[PATCH v2] x86: do away with HAVE_AS_NEGATIVE_TRUE

Jan Beulich posted 1 patch 11 months, 1 week ago
Failed in applying to current master (apply log)
[PATCH v2] x86: do away with HAVE_AS_NEGATIVE_TRUE
Posted by Jan Beulich 11 months, 1 week ago
There's no real need for the associated probing - we can easily convert
to a uniform value without knowing the specific behavior (note also that
the respective comments weren't fully correct and have gone stale).

No difference in generated code.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
v2: Use "& 1".

--- a/xen/arch/x86/arch.mk
+++ b/xen/arch/x86/arch.mk
@@ -26,10 +26,6 @@ $(call as-option-add,CFLAGS,CC,"invpcid
 $(call as-option-add,CFLAGS,CC,"movdiri %rax$(comma)(%rax)",-DHAVE_AS_MOVDIR)
 $(call as-option-add,CFLAGS,CC,"enqcmd (%rax)$(comma)%rax",-DHAVE_AS_ENQCMD)
 
-# GAS's idea of true is -1.  Clang's idea is 1
-$(call as-option-add,CFLAGS,CC,\
-    ".if ((1 > 0) < 0); .error \"\";.endif",,-DHAVE_AS_NEGATIVE_TRUE)
-
 # Check to see whether the assmbler supports the .nop directive.
 $(call as-option-add,CFLAGS,CC,\
     ".L1: .L2: .nops (.L2 - .L1)$(comma)9",-DHAVE_AS_NOPS_DIRECTIVE)
--- a/xen/arch/x86/include/asm/alternative.h
+++ b/xen/arch/x86/include/asm/alternative.h
@@ -35,19 +35,19 @@ extern void alternative_branches(void);
 #define alt_repl_e(num)    ".LXEN%=_repl_e"#num
 #define alt_repl_len(num)  "(" alt_repl_e(num) " - " alt_repl_s(num) ")"
 
-/* GAS's idea of true is -1, while Clang's idea is 1. */
-#ifdef HAVE_AS_NEGATIVE_TRUE
-# define AS_TRUE "-"
-#else
-# define AS_TRUE ""
-#endif
+/*
+ * GAS's idea of true is sometimes 1 and sometimes -1, while Clang's idea
+ * was consistently 1 up to 6.x (it matches GAS's now).  Transform it to
+ * uniformly 1.
+ */
+#define AS_TRUE(x) "((" x ") & 1)"
 
-#define as_max(a, b) "(("a") ^ ((("a") ^ ("b")) & -("AS_TRUE"(("a") < ("b")))))"
+#define as_max(a, b) "(("a") ^ ((("a") ^ ("b")) & -("AS_TRUE("("a") < ("b")")")))"
 
 #define OLDINSTR(oldinstr, padding)                              \
     ".LXEN%=_orig_s:\n\t" oldinstr "\n .LXEN%=_orig_e:\n\t"      \
     ".LXEN%=_diff = " padding "\n\t"                             \
-    "mknops ("AS_TRUE"(.LXEN%=_diff > 0) * .LXEN%=_diff)\n\t"    \
+    "mknops ("AS_TRUE(".LXEN%=_diff > 0")" * .LXEN%=_diff)\n\t"  \
     ".LXEN%=_orig_p:\n\t"
 
 #define OLDINSTR_1(oldinstr, n1)                                 \
--- a/xen/arch/x86/include/asm/alternative-asm.h
+++ b/xen/arch/x86/include/asm/alternative-asm.h
@@ -29,12 +29,12 @@
 #endif
 .endm
 
-/* GAS's idea of true is -1, while Clang's idea is 1. */
-#ifdef HAVE_AS_NEGATIVE_TRUE
-# define as_true(x) (-(x))
-#else
-# define as_true(x) (x)
-#endif
+/*
+ * GAS's idea of true is sometimes 1 and sometimes -1, while Clang's idea
+ * was consistently 1 up to 6.x (it matches GAS's now).  Transform it to
+ * uniformly 1.
+ */
+#define as_true(x) ((x) & 1)
 
 #define decl_orig(insn, padding)                  \
  .L\@_orig_s: insn; .L\@_orig_e:                  \
Re: [PATCH v2] x86: do away with HAVE_AS_NEGATIVE_TRUE
Posted by Andrew Cooper 11 months, 1 week ago
On 23/05/2023 1:37 pm, Jan Beulich wrote:
> There's no real need for the associated probing - we can easily convert
> to a uniform value without knowing the specific behavior (note also that
> the respective comments weren't fully correct and have gone stale).
>
> No difference in generated code.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>

Thanks.  I do think this form is easier to follow.

~Andrew