linux-kernel - Re: [PATCH 8/9] x86: kvm_hv_set_msr(): use __put_user() instead of 32bit __clear

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, 30 May 2020 20:27:14 +0100
From:   Al Viro <viro@...iv.linux.org.uk>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Paolo Bonzini <pbonzini@...hat.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>,
        KVM list <kvm@...r.kernel.org>
Subject: Re: [PATCH 8/9] x86: kvm_hv_set_msr(): use __put_user() instead of
 32bit __clear_user()

On Sat, May 30, 2020 at 08:19:40PM +0100, Al Viro wrote:
> On Sat, May 30, 2020 at 11:52:44AM -0700, Linus Torvalds wrote:
> 
> > And I don't understand why you mention set_fs() vs access_ok(). None
> > of this code has anything that messes with set_fs(). The access_ok()
> > is garbage and shouldn't exist, and those user accesses should all use
> > the checking versions and the double underscores are wrong.
> > 
> > I have no idea why you think the double underscores could _possibly_
> > be worth defending.
> 
> I do not.  What I'm saying is that this just might be a beast different
> from *both* __... and the normal ones.  I'm not saying that this
> __put_user() (or __clear_user(), etc.) is the right primitive here.
> If anything, it's closer to the situation for (x86) copy_stack_trace().

... and no, I'm not saying that copy_stack_trace() should stay with
__get_user() either.  It feels like we are lacking primitives needed
to express that cleanly and copy_stack_trace() currently cobbles something
up out of what we have.  Which works for arch-specific code, but yes,
that kind of thing is brittle for arch-independent places like virt/kvm;
I wonder if e.g. s390 is really OK there.