[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] [PATCH v3 2/2] x86/xstate: Make errors in xstate calculations more obvious by crashing the domain



If xcr0_max exceeds xfeature_mask, then something is broken with the CPUID
policy derivation or auditing logic.  If hardware rejects new_bv, then
something is broken with Xen's xstate logic.

In both cases, crash the domain with an obvious error message, to help
highlight the issues.

Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
---
CC: Jan Beulich <JBeulich@xxxxxxxx>

v2:
 * Rebase over changes to patch 1.
v3:
 * Check xcr0_max against xfeature_mask, rather than new_bv.
 * Additional comments explaining what is going on.
---
 xen/arch/x86/xstate.c | 26 +++++++++++++++++++++++---
 1 file changed, 23 insertions(+), 3 deletions(-)

diff --git a/xen/arch/x86/xstate.c b/xen/arch/x86/xstate.c
index 1fbb087..c8197d2 100644
--- a/xen/arch/x86/xstate.c
+++ b/xen/arch/x86/xstate.c
@@ -707,12 +707,32 @@ int handle_xsetbv(u32 index, u64 new_bv)
     if ( index != XCR_XFEATURE_ENABLED_MASK )
         return -EOPNOTSUPP;
 
-    if ( (new_bv & ~xcr0_max) ||
-         (new_bv & ~xfeature_mask) || !valid_xcr0(new_bv) )
+    /*
+     * The CPUID logic shouldn't be able to hand out an XCR0 exceeding Xen's
+     * maximum features, but keep the check for robustness.
+     */
+    if ( unlikely(xcr0_max & ~xfeature_mask) )
+    {
+        gprintk(XENLOG_ERR,
+                "xcr0_max %016" PRIx64 " exceeds hardware max %016" PRIx64 
"\n",
+                new_bv, xfeature_mask);
+        domain_crash(curr->domain);
+
+        return -EINVAL;
+    }
+
+    if ( (new_bv & ~xcr0_max) || !valid_xcr0(new_bv) )
         return -EINVAL;
 
-    if ( !set_xcr0(new_bv) )
+    /* By this point, new_bv really should be accepted by hardware. */
+    if ( unlikely(!set_xcr0(new_bv)) )
+    {
+        gprintk(XENLOG_ERR, "new_bv %016" PRIx64 " rejected by hardware\n",
+                new_bv);
+        domain_crash(curr->domain);
+
         return -EFAULT;
+    }
 
     mask = new_bv & ~curr->arch.xcr0_accum;
     curr->arch.xcr0 = new_bv;
-- 
2.1.4


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.