All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eric Biggers <ebiggers3@gmail.com>
To: keyrings@vger.kernel.org
Subject: Re: [PATCH] keyctl: use keyctl_read_alloc() in dump_key_tree_aux()
Date: Tue, 20 Feb 2018 19:43:49 +0000	[thread overview]
Message-ID: <20180220194349.GD70958@gmail.com> (raw)
In-Reply-To: <20171026210008.106248-1-ebiggers3@gmail.com>

On Thu, Oct 26, 2017 at 02:00:08PM -0700, Eric Biggers wrote:
> From: Eric Biggers <ebiggers@google.com>
> 
> dump_key_tree_aux() (part of 'keyctl show') was racy: it allocated a
> buffer for the keyring contents, then read the keyring.  But it's
> possible that keys are added to the keyring concurrently.  This is
> problematic for two reasons.  First, when keyctl_read() is passed a
> buffer that is too small, it is unspecified whether it is filled or not.
> Second, even if the buffer is filled, some keys (not necessarily even
> the newest ones) would be omitted from the listing.
> 
> Switch to keyctl_read_alloc() which handles the "buffer too small" case
> correctly by retrying the read.
> 
> Signed-off-by: Eric Biggers <ebiggers@google.com>
> ---
>  keyctl.c | 22 ++++++----------------
>  1 file changed, 6 insertions(+), 16 deletions(-)
> 
> diff --git a/keyctl.c b/keyctl.c
> index 801a864..65a7397 100644
> --- a/keyctl.c
> +++ b/keyctl.c
> @@ -1808,29 +1808,18 @@ static int dump_key_tree_aux(key_serial_t key, int depth, int more, int hex_key_
>  	/* if it's a keyring then we're going to want to recursively
>  	 * display it if we can */
>  	if (strcmp(type, "keyring") = 0) {
> -		/* find out how big the keyring is */
> -		ret = keyctl_read(key, NULL, 0);
> -		if (ret < 0)
> -			error("keyctl_read");
> -		if (ret = 0)
> -			return 0;
> -		ringlen = ret;
>  
>  		/* read its contents */
> -		payload = malloc(ringlen);
> -		if (!payload)
> -			error("malloc");
> -
> -		ret = keyctl_read(key, payload, ringlen);
> +		ret = keyctl_read_alloc(key, &payload);
>  		if (ret < 0)
> -			error("keyctl_read");
> +			error("keyctl_read_alloc");
>  
> -		ringlen = ret < ringlen ? ret : ringlen;
> +		ringlen = ret;
>  		kcount = ringlen / sizeof(key_serial_t);
>  
>  		/* walk the keyring */
>  		pk = payload;
> -		do {
> +		while (ringlen >= sizeof(key_serial_t)) {
>  			key = *pk++;
>  
>  			/* recurse into nexted keyrings */
> @@ -1858,7 +1847,8 @@ static int dump_key_tree_aux(key_serial_t key, int depth, int more, int hex_key_
>  							    hex_key_IDs);
>  			}
>  
> -		} while (ringlen -= 4, ringlen >= sizeof(key_serial_t));
> +			ringlen -= sizeof(key_serial_t);
> +		}
>  
>  		free(payload);
>  	}
> -- 
> 2.15.0.rc2.357.g7e34df9404-goog
> 

Ping.

  parent reply	other threads:[~2018-02-20 19:43 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-26 21:00 [PATCH] keyctl: use keyctl_read_alloc() in dump_key_tree_aux() Eric Biggers
2017-10-27  8:06 ` James Morris
2018-02-20 19:43 ` Eric Biggers [this message]
2018-05-08 18:09 ` Eric Biggers

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=20180220194349.GD70958@gmail.com \
    --to=ebiggers3@gmail.com \
    --cc=keyrings@vger.kernel.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 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.