linux-kernel - [PATCH] x86/xen: init %gs very early to avoid page faults with stack protector

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-next>] [day] [month] [year] [list]

Message-Id: <20180201121601.1268-1-jgross@suse.com>
Date:   Thu,  1 Feb 2018 13:16:01 +0100
From:   Juergen Gross <jgross@...e.com>
To:     linux-kernel@...r.kernel.org, xen-devel@...ts.xenproject.org
Cc:     boris.ostrovsky@...cle.com, Juergen Gross <jgross@...e.com>,
        stable@...r.kernel.org
Subject: [PATCH] x86/xen: init %gs very early to avoid page faults with stack protector

When running as Xen pv guest %gs is initialized some time after
C code is started. Depending on stack protector usage this might be
too late, resulting in page faults.

So setup %gs and MSR_GS_BASE in assembly code already.

Cc: stable@...r.kernel.org
Signed-off-by: Juergen Gross <jgross@...e.com>
---
 arch/x86/xen/xen-head.S | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/arch/x86/xen/xen-head.S b/arch/x86/xen/xen-head.S
index 497cc55a0c16..b47d87076efb 100644
--- a/arch/x86/xen/xen-head.S
+++ b/arch/x86/xen/xen-head.S
@@ -9,7 +9,9 @@
 
 #include <asm/boot.h>
 #include <asm/asm.h>
+#include <asm/msr.h>
 #include <asm/page_types.h>
+#include <asm/percpu.h>
 #include <asm/unwind_hints.h>
 
 #include <xen/interface/elfnote.h>
@@ -35,6 +37,18 @@ ENTRY(startup_xen)
 	mov %_ASM_SI, xen_start_info
 	mov $init_thread_union+THREAD_SIZE, %_ASM_SP
 
+	/* Set up %gs.
+	 *
+	 * The base of %gs always points to the bottom of the irqstack
+	 * union.  If the stack protector canary is enabled, it is
+	 * located at %gs:40.  Note that, on SMP, the boot cpu uses
+	 * init data section till per cpu areas are set up.
+	 */
+	movl	$MSR_GS_BASE,%ecx
+	movq	$INIT_PER_CPU_VAR(irq_stack_union),%rax
+	cdq
+	wrmsr
+
 	jmp xen_start_kernel
 END(startup_xen)
 	__FINIT
-- 
2.13.6