linux-kernel - Re: [PATCH v9 2/3] mm: add a field to store names for private anonymous memory

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <d8619a98-2380-ca96-001e-60fe9c6204a6@rasmusvillemoes.dk>
Date:   Fri, 1 Oct 2021 09:01:23 +0200
From:   Rasmus Villemoes <linux@...musvillemoes.dk>
To:     Suren Baghdasaryan <surenb@...gle.com>, akpm@...ux-foundation.org
Cc:     ccross@...gle.com, sumit.semwal@...aro.org, mhocko@...e.com,
        dave.hansen@...el.com, keescook@...omium.org, willy@...radead.org,
        kirill.shutemov@...ux.intel.com, vbabka@...e.cz,
        hannes@...xchg.org, corbet@....net, viro@...iv.linux.org.uk,
        rdunlap@...radead.org, kaleshsingh@...gle.com, peterx@...hat.com,
        rppt@...nel.org, peterz@...radead.org, catalin.marinas@....com,
        vincenzo.frascino@....com, chinwen.chang@...iatek.com,
        axelrasmussen@...gle.com, aarcange@...hat.com, jannh@...gle.com,
        apopple@...dia.com, jhubbard@...dia.com, yuzhao@...gle.com,
        will@...nel.org, fenghua.yu@...el.com, thunder.leizhen@...wei.com,
        hughd@...gle.com, feng.tang@...el.com, jgg@...pe.ca, guro@...com,
        tglx@...utronix.de, krisman@...labora.com, chris.hyser@...cle.com,
        pcc@...gle.com, ebiederm@...ssion.com, axboe@...nel.dk,
        legion@...nel.org, eb@...ix.com, gorcunov@...il.com,
        songmuchun@...edance.com, viresh.kumar@...aro.org,
        thomascedeno@...gle.com, sashal@...nel.org, cxfcosmos@...il.com,
        linux@...musvillemoes.dk, linux-kernel@...r.kernel.org,
        linux-fsdevel@...r.kernel.org, linux-doc@...r.kernel.org,
        linux-mm@...ck.org, kernel-team@...roid.com
Subject: Re: [PATCH v9 2/3] mm: add a field to store names for private
 anonymous memory

On 03/09/2021 01.18, Suren Baghdasaryan wrote:
> From: Colin Cross <ccross@...gle.com>
> 
> 
> changes in v9
> - Changed max anon vma name length from 64 to 256 (as in the original patch)
> because I found one case of the name length being 139 bytes. If anyone is
> curious, here it is:
> dalvik-/data/dalvik-cache/arm64/apex@....android.permission@...v-app@...glePermissionController@...glePermissionController.apk@...sses.art

I'm not sure that's a very convincing argument. We don't add code
arbitrarily just because some userspace code running on some custom
kernel (ab)uses something in that kernel. Surely that user can come up
with a name that doesn't contain GooglePermissionController twice.

The argument for using strings and not just a 128 bit uuid was that it
should (also) be human readable, and 250-byte strings are not that.
Also, there's no natural law forcing this to be some power-of-two, and
in fact the implementation means that it's actually somewhat harmful
(give it a 256 char name, and we'll do a 260 byte alloc, which becomes a
512 byte alloc). So just make the limit 80, the kernel's definition of a
sane line length. As for the allowed chars, it can be relaxed later if
convincing arguments can be made.


> +/* mmap_lock should be read-locked */
> +static inline bool is_same_vma_anon_name(struct vm_area_struct *vma,
> +					 const char *name)
> +{
> +	const char *vma_name = vma_anon_name(vma);
> +
> +	if (likely(!vma_name))
> +		return name == NULL;
> +
> +	return name && !strcmp(name, vma_name);

It's probably preferable to spell this

  /* either both NULL, or pointers to same refcounted string */
  if (vma_name == name)
      return true;

  return name && vma_name && !strcmp(name, vma_name);

so you have one less conditional in the common case.

Rasmus