linux-kernel - Re: [PATCH v16 18/22] platform/x86: Intel SGX driver

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]

Message-ID: <20181115201625.GA25953@linux.intel.com>
Date:   Thu, 15 Nov 2018 22:16:25 +0200
From:   Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>
To:     Sean Christopherson <sean.j.christopherson@...el.com>
Cc:     x86@...nel.org, platform-driver-x86@...r.kernel.org,
        linux-sgx@...r.kernel.org, dave.hansen@...el.com,
        nhorman@...hat.com, npmccallum@...hat.com, serge.ayoun@...el.com,
        shay.katz-zamir@...el.com, haitao.huang@...el.com,
        andriy.shevchenko@...ux.intel.com, tglx@...utronix.de,
        kai.svahn@...el.com, Suresh Siddha <suresh.b.siddha@...el.com>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>,
        Darren Hart <dvhart@...radead.org>,
        Andy Shevchenko <andy@...radead.org>,
        "open list:X86 ARCHITECTURE (32-BIT AND 64-BIT)" 
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v16 18/22] platform/x86: Intel SGX driver

On Thu, Nov 15, 2018 at 10:04:06PM +0200, Jarkko Sakkinen wrote:
> On Thu, Nov 15, 2018 at 10:00:02PM +0200, Jarkko Sakkinen wrote:
> > On Thu, Nov 08, 2018 at 04:46:03PM +0200, Jarkko Sakkinen wrote:
> > > On Wed, Nov 07, 2018 at 10:00:57AM -0800, Sean Christopherson wrote:
> > > > What do we gain by a single buffer vs. separate buffers?  The ioctl()
> > > > would be slightly smaller but it seems like the actual code would be
> > > > more complex.
> > > 
> > > I'm fine with either. It was just a suggestion.
> > > 
> > > > The enclave build process also utilizes the backing as temp storage
> > > > to avoid having to alloc kernel memory when queueing pages to be added
> > > > by the worker thread (which reminds me that I wanted to document why a
> > > > worker thread is used).  Keeping this behavior would effectively make
> > > > providing backing mandatory.
> > > 
> > > Would it be a problem just allocate those pages with alloc_page() and
> > > free them in the worker thread?
> > > 
> > > > Are there any potential complications with ENCLS consuming userspace
> > > > pointers?  We'd have to wrap them with user_access_{begin,end}() and
> > > > probably tweak the fixup, but I assume having the fixup handler means
> > > > we're generally ok?
> > > 
> > > Last time I did it I used get_user_pages() for pinning. I'm not sure
> > > why I should do anything but just re-use that.
> > 
> > What about VA page swapping? Not saying that it'd have to be done right
> > now but we need to answer whether it is enclave local or a global asset.
> > If it is local it would also require an argument.
> > 
> > I will most likely won't fix this for v17 because this detail needs
> > careful consideration.
> 
> I wonder if you can map shmem file to process address space so that you
> get it accounted for the process? That would be optimal for us. This way
> this won't become an API issue.
> 
> Yeah, as I started to implement this I realized these issues with the
> API side that will arise. Even doing vm_mmap() in the kernel code would
> be better than taking addresses through the ioctl. That is another
> option.

This is how strongly think. I can go with VMAs for swapping *but* if
that is the route I would recommend using vm_mmap() instead of taking
pointer from user space. This way the way swapping is done can be always
changed. It will be a huge lock-in to do it otherwise.

/Jarkko