netdev - Re: [PATCH net-next v6 2/5] bitops: add non-atomic bitops for pointers

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: <20210702161109.q2cep7hmxyvfelh5@apollo>
Date:   Fri, 2 Jul 2021 21:41:09 +0530
From:   Kumar Kartikeya Dwivedi <memxor@...il.com>
To:     David Laight <David.Laight@...lab.com>
Cc:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Toke Høiland-Jørgensen <toke@...hat.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>,
        Jesper Dangaard Brouer <brouer@...hat.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        John Fastabend <john.fastabend@...il.com>,
        Martin KaFai Lau <kafai@...com>,
        Eric Leblond <eric@...it.org>,
        "bpf@...r.kernel.org" <bpf@...r.kernel.org>
Subject: Re: [PATCH net-next v6 2/5] bitops: add non-atomic bitops for
 pointers

On Fri, Jul 02, 2021 at 08:17:47PM IST, David Laight wrote:
> From: Kumar Kartikeya Dwivedi
> > Sent: 02 July 2021 12:18
> >
> > cpumap needs to set, clear, and test the lowest bit in skb pointer in
> > various places. To make these checks less noisy, add pointer friendly
> > bitop macros that also do some typechecking to sanitize the argument.
>
> Would this work?
> #define BIT_OP(val, op) ((typeof (val))((unsigned long)(val) op))
>
> Should let you do:
> 	ptr = BIT_OP(ptr, | 1);
> 	ptr = BIT_OP(ptr, & ~1);
> 	if (BIT_OPT(ptr, & 1))
> 		...
>
> See https://godbolt.org/z/E57aGK4js
>

This certainly works, but my preference for keeping it this way was reusing the
existing infrastructure (which also has KASAN/KCSAN instrumentation) and avoids
UB while shifting. Ofcourse for this particular case, anything works, but if
putting this in bitops, I thought we should keep it as simple wrappers over
__set_bit/__clear_bit/test_bit.

Also, I compared codegen for both, and it looks pretty much the same to me...

See https://godbolt.org/z/s9cjEnYKj

Let me know if I'm missing something.

> 	David
>
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)

--
Kartikeya