[syzbot] upstream test error: KASAN: invalid-access Read in __entry_tramp_text_end

[Josh Poimboeuf]

On Tue, Sep 28, 2021 at 11:35:43AM +0100, Mark Rutland wrote:
> > In the other x86 thread Josh Poimboeuf suggested to use asm goto to a
> > cold part of the function instead of .fixup:
> > https://lore.kernel.org/lkml/20210927234543.6waods7rraxseind@treble/
> > This sounds like a more reliable solution that will cause less
> > maintenance burden. Would it work for arm64 as well?
> 
> Maybe we can use that when CC_HAS_ASM_GOTO_OUTPUT is avaiable, but in
> general we can't rely on asm goto supporting output arguments (and IIRC
> GCC doesn't support that at all), so we'd still have to support the
> current fixup scheme.

Even without CC_HAS_ASM_GOTO_OUTPUT it should still be possible to hack
something together if you split the original insn asm and the extable
asm into separate statements, like:

diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h
index 6b52182e178a..8f62469f2027 100644
--- a/arch/x86/include/asm/msr.h
+++ b/arch/x86/include/asm/msr.h
@@ -137,20 +139,21 @@ static inline unsigned long long native_read_msr_safe(unsigned int msr,
 {
 	DECLARE_ARGS(val, low, high);
 
-	asm volatile("2: rdmsr ; xor %[err],%[err]\n"
-		     "1:\n\t"
-		     ".section .fixup,\"ax\"\n\t"
-		     "3: mov %[fault],%[err]\n\t"
-		     "xorl %%eax, %%eax\n\t"
-		     "xorl %%edx, %%edx\n\t"
-		     "jmp 1b\n\t"
-		     ".previous\n\t"
-		     _ASM_EXTABLE(2b, 3b)
-		     : [err] "=r" (*err), EAX_EDX_RET(val, low, high)
-		     : "c" (msr), [fault] "i" (-EIO));
+	*err = 0;
+	asm volatile("417: rdmsr\n"
+		     : EAX_EDX_RET(val, low, high)
+		     : "c" (msr));
+	asm_volatile_goto(_ASM_EXTABLE(417b, %l[Efault]) :::: Efault);
+
+done:
 	if (tracepoint_enabled(read_msr))
 		do_trace_read_msr(msr, EAX_EDX_VAL(val, low, high), *err);
 	return EAX_EDX_VAL(val, low, high);
+
+Efault:
+	*err = -EIO;
+	ZERO_ARGS(val, low, high);
+	goto done;
 }
 
 /* Can be uninlined because referenced by paravirt */