linux-kernel - Re: [PATCH entry v2 3/6] x86/entry/xen: Route #DB correctly on Xen PV

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, 6 Jul 2020 10:41:55 +0200
From:   Michal Kubecek <mkubecek@...e.cz>
To:     Andy Lutomirski <luto@...nel.org>
Cc:     x86@...nel.org, Andrew Cooper <andrew.cooper3@...rix.com>,
        Juergen Gross <jgross@...e.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH entry v2 3/6] x86/entry/xen: Route #DB correctly on Xen PV

On Fri, Jul 03, 2020 at 10:02:55AM -0700, Andy Lutomirski wrote:
> On Xen PV, #DB doesn't use IST.  We still need to correctly route it
> depending on whether it came from user or kernel mode.
> 
> This patch gets rid of DECLARE/DEFINE_IDTENTRY_XEN -- it was too
> hard to follow the logic.  Instead, route #DB and NMI through
> DECLARE/DEFINE_IDTENTRY_RAW on Xen, and do the right thing for #DB.
> Also add more warnings to the exc_debug* handlers to make this type
> of failure more obvious.
> 
> This fixes various forms of corruption that happen when usermode
> triggers #DB on Xen PV.
> 
> Fixes: 4c0dcd8350a0 ("x86/entry: Implement user mode C entry points for #DB and #MCE")
> Signed-off-by: Andy Lutomirski <luto@...nel.org>
> ---
>  arch/x86/include/asm/idtentry.h | 24 ++++++------------------
>  arch/x86/kernel/traps.c         | 12 ++++++++++++
>  arch/x86/xen/enlighten_pv.c     | 28 ++++++++++++++++++++++++----
>  arch/x86/xen/xen-asm_64.S       |  5 ++---
>  4 files changed, 44 insertions(+), 25 deletions(-)
> 
> diff --git a/arch/x86/include/asm/idtentry.h b/arch/x86/include/asm/idtentry.h
> index cf51c50eb356..94333ac3092b 100644
> --- a/arch/x86/include/asm/idtentry.h
> +++ b/arch/x86/include/asm/idtentry.h
[...]
> @@ -570,11 +554,15 @@ DECLARE_IDTENTRY_MCE(X86_TRAP_MC,	exc_machine_check);
>  
>  /* NMI */
>  DECLARE_IDTENTRY_NMI(X86_TRAP_NMI,	exc_nmi);
> -DECLARE_IDTENTRY_XEN(X86_TRAP_NMI,	nmi);
> +#ifdef CONFIG_XEN_PV
> +DECLARE_IDTENTRY_RAW(X86_TRAP_NMI,	xenpv_exc_nmi);
> +#endif
>  
>  /* #DB */
>  DECLARE_IDTENTRY_DEBUG(X86_TRAP_DB,	exc_debug);
> -DECLARE_IDTENTRY_XEN(X86_TRAP_DB,	debug);
> +#ifdef CONFIG_XEN_PV
> +DECLARE_IDTENTRY_RAW(X86_TRAP_DB,	xenpv_exc_debug);
> +#endif
>  
>  /* #DF */
>  DECLARE_IDTENTRY_DF(X86_TRAP_DF,	exc_double_fault);

Hello,

this patch - now in mainline as commit 13cbc0cd4a30 ("x86/entry/32: Fix
#MC and #DB wiring on x86_32") - seems to break i586 builds with
CONFIG_XEN_PV=y as xenpv_exc_nmi and xenpv_exc_debug are only defined
with CONFIG_X86_64:

[ 1279s] ld: arch/x86/entry/entry_32.o: in function `asm_xenpv_exc_nmi':
[ 1279s] /home/abuild/rpmbuild/BUILD/kernel-pae-5.8.rc4/linux-5.8-rc4/linux-obj/../arch/x86/include/asm/idtentry.h:557: undefined reference to `xenpv_exc_nmi'
[ 1279s] ld: arch/x86/entry/entry_32.o: in function `asm_xenpv_exc_debug':
[ 1279s] /home/abuild/rpmbuild/BUILD/kernel-pae-5.8.rc4/linux-5.8-rc4/linux-obj/../arch/x86/include/asm/idtentry.h:567: undefined reference to `xenpv_exc_debug'

Michal Kubecek