mailing list of musl libc
 help / color / mirror / code / Atom feed
From: Rich Felker <dalias@libc.org>
To: James Raphael Tiovalen <jamestiotio@gmail.com>
Cc: musl@lists.openwall.com
Subject: Re: [musl] [PATCH] Add a safe dequeue integrity check for mallocng
Date: Fri, 8 Sep 2023 20:48:35 -0400	[thread overview]
Message-ID: <20230909004835.GD4163@brightrain.aerifal.cx> (raw)
In-Reply-To: <20230908174939.80579-1-jamestiotio@gmail.com>

On Sat, Sep 09, 2023 at 01:49:39AM +0800, James Raphael Tiovalen wrote:
> This commit adds an integrity check to allow for safer dequeuing of meta
> structs in mallocng. If the unlikely condition is true due to some sort
> of heap corruption, we print an error message and abort.
> 
> This approach is similar to the safe unlinking check performed by glibc.
> 
> While this check would not prevent more sophisticated attacks in more
> specific scenarios, as shown by the historical exploitation efforts on
> glibc, this check would prevent more basic heap corruption attacks from
> being successfully executed.
> ---
>  src/malloc/mallocng/meta.h | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/src/malloc/mallocng/meta.h b/src/malloc/mallocng/meta.h
> index 61ec53f9..57946d01 100644
> --- a/src/malloc/mallocng/meta.h
> +++ b/src/malloc/mallocng/meta.h
> @@ -2,9 +2,11 @@
>  #define MALLOC_META_H
>  
>  #include <stdint.h>
> +#include <stdio.h>
>  #include <errno.h>
>  #include <limits.h>
>  #include "glue.h"
> +#include "libm.h"
>  
>  __attribute__((__visibility__("hidden")))
>  extern const uint16_t size_classes[];
> @@ -90,6 +92,10 @@ static inline void queue(struct meta **phead, struct meta *m)
>  static inline void dequeue(struct meta **phead, struct meta *m)
>  {
>  	if (m->next != m) {
> +		if (predict_false(m->prev->next != m || m->next->prev != m)) {
> +			fprintf(stderr, "Corrupted doubly-linked meta list\n");
> +			abort();
> +		}
>  		m->prev->next = m->next;
>  		m->next->prev = m->prev;
>  		if (*phead == m) *phead = m->next;
> -- 
> 2.42.0

This could and should be written with the assert macro, like all the
other safety assertions in mallocng, not pulling in stdio and abort.
But I think you're over-estimating the value of the check here. The
pointers in question are not part of "the heap" but are out-of-band,
intended not to be reachable except by an attacker who already has
arbitrary code execution or at least strong gadgets for modifying
memory they shouldn't with multiple levels of offsetting and
indirection, which could generally be used in lots of other ways to
obtain arbitrary code execution.

Rich

  reply	other threads:[~2023-09-09  0:48 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-08 17:49 James Raphael Tiovalen
2023-09-09  0:48 ` Rich Felker [this message]
2023-09-14  5:13   ` James R T
2023-09-14  9:23     ` Joakim Sindholt
2023-09-14 12:18       ` Rich Felker
2023-09-16  6:53         ` James R T

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=20230909004835.GD4163@brightrain.aerifal.cx \
    --to=dalias@libc.org \
    --cc=jamestiotio@gmail.com \
    --cc=musl@lists.openwall.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.
Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/musl/

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).