linux-kernel - Re: Unsigned widening casts of binary "not" operations..

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: <CA+55aFwogFL=nm=YHwARcz9OTRBdWEDPG0xcqM-H3k-VZTOZPQ@mail.gmail.com>
Date:	Tue, 23 Apr 2013 07:29:00 -0700
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	David Laight <David.Laight@...lab.com>
Cc:	Ingo Molnar <mingo@...nel.org>, "H. Peter Anvin" <hpa@...or.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	David Miller <davem@...emloft.net>,
	"Theodore Ts'o" <tytso@....edu>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	"the arch/x86 maintainers" <x86@...nel.org>,
	Network Development <netdev@...r.kernel.org>,
	"linux-ext4@...r.kernel.org" <linux-ext4@...r.kernel.org>
Subject: Re: Unsigned widening casts of binary "not" operations..

On Tue, Apr 23, 2013 at 1:59 AM, David Laight <David.Laight@...lab.com> wrote:
>
> If the narrower type is signed it is probably even more confusing!
> The high bits will be preserved unless you are masking off bit 31.

Yes. However, that case doesn't trigger with the normal case of small
values. So "~4" works fine with widening, in a way that "~4u" does
not.

Which doesn't mean that you aren't right, it only means that it's
harder to check for in sparse. The hacky little patch I sent out with
already resulted in a lot of noise for things like "~0u" (UINT_MAX)
and the "~4u" use in NLMSG_ALIGNTO, I'd dread to do the same for
signed values.

That said, with the most minimal value analysis (ie only looking at
constants), maybe it wouldn't be too bad. I started out just worrying
about the PAGE_MASK case, though, where we're talking about (somewhat
complicated) generated constants, and then the signed case is largely
irrelevant (although a signed "1 << 31" would - as you say - trigger
this same thing too).

                         Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/