linux-kernel - Re: [PATCH] mm/mmap: replace if (cond) BUG() with BUG

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, 7 Jan 2021 18:28:29 +0100
From:   Vlastimil Babka <vbabka@...e.cz>
To:     Hugh Dickins <hughd@...gle.com>,
        Andrea Arcangeli <aarcange@...hat.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Alex Shi <alex.shi@...ux.alibaba.com>,
        Minchan Kim <minchan@...nel.org>,
        Michal Hocko <mhocko@...e.com>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm/mmap: replace if (cond) BUG() with BUG_ON()

On 1/6/21 9:18 PM, Hugh Dickins wrote:
> On Wed, 6 Jan 2021, Andrea Arcangeli wrote:
>> 
>> I'd be surprised if the kernel can boot with BUG_ON() defined as "do
>> {}while(0)" so I guess it doesn't make any difference.
> 
> I had been afraid of that too, when CONFIG_BUG is not set:
> but I think it's actually "if (cond) do {} while (0)".

It's a maze of configs and arch-specific vs generic headers, but I do see this
in include/asm-generic/bug.h:

#else /* !CONFIG_BUG */
#ifndef HAVE_ARCH_BUG
#define BUG() do {} while (1)
#endif

So seems to me there *are* configurations possible where side-effects are indeed
thrown away, right?

WARN_ON is different as the result of the "inner" condition should be further
usable for constructing "outer" conditions:

(still in !CONFIG_BUG section)
#ifndef HAVE_ARCH_WARN_ON
#define WARN_ON(condition) ({
        int __ret_warn_on = !!(condition);
        unlikely(__ret_warn_on);
})
#endif

For completeness let's look at our own extensions when VM_DEBUG is disabled,
which is quite analogical to disabling CONFIG_BUG and thus it should better be
consistent with the generic stuff.

#define VM_BUG_ON(cond) BUILD_BUG_ON_INVALID(cond)

where BUILD_BUG_ON_INVALID generates no code, so it's consistent with BUG_ON()
and !CONFIG_BUG.

#define VM_WARN_ON(cond) BUILD_BUG_ON_INVALID(cond)

... well that's not consistent with WARN_ON. Hmm if you have asked me before I
checked, I would have said that it is, that I checked it already in the past
and/or there was some discussion already about it. Memory is failing me it
seems. We should better fix this?