linux-kernel - Re: [PATCH 02/13] x86/fault: Check user_mode(regs) when validating a stack extension

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: <20181120073925.GC79825@gmail.com>
Date:   Tue, 20 Nov 2018 08:39:25 +0100
From:   Ingo Molnar <mingo@...nel.org>
To:     Andy Lutomirski <luto@...nel.org>
Cc:     x86@...nel.org, LKML <linux-kernel@...r.kernel.org>,
        Yu-cheng Yu <yu-cheng.yu@...el.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Borislav Petkov <bp@...en8.de>
Subject: Re: [PATCH 02/13] x86/fault: Check user_mode(regs) when validating a
 stack extension


* Andy Lutomirski <luto@...nel.org> wrote:

> The fault handling code tries to validate that a page fault from
> user mode that would extend the stack is within a certain range of
> the user SP.  regs->sp is only equal to the user SP if
> user_mode(regs).  In the extremely unlikely event that that
> sw_error_code had the USER bit set but the faulting instruction was
> in the kernel (i.e. the faulting instruction was WRUSS), then the
> *kernel* stack pointer would have been checked, which would be an
> info leak.
> 
> Note to -stable maintainers: don't backport this unless you backport
> CET.  The bug it fixes is unobservable in current kernels.
> 
> Signed-off-by: Andy Lutomirski <luto@...nel.org>
> ---
>  arch/x86/mm/fault.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
> index 91d4d2722f2e..eae7ee3ce89b 100644
> --- a/arch/x86/mm/fault.c
> +++ b/arch/x86/mm/fault.c
> @@ -1377,7 +1377,7 @@ void do_user_addr_fault(struct pt_regs *regs,
>  		bad_area(regs, sw_error_code, address);
>  		return;
>  	}
> -	if (sw_error_code & X86_PF_USER) {
> +	if (user_mode(regs)) {
>  		/*
>  		 * Accessing the stack below %sp is always a bug.
>  		 * The large cushion allows instructions like enter

Note that this check is gone now due to:

  1d8ca3be86eb: x86/mm/fault: Allow stack access below %rsp

Thanks,

	Ingo