All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: Christoph Hellwig <hch@lst.de>
Cc: Jens Axboe <axboe@kernel.dk>,
	"Rafael J. Wysocki" <rafael@kernel.org>,
	Denis Efremov <efremov@linux.com>,
	"David S. Miller" <davem@davemloft.net>,
	Song Liu <song@kernel.org>, Al Viro <viro@zeniv.linux.org.uk>,
	linux-block@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-ide@vger.kernel.org, linux-raid@vger.kernel.org,
	linux-scsi@vger.kernel.org, linux-m68k@lists.linux-m68k.org
Subject: Re: [PATCH 01/19] char_dev: replace cdev_map with an xarray
Date: Wed, 26 Aug 2020 10:19:05 +0200	[thread overview]
Message-ID: <20200826081905.GB1796103@kroah.com> (raw)
In-Reply-To: <20200826062446.31860-2-hch@lst.de>

On Wed, Aug 26, 2020 at 08:24:28AM +0200, Christoph Hellwig wrote:
> None of the complicated overlapping regions bits of the kobj_map are
> required for the character device lookup, so just a trivial xarray
> instead.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Really?  This is ok to use and just as fast?  If so, wonderful, it would
be great to clean up kobj_map users.

But I don't think you got all of the needed bits here:

> ---
>  fs/char_dev.c | 94 +++++++++++++++++++++++++--------------------------
>  fs/dcache.c   |  1 -
>  fs/internal.h |  5 ---
>  3 files changed, 46 insertions(+), 54 deletions(-)
> 
> diff --git a/fs/char_dev.c b/fs/char_dev.c
> index ba0ded7842a779..6c4d6c4938f14b 100644
> --- a/fs/char_dev.c
> +++ b/fs/char_dev.c
> @@ -17,7 +17,6 @@
>  #include <linux/seq_file.h>
>  
>  #include <linux/kobject.h>
> -#include <linux/kobj_map.h>
>  #include <linux/cdev.h>
>  #include <linux/mutex.h>
>  #include <linux/backing-dev.h>
> @@ -25,8 +24,7 @@
>  
>  #include "internal.h"
>  
> -static struct kobj_map *cdev_map;
> -
> +static DEFINE_XARRAY(cdev_map);
>  static DEFINE_MUTEX(chrdevs_lock);
>  
>  #define CHRDEV_MAJOR_HASH_SIZE 255
> @@ -367,6 +365,29 @@ void cdev_put(struct cdev *p)
>  	}
>  }
>  
> +static struct cdev *cdev_lookup(dev_t dev)
> +{
> +	struct cdev *cdev;
> +
> +retry:
> +	mutex_lock(&chrdevs_lock);
> +	cdev = xa_load(&cdev_map, dev);
> +	if (!cdev) {
> +		mutex_unlock(&chrdevs_lock);
> +
> +		if (request_module("char-major-%d-%d",
> +				   MAJOR(dev), MINOR(dev)) > 0)
> +			/* Make old-style 2.4 aliases work */
> +			request_module("char-major-%d", MAJOR(dev));
> +		goto retry;
> +	}
> +
> +	if (!cdev_get(cdev))
> +		cdev = NULL;
> +	mutex_unlock(&chrdevs_lock);
> +	return cdev;
> +}
> +
>  /*
>   * Called every time a character special file is opened
>   */
> @@ -380,13 +401,10 @@ static int chrdev_open(struct inode *inode, struct file *filp)
>  	spin_lock(&cdev_lock);
>  	p = inode->i_cdev;
>  	if (!p) {
> -		struct kobject *kobj;
> -		int idx;
>  		spin_unlock(&cdev_lock);
> -		kobj = kobj_lookup(cdev_map, inode->i_rdev, &idx);
> -		if (!kobj)
> +		new = cdev_lookup(inode->i_rdev);
> +		if (!new)
>  			return -ENXIO;
> -		new = container_of(kobj, struct cdev, kobj);
>  		spin_lock(&cdev_lock);
>  		/* Check i_cdev again in case somebody beat us to it while
>  		   we dropped the lock. */
> @@ -454,18 +472,6 @@ const struct file_operations def_chr_fops = {
>  	.llseek = noop_llseek,
>  };
>  
> -static struct kobject *exact_match(dev_t dev, int *part, void *data)
> -{
> -	struct cdev *p = data;
> -	return &p->kobj;
> -}
> -
> -static int exact_lock(dev_t dev, void *data)
> -{
> -	struct cdev *p = data;
> -	return cdev_get(p) ? 0 : -1;
> -}
> -
>  /**
>   * cdev_add() - add a char device to the system
>   * @p: the cdev structure for the device
> @@ -478,7 +484,7 @@ static int exact_lock(dev_t dev, void *data)
>   */
>  int cdev_add(struct cdev *p, dev_t dev, unsigned count)
>  {
> -	int error;
> +	int error, i;
>  
>  	p->dev = dev;
>  	p->count = count;
> @@ -486,14 +492,22 @@ int cdev_add(struct cdev *p, dev_t dev, unsigned count)
>  	if (WARN_ON(dev == WHITEOUT_DEV))
>  		return -EBUSY;
>  
> -	error = kobj_map(cdev_map, dev, count, NULL,
> -			 exact_match, exact_lock, p);
> -	if (error)
> -		return error;
> +	mutex_lock(&chrdevs_lock);
> +	for (i = 0; i < count; i++) {
> +		error = xa_insert(&cdev_map, dev + i, p, GFP_KERNEL);
> +		if (error)
> +			goto out_unwind;
> +	}
> +	mutex_unlock(&chrdevs_lock);
>  
>  	kobject_get(p->kobj.parent);

Can't you drop this kobject_get() too?

And also the "struct kobj" in struct cdev can be gone as well, as the
kobj_map was the only "real" user of this structure.  I know some
drivers liked to touch that field as well, but it never actually did
anything for them, so it was pointless, but it will take some 'make
allmodconfig' builds to flush them out.

thanks,

greg k-h

  reply	other threads:[~2020-08-26  8:18 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-26  6:24 simplify gendisk lookup and remove struct block_device aliases Christoph Hellwig
2020-08-26  6:24 ` [PATCH 01/19] char_dev: replace cdev_map with an xarray Christoph Hellwig
2020-08-26  8:19   ` Greg Kroah-Hartman [this message]
2020-08-27  8:53     ` Christoph Hellwig
2020-08-27  9:18       ` Greg Kroah-Hartman
2020-08-27  9:39         ` Christoph Hellwig
2020-08-27  7:25   ` Hannes Reinecke
2020-08-27  8:55     ` Christoph Hellwig
2020-08-26  6:24 ` [PATCH 02/19] block: merge drivers/base/map.c into block/genhd.c Christoph Hellwig
2020-08-26  8:19   ` Greg Kroah-Hartman
2020-08-26  6:24 ` [PATCH 03/19] block: cleanup del_gendisk a bit Christoph Hellwig
2020-08-26  6:24 ` [PATCH 04/19] block: split block_class_lock Christoph Hellwig
2020-08-26  6:24 ` [PATCH 05/19] block: rework requesting modules for unclaimed devices Christoph Hellwig
2020-08-26  6:24 ` [PATCH 06/19] block: add an optional probe callback to major_names Christoph Hellwig
2020-08-26  6:24 ` [PATCH 07/19] ide: remove ide_{,un}register_region Christoph Hellwig
2020-08-27  7:25   ` Hannes Reinecke
2020-08-26  6:24 ` [PATCH 08/19] swim: don't call blk_register_region Christoph Hellwig
2020-08-26  6:24 ` [PATCH 09/19] sd: use __register_blkdev to avoid a modprobe for an unregistered dev_t Christoph Hellwig
2020-08-26  6:24 ` [PATCH 10/19] brd: use __register_blkdev to allocate devices on demand Christoph Hellwig
2020-08-26  6:24 ` [PATCH 11/19] loop: " Christoph Hellwig
2020-08-26  6:24 ` [PATCH 12/19] md: " Christoph Hellwig
2020-08-28  7:08   ` Song Liu
2020-08-26  6:24 ` [PATCH 13/19] ide: switch to __register_blkdev for command set probing Christoph Hellwig
2020-08-26  6:24 ` [PATCH 14/19] floppy: use a separate gendisk for each media format Christoph Hellwig
2020-08-26  6:24 ` [PATCH 15/19] amiflop: use separate gendisks for Amiga vs MS-DOS mode Christoph Hellwig
2020-08-26  6:24 ` [PATCH 16/19] ataflop: use a separate gendisk for each media format Christoph Hellwig
2020-08-26  6:24 ` [PATCH 17/19] z2ram: reindent Christoph Hellwig
2020-08-26  9:12   ` Joe Perches
2020-08-26  9:49     ` John Paul Adrian Glaubitz
2020-08-26 14:51       ` Joe Perches
2020-08-27 23:21         ` Finn Thain
2020-08-28  0:01           ` Joe Perches
2020-08-28  0:57             ` Finn Thain
2020-08-28  1:11               ` Joe Perches
2020-08-28  4:25               ` Christoph Hellwig
2020-08-26  6:24 ` [PATCH 18/19] z2ram: use separate gendisk for the different modes Christoph Hellwig
2020-08-26  6:24 ` [PATCH 19/19] block: switch gendisk lookup to a simple xarray Christoph Hellwig
2020-08-30  6:24 simplify gendisk lookup and remove struct block_device aliases v2 Christoph Hellwig
2020-08-30  6:24 ` [PATCH 01/19] char_dev: replace cdev_map with an xarray Christoph Hellwig
2020-08-30  7:31   ` Greg Kroah-Hartman
2020-09-03  8:01 simplify gendisk lookup and remove struct block_device aliases v3 Christoph Hellwig
2020-09-03  8:01 ` [PATCH 01/19] char_dev: replace cdev_map with an xarray Christoph Hellwig
2020-09-04 10:16   ` Hannes Reinecke

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=20200826081905.GB1796103@kroah.com \
    --to=gregkh@linuxfoundation.org \
    --cc=axboe@kernel.dk \
    --cc=davem@davemloft.net \
    --cc=efremov@linux.com \
    --cc=hch@lst.de \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-ide@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-m68k@lists.linux-m68k.org \
    --cc=linux-raid@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=rafael@kernel.org \
    --cc=song@kernel.org \
    --cc=viro@zeniv.linux.org.uk \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.