Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 23 Jan 2019 10:33:22 -0800
From:   Tom Roeder <tmroeder@google.com>
To:     Sean Christopherson <sean.j.christopherson@intel.com>
Cc:     Paolo Bonzini <pbonzini@redhat.com>,
        Radim =?utf-8?B?S3LEjW3DocWZ?= <rkrcmar@redhat.com>,
        Liran Alon <liran.alon@oracle.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        "H . Peter Anvin" <hpa@zytor.com>, x86@kernel.org,
        kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
        syzbot+ded1696f6b50b615b630@syzkaller.appspotmail.com
Subject: Re: [RFC PATCH] kvm: x86/vmx: Use kzalloc for cached_vmcs12
Message-ID: <20190123183322.GB160275@google.com>
References: <6f79d9be-fa76-3a06-2612-f44f3a18ece7@redhat.com>
 <20190114234728.49239-1-tmroeder@google.com>
 <20190115024304.GD5141@linux.intel.com>
 <20190115175111.GB68985@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190115175111.GB68985@google.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Jan 15, 2019 at 09:51:11AM -0800, Tom Roeder wrote:
> On Mon, Jan 14, 2019 at 06:43:04PM -0800, Sean Christopherson wrote:
> > On Mon, Jan 14, 2019 at 03:47:28PM -0800, Tom Roeder wrote:
> > > This changes the allocation of cached_vmcs12 to use kzalloc instead of
> > > kmalloc. This removes the information leak found by Syzkaller (see
> > > Reported-by) in this case and prevents similar leaks from happening
> > > based on cached_vmcs12.
> > 
> > Is the leak specific to vmx_set_nested_state(), e.g. can we zero out
> > the memory if copy_from_user() fails instead of taking the hit on every
> > allocation?
> 
> I don't know if the leak is specific to vmx_set_nested_state.

I've looked at the code more now, and it looks to me like there might be
other cases where the memory could leak. But I don't know enough of the
flows to be sure. The enter_vmx_operation function is called in
handle_vmon, and no data is copied from the guest immediately after
that. So, it depends on what happens after VMXON.

Even in vmx_set_nested_state, there are about 30 lines of code in
between enter_vmx_operation and copy_from_user, and there are a couple
of cases that cause vmx_set_nested_state to return with an error. So if
we want to fix this by handling all the error paths, I think it might be
cleanest to convert vmx_set_nested_state to use goto error handling,
since that would allow us to clear the allocated memory in one place.

What do you think?