctl,
> -	.open = sbu_open,
> -	.close = sbu_close,
> -	.flush = sbu_flush,
> -	.erase = sbu_erase,
> -	.protect = sbu_protect,
> -	.discard_range = sbu_discard_range,
> -	.memmap = sbu_memmap,
> -	.truncate = sbu_truncate,
> +static struct cdev_operations sba_ops = {
> +	.read = sba_read,
> +	.write = sba_write,
> +	.ioctl = sba_ioctl,
> +	.open = sba_open,
> +	.close = sba_close,
> +	.flush = sba_flush,
> +	.erase = sba_erase,
> +	.protect = sba_protect,
> +	.discard_range = sba_discard_range,
> +	.memmap = sba_memmap,
> +	.truncate = sba_truncate,
>  };
>  
> -static void storage_by_uuid_add_partitions(struct sbu *sbu, struct cdev *rcdev)
> +static void sba_add_partitions(struct sba *sba, struct cdev *rcdev)
>  {
>  	int ret;
>  
> -	if (sbu->rcdev)
> +	if (sba->rcdev)
>  		return;
>  
> -	sbu->rcdev = rcdev;
> -	sbu->cdev.name = sbu->uuid;
> -	sbu->cdev.size = rcdev->size;
> -	sbu->cdev.ops = &sbu_ops;
> -	sbu->cdev.dev = sbu->dev;
> -	sbu->cdev.priv = sbu;
> +	sba->rcdev = rcdev;
> +	sba->cdev.name = sba->alias;
> +	sba->cdev.size = rcdev->size;
> +	sba->cdev.ops = &sba_ops;
> +	sba->cdev.dev = sba->dev;
> +	sba->cdev.priv = sba;
>  
>  	if (rcdev->flags & DEVFS_WRITE_AUTOERASE)
> -		sbu->cdev.flags |= DEVFS_WRITE_AUTOERASE;
> +		sba->cdev.flags |= DEVFS_WRITE_AUTOERASE;
>  
> -	ret = devfs_create(&sbu->cdev);
> +	ret = devfs_create(&sba->cdev);
>  	if (ret) {
> -		dev_err(sbu->dev, "Failed to create cdev: %s\n", strerror(-ret));
> +		dev_err(sba->dev, "Failed to create cdev: %s\n", strerror(-ret));
>  		return;
>  	}
>  
> -	of_parse_partitions(&sbu->cdev, sbu->dev->of_node);
> +	of_parse_partitions(&sba->cdev, sba->dev->of_node);
>  }
>  
> -static void check_exist(struct sbu *sbu)
> +static void check_exist(struct sba *sba)
>  {
>  	struct cdev *cdev;
>  
>  	for_each_cdev(cdev) {
>  		if (cdev_is_partition(cdev))
>  			continue;
> -		if (strcmp(cdev->diskuuid, sbu->uuid))
> +		if (strcmp(cdev->diskuuid, sba->alias))
>  			continue;
>  
> -		dev_dbg(sbu->dev, "Found %s %s\n", cdev->name, cdev->diskuuid);
> -		storage_by_uuid_add_partitions(sbu, cdev);
> +		dev_dbg(sba->dev, "Found %s %s\n", cdev->name, cdev->diskuuid);
> +		sba_add_partitions(sba, cdev);
>  		return;
>  	}
>  }
>  
> -static int sbu_detect(struct device *dev)
> +static int sba_detect(struct device *dev)
>  {
> -	struct sbu *sbu = dev->priv;
> +	struct sba *sba = dev->priv;
>  
>  	dev_dbg(dev, "%s\n", __func__);
>  
> -	check_exist(sbu);
> +	check_exist(sba);
>  
>  	return 0;
>  }
>  
> -static int storage_by_uuid_probe(struct device *dev)
> +static int storage_by_uuid_init(struct sba *sba)
>  {
> -	struct sbu *sbu;
> -	int ret;
>  	const char *uuid;
> +	int ret;
>  
> -	sbu = xzalloc(sizeof(*sbu));
> -
> -	ret = of_property_read_string(dev->of_node, "uuid", &uuid);
> +	ret = of_property_read_string(sba->dev->of_node, "uuid", &uuid);
>  	if (ret)
>  		return ret;
>  
> -	sbu->dev = dev;
> -	sbu->uuid = xstrdup(uuid);
> -
> -	list_add_tail(&sbu->list, &sbu_list);
> -
> -	check_exist(sbu);
> -	dev->priv = sbu;
> -	dev->detect = sbu_detect;
> +	sba->alias = xstrdup(uuid);
>  
>  	return 0;
>  }
>  
> -static struct of_device_id storage_by_uuid_dt_ids[] = {
> +static int sba_probe(struct device *dev)
> +{
> +	int (*init)(struct sba *);

This does the job but a proper driver_data struct would be the more
aligned to all other drivers.

