linux-kernel - Re: [PATCH v2 3/3] fork: lock VMAs of the parent process when forking

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:   Sat, 5 Aug 2023 01:25:48 +0200
From:   Mateusz Guzik <mjguzik@...il.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Suren Baghdasaryan <surenb@...gle.com>, akpm@...ux-foundation.org,
        regressions@...mhuis.info, bagasdotme@...il.com,
        jacobly.alt@...il.com, willy@...radead.org,
        liam.howlett@...cle.com, david@...hat.com, peterx@...hat.com,
        ldufour@...ux.ibm.com, linux-kernel@...r.kernel.org,
        linux-mm@...ck.org, linuxppc-dev@...ts.ozlabs.org,
        linux-arm-kernel@...ts.infradead.org, gregkh@...uxfoundation.org,
        regressions@...ts.linux.dev, Jiri Slaby <jirislaby@...nel.org>,
        Holger Hoffstätte <holger@...lied-asynchrony.com>,
        stable@...r.kernel.org
Subject: Re: [PATCH v2 3/3] fork: lock VMAs of the parent process when forking

On 8/5/23, Linus Torvalds <torvalds@...ux-foundation.org> wrote:
> On Fri, 4 Aug 2023 at 14:46, Mateusz Guzik <mjguzik@...il.com> wrote:
>>
>> I don't see it mentioned in the discussion, so at a risk of ruffling
>> feathers or looking really bad I'm going to ask: is the locking of any
>> use if the forking process is single-threaded? T
>
> Sadly, we've always been able to access the mm from other processes,
> so the locking is - I think - unavoidable.
>
> And some of those "access from other processes" aren't even uncommon
> or special. It's things like "ps" etc, that do it just to see the
> process name and arguments.
>

I know of these guys, I think they are excluded as is -- they go
through access_remote_vm, starting with:
        if (mmap_read_lock_killable(mm))
                return 0;

while dup_mmap already write locks the parent's mm.

I don't see any surprise relocks of the semaphore.

Granted, should someone *bypass* this mechanism the above would be moot.

-- 
Mateusz Guzik <mjguzik gmail.com>