linux-kernel - Re: [PATCH] x86_64: uninline TASK

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] [day] [month] [year] [list]

Message-ID: <20190424103833.GB9642@gmail.com>
Date:   Wed, 24 Apr 2019 12:38:33 +0200
From:   Ingo Molnar <mingo@...nel.org>
To:     Andy Lutomirski <luto@...nel.org>
Cc:     Alexey Dobriyan <adobriyan@...il.com>,
        "H. Peter Anvin" <hpa@...or.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>,
        Peter Zijlstra <a.p.zijlstra@...llo.nl>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Al Viro <viro@...iv.linux.org.uk>
Subject: Re: [PATCH] x86_64: uninline TASK_SIZE


* Andy Lutomirski <luto@...nel.org> wrote:

> > Or am I missing some complication?
> 
> Seems like a great idea to me.
> 
> BTW, what the heck is up with get_gate_page()?  I'm struggling to
> understand what it's even trying to do.  If there's an architecture
> that allows a user program to mremap() or otherwise position its gate
> VMA between TASK_SIZE and TASK_SIZE_MAX, then that code is going to
> explode horribly.

I believe it was an old attempt from the times when the vsyscall area 
*didn't* have a vma, at all, and only get_gate_page() kept the mmap 
allocator from overlapping it with a user vma?

Should IMHO be entirely solved by the vma-ification of all things 
vsyscall and vdso, and we can remove this remnant.

> A whole bunch of work in this direction is here:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/luto/linux.git/log/?h=x86/fixes
> 
> It's almost entirely untested.

Please post it as patches once you are somewhat confident in the outcome 
and general direction.

Thanks,

	Ingo