Regards,
  Marco


> +	struct sba *sba;
> +	int ret;
> +
> +	sba = xzalloc(sizeof(*sba));
> +	sba->dev = dev;
> +
> +	init = device_get_match_data(dev);
> +	if (!init)
> +		return -ENODEV;
> +
> +	ret = init(sba);
> +	if (ret)
> +		return ret;
> +
> +	list_add_tail(&sba->list, &sba_list);
> +
> +	check_exist(sba);
> +	dev->priv = sba;
> +	dev->detect = sba_detect;
> +
> +	return 0;
> +}
> +
> +static struct of_device_id sba_dt_ids[] = {
>  	{
>  		.compatible = "barebox,storage-by-uuid",
> +		.data = storage_by_uuid_init
>  	}, {
>  		/* sentinel */
>  	}
>  };
> -MODULE_DEVICE_TABLE(of, storage_by_uuid_dt_ids);
> +MODULE_DEVICE_TABLE(of, sba_dt_ids);
>  
> -static struct driver storage_by_uuid_driver = {
> -	.name		= "storage-by-uuid",
> -	.probe		= storage_by_uuid_probe,
> -	.of_compatible	= storage_by_uuid_dt_ids,
> +static struct driver sba_driver = {
> +	.name		= "storage-by-alias",
> +	.probe		= sba_probe,
> +	.of_compatible	= sba_dt_ids,
>  };
> -device_platform_driver(storage_by_uuid_driver);
> +device_platform_driver(sba_driver);
> -- 
> 2.39.5
> 
> 
> 



  reply	other threads:[~2025-04-14 10:43 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-04-14  6:31 [PATCH 00/18] boot: improve cdev alias and blspec support Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 01/18] cdev: make cdev_find_child_by_gpt_typeuuid external Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 02/18] cdev: constify cdev_find_child_by_gpt_typeuuid's guid_t pointer argument Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 03/18] partition: efi: respect DPS_TYPE_FLAG_NO_AUTO flag Ahmad Fatoum
2025-04-14 11:38   ` Sascha Hauer
2025-04-14  6:31 ` [PATCH 04/18] blspec: iterate over all XBOOTLDR partitions Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 05/18] blspec: support GPT XBOOTLDR partition Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 06/18] blspec: collect bootloader spec entries from ESP Ahmad Fatoum
2025-04-14 10:14   ` Marco Felsch
2025-04-14 10:19     ` Ahmad Fatoum
2025-04-14 10:41       ` Marco Felsch
2025-04-14  6:31 ` [PATCH 07/18] cdev: alias: add support for partition links Ahmad Fatoum
2025-04-14 10:17   ` Marco Felsch
2025-04-14 10:20     ` Ahmad Fatoum
2025-04-14 10:49       ` Marco Felsch
2025-04-14 13:28         ` Sascha Hauer
2025-04-14  6:31 ` [PATCH 08/18] cdev: do not dereference partname for non-partitions Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 09/18] cdev: use common cdev_free helper internally Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 10/18] cdev: implement devfs_create_link in terms of devfs_create Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 11/18] misc: storage-by-uuid: consider only first diskuuid match Ahmad Fatoum
2025-04-14 10:24   ` Marco Felsch
2025-04-14 10:31     ` Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 12/18] misc: storage-by-uuid: prepare for generalizing to different aliases Ahmad Fatoum
2025-04-14 10:31   ` Marco Felsch [this message]
2025-04-14 10:35     ` Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 13/18] misc: storage-by-uuid: rename to storage-by-alias Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 14/18] misc: storage-by-alias: switch over to using cdev aliases Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 15/18] misc: storage-by-alias: add support barebox,bootsource Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 16/18] misc: storage-by-alias: register as link to main cdev Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 17/18] cdev: constify cdev_readlink argument Ahmad Fatoum
2025-04-14  6:31 ` [PATCH 18/18] cdev: follow links in cdev_find_child_by_gpt_typeuuid Ahmad Fatoum
2025-04-14 10:51 ` [PATCH 00/18] boot: improve cdev alias and blspec support Marco Felsch
2025-04-14 13:25 ` Sascha Hauer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250414103121.tak6wi5ze7k3qy5g@pengutronix.de \
    --to=m.felsch@pengutronix.de \
    --cc=a.fatoum@pengutronix.de \
    --cc=barebox@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox