Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1472519 > unrolled thread

Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping Oleg Nesterov <oleg@redhat.com> - 2016-08-30 16:40 +0200
    Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping Dmitry Safonov <0x7f454c46@gmail.com> - 2016-08-30 17:00 +0200
    Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping Al Viro <viro@ZenIV.linux.org.uk> - 2016-09-03 02:20 +0200

#1472519 — Re: [RFC 1/3] x86/vdso: create vdso file, use it for mapping

On 08/25, Dmitry Safonov wrote:
>
> +static __init struct file *init_vdso_file(const struct vdso_image *vdso_image,
> +					const char *name)
> +{
> +	struct super_block *sb;
> +	struct qstr name_str;
> +	struct inode *inode;
> +	struct path path;
> +	struct file *res;
> +
> +	if (IS_ERR(vdso_mnt))
> +		return ERR_CAST(vdso_mnt);
> +	sb = vdso_mnt->mnt_sb;
> +
> +	name_str.hash = 0;
> +	name_str.len = strlen(name);
> +	name_str.name = name;
> +
> +	res = ERR_PTR(-ENOMEM);
> +	path.mnt = mntget(vdso_mnt);
> +	path.dentry = d_alloc_pseudo(sb, &name_str);
> +	if (!path.dentry)
> +		goto put_path;
> +	d_set_d_op(path.dentry, &vdso_dops);
> +
> +	res = ERR_PTR(-ENOSPC);
> +	inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);
> +	if (!inode)
> +		goto put_path;
> +
> +	inode->i_flags |= S_PRIVATE;
> +	d_instantiate(path.dentry, inode);
> +	inode->i_size = vdso_image->size;
> +
> +	res = ERR_PTR(add_vdso_pages_to_page_cache(vdso_image, inode));
> +	if (IS_ERR(res))
> +		goto put_path;
> +
> +	res = alloc_file(&path, FMODE_READ, &ramfs_file_operations);
> +	if (!IS_ERR(res))
> +		return res;
> +
> +put_path:
> +	path_put(&path);
> +	return res;
> +}

Not sure... I think alloc_anon_inode() makes more sense.

Perhaps you can look at fs/aio.c and extract some bits of code into
the new helpers which could be used by both aio_fs and vdso_fs?

Oleg.

[toc] | [next] | [standalone]

#1472532

2016-08-30 17:33 GMT+03:00 Oleg Nesterov <oleg@redhat.com>:
> On 08/25, Dmitry Safonov wrote:
>>
>> +static __init struct file *init_vdso_file(const struct vdso_image *vdso_image,
>> +                                     const char *name)
>> +{
>> +     struct super_block *sb;
>> +     struct qstr name_str;
>> +     struct inode *inode;
>> +     struct path path;
>> +     struct file *res;
>> +
>> +     if (IS_ERR(vdso_mnt))
>> +             return ERR_CAST(vdso_mnt);
>> +     sb = vdso_mnt->mnt_sb;
>> +
>> +     name_str.hash = 0;
>> +     name_str.len = strlen(name);
>> +     name_str.name = name;
>> +
>> +     res = ERR_PTR(-ENOMEM);
>> +     path.mnt = mntget(vdso_mnt);
>> +     path.dentry = d_alloc_pseudo(sb, &name_str);
>> +     if (!path.dentry)
>> +             goto put_path;
>> +     d_set_d_op(path.dentry, &vdso_dops);
>> +
>> +     res = ERR_PTR(-ENOSPC);
>> +     inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);
>> +     if (!inode)
>> +             goto put_path;
>> +
>> +     inode->i_flags |= S_PRIVATE;
>> +     d_instantiate(path.dentry, inode);
>> +     inode->i_size = vdso_image->size;
>> +
>> +     res = ERR_PTR(add_vdso_pages_to_page_cache(vdso_image, inode));
>> +     if (IS_ERR(res))
>> +             goto put_path;
>> +
>> +     res = alloc_file(&path, FMODE_READ, &ramfs_file_operations);
>> +     if (!IS_ERR(res))
>> +             return res;
>> +
>> +put_path:
>> +     path_put(&path);
>> +     return res;
>> +}
>
> Not sure... I think alloc_anon_inode() makes more sense.
>
> Perhaps you can look at fs/aio.c and extract some bits of code into
> the new helpers which could be used by both aio_fs and vdso_fs?

Thanks, will do for the next version.

-- 
             Dmitry

[toc] | [prev] | [next] | [standalone]

#1475508

On Tue, Aug 30, 2016 at 04:33:12PM +0200, Oleg Nesterov wrote:
> > +	inode = ramfs_get_inode(sb, NULL, S_IFREG | S_IRUGO | S_IXUGO, 0);

> Not sure... I think alloc_anon_inode() makes more sense.

	Compared to this, you mean?  It's going to give you the wrong
permissions/i_op/a_ops, and you'll have to assign them manually.  ramfs
already gives the right ones for what's wanted, AFAICS.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web