x86: handle SVR4 escaped binary operators

PR gas/4572

When / is a comment character, its use as binary "divide" operator needs
escaping by a backslash. Besides the scrubber needing to support this
(addressed in an earlier change), there are also a few provisions needed
in target specific operator handling.

As the spec calls for % and * to also be escaped because of being
"overloaded", also recognize these, despite the overloading there not
really preventing their use as operators in most (%) or all (*) cases,
given the way how the rest of the assembler works.

To bring source and testsuite in line, also drop the TE_I386AIX part of
the respective conditional, as i?86-*-aix* support had been removed a
while ago.

6 files changed, 95 insertions, 7 deletions

diff --git a/gas/ChangeLog b/gas/ChangeLog
index 67fabb9..2aed48f 100644
--- a/gas/ChangeLog
+++ b/gas/ChangeLog
@@ -1,6 +1,18 @@
 2020-07-20  Jan Beulich  <jbeulich@suse.com>
 
 	PR gas/4572
+	* config/tc-i386.c (i386_comment_chars): Drop TE_I386AIX from
+	conditional around it.
+	(md_begin): Insert backslash into operand_chars[] when slash is
+	a comment character.
+	* config/tc-i386-intel.c (i386_operator): Recognize \/, \%, and
+	\* as operators when / may be a comment character.
+	* testsuite/gas/i386/svr4.s, testsuite/gas/i386/svr4.d: New.
+	* testsuite/gas/i386/i386.exp: Run new test.
+
+2020-07-20  Jan Beulich  <jbeulich@suse.com>
+
+	PR gas/4572
 	* app.c (last_char): Drop TC_ARM conditional around it.
 	(struct app_save): Drop TC_ARM conditional around last_char.
 	(app_push, app_pop): Drop TC_ARM conditional from last_char
diff --git a/gas/config/tc-i386-intel.c b/gas/config/tc-i386-intel.c
index b326d30..a98771b 100644
--- a/gas/config/tc-i386-intel.c
+++ b/gas/config/tc-i386-intel.c
@@ -123,6 +123,16 @@ operatorT i386_operator (const char *name, unsigned int operands, char *pc)
 {
   unsigned int j;
 
+#ifdef SVR4_COMMENT_CHARS
+  if (!name && operands == 2 && *input_line_pointer == '\\')
+    switch (input_line_pointer[1])
+      {
+      case '/': input_line_pointer += 2; return O_divide;
+      case '%': input_line_pointer += 2; return O_modulus;
+      case '*': input_line_pointer += 2; return O_multiply;
+      }
+#endif
+
   if (!intel_syntax)
     return O_absent;
 
diff --git a/gas/config/tc-i386.c b/gas/config/tc-i386.c
index 5b0a014e..9ab8413 100644
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -480,13 +480,12 @@ const char extra_symbol_chars[] = "*%-([{}"
 #endif
 	;
 
-#if (defined (TE_I386AIX)				\
-     || ((defined (OBJ_ELF) || defined (OBJ_MAYBE_ELF))	\
-	 && !defined (TE_GNU)				\
-	 && !defined (TE_LINUX)				\
-	 && !defined (TE_FreeBSD)			\
-	 && !defined (TE_DragonFly)			\
-	 && !defined (TE_NetBSD)))
+#if ((defined (OBJ_ELF) || defined (OBJ_MAYBE_ELF))	\
+     && !defined (TE_GNU)				\
+     && !defined (TE_LINUX)				\
+     && !defined (TE_FreeBSD)				\
+     && !defined (TE_DragonFly)				\
+     && !defined (TE_NetBSD))
 /* This array holds the chars that always start a comment.  If the
    pre-processor is disabled, these aren't very useful.  The option
    --divide will remove '/' from this list.  */
@@ -3116,6 +3115,10 @@ md_begin (void)
 	    mnemonic_chars[c] = c;
 	    operand_chars[c] = c;
 	  }
+#ifdef SVR4_COMMENT_CHARS
+	else if (c == '\\' && strchr (i386_comment_chars, '/'))
+	  operand_chars[c] = c;
+#endif
 
 	if (ISALPHA (c) || ISDIGIT (c))
 	  identifier_chars[c] = c;
diff --git a/gas/testsuite/gas/i386/i386.exp b/gas/testsuite/gas/i386/i386.exp
index f401732..a91ad79 100644
--- a/gas/testsuite/gas/i386/i386.exp
+++ b/gas/testsuite/gas/i386/i386.exp
@@ -692,6 +692,14 @@ if [expr [istarget "i*86-*-*"] || [istarget "x86_64-*-*"]] then {
 	run_dump_test "dw2-compressed-1"
 	run_dump_test "dw2-compressed-3a"
 	run_dump_test "dw2-compressed-3b"
+
+	if {![istarget "*-*-dragonfly*"]
+	    && ![istarget "*-*-gnu*"]
+	    && ![istarget "*-*-freebsd*"]
+	    && ![istarget "*-*-linux*"]
+	    && ![istarget "*-*-netbsd*"]} then {
+	    run_dump_test "svr4"
+	}
     }
 }
 
diff --git a/gas/testsuite/gas/i386/svr4.d b/gas/testsuite/gas/i386/svr4.d
new file mode 100644
index 0000000..3a46354
--- /dev/null
+++ b/gas/testsuite/gas/i386/svr4.d
@@ -0,0 +1,24 @@
+#objdump: -dtw
+#name: SVR4 comment char escape handling
+
+.*: +file format .*
+
+SYMBOL TABLE:
+0+00 .* \.text[ 	]+0+ \.text
+0+00 .* \.data[ 	]+0+ \.data
+0+00 .* \.bss[ 	]+0+ \.bss
+0+00 .* \.text[ 	]+0+ svr4
+0+04 .* \*ABS\*[ 	]+0+ a
+0+03 .* \*ABS\*[ 	]+0+ b
+0+4c .* \*ABS\*[ 	]+0+ c
+
+Disassembly of section .text:
+
+0+0 <svr4>:
+[	 ]*[0-9a-f]+:[	 ]+b0 07[	 ]+mov    \$0x7,%al
+[	 ]*[0-9a-f]+:[	 ]+b0 01[	 ]+mov    \$0x1,%al
+[	 ]*[0-9a-f]+:[	 ]+b0 1e[	 ]+mov    \$0x1e,%al
+[	 ]*[0-9a-f]+:[	 ]+b0 05[	 ]+mov    \$0x5,%al
+[	 ]*[0-9a-f]+:[	 ]+b0 02[	 ]+mov    \$0x2,%al
+[	 ]*[0-9a-f]+:[	 ]+b0 33[	 ]+mov    \$0x33,%al
+#pass
diff --git a/gas/testsuite/gas/i386/svr4.s b/gas/testsuite/gas/i386/svr4.s
new file mode 100644
index 0000000..4565e01
--- /dev/null
+++ b/gas/testsuite/gas/i386/svr4.s
@@ -0,0 +1,31 @@
+	.text
+
+	.if 1 / 2
+	.else
+	.error
+	.endif
+
+	.if 1 \/ 2
+	.error
+	.endif
+
+	.if 4 \% 2
+	.error
+	.endif
+
+	.if 1 \* 0
+	.error
+	.endif
+
+svr4:
+	mov	$(15 \/ 2), %al
+	mov	$(15 \% 2), %al
+	mov	$(15 \* 2), %al
+
+	.byte	0xb0, 17 \/ 3
+	.byte	0xb0, 17 \% 3
+	.byte	0xb0, 17 \* 3
+
+	.equiv	a, 19 \/ 4
+	.equiv	b, 19 \% 4
+	.equiv	c, 19 \* 4