linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Hugh Dickins <hughd@google.com>
To: Oleksandr Tymoshenko <ovt@google.com>
Cc: Jonathan Corbet <corbet@lwn.net>, Hugh Dickins <hughd@google.com>,
	 Andrew Morton <akpm@linux-foundation.org>,
	 Christian Brauner <brauner@kernel.org>,
	linux-doc@vger.kernel.org,  linux-kernel@vger.kernel.org,
	linux-mm@kvack.org
Subject: Re: [PATCH] shmem: add support for user extended attributes
Date: Thu, 20 Jul 2023 09:56:51 -0700 (PDT)	[thread overview]
Message-ID: <9b8d38f0-fd22-3f98-d070-16baf976ecb5@google.com> (raw)
In-Reply-To: <20230720065430.2178136-1-ovt@google.com>

On Thu, 20 Jul 2023, Oleksandr Tymoshenko wrote:

> User extended attributes are not enabled in tmpfs because
> the size of the value is not limited and the memory allocated
> for it is not counted against any limit. Malicious
> non-privileged user can exhaust kernel memory by creating
> user.* extended attribute with very large value.
> 
> There are still situations when enabling suport for extended
> user attributes on tmpfs is required and the attack vector
> is not applicable, for instance batch jobs with trusted binaries.
> 
> This patch introduces two mount options to enable/disable
> support for user.* extended attributes on tmpfs:
> 
> user_xattr    enable support for user extended aatributes
> nouser_xattr  disable support for user extended attributes
> 
> The default behavior of the filesystem is not changed.
> 
> Signed-off-by: Oleksandr Tymoshenko <ovt@google.com>

Thanks, but no.

This is not something we want mount options for:
we just want to limit the memory usage of tmpfs user xattrs.

I've had the patch to do that limiting (taking it out of the inode
space already limited by nr_inodes) in my test tree for 2.5 years now:
waiting to reach the top of the heap to pull together and submit.

Your sending this patch does help to raise the priority for my
sending that patch: thank you; but I cannot promise when that will be.

(And the way mm/shmem.c is conflicted between vfs and mm trees
is rather discouraging development there at the moment: I'm hoping
it can be mostly wrested back into the mm tree in the next cycle.)

