[PATCH 2/6] x86/shadow: Fixes to hvm_emulate_insn_fetch()

[Andrew Cooper <andrew.cooper3@citrix.com>]

Zero-legnth reads are jump-target segmentation checks; never serve them from
the cache.

Force insn_off to a single byte, as offset can wrap around or truncate with
respect to sh_ctxt->insn_buf_eip under a number of normal circumstances.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Tim Deegan <tim@xen.org>
CC: Jan Beulich <JBeulich@suse.com>
---
 xen/arch/x86/mm/shadow/common.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/xen/arch/x86/mm/shadow/common.c b/xen/arch/x86/mm/shadow/common.c
index 2e64a77..deea03a 100644
--- a/xen/arch/x86/mm/shadow/common.c
+++ b/xen/arch/x86/mm/shadow/common.c
@@ -235,12 +235,16 @@ hvm_emulate_insn_fetch(enum x86_segment seg,
 {
     struct sh_emulate_ctxt *sh_ctxt =
         container_of(ctxt, struct sh_emulate_ctxt, ctxt);
-    unsigned int insn_off = offset - sh_ctxt->insn_buf_eip;
+    /* Careful, as offset can wrap or truncate WRT insn_buf_eip. */
+    uint8_t insn_off = offset - sh_ctxt->insn_buf_eip;
 
     ASSERT(seg == x86_seg_cs);
 
-    /* Fall back if requested bytes are not in the prefetch cache. */
-    if ( unlikely((insn_off + bytes) > sh_ctxt->insn_buf_bytes) )
+    /*
+     * Fall back if requested bytes are not in the prefetch cache, but always
+     * perform the zero-length read for segmentation purposes.
+     */
+    if ( !bytes || unlikely((insn_off + bytes) > sh_ctxt->insn_buf_bytes) )
         return hvm_read(seg, offset, p_data, bytes,
                         hvm_access_insn_fetch, sh_ctxt);
 
-- 
2.1.4


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel