This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[PATCH] [ARC] Fix handling of cpu=... disassembler option value


There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'.  In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This could have been fixed by using strncasecmp and passing length of
cpu_types[i].name as size argument, or using strcasestr, but that would cause
another problem as value `em4,dsp' would be equal to `em' cpy_type.name in
those cases.  Therefore the right solution is to extract value of cpu option
fully and pass it as an argument to parse_cpu_option.

opcodes/ChangeLog:

yyyy-mm-dd  Anton Kolesov  <Anton.Kolesov@synopsys.com>

	* arc-dis.c (parse_disassembler_options): Properly handle cpu=...
	  option when there is an option after it.
---
 opcodes/arc-dis.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/opcodes/arc-dis.c b/opcodes/arc-dis.c
index edd0c07..0121682 100644
--- a/opcodes/arc-dis.c
+++ b/opcodes/arc-dis.c
@@ -844,8 +844,12 @@ parse_disassembler_options (const char *options)
       /* A CPU option?  Cannot use STRING_COMMA_LEN because strncmp is also a
 	 preprocessor macro.  */
       if (strncmp (options, "cpu=", 4) == 0)
-	/* Strip leading `cpu=`.  */
-	enforced_isa_mask = parse_cpu_option (options + 4);
+	{
+	  /* Strip leading `cpu=`.  */
+	  char *options_copy = strdup (options + 4);
+	  enforced_isa_mask = parse_cpu_option (strtok (options_copy, ","));
+	  free (options_copy);
+	}
       else
 	parse_option (options);
 
-- 
2.8.3


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]