Hugh

> ---
>  Documentation/filesystems/tmpfs.rst | 12 ++++++++
>  include/linux/shmem_fs.h            |  1 +
>  mm/shmem.c                          | 45 +++++++++++++++++++++++++++++
>  3 files changed, 58 insertions(+)
> 
> diff --git a/Documentation/filesystems/tmpfs.rst b/Documentation/filesystems/tmpfs.rst
> index f18f46be5c0c..5700ba72d095 100644
> --- a/Documentation/filesystems/tmpfs.rst
> +++ b/Documentation/filesystems/tmpfs.rst
> @@ -215,6 +215,16 @@ will give you tmpfs instance on /mytmpfs which can allocate 10GB
>  RAM/SWAP in 10240 inodes and it is only accessible by root.
>  
>  
> +tmpfs, when compiled with CONFIG_TMPFS_XATTR, does not support
> +Extended User Attributes for security reasons. The support can be
> +enabled/disabled by two mount options:
> +
> +============  ===========================================
> +user_xattr    Enable support for Extended User Attributes
> +nouser_xattr  Disable upport for Extended User Attributes
> +============  ===========================================
> +
> +
>  :Author:
>     Christoph Rohland <cr@sap.com>, 1.12.01
>  :Updated:
> @@ -223,3 +233,5 @@ RAM/SWAP in 10240 inodes and it is only accessible by root.
>     KOSAKI Motohiro, 16 Mar 2010
>  :Updated:
>     Chris Down, 13 July 2020
> +:Updated:
> +   Oleksandr Tymoshenko, 19 July 2023
> diff --git a/include/linux/shmem_fs.h b/include/linux/shmem_fs.h
> index 9029abd29b1c..f06d18b9041c 100644
> --- a/include/linux/shmem_fs.h
> +++ b/include/linux/shmem_fs.h
> @@ -53,6 +53,7 @@ struct shmem_sb_info {
>  	spinlock_t shrinklist_lock;   /* Protects shrinklist */
>  	struct list_head shrinklist;  /* List of shinkable inodes */
>  	unsigned long shrinklist_len; /* Length of shrinklist */
> +	bool user_xattr;	      /* user.* xattrs are allowed */
>  };
>  
>  static inline struct shmem_inode_info *SHMEM_I(struct inode *inode)
> diff --git a/mm/shmem.c b/mm/shmem.c
> index 2f2e0e618072..4f7d46d65494 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -85,6 +85,7 @@ static struct vfsmount *shm_mnt;
>  
>  #define BLOCKS_PER_PAGE  (PAGE_SIZE/512)
>  #define VM_ACCT(size)    (PAGE_ALIGN(size) >> PAGE_SHIFT)
> +#define TMPFS_USER_XATTR_INDEX 1
>  
>  /* Pretend that each entry is of this size in directory's i_size */
>  #define BOGO_DIRENT_SIZE 20
> @@ -116,11 +117,13 @@ struct shmem_options {
>  	int huge;
>  	int seen;
>  	bool noswap;
> +	bool user_xattr;
>  #define SHMEM_SEEN_BLOCKS 1
>  #define SHMEM_SEEN_INODES 2
>  #define SHMEM_SEEN_HUGE 4
>  #define SHMEM_SEEN_INUMS 8
>  #define SHMEM_SEEN_NOSWAP 16
> +#define SHMEM_SEEN_USER_XATTR 32
>  };
>  
>  #ifdef CONFIG_TMPFS
> @@ -3447,6 +3450,16 @@ static int shmem_xattr_handler_get(const struct xattr_handler *handler,
>  				   const char *name, void *buffer, size_t size)
>  {
>  	struct shmem_inode_info *info = SHMEM_I(inode);
> +	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
> +
> +	switch (handler->flags) {
> +	case TMPFS_USER_XATTR_INDEX:
> +		if (!sbinfo->user_xattr)
> +			return -EOPNOTSUPP;
> +		break;
> +	default:
> +		break;
> +	}
>  
>  	name = xattr_full_name(handler, name);
>  	return simple_xattr_get(&info->xattrs, name, buffer, size);
> @@ -3459,8 +3472,18 @@ static int shmem_xattr_handler_set(const struct xattr_handler *handler,
>  				   size_t size, int flags)
>  {
>  	struct shmem_inode_info *info = SHMEM_I(inode);
> +	struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
>  	int err;
>  
> +	switch (handler->flags) {
> +	case TMPFS_USER_XATTR_INDEX:
> +		if (!sbinfo->user_xattr)
> +			return -EOPNOTSUPP;
> +		break;
> +	default:
> +		break;
> +	}
> +
>  	name = xattr_full_name(handler, name);
>  	err = simple_xattr_set(&info->xattrs, name, value, size, flags, NULL);
>  	if (!err) {
> @@ -3482,9 +3505,17 @@ static const struct xattr_handler shmem_trusted_xattr_handler = {
>  	.set = shmem_xattr_handler_set,
>  };
>  
> +static const struct xattr_handler shmem_user_xattr_handler = {
> +	.prefix = XATTR_USER_PREFIX,
> +	.flags = TMPFS_USER_XATTR_INDEX,
> +	.get = shmem_xattr_handler_get,
> +	.set = shmem_xattr_handler_set,
> +};
> +
>  static const struct xattr_handler *shmem_xattr_handlers[] = {
>  	&shmem_security_xattr_handler,
>  	&shmem_trusted_xattr_handler,
> +	&shmem_user_xattr_handler,
>  	NULL
>  };
>  
> @@ -3604,6 +3635,8 @@ enum shmem_param {
>  	Opt_inode32,
>  	Opt_inode64,
>  	Opt_noswap,
> +	Opt_user_xattr,
> +	Opt_nouser_xattr,
>  };
>  
>  static const struct constant_table shmem_param_enums_huge[] = {
> @@ -3626,6 +3659,8 @@ const struct fs_parameter_spec shmem_fs_parameters[] = {
>  	fsparam_flag  ("inode32",	Opt_inode32),
>  	fsparam_flag  ("inode64",	Opt_inode64),
>  	fsparam_flag  ("noswap",	Opt_noswap),
> +	fsparam_flag  ("user_xattr",	Opt_user_xattr),
> +	fsparam_flag  ("nouser_xattr",	Opt_nouser_xattr),
>  	{}
>  };
>  
> @@ -3717,6 +3752,14 @@ static int shmem_parse_one(struct fs_context *fc, struct fs_parameter *param)
>  		ctx->noswap = true;
>  		ctx->seen |= SHMEM_SEEN_NOSWAP;
>  		break;
> +	case Opt_user_xattr:
> +		ctx->user_xattr = true;
> +		ctx->seen |= SHMEM_SEEN_USER_XATTR;
> +		break;
> +	case Opt_nouser_xattr:
> +		ctx->user_xattr = false;
> +		ctx->seen |= SHMEM_SEEN_USER_XATTR;
> +		break;
>  	}
>  	return 0;
>  
> @@ -3834,6 +3877,8 @@ static int shmem_reconfigure(struct fs_context *fc)
>  		sbinfo->max_inodes  = ctx->inodes;
>  		sbinfo->free_inodes = ctx->inodes - inodes;
>  	}
> +	if (ctx->seen & SHMEM_SEEN_USER_XATTR)
> +		sbinfo->user_xattr = ctx->user_xattr;
>  
>  	/*
>  	 * Preserve previous mempolicy unless mpol remount option was specified.
> -- 
> 2.41.0.255.g8b1d071c50-goog
> 
> 


  reply	other threads:[~2023-07-20 16:57 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-20  6:54 [PATCH] shmem: add support for user extended attributes Oleksandr Tymoshenko
2023-07-20 16:56 ` Hugh Dickins [this message]
2023-07-20 17:09   ` Oleksandr Tymoshenko
2023-07-20 17:42     ` Hugh Dickins
2023-08-14  8:23   ` Snaipe
2023-08-15  3:51     ` Hugh Dickins
2023-08-15  7:46       ` Franklin “Snaipe” Mathieu
2023-08-15  8:23         ` Christian Brauner
2023-08-21 17:52           ` Hugh Dickins
2023-08-22  9:01             ` Christian Brauner
2023-08-22 17:50             ` Oleksandr Tymoshenko

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=9b8d38f0-fd22-3f98-d070-16baf976ecb5@google.com \
    --to=hughd@google.com \
    --cc=akpm@linux-foundation.org \
    --cc=brauner@kernel.org \
    --cc=corbet@lwn.net \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=ovt@google.com \
    /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;
as well as URLs for NNTP newsgroup(s).