x86/tdx: Fix "in-kernel MMIO" check

TDX only supports kernel-initiated MMIO operations. The handle_mmio()
function checks if the #VE exception occurred in the kernel and rejects
the operation if it did not.

However, userspace can deceive the kernel into performing MMIO on its
behalf. For example, if userspace can point a syscall to an MMIO address,
syscall does get_user() or put_user() on it, triggering MMIO #VE. The
kernel will treat the #VE as in-kernel MMIO.

Ensure that the target MMIO address is within the kernel before decoding
instruction.

Fixes: 31d58c4 ("x86/tdx: Handle in-kernel MMIO")
Signed-off-by: Alexey Gladkov (Intel) <[email protected]>
Signed-off-by: Dave Hansen <[email protected]>
Reviewed-by: Kirill A. Shutemov <[email protected]>
Acked-by: Dave Hansen <[email protected]>
Cc:[email protected]
Link: https://lore.kernel.org/all/565a804b80387970460a4ebc67c88d1380f61ad1.1726237595.git.legion%40kernel.org

Author: legionus

arch/x86/coco/tdx/tdx.c

@@ -16,6 +16,7 @@
 #include <asm/insn-eval.h>
 #include <asm/pgtable.h>
 #include <asm/set_memory.h>
+#include <asm/traps.h>
 
 /* MMIO direction */
 #define EPT_READ	0
@@ -433,6 +434,11 @@ static int handle_mmio(struct pt_regs *regs, struct ve_info *ve)
 			return -EINVAL;
 	}
 
+	if (!fault_in_kernel_space(ve->gla)) {
+		WARN_ONCE(1, "Access to userspace address is not supported");
+		return -EINVAL;
+	}
+
 	/*
 	 * Reject EPT violation #VEs that split pages.
 	 *