linux-kernel - Re: [RFC PATCH] <linux/taggedptr.h>: Introduce tagged pointer

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]

Date:   Thu, 28 Jun 2018 17:40:36 +0800
From:   Gao Xiang <gaoxiang25@...wei.com>
To:     Matthew Wilcox <willy@...radead.org>
CC:     <viro@...iv.linux.org.uk>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Kate Stewart <kstewart@...uxfoundation.org>,
        Philippe Ombredanne <pombredanne@...b.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Chao Yu <yuchao0@...wei.com>, Miao Xie <miaoxie@...wei.com>,
        <linux-fsdevel@...r.kernel.org>, <linux-erofs@...ts.ozlabs.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH] <linux/taggedptr.h>: Introduce tagged pointer

Hi Matthew,

On 2018/6/28 17:23, Matthew Wilcox wrote:
> On Thu, Jun 28, 2018 at 05:06:29PM +0800, Gao Xiang wrote:
>> Therefore, this patch introduces simple generic methods to fold
>> tags into a pointer integer. It currently reuses the last 2 bits
>> of the pointer for tags, which are safely for all modern platforms.
> 
> The m68k people will have your head.  alignof(unsigned long) == 2 on
> m68k.  Now, kmalloc always returns 8-byte aligned quantities, but
> if you have:
> 
> static unsigned long foo;
> 
> then ((unsigned long)&foo & 2) may be non-zero.
> 

Oh.. I missed it. How about covering dynamic allocation scenario only?
(since 1 bit is too limited to be used for generic purposes... :( )

Or generate different bit mask versions by using macro? --- I could try in the next patch.

Anyway, I think it is depended on specific use cases and corresponding architectures...

>> +/*
>> + * mark these special integers as another type
>> + * in order to highlight the tagged pointer usage.
>> + */
>> +typedef uintptr_t	taggedptr_t;
> 
> I find this a bit verbose.  How about tagptr_t ?
> 

Boths are ok for me :)

Thanks,
Gao Xiang