x86/asm: Use "m" operand constraint in WRUSSQ asm template

The WRUSSQ instruction uses a memory operand, so use "m"
operand constraint instead of forcing usage of pointer
register using "r" constraint. The generated assembly
code improves from:

   6ece3:	48 8d 43 f8          	lea    -0x8(%rbx),%rax
   ...
   6eceb:	66 48 0f 38 f5 18    	wrussq %rbx,(%rax)

to:

   6ecea:	66 48 0f 38 f5 43 f8 	wrussq %rax,-0x8(%rbx)

Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Link: https://lore.kernel.org/r/20240319171714.76342-1-ubizjak@gmail.com

diff --git a/arch/x86/include/asm/special_insns.h b/arch/x86/include/asm/special_insns.h
index 0ee2ba589492b9b3ebf85e75b70d9a74ae279a1a..aec6e2d3aa1d52e5c8f513e188015a45e9eeaeb2 100644 (file)
--- a/arch/x86/include/asm/special_insns.h
+++ b/arch/x86/include/asm/special_insns.h
@@ -205,9 +205,9 @@ static inline void clwb(volatile void *__p)
 #ifdef CONFIG_X86_USER_SHADOW_STACK
 static inline int write_user_shstk_64(u64 __user *addr, u64 val)
 {
-       asm goto("1: wrussq %[val], (%[addr])\n"
+       asm goto("1: wrussq %[val], %[addr]\n"
                          _ASM_EXTABLE(1b, %l[fail])
-                         :: [addr] "r" (addr), [val] "r" (val)
+                         :: [addr] "m" (*addr), [val] "r" (val)
                          :: fail);
        return 0;
 fail: