linux-kernel - Re: [RESEND PATCH v10 06/10] vmx: spp: Set up SPP paging table at vmentry/vmexit

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:   Mon, 13 Jan 2020 13:47:13 -0800
From:   Sean Christopherson <sean.j.christopherson@...el.com>
To:     Adalbert Lazăr <alazar@...defender.com>
Cc:     Yang Weijiang <weijiang.yang@...el.com>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org, pbonzini@...hat.com,
        jmattson@...gle.com, yu.c.zhang@...ux.intel.com,
        edwin.zhai@...el.com, tamas@...engyel.com,
        mathieu.tarral@...tonmail.com
Subject: Re: [RESEND PATCH v10 06/10] vmx: spp: Set up SPP paging table at
 vmentry/vmexit

On Mon, Jan 13, 2020 at 08:55:46PM +0200, Adalbert Lazăr wrote:
> On Mon, 13 Jan 2020 09:33:58 -0800, Sean Christopherson <sean.j.christopherson@...el.com> wrote:
> > On Mon, Jan 13, 2020 at 04:10:50PM +0800, Yang Weijiang wrote:
> > > On Fri, Jan 10, 2020 at 10:04:59AM -0800, Sean Christopherson wrote:
> > > > On Thu, Jan 02, 2020 at 02:13:15PM +0800, Yang Weijiang wrote:
> > > > > @@ -3585,7 +3602,30 @@ static bool fast_page_fault(struct kvm_vcpu *vcpu, gva_t gva, int level,
> > > > >  		if ((error_code & PFERR_WRITE_MASK) &&
> > > > >  		    spte_can_locklessly_be_made_writable(spte))
> > > > >  		{
> > > > > -			new_spte |= PT_WRITABLE_MASK;
> > > > > +			/*
> > > > > +			 * Record write protect fault caused by
> > > > > +			 * Sub-page Protection, let VMI decide
> > > > > +			 * the next step.
> > > > > +			 */
> > > > > +			if (spte & PT_SPP_MASK) {
> > > > > +				int len = kvm_x86_ops->get_inst_len(vcpu);
> > > > 
> > > > There's got to be a better way to handle SPP exits than adding a helper
> > > > to retrieve the instruction length.
> > > >
> > > The fault instruction was skipped by kvm_skip_emulated_instruction()
> > > before, but Paolo suggested leave the re-do or skip option to user-space
> > > to make it flexible for write protection or write tracking, so return
> > > length to user-space.
> > 
> > Sorry, my comment was unclear.  I have no objection to punting the fault
> > to userspace, it's the mechanics of how it's done that I dislike.
> > 
> > Specifically, (a) using run->exit_reason to propagate the SPP exit up the
> > stack, e.g. instead of modifying affected call stacks to play nice with
> > any exit to userspace, (b) assuming ->get_insn_len() will always be
> > accurate, e.g. see the various caveats in skip_emulated_instruction() for
> > both VMX and SVM, and (c) duplicating the state capture code in every
> > location that can encounter a SPP fault.
> > 
> > What I'm hoping is that it's possible to modify the call stacks to
> > explicitly propagate an exit to userspace and/or SPP fault, and shove all
> > the state capture into a common location, e.g. handle_ept_violation().
> > 
> > Side topic, assuming the userspace VMI is going to be instrospecting the
> > faulting instruction, won't it decode the instruction?  I.e. calculate
> > the instruction length anyways?
> 
> Indeed, we decode the instruction from userspace. I don't know if the
> instruction length helps other projects. Added Tamas and Mathieu.
> 
> In our last VMI API proposal, the breakpoint event had the instruction
> length sent to userspace, but I can't remember why.

INT3 is trap-like, i.e. the VM-Exit occurs after the instruction retires.
It's impossible for software to know how far to unwind RIP without the
instruction length being provided by hardware/KVM, e.g. if the guest is
being silly and prepends ignored prefixes on the INT3.

Self-aware software has a priori knowledge of what's being patched in,
and practically speaking I don't any well-behaved sane software uses
prefixes with INT3, but from a VMM's perspective it's legal and possible.

> 
> https://lore.kernel.org/kvm/20190809160047.8319-62-alazar@bitdefender.com/