linux-kernel - Re: [for-linus][PATCH 01/10] tracing: Suppress sparse warnings triggered by is_signed

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:   Tue, 23 Aug 2022 15:05:21 -0700
From:   Bart Van Assche <bvanassche@....org>
To:     Rasmus Villemoes <linux@...musvillemoes.dk>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Steven Rostedt <rostedt@...dmis.org>, linux-kernel@...r.kernel.org,
        Ingo Molnar <mingo@...nel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Christoph Hellwig <hch@....de>,
        Luc Van Oostenryck <luc.vanoostenryck@...il.com>,
        Jens Axboe <axboe@...nel.dk>
Subject: Re: [for-linus][PATCH 01/10] tracing: Suppress sparse warnings
 triggered by is_signed_type()

On 8/23/22 00:06, Rasmus Villemoes wrote:
> On 22/08/2022 20.45, Linus Torvalds wrote:
>> But we did have a sparse fix for it, didn't we? That fix required that
>> the '< (type)1' cast be changed to '<= (type)0' iirc, and a patch to
>> sparse, but it at least avoided the problem.
> 
> Heh. I originally wrote the comparison "< (t)1" instead of "< (t)0" to
> avoid a -Wtype-limits warning when applied to unsigned types - yeah
> yeah, the kernel isn't built with that, but it's a nice macro to
> copy-paste to other projects, and sometimes people do explicitly enable
> -Wtype-limits to manually go through some, and then it's nice to not
> have tons of false positives from this macro.
> 
> But of course <1 is the same as <=0, and we can indeed spell it that way
> without triggering Wtype-limits. So if that can help with also silencing
> sparse, ack from me on that part.

Thank you Rasmus for having shared this information. Since sparse will 
have to be modified anyway, how about extending it such that the bitwise 
attribute can be removed from a type, e.g. via a new no_bitwise 
attribute? Wouldn't that be a more generic solution than only 
suppressing sparse complaints when comparing compile-time constants and 
when the left-hand-side and right-hand-side have different bitwise 
attributes? For reference purposes, this is how __bitwise is defined:

#ifdef __CHECKER__
#define __bitwise	__attribute__((bitwise))
#else
#define __bitwise
#endif

Thanks,

Bart.