From: Andrew Cooper <andrew.cooper3@citrix.com>
Subject: x86/traps: Fix %dr6 handing in #DB handler

Most bits in %dr6 accumulate, rather than being set directly based on the
current source of #DB.  Have the handler follow the manuals guidance, which
avoids leaking hypervisor debugging activities into guest context.

This is part of XSA-260 / CVE-2018-8897.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/traps.c	2018-04-13 15:29:36.006747135 +0200
+++ b/xen/arch/x86/traps.c	2018-04-13 15:44:57.015516185 +0200
@@ -1761,11 +1761,36 @@ static void ler_enable(void)
 
 void do_debug(struct cpu_user_regs *regs)
 {
+    unsigned long dr6;
     struct vcpu *v = current;
 
+    /* Stash dr6 as early as possible. */
+    dr6 = read_debugreg(6);
+
     if ( debugger_trap_entry(TRAP_debug, regs) )
         return;
 
+    /*
+     * At the time of writing (March 2018), on the subject of %dr6:
+     *
+     * The Intel manual says:
+     *   Certain debug exceptions may clear bits 0-3. The remaining contents
+     *   of the DR6 register are never cleared by the processor. To avoid
+     *   confusion in identifying debug exceptions, debug handlers should
+     *   clear the register (except bit 16, which they should set) before
+     *   returning to the interrupted task.
+     *
+     * The AMD manual says:
+     *   Bits 15:13 of the DR6 register are not cleared by the processor and
+     *   must be cleared by software after the contents have been read.
+     *
+     * Some bits are reserved set, some are reserved clear, and some bits
+     * which were previously reserved set are reused and cleared by hardware.
+     * For future compatibility, reset to the default value, which will allow
+     * us to spot any bit being changed by hardware to its non-default value.
+     */
+    write_debugreg(6, X86_DR6_DEFAULT);
+
     if ( !guest_mode(regs) )
     {
         if ( regs->eflags & X86_EFLAGS_TF )
@@ -1798,7 +1823,8 @@ void do_debug(struct cpu_user_regs *regs
     }
 
     /* Save debug status register where guest OS can peek at it */
-    v->arch.debugreg[6] = read_debugreg(6);
+    v->arch.debugreg[6] |= (dr6 & ~X86_DR6_DEFAULT);
+    v->arch.debugreg[6] &= (dr6 | ~X86_DR6_DEFAULT);
 
     ler_enable();
     pv_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
--- a/xen/include/asm-x86/debugreg.h	2015-02-11 09:36:29.000000000 +0100
+++ b/xen/include/asm-x86/debugreg.h	2018-04-13 15:44:57.015516185 +0200
@@ -24,6 +24,8 @@
 #define DR_STATUS_RESERVED_ZERO (~0xffffeffful) /* Reserved, read as zero */
 #define DR_STATUS_RESERVED_ONE  0xffff0ff0ul /* Reserved, read as one */
 
+#define X86_DR6_DEFAULT 0xffff0ff0ul    /* Default %dr6 value. */
+
 /* Now define a bunch of things for manipulating the control register.
    The top two bytes of the control register consist of 4 fields of 4
    bits - each field corresponds to one of the four debug registers,