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: <20190115190846.GE21622@linux.intel.com>
Date:   Tue, 15 Jan 2019 11:08:46 -0800
From:   Sean Christopherson <sean.j.christopherson@...el.com>
To:     Qian Cai <cai@....pw>
Cc:     Paolo Bonzini <pbonzini@...hat.com>, rkrcmar@...hat.com,
        tglx@...utronix.de, mingo@...hat.com, bp@...en8.de, hpa@...or.com,
        x86@...nel.org, kvm@...r.kernel.org, linux-kernel@...r.kernel.org,
        Josh Poimboeuf <jpoimboe@...hat.com>
Subject: Re: [PATCH] kvm: add proper frame pointer logic for vmx

+cc Josh, for real this time.

On Tue, Jan 15, 2019 at 11:06:17AM -0800, Sean Christopherson wrote:
> +cc Josh
> 
> Josh,
> 
> The issue we're encountering is that vmx_vcpu_run() is marked with
> STACK_FRAME_NON_STANDARD because %rbp is saved/modified/restored when
> transitioning to/from the guest.  But vmx_vcpu_run() also happens to
> be a giant blob of code that gcc will sometime split into separate
> .part.* functions, which means that objtool's symbol lookup to whitelist
> the function doesn't work.
> 
> On Tue, Jan 15, 2019 at 11:43:20AM -0500, Qian Cai wrote:
> > 
> > 
> > On 1/15/19 2:13 AM, Paolo Bonzini wrote:
> > > Hmm, maybe like this:
> > > 
> > > diff --git a/arch/x86/kvm/vmx/vmenter.S b/arch/x86/kvm/vmx/vmenter.S
> > > index bcef2c7e9bc4..33122fa9d4bd 100644
> > > --- a/arch/x86/kvm/vmx/vmenter.S
> > > +++ b/arch/x86/kvm/vmx/vmenter.S
> > > @@ -26,19 +26,17 @@ ENTRY(vmx_vmenter)
> > >  	ret
> > > 
> > >  2:	vmlaunch
> > > +3:
> > >  	ret
> > > 
> > > -3:	cmpb $0, kvm_rebooting
> > > -	jne 4f
> > > -	call kvm_spurious_fault
> > > -4:	ret
> > > -
> > >  	.pushsection .fixup, "ax"
> > > -5:	jmp 3b
> > > +4:	cmpb $0, kvm_rebooting
> > > +	jne 3b
> > > +	jmp kvm_spurious_fault
> > >  	.popsection
> > > 
> > > -	_ASM_EXTABLE(1b, 5b)
> > > -	_ASM_EXTABLE(2b, 5b)
> > > +	_ASM_EXTABLE(1b, 4b)
> > > +	_ASM_EXTABLE(2b, 4b)
> > > 
> > >  ENDPROC(vmx_vmenter)
> > 
> > No, that will not work. The problem is in vmx.o where I just sent another patch
> > for it.
> > 
> > I can see there are five options to solve it.
> > 
> > 1) always inline vmx_vcpu_run()
> > 2) always noinline vmx_vcpu_run()
> > 3) add -fdiable-ipa-fnsplit option to Makefile for vmx.o
> > 4) let STACK_FRAME_NON_STANDARD support part.* syntax.
> > 5) trim-down vmx_vcpu_run() even more to not causing splitting by GCC.
> > 
> > Option 1) and 2) seems give away the decision for user with
> > CONFIG_CC_OPTIMIZE_FOR_(PERFORMANCE/SIZE).
> > 
> > Option 3) prevents other functions there for splitting for optimization.
> > 
> > Option 4) and 5) seems tricky to implement.
> > 
> > I am not more leaning to 3) as only other fuction will miss splitting is
> > vmx_segment_access_rights().
> 
> Option 4) is the most correct, but "tricky" is an understatement.  Unless
> Josh is willing to pick up the task it'll likely have to wait.
> 
> There's actually a few more options:
> 
>  6) Replace "pop %rbp" in the vmx_vmenter() asm blob with an open-coded
>     equivalent, e.g. "mov [%rsp], %rbp; add $8, %rsp".  This runs an end-
>     around on objtool since objtool explicitly keys off "pop %rbp" and NOT
>     "mov ..., %rbp" (which is probably an objtool checking flaw?").
> 
>  7) Move the vmx_vmenter() asm blob and a few other lines of code into a
>     separate helper, e.g. __vmx_vcpu_run(), and mark that as having a
>     non-standard stack frame.
> 
> Option 6) is an ugly (but amusing) hack.
> 
> Arguably we should do option 7) regardless of whether or not objtool gets
> updated to support fnsplit behavior, as it it allows objtool to do proper
> stack checking on the bulk of vmx_vcpu_run().  And it's a pretty trivial
> change.
> 
> I'll send a patch for option 7), which in theory should fix the warning
> for all .configs.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