All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrey Ryabinin <aryabinin@virtuozzo.com>
To: Wengang Wang <wen.gang.wang@oracle.com>, linux-mm@kvack.org
Cc: glider@google.com, dvyukov@google.com
Subject: Re: [PATCH 0/5] mm/kasan: advanced check
Date: Wed, 22 Nov 2017 15:04:26 +0300	[thread overview]
Message-ID: <08db0958-220a-f31e-0ddb-273d7126150e@virtuozzo.com> (raw)
In-Reply-To: <20171117223043.7277-1-wen.gang.wang@oracle.com>

On 11/18/2017 01:30 AM, Wengang Wang wrote:
> Kasan advanced check, I'm going to add this feature.
> Currently Kasan provide the detection of use-after-free and out-of-bounds
> problems. It is not able to find the overwrite-on-allocated-memory issue.
> We sometimes hit this kind of issue: We have a messed up structure
> (usually dynamially allocated), some of the fields in the structure were
> overwritten with unreasaonable values. And kernel may panic due to those
> overeritten values. We know those fields were overwritten somehow, but we
> have no easy way to find out which path did the overwritten. The advanced
> check wants to help in this scenario.
> 
> The idea is to define the memory owner. When write accesses come from
> non-owner, error should be reported. Normally the write accesses on a given
> structure happen in only several or a dozen of functions if the structure
> is not that complicated. We call those functions "allowed functions".
> The work of defining the owner and binding memory to owner is expected to
> be done by the memory consumer. In the above case, memory consume register
> the owner as the functions which have write accesses to the structure then
> bind all the structures to the owner. Then kasan will do the "owner check"
> after the basic checks.
> 
> As implementation, kasan provides a API to it's user to register their
> allowed functions. The API returns a token to users.  At run time, users
> bind the memory ranges they are interested in to the check they registered.
> Kasan then checks the bound memory ranges with the allowed functions.
> 

NAK. We don't add APIs with no users in the kernel.
If nothing in the kernel uses this API than there is no way to tell if this works or not.

Besides, I'm bit skeptical about usefulness of this feature. Those kinds of issues that
advanced check is supposed to catch, is almost always is just some sort of longstanding
use after free, which eventually should be caught by kasan.



--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2017-11-22 12:01 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-17 22:30 [PATCH 0/5] mm/kasan: advanced check Wengang Wang
2017-11-17 22:30 ` [PATCH 1/5] mm/kasan: make space in shadow bytes for " Wengang Wang
2017-11-17 22:30 ` [PATCH 2/5] mm/kasan: pass access mode to poison check functions Wengang Wang
2017-11-17 22:30 ` [PATCH 3/5] mm/kasan: do advanced check Wengang Wang
2017-11-17 22:30 ` [PATCH 4/5] mm/kasan: register check and bind it to memory Wengang Wang
2017-11-17 22:30 ` [PATCH 5/5] mm/kasan: add advanced check test case Wengang Wang
2017-11-17 22:32 ` [PATCH 0/5] mm/kasan: advanced check Wengang Wang
2017-11-17 22:56 ` Dmitry Vyukov
2017-11-20  1:50   ` Joonsoo Kim
2017-11-20  8:41     ` Dmitry Vyukov
2017-11-20 20:05       ` Wengang
2017-11-20 20:20         ` Dmitry Vyukov
2017-11-20 20:29           ` Wengang
2017-11-21  9:54             ` Dmitry Vyukov
2017-11-21 19:17               ` Wengang Wang
2017-11-22  8:48                 ` Dmitry Vyukov
2017-11-22 21:09                   ` Wengang Wang
2017-11-20 19:56     ` Wengang
2017-11-22  4:30       ` Joonsoo Kim
2017-11-22  8:51         ` Dmitry Vyukov
2017-11-23  6:07           ` Joonsoo Kim
2017-11-22 19:43         ` Wengang Wang
2017-11-23  6:23           ` Joonsoo Kim
2017-11-23  6:35             ` Joonsoo Kim
2017-11-22 12:04     ` Andrey Ryabinin
2017-11-23  5:57       ` Joonsoo Kim
2017-11-22 12:04 ` Andrey Ryabinin [this message]
2017-11-22 19:29   ` Wengang Wang
2017-11-26 19:37     ` Wengang Wang

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=08db0958-220a-f31e-0ddb-273d7126150e@virtuozzo.com \
    --to=aryabinin@virtuozzo.com \
    --cc=dvyukov@google.com \
    --cc=glider@google.com \
    --cc=linux-mm@kvack.org \
    --cc=wen.gang.wang@oracle.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 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.