It is possible to utilize 32-bit address encodings in virtual-8086 mode via
an address override instruction prefix. However, the range of the
effective address is still limited to [0x-0xffff]. In such a case, return
error.
Also, linear addresses in virtual-8086 mode are limited to 20 bits. Enforce
such limit by truncating the most significant bytes of the computed linear
address.
Cc: Dave Hansen <dave.han...@linux.intel.com>
Cc: Adam Buchbinder <adam.buchbin...@gmail.com>
Cc: Colin Ian King <colin.k...@canonical.com>
Cc: Lorenzo Stoakes <lstoa...@gmail.com>
Cc: Qiaowei Ren <qiaowei....@intel.com>
Cc: Arnaldo Carvalho de Melo <a...@redhat.com>
Cc: Masami Hiramatsu <mhira...@kernel.org>
Cc: Adrian Hunter <adrian.hun...@intel.com>
Cc: Kees Cook <keesc...@chromium.org>
Cc: Thomas Garnier <thgar...@google.com>
Cc: Peter Zijlstra <pet...@infradead.org>
Cc: Borislav Petkov <b...@suse.de>
Cc: Dmitry Vyukov <dvyu...@google.com>
Cc: Ravi V. Shankar <ravi.v.shan...@intel.com>
Cc: x...@kernel.org
Signed-off-by: Ricardo Neri <ricardo.neri-calde...@linux.intel.com>
---
arch/x86/lib/insn-eval.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/arch/x86/lib/insn-eval.c b/arch/x86/lib/insn-eval.c
index 1d510a6..d43808c 100644
--- a/arch/x86/lib/insn-eval.c
+++ b/arch/x86/lib/insn-eval.c
@@ -862,12 +862,23 @@ static void __user *get_addr_ref_32(struct insn *insn,
struct pt_regs *regs)
goto out;
/*
+ * Even though 32-bit address encodings are allowed in virtual-8086
+ * mode, the address range is still limited to [0x-0xffff].
+ */
+ if (v8086_mode(regs) && (eff_addr & ~0xffff))
+ goto out;
+
+ /*
* Data type long could be 64 bits in size. Ensure that our 32-bit
* effective address is not sign-extended when computing the linear
* address.
*/
linear_addr = (unsigned long)(eff_addr & 0xffffffff) + seg_base_addr;
+ /* Limit linear address to 20 bits */
+ if (v8086_mode(regs))
+ linear_addr &= 0xfffff;
+
out:
return (void __user *)linear_addr;
}
--
2.7.4